RELATIONSHIPS BETWEEN LANDSCAPE STRUCTURE, COCCINELLID ACTIVITY
DENSITY, AND APHID SUPPRESSION IN SOYBEAN
By
JESSICA MEGAN WOLTZ

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Entomology – Doctor of Philosophy
Ecology, Evolutionary Biology, and Behavior – Dual Major
2013

i

ABSTRACT
RELATIONSHIPS BETWEEN LANDSCAPE STRUCTURE, COCCINELLID ACTIVITY
DENSITY, AND APHID SUPPRESSION IN SOYBEAN
By
JESSICA MEGAN WOLTZ

Arthropod natural enemies provide valuable pest suppression in agricultural landscapes,
and numerous studies have demonstrated the effects of landscape structure on natural enemy
abundance and diversity. Coccinellid beetles are important generalist predators in a variety of
agricultural landscapes, and have been found to be particularly important for suppression of the
soybean aphid, Aphis glycines, Matsumura. I assessed the relationship between multiple
landscape metrics and coccinellid abundances in soybean fields using Partial Least Squares
regression. Coccinellid abundances were lowest in landscapes dominated by annual crops grown
in large fields and highest in landscapes dominated by semi-natural habitat, with a higher
diversity of crops grown in smaller fields. In addition to landscape-scale drivers, a large body of
research examines the ability of local habitat management to enhance natural enemy efficacy. I
examined the influence of landscape structure, local habitat management, and their potential
interaction, on coccinellid abundances and A. glycines suppression in soybean. I selected pairs of
soybean fields in landscapes of varying composition and planted buckwheat, Fagopyrum
esculentum, strips adjacent to one field in each pair. A. glycines suppression was high in all
contexts. Coccinellid abundance was higher in buckwheat than in control margins in all
landscapes. Additionally, coccinellid abundance in soybean was positively related to amount of
semi-natural vegetation in the landscape. I found no evidence of an interaction between

ii

landscape and local variables, suggesting that landscape factors drive coccinellid potential to
suppress soybean aphid.
Aphis glycines colonies are patchily distributed within soybean fields. To examine how
coccinellid arrival rate at aphid-infested patches influences A. glycines control, I experimentally
2

varied coccinellid immigration to 1m soybean patches using selective barriers. In a year with
low ambient aphid pressure, naturally-occurring levels of coccinellid immigration to host patches
were sufficient to suppress aphid population growth. In contrast, A. glycines populations
increased dramatically in patches with reduced coccinellid immigration. In a year with higher
ambient aphid colonization, aphid suppression was incomplete. In a system in which natural
enemy populations cannot track pest populations through reproduction alone, natural enemy
immigration to infested patches can result in pest control. Finally, nocturnal predators have been
shown to be responsible for significant predation of soybean herbivores in some regions, yet
research on A. glycines natural enemies has focused on members of the diurnal predator
community. I assessed diel variation of the predator community and A. glycines predation events.
Anthocorids were responsible for the majority of observed predation, which occurred during
daylight hours. Anthocorids and nabids were more active in the afternoon and carabids were
more active at night. Different taxa were observed most frequently on video and within vacuum
samples. Vacuum samples may represent predator abundance more accurately, while video data
can be used to better assess relative time spent foraging. Overall, A. glycines population
dynamics are influenced by a combination of local and landscape effects that structure predator
communities and influence the spatial and temporal interactions between predators this aphid.
Factors that influence early season coccinellid immigration into infested patches show the
greatest potential to improve A. glycines suppression.
iii

Copyright by
JESSICA MEGAN WOLTZ
2013

iv

For Mama, who always wanted me to do whatever would make me happy, despite the fact that
she would have found this whole process to be ridiculous.

v

ACKNOWLEDGEMENTS

I would like to thank Doug Landis for always providing opportunities to develop as a
scientist and for being supportive of my varied academic interests. I would like to thank
members of my graduate committee, Chris DiFonzo, Rufus Isaacs, Kay Gross and Jack Liu for
their guidance and advice. Chris, you will never know how much I appreciate having such a
down-to-earth role model. I am indebted to the Michigan farmers who allowed me to conduct
this research in their fields: Bill Westrate, Jerry & Terry Voyce, John Frayer, Rob Cook, Tom
Betz, Rick Kreeger, Pete Phelps, John Coy, Jerry Grigar, Aaron Keilen, Gary Fisher, Robert
Adamic, Bill Hunt, Rob Richardson, Jack Ringewold, Rick Lawrence, and Mike Scott. I am also
indebted to Paul Horny of the Saginaw Valley Research and Extension Center, Jim Bronson of
the Kellogg Farm and the Kellogg Biological Station LTER Staff, and Brian Graff of the MSU
Agronomy Farm for allowing us to use their fields. I whole-heartedly thank John Davenport for
planting my buckwheat strips and his expertise backing up a tractor and seed-drill onto a trailer. I
am grateful to Marilyn Thelen, Dan Rossman, and Dan Hudson of Michigan State University
Extension and Chris DiFonzo for assistance locating field sites. I could have never collected or
processed all of my data without help from Amanda Cheeseman, Grace Hirzel, Julia Perrone,
Emily Pochubay, Luke Tomczak, Catherine Spencer, Emily Hall, Courtney Rochenbach, Luke
Petersen, Heather Russon, Emily Franklin, Carly Eakin, Chad Eakin, Courtney Weatherbee, Jeff
Shoemaker, Liz Racey, Aaron Adams, Laurel Lindeman, Jordan Shelley, Patrick Kelly, Ian
Lane, and Erica Luu. My graduate degree and research were supported by a University
Distinguished Fellowship from the MSU Graduate School, a KBS LTER Graduate Summer
Research Fellowship, a Pioneer Hi-Bred International Graduate Research Fellowship, and a
USDA AFRI grant. My student Luke Petersen was supported by KBS LTER REU funds. I am
vi

forever grateful to Nate Walton for being my personal electronics expert, mechanical engineer,
and scientific sounding board. I also thank Nick Haddad for always understanding and for giving
the best advice. Thank you to current and previous members of the Landis lab for support. I am
especially thankful to Mary Gardiner for helping me get my first field project up and running,
Ben Werling for eternal patience and good humor in the face of endless statistical questions, and
Julia Perrone for always being willing to jump in with field work, lab samples and copy-editing.
Thank-you so much to Heather Lenartson-Kluge, Carolyn Devereaux, Linda Gallagher, Brooke
Gallagher, Courtneay Smith and the rest of the Entomology Department office staff for always
being on top of things and helping to keep my life running smoothly.
Since coming to MSU I have fallen in love with teaching and outreach. Thank you Barb
Stinnett, Gary Parsons, and Courtneay Smith for keeping the Bug House running and making this
wonderful learning experience available. I am grateful to Chris DiFonzo for allowing me to get
my hands dirty teaching in her class; to Walter Pett for giving me free rein to turn his lesson
plans upside-down; to Diane Ebert-May for helping me develop my very first active-learning
exercises; to Stephen Thomas for getting me excited about educational technology and for giving
me the opportunity to participate in the Foundations of Science MOOC; and to Joyce Parker and
the MSU Graduate School for the invaluable learning and experience I gained from the
Certification in Teaching College Science Program.
I am thankful to Gramma, Bert, Daddy, Adam, Sarah, and Rebecca for unconditional love
and support. I am eternally grateful to Lina Quesada & Jarrod Morrice, Dan Hulbert & Rachel
Naegele, Beth Brisco, Leah Granke, Joseph Tychonievich, and Julia Perrone for costume parties,
Renaissance festivals, Avatar nights, puppy play dates, endless book discussions and generally
making this a wonderful time in my life, as well as for taking care of my furry babies every time

vii

I went out of town. Thanks to Jeff Evans and Aaron Smith for lots of laughs the first year. To
Colin Kremer for being mathemagical. To Jane Smith, for laughs, helping me keep things in
perspective, and always knowing exactly what I mean. To Molly McCullough, Christopher
Sanchez and Alison Grantham for love, support, advice and appreciating my sense of humor. To
Happy, my favorite, who has always been perceptive to my moods and needs; to Scupper, who’s
so irresistible she melts even the coldest Nodak hearts; and to Lucy, for filling my days with
smiles, laughs, cuddles and ears. And finally, to Rob, my darling Tree Beard, who has seen me
lose all my charm on more than one occasion and still keeps hanging around. Thanks for giving
me a reason to learn how to cook, for making me eat all my vegetables, for putting up with fur
and furries, for calmly absorbing emotional storms, and for always being ready for an impromptu
classic country sing-a-long in the living room. You deserve a medal, but I will probably just
make you mashed potatoes and gravy instead.

viii

TABLE OF CONTENTS

LIST OF TABLES

xiii

LIST OF FIGURES

xvi

LITERATURE REVIEW
Introduction
Landscape effects on natural enemies and biocontrol services
Landscape composition
Seminatural habitats
Perennial crop habitats
Habitat and crop diversity
Landscape configuration
Differential responses by habitat generalists and specialists
Natural enemy prey specialization and immigration
Natural enemy dispersal ability and response to landscape
Local habitat effects on insect communities and biocontrol services
Habitat management
Interactions between landscape and local effects on natural enemies
Study system
Coccinellids and biological control of soybean aphid
Coccinellid responses to landscape and habitat management
Conclusions
LITERATURE CITED

1
1
3
3
3
4
5
5
6
7
8
9
10
11
12
14
15
16
18

CHAPTER I: Landscape structure and habitat management differentially influence
coccinellids in an agricultural landscape
Abstract
Introduction
Methods
Site design
Buckwheat establishment and control margin characteristics
Aphid counts
Coccinellid counts
Biocontrol study
Landscape analysis
Statistical analysis
Coccinellid abundance
Biocontrol study
Results
Buckwheat establishment
Landscape analysis
Aphid counts
Coccinellid abundance

29
29
30
33
33
34
34
35
35
37
39
39
40
40
40
41
41
42

ix

Biocontrol study
Discussion
Landscape effects on local management to improve services
Effects of local management on coccinellid beetles
Landscape effects on coccinellid beetles
Effects of predators on soybean aphid density
Conclusions
APPENDIX
LITERATURE CITED

43
44
44
45
46
47
49
51
74

CHAPTER II: Coccinellid response to landscape structure across an agricultural
intensification gradient.
Abstract
Introduction
Methods
Site design
Coccinellid counts
Landscape data
Measures of landscape composition
Abundance of primary land-covers
Cover diversity
Measures of landscape configuration
Crop mesh size
Landscape shape index
Statistical analysis
Results
Coccinellid captures
Landscape variables
PLSR model
Discussion
Conclusions
APPENDIX
LITERATURE CITED

80
80
80
84
84
84
84
85
85
85
86
86
86
87
89
89
90
90
92
94
96
101

CHAPTER III: Coccinellid immigration to infested host patches influences suppression
of Aphis glycines in soybean.
Abstract
Introduction
Methods
Site
Plot establishment
Predator monitoring
Aphid monitoring
Statistical analysis
Predator counts
Aphid abundance

106
106
107
110
110
110
112
113
114
114
115

x

Results
Adult coccinellid captures
Ground-dwelling predators
Foliar- and aerially-dispersing predators
Aphid abundance
Discussion
APPENDIX
LITERATURE CITED
CHAPTER IV: Diel variation in natural enemy community of Aphis glycines in soybean
Abstract
Introduction
Methods: Direct observation study
Study locations
Plot establishment
Aphid infestation
Aphid surveys
Predator sampling
Temperature
Statistical analysis
Aphid surveys
Vacuum samples
Methods: Video observation study
Study location
Aphid surveys
Video observations
Video processing
Statistical analysis
Aphid surveys
Vacuum samples
Video data
Results : Direct observation study
Colony establishment
Aphid surveys
Temperature
Predator observations
Vacuum samples
Results: Video observation study
Aphid surveys
Temperature
Vacuum samples
Video data
Discussion
Aphid colony establishment
Predator sampling method comparison
Use of video data in natural enemy surveys
xi

115
115
116
118
118
119
124
135
140
140
140
142
142
142
143
144
144
145
145
145
146
147
147
147
147
148
148
148
149
149
149
149
149
150
150
151
151
151
151
151
152
152
152
153
155

APPENDIX
LITERATURE CITED

157
172

CHAPTER V: Conclusion
Implications for managing landscapes for ecosystem services
Context dependence of landscape effects
Expanding to underrepresented regions and agroecosystems
LITERATURE CITED

xii

177
177
179
182
185

LIST OF TABLES

Table 1.1: Township, county and coordinates from soybean treatment and control fields
used in 2008.
53
Table 1.2: Township, county and coordinates from soybean treatment and control fields
used in 2009.
54
Table 1.3: Buckwheat strip data for 2008, including date each strip was planted and
2
subsequently reached full bloom. Density (stems/10 cm ) and height (cm) measurements
were taken at full bloom.
55
Table 1.4: Buckwheat strip data for 2009, including date each strip was planted and
2
subsequently reached full bloom. Density (stems/10 cm ) and height (cm) measurements
were taken at full bloom.
56
Table 1.5: Record of deposition of voucher specimens for Chapter I. The specimens listed
below have been deposited in the Albert J. Cook Arthropod Research Collection,
Michigan State University (MSU) as samples of those species or other taxa, which were
used in this research. Voucher recognition labels bearing the voucher number 2013-08
have been attached or included in fluid preserved specimens.
57
Table 1.6: Dates soybean aphids were first detected in each treatment and control field,
and dates cage study was conducted in each site in 2008.

58

Table 1.7: Dates soybean aphids were first detected in each treatment and control field,
and date cage study was conducted in each site in 2009.

59

Table 1.8: A) Crop and B) non-crop land covers found in landscapes surrounding study
fields. All land covers found were used to calculated Simpson’s diversity metric (-lnD).
The categories of grassland and forest land covers were subsequently combined into
“semi-natural habitat” for landscape analyses.
60
Table 1.9: Simpson’s diversity metric and proportion of each major land cover type
surrounding study fields at the 2 km scale in 2008.

61

Table 1.10: Simpson’s diversity metric and proportion of each major land cover type
surrounding study fields at the 2 km scale in 2009.

63

Table 1.11: Effects of landscape and local habitat management on coccinellid abundance
in 2008 soybean fields, 2009 soybean fields, and 2009 field margins. Separate analyses
were run with A) Simpson’s diversity index –lnD, and with B) proportion of semi-natural
habitat in the landscape as the landscape descriptor variables. Results are shown for
landscape values at the 2.0 km scale, but were consistent at all scales
65
xiii

Table 1.12: Mean ± S.E. of coccinellid beetles of each species found in soybean fields or
field margins during 2008. No comparisons between treatment and control fields were
significant in this year.
66
Table 1.13: Mean ± SE of coccinellid beetle species found in each field or field margin
during 2009. Individual counts were rank transformed and compared using ANOVA.
Asterisks (*) indicate a significant difference in abundance between buckwheat strips
and control field margins, or between treatment and control fields
(*** p < 0.01, ** p < 0.05, * p < 0.1).

67

Table 1.14: Biocontrol Services Index values for aphid suppression by natural enemies
for each soybean field in 2008 and 2009.

68

Table 2.1: Landscape variables used in PLS analysis organized by variable type.
Variables analyzed are shown in italics. Minimum and maximum values for each
variable within the studied landscapes are shown, along with variable influence for
2
project (VIP) values. Rx indicates the amount of variation in each predictor variable
explained by Factor 1.

98

Table 3.1: Means ± standard errors for all adult coccinellid species captured on sticky
traps in 2010.

126

Table 3.2: Means ± standard errors for all adult coccinellid species captured on sticky
traps in 2011.

127

Table 3.3: Means ± standard errors for the most common ground-dwelling predator
species captured in pitfall traps in 2010 and 2011.

128

Table 3.4: Means ± standard errors for foliar-dispersing predators observed on soybean
plants in 2010 and 2011.

129

Table 3.5: Means ± standard errors for aerially-dispersing predator species captured on
sticky traps in 2010 and 2011.

130

Table 3.6: Average number of alate aphids observed per plant in each week in each plot
in 2010 and 2011.
131
Table 4.1: Field locations and characteristics for direct and video observation of
A. glycines predator communities and predation events.

159

Table 4.2: Record of deposition of voucher specimens for Chapter IV. The specimens
listed below have been deposited in the Albert J. Cook Arthropod Research Collection,
Michigan State University (MSU) museum as samples of those species or other taxa,
which were used in this research. Voucher recognition labels bearing the voucher
number 2013-08 have been attached or included in fluid preserved specimens.

160

xiv

Table 4.3: Means ± S.E. of aphid change during each time period of the direct
observation study. Bold indicates means that were significantly different from zero.
Different letters indicate means which are significantly different from one another.

161

Table 4.4: Predators directly observed during 24 h aphid surveys. Observations were
made 180 times at 0:00 and 3:00 h, and 360 times at all other time periods.

162

Table 4.5: Diel comparisons in predatory taxa from vacuum and video samples.

163

Table 4.6: Means ± S.E. of predators collected in vacuum samples of the direct
observation study at morning (10:00), afternoon (16:00), and night (22:00).

164

Table 4.7: Means ± S.E. of aphid change during each time period of the video observation
study. Means shown in bold are significantly different form zero. Different letters indicate
means which are significantly different from one another.
165
Table 4.8: Means ± S.E. of predators collect in vacuum samples of the video observation
samples at morning (9:00), afternoon (15:00) and night (0:00).
166
Table 4.9: Mean ± S.E. number of predator individuals observed in daylight hours during
each three hour period of video surveillance.
167
Table 4.10: Mean ± S.E. number of predator individuals observed in night hours during
each three hour period of video surveillance.
168

xv

LIST OF FIGURES

Figure 1.1: Site locations of paired soybean fields across southern Michigan sampled for
aphids and natural enemies in 2008 and 2009. Sites on the same or nearby farms from
year to year appear to overlap at this scale (crossed circles), but samples were collected
in a different set of fields in each year. For interpretation of the references to color in this
and all other figures, the reader is referred to the electronic version of this dissertation. 69
Figure 1.2: Schematic of sampling design showing relative locations of plots and
buckwheat strip or control field margin. Inset: a yellow sticky trap and a pair of caged
and uncaged soybean plants for the biocontrol experiment were located in the center of
each plot.

70

Figure 1.3: An example buckwheat strip in full bloom in July 2009.

71

Figure 1.4: A) Coccinellid abundance in field margins was consistently and significantly
higher in buckwheat strips than in control field margins across a gradient of semi-natural
habitat within landscapes. B) Coccinellid abundance in soybean fields in 2009 increased
significantly with the proportion of seminatural habitat in the surrounding landscape at
2.0 km. The effect did not differ between fields adjacent to or without buckwheat strips. 72
Figure 1.5: Average number of aphids per soybean plant exposed to and excluding
predators in treatment and control fields, at 7 and 14 days after infestation in A) 2008
and B) 2009. The horizontal dashed line indicates the economic threshold for soybean
aphid of 250 aphids/ plant. In both years, natural enemies dramatically suppressed
soybean aphid populations, and the differences in aphid populations between caged and
uncaged plants were greater at 14 days than at 7 days. However, buckwheat treatment did
not significantly affect the size of the difference in aphid populations on caged and
uncaged plants in either year.
73
Figure 2.1: Relationship between coccinellid abundance in soybean fields in 2009 and
landscape variable scores from PLS analysis. Predictor and explanatory variables show a
positive correlation within factor 1, suggesting that this factor captures a real relationship
between these variables.
99
Figure 2.2: Loadings of response (coccinellids, in box) and predictor (landscape) variables
from PLS regression. The abundance of coccinellids loads positively along factor one,
along with non-crop habitats like grasslands, residential areas, and forests, and with crop
diversity (SDIcrop). In contrast, annual crops (maize, soybean, wheat) and crop field size
(MESH) load negatively. The numbers 1 through 17, representing the landscapes, are
generally spread-out and show no obvious outliers.
100
2

Figure 3.1: Predator activity-density and aphid abundance were measured in 1 m plots.
Barriers to coccinellid immigration were constructed with 2 mm opening mesh netting and
were raised 2 cm above the ground to allow full access by small and ground-dwelling
xvi

predators. Open plots (A & D) had no barriers to immigration. 2-sided plots (B & E) had
barriers erected on the east and west sides of plot. 5-sided plots (C & F) had barriers on all
sides. Each field contained two sets of plots. In sampling plots (A-C), predators were
sampled without replacement. In observation plots (D-F), aphid populations were
monitored without disturbing predator populations.
132
Figure 3.2: Mean number of adult coccinellids captured in treatment plots in 2010 and 2011.
In 2010, coccinellid immigration was higher in open and 2-sided than in 5-sided treatment
plots. In 2011, coccinellid immigration rates were different in all treatments.
133
Figure 3.3: Mean number of A. glycines observed per plant each week in soybean fields
outside of plots and in open, 2- and 5-sided treatment plots in 2010 and 2011. Symbols
indicate significantly more aphids in (+) 5-sided and 2-sided than in open plots, (*)
between all treatments and (x) in 5-sided than in open or 2-sided plots.
134
Figure 4.1: Diagram of plot set-up for observation study. A) Two exterior plots were
situated from 0 – 10 m into the field. B) Two interior plots were situated from 50 – 60 m
into the field. All plots were 50 m away from all other field edges. C) Expanded view of
plot demonstrated random selection of plants for aphid infestation. D) Caged plants
located in between plots were used to account for mortality in the absence of predation. 169
Figure 4.2: Video observation set-up. A) Aphids were recorded with video surveillance
cameras suspended form PVC frames. B) 8 cameras were arranged radially around a DVR.
C) The DVR was powered by 2 marine batteries and connected to a removable hard drive
rack that allowed video files to be taken to the lab for uploading and viewing.
170
Figure 4.3: The percentage of plants on which aphid colonies decreased, remained steady,
or increased within 6 h of infestation. Plants were infested prior to dusk or dawn and were
caged to exclude predators or uncaged to be exposed to predators.
171

xvii

LITERATURE REVIEW
Introduction
Agricultural landscapes cover at least one quarter of terrestrial surfaces on our planet
(Millennium Ecosystem Assessment, 2005) and are critical for providing food, fuel and fiber for
our growing global population. In addition to these provisioning services, agricultural landscapes
also contribute to supporting, regulating and cultural services. For example, agricultural
landscapes can be managed to support the biodiversity that underpins many ecosystem services,
as well as the regulation of water quality, soil fertility, and greenhouse gases (Swinton et al.
2007, Power 2010). They can also be managed to minimize disservices like soil and nutrient
runoff and exposure to pesticides (Power, 2010). By altering the interactions of the organisms
that live there, landscape structure, including the identity, relative abundance, and arrangement
of various habitat types, affects ecosystem functions and thus the provision of ecosystem services
in agricultural landscapes (Altieri 1999, Schlapfer et al. 1999, Tilman et al. 2002, Kremen 2005).
As such, understanding how to best manage these landscapes to provide a sustainable supply of
these ecosystem services is critical.
Within agricultural landscapes, insects provide vital ecosystem services like pollination,
decomposition and biological control of crop pests by predator and parasitoid natural enemies
(Losey and Vaughan 2006). These insect communities, and thus the services they provide, are
also influenced by the landscape context. For example, landscape structure affects the
composition of insect communities and the abundance of individual species, which in turn
influences biocontrol (Tscharntke et al. 2005). Local habitats, or those adjacent to crop fields,
can also affect insect communities and biocontrol services in their vicinity. Herbaceous and
woody vegetation adjacent to crop fields provide resources to natural enemies, contributing to

1

population increases and greater pest control (Bianchi et al. 2006). In many cases, local habitat
can be manipulated to provide natural enemies with resources like alternative prey and refuge
habitat to optimize their performance, a practice known as habitat management (Landis et al.
2000).
Because landscape structure affects insect communities at a variety of scales, the ability
of habitat management to improve biocontrol services can be expected to depend on landscape
complexity and composition. Tscharntke et al. (2005) found that local management practices
only had an impact on biocontrol services in simplified landscapes, presumably because complex
landscapes already supply natural enemies with necessary resources. On the other hand,
extremely simplified landscapes may not support sufficient natural enemy populations to
measure significant effects of habitat management. Furthermore, natural enemies may respond to
multiple aspects of landscape structure, including heterogeneity of semi-natural areas,
heterogeneity of crop habitats, and the configuration or spatial arrangement of land covers
(Fahrig et al. 2011). Additionally, the diversity of agronomic practices (e.g. forms of tillage,
fertilization regimes and pesticide usage, timing of practices) can create environments with
different suitability for natural enemies in multiple fields of the same crop type (Vasseur et al.
2013). However, in spite of the complexity of possible interactions between natural enemies and
landscape structure, studies examining these effects routinely characterize landscapes solely by
the amount of semi-natural habitat present (Burel et al. 2013). In light of the fact that farmers are
likely to be more willing and able to manage some aspects of landscape structure (like crop
heterogeneity) than others (like natural areas), a more detailed examination of the interaction
between landscape structure and insect communities could lead to more practical management
recommendations (Fahrig et al. 2011). Finally, in addition to community effects, landscapes

2

could influence the timing and degree of natural enemy dispersal (Bianchi and van der Werf
2003), which could have implications for pest suppression in agroecosystems (Landis and van
der Werf 1997).
Landscape effects on natural enemies and biocontrol services
Agricultural landscapes exist as a mosaic of crop and non-crop habitats, which organisms
may use in a variety of ways (Fahrig et al. 2011). Landscape structure itself can be quantified in
multiple ways, including composition metrics that describe which and how many habitats are
present and their relative abundances, and configuration metrics that indicate how those habitats
are spatially arranged (Turner et al. 2001). Cummings and Spiesman (2006) point out the
importance of understanding the effects of landscape structure on ecosystem services like pest
management. Natural enemies are thought to control most potential crop pests much of the time
(Losey and Vaughan 2006), yet the success of biological control is dependent upon the provision
of natural enemies by the landscape. Landscape structure could impact the success of biocontrol
by affecting the ability of natural enemies to locate and aggregate in areas of high pest density
(Bommarco 1998).
Landscape composition. Semi-natural habitats. Non-crop habitats like forests and
grasslands can play an important role in agricultural landscapes by providing natural enemies
with food sources like nectar, pollen, and prey (Thies et al. 2003), as well as refuge from
insecticide use (Lee et al. 2001), tillage, and harvest (Landis et al. 2000). Having access to these
resources can lead to increased populations of natural enemies like coccinellids (Coleoptera:
Coccinellidae) (Gardiner et al. 2009a), spiders (Araneae) (Schmidt and Tscharntke 2005, Drapela
et al. 2008, Gardiner et al. 2010, Pluess et al. 2010), carabids (Coleoptera: Carabidae) (Purtauf et
al. 2005, Batáry et al. 2007), and Opiliones (Prasifka et al. 2004, Gardiner et al. 2010). These

3

perennial non-crop habitats can also support higher species richness of many natural enemy taxa,
including spiders (Clough et al. 2005, Schmidt et al. 2005, Drapela et al. 2008, Schmidt et al.
2008), coccinellids (Colunga-Garcia et al. 1997), and carabids (Purtauf et al. 2005, Batáry et al.
2007, Werling and Gratton 2008). The increases in natural enemy abundance and richness in
landscapes with more semi-natural habitat may contribute to better control of crop pests, as these
landscapes have also been found to have higher suppression rates of rape pollen beetle,
Meligethes aeneus (F.) (Coleoptera: Nitidulidae) (Thies and Tscharntke 1999, Thies et al. 2003),
cabbage moth, Mamestra brassicae L. (Lepidoptera: Noctuidae) (Bianchi et al. 2005), cereal
aphids (Hemiptera: Aphididae), Sitobion avenae (F.), Metopolophium dirhodum (Walker), and
Rhopalosiphum padi (L.), (Roschewitz et al. 2005, Thies et al. 2005), diamond back moth,
Plutella xylostella (Lepidoptera; Plutellidae) (Bianchi et al. 2008), and soybean aphid, Aphis
glycines Matsumura (Gardiner et al. 2009a). In fact, a meta-analysis of landscape studies showed
that natural enemy activity was enhanced by landscape heterogeneity in 70% of cases, and
landscape composition measured as proportion of woody and herbaceous habitat in 71% and
80% of cases, respectively (Bianchi et al. 2006).
Perennial crop habitats. Studies of landscape composition effects on insects typically
consider only semi-natural habitats (Burel et al. 2013). However, other types of land covers
could also influence natural enemy communities. For example, perennial crops may provide
many of the same benefits as semi-natural habitat, including reduced disturbance and alternative
prey (Landis et al. 2000). Although studies examining perennial crops have been few in number,
landscapes which contain more perennial crops have been found to have higher natural enemy
activity density (Oberg et al. 2007) and diversity (Elliott et al. 1999), and lower pest rates
(Östman et al. 2001). Additionally, natural enemies may have higher fitness and fecundity in

4

landscapes with more perennial crops (Bommarco 1998). Perennial crops may also allow
populations of natural enemies to persist year to year due to their lower levels of disturbance
(Landis et al. 2000).
Habitat and crop diversity. Habitat diversity, another metric of landscape composition
(Turner et al. 2001, McGarigal et al. 2012), may also be important for naturally occurring
biological control. Multiple land covers in a local area may provide different resources needed
by a single organism (Dunning et al. 1992), which can in turn influence communities in
agricultural landscapes (Fahrig et al. 2011). For example, a predatory insect may require one
type of habitat for overwintering and another for foraging and egg-laying. High habitat diversity
in the landscape can also support greater predator density by allowing cross-foraging (Cumming
and Spiesman 2006). Predators receiving prey “subsidies” from other habitats may increase in
abundance and more effectively suppress pests in focal areas (e.g. crop fields) (Polis et al. 1997).
Additionally, because different habitat types support different natural enemies, landscapes with
higher habitat diversity may support higher natural enemy diversity (Elliott et al. 1999, Bianchi
et al. 2006), which can facilitate recovery of pest suppression services after disturbances
(Tscharntke et al. 2005). Conversely, simplified landscapes of predominantly crop land support
lower biodiversity (Bianchi et al. 2006), and may lead to a loss of ecosystem services like pest
control (Andow 1983, Kareiva 1987, Thies and Tscharntke 1999, Symondson et al. 2002,
Tscharntke et al. 2005). Crop diversity may provide many of the same benefits as overall habitat
diversity, as different crops with different phenologies may provide complementary resources at
different times of the year (Vasseur et al. 2013).
Landscape configuration. In contrast to measures of landscape composition, the
influence of landscape configuration on natural enemies has been largely overlooked (Burel et al.

5

2013). Yet the configuration of habitats can affect the rate of arrival of natural enemies into crop
fields (Bianchi et al. 2006), which can in turn influence natural enemy efficacy (Hogg and Daane
2010). In particular, modeling has shown that an even distribution of non-crop habitats within a
landscape should facilitate colonization of fields by natural enemies (Bianchi and van der Werf
2003), suggesting that landscapes with small crop fields and greater dispersion of non-crop
habitat could promote higher natural enemy abundances in crop fields. Additionally, researchers
have found greater pest control in small fields adjacent to larger non-crop borders than in larger
fields with less border area (Marino and Landis 1996, Werling and Gratton 2010). Furthermore,
landscapes with smaller average patch sizes should promote early colonization of crop fields by
natural enemies (Bianchi et al. 2009), which could lead to higher biocontrol in landscapes with
smaller patch size (Bianchi et al. 2006).
Differential responses by habitat generalists and specialists
A natural enemy taxon’s degree of habitat specialization affects how it will perceive and
be influenced by the landscape. Habitat specialists may require a high degree of landscape
diversity to fulfill their needs, while generalists may be more dependent on overall area of
appropriate habitat (Ricklefs and Lovette 1999). Landscapes may appear more fragmented to
specialized species, which require resources from specific habitat types, and more connected for
generalists because they are able to use more of the habitat types on the landscape (Kruess and
Tscharntke 1994, Tscharntke et al. 2002). Alternately, landscape heterogeneity may be important
to habitat generalists due to provision of diverse resources, while habitat specialists depend more
on connectivity and total area of specific habitat types (Steffan-Dewenter 2003).

6

Natural enemy prey specialization and immigration
In addition to variation in habitat specialization, natural enemies of crop pests also exhibit
a wide range of prey specificity. Modeling studies suggest that effective pest suppression may be
obtained from either specialist natural enemies inflicting high mortality rates, or from generalist
natural enemies with high immigration rates into an infested area (Chang and Kareiva 1999).
Additionally, each of these strategies may be better adapted to particular crop types. For
example, specialist natural enemies may be well-adapted for perennial crops in which temporal
stability allows equilibrium to be achieved between natural enemy and pest populations, while
generalists may be better suited for annual crops which both pests and natural enemies have to
recolonize each growing season (Murdoch et al. 2006). Furthermore, because generalist natural
enemies like coccinellids tend to have long generation times relative to those of their prey and
thus cannot rapidly respond numerically to pest population dynamics (Dixon et al. 1997,
Kindlmann and Dixon 1999b, a, 2001), generalists must exhibit high immigration rates to be
effective (Murdoch et al. 2006). In fact, higher immigration rates have been demonstrated to
decrease the predator/prey ratios necessary for effective pest suppression (van der Werf 1995),
and manipulating immigration rates of generalist predators has resulted in lower pest abundance
and higher yields (Snyder and Wise 1999).
Natural enemy immigration occurs at different temporal and spatial scales. In annual
agriculture, many natural enemies overwinter outside of the crop and must first immigrate into
crop fields after their establishment. Once in the crop habitat, they must then disperse and locate
prey-infested patches within the field. In situations in which natural enemies must immigrate into
fields, the timing of immigration is important for pest suppression. Natural enemies must arrive
prior to pest populations reaching high densities and growing exponentially. Reducing natural

7

enemies in the early season or preventing them from colonizing pest-infested patches can result
in higher pest densities later in the season (Settle et al. 1996, Landis and van der Werf 1997).
Natural enemy dispersal ability and response to landscape
Dispersal ability influences the scale at which landscape structure influences insect
communities and individual insect success (Roland and Taylor 1997, Bommarco 1998). Habitat
connectivity is less important for vagile species than for poor dispersers, with the result being
that landscapes appear more fragmented to poor dispersers (deVries et al. 1996, Tscharntke et al.
2002). For example, in a study of coccinellid ability to track aphid populations in fragmented
landscapes, distributions of the native coccinellid Coleomegilla maculata (DeGeer) were more
affected by habitat fragmentation than those of the exotic coccinellid Harmonia axyridis (Pallas),
which is a better disperser (With et al. 2002). Several studies have estimated the spatial scale at
which particular insects respond to landscape features by first quantifying the landscape
composition at various radii around a study plot and then determining the scale at which
landscape composition and insect density was most highly correlated (Steffan-Dewenter et al.
2001, Steffan-Dewenter et al. 2002, Krauss et al. 2003, Thies et al. 2003, Westphal et al. 2003,
Drapela et al. 2008, Pluess et al. 2010). These studies have shown different scales of response for
different types of insects. Overall, natural enemy diversity and abundance and pest suppression
have been related to the structure of the surrounding landscape at scales between hundreds of
meters to 10 km (Woltz et al. 2012). Because dispersal ability of insects affects the scale at
which they respond to landscape structure, parasitoids with small dispersal ranges are expected
to respond at smaller scales, while generalist predators like beetles are affected at larger scales
(Tscharntke et al. 2005). Even among coccinellids, the different life histories and dispersal
abilities of various species cause them to respond to landscape features at different scales,

8

although, in general, landscape factors have greater impacts than within field variables on
coccinellid abundance and diversity (Elliott et al. 1998). Finally, the dispersal abilities of
different natural enemy taxa determine which species coexist in crop fields (Bianchi et al. 2009).
Therefore, natural enemy responses to landscape structure are primarily driven by highly mobile
taxa (Tscharntke et al. 2005).
Local habitat effects on insect communities and biocontrol services
Local habitat characteristics, those of the crop field of interest and its immediate
surroundings, also affect insect communities in agro-ecosystems. Natural enemies have several
requirements, including appropriate microclimates, constant food supplies, and overwintering
habitats (Jonsson et al. 2008) that are not continuously available in crop habitats. Crop fields are
frequently disturbed by tillage, planting, pesticide application, and harvest, making them
temporarily unsuitable for natural enemies to inhabit (Landis et al. 2000). Thus, the suitability of
adjacent habitats for natural enemies affects their ability to survive crop disturbances and recolonize crop fields. Insects, including natural enemies of crop pests, often survive in field edges
and non-crop habitats such as woodlots and hedgerows during crop disturbance (Duelli et al.
1999, Bianchi et al. 2006). Natural enemies also overwinter in nearby non-crop habitat, from
which they subsequently recolonize crops during the growing season (Bianchi et al. 2006,
Benjamin et al. 2008). Herbaceous habitats may provide especially important overwintering
habitats for generalist predators (Geiger et al. 2009). Local non-crop habitat can also provide
nutritional resources for natural enemies. Herbaceous and woody vegetation adjacent to crop
fields can support alternative hosts for parasitoids and prey for predators, as well as nectar and
pollen sources (Meek et al. 2002). The presence of alternate prey may allow natural enemies to

9

establish before populations of pests do so (Landis et al. 2000), providing greater pest control
(Bianchi et al. 2006).
Habitat management
Conservation biological control refers to practices aimed at increasing the effectiveness
of natural enemies already present in the landscape (Landis et al. 2000). Habitat management is a
type of conservation biological control that involves altering habitats within or adjacent to crops
to provide natural enemies with resources necessary to optimize their performance (Landis et al.
2000). Such resources could include alternative prey or hosts (Settle et al. 1996), nectar and
pollen (Treacy et al. 1987, Idris and Grafius 1995, Hickman and Wratten 1996), overwintering
habitat (Thomas et al. 1992), and refuge from field disturbances (Landis et al. 2000).
Habitat management often involves increasing the diversity of plants within an agroecosystem,
whether within a field, along a crop border, or in the broader landscape (Gurr et al. 2003). The
use of field borders for habitat management often takes the form of linear strips such as beetle
banks or floral resource strips. Floral resource strips are used to increase insect diversity and to
support pollinators and natural enemies of crop pests (Haaland et al. 2011). Floral plantings may
increase natural enemy efficacy by providing them with resources such as nectar, pollen, or
alternative prey (Landis et al. 2000). For generalist predators that are less reliant on floral
resources, the vegetative structure of plantings can be important (Haaland et al. 2011), and
grassy strips may shelter greater abundances than adjacent fields (Smith et al. 2008). Vegetation
strips can serve as a refuge from crop disturbance allowing natural enemies to recolonize fields
after insecticide application (Lee et al. 2001). Moderated microclimate and carbohydrate
resources can increase parasitoid longevity, resulting in greater lifetime parasitism rates
(Costamagna and Landis 2004). However, in some cases, field margin plantings may harbor

10

natural enemies, but not result in higher natural enemy densities within the adjacent crop (Koji et
al. 2007).
In general, floral plantings support higher abundances of natural enemies than crop edges
or fields (Haaland et al. 2011). Non-crop flowers can increase abundance of ground-dwelling
generalist predators like spiders and carabid beetles (Ditner et al. 2013) probably by providing
more favorable microclimates. Floral resource strips have been shown to result in higher
abundances of beneficial insects (Walton and Isaacs 2011), higher aphid parasitism rates (Bigger
and Chaney 1998, Langer and Hance 2004), greater egg-laying by syrphids and lower aphid
densities (Hickman and Wratten 1996), and higher pest predation rates (Frank and Shrewsbury
2004) in adjacent crop fields. However, not all natural enemy taxa respond in the same way, and
some that are found in great abundance in the strips are less abundant in adjacent crops (Bigger
and Chaney 1998). Additionally, effects may be idiosyncratic, impacting some pests but not
others, or working in some sites but not others (Pfiffner et al. 2009).
Interaction between landscape and local effects on natural enemies
Non-crop habitats, provided locally within the landscape, provide important resources for
natural enemies, yet the natural enemies that respond to local resources ultimately come from the
broader landscape. Clough et al. (2007) concluded that “the landscape determines potential insect
diversity that may be realized or not depending on local conditions.” This can create
complicated dynamics between landscape and local effects on natural enemies. Landscape
structure has been shown to have more influence on natural enemy communities than local
adjacent habitats in cotton (Prasifka et al. 2004) and potatoes (Werling and Gratton 2008),
although Werling and colleagues (2010) found that both landscape conditions and local adjacent
habitats influenced pest predation in potato fields. Others have simultaneously examined the

11

impact of landscape structure and conventional vs. organic management. In some cases,
landscape characteristics have been shown to have greater influences on natural enemies than
field management (Clough et al. 2005, Purtauf et al. 2005), while in others, both landscape and
field management influenced natural enemies (Schmidt et al. 2005, Isaia et al. 2006), and pests
(Östman et al. 2001, Roschewitz et al. 2005). In some cases, landscape structure is even more
influential on natural enemy communities than the particular habitat in which natural enemies are
sampled (Aviron et al. 2005) while others saw influences of both (Wilby et al. 2006).
Together, these studies suggest that while in some systems landscape factors are
overriding, in others both scales may simultaneously influence natural enemies. In some cases,
landscape and local effects may even be interactive. For example, Haenke and others (Haenke et
al. 2009) found that syrphid densities were higher in flower strips than in grassy strips, but within
flower strips syprhid densities also increased with amount of crop in the landscape, perhaps
suggesting that the syrphids aggregated in the flower strips in the more resource-poor landscapes.
Conversely, others have found that spider density (Schmidt et al. 2005) and carabid richness
(Weibull et al. 2003) were related to landscape conditions in conventional but not organic farms,
indicating that landscape conditions were more important for natural enemies on resource-poor
farms.
Study System
Soybean, Glycine max, is an ideal system to compare the effects of landscape complexity
and local habitat management on biocontrol services because it is attacked by a common insect
pest which is in turn preyed upon by a diverse guild of broadly occurring generalist predators
(Ragsdale et al. 2004, Costamagna and Landis 2007). Aphids are common crop pests worldwide, feeding on crops as diverse as limes, cabbage, and beans (Dixon 1985). Soybean aphid,

12

Aphis glycines Matsumura, specializes on soybean, causing heavy damage (Wu et al. 2004).
Soybean aphid natural enemy abundance is correlated with landscape features (Gardiner et al.
2009a) and may respond to floral resources (Harmon et al. 2000).
Aphis glycines was first detected in the U.S. in Wisconsin in July 2000 (Ragsdale et al.
2004) and has since been reported in 22 states and 3 Canadian provinces (Costamagna and
Landis 2007). Aphids are dispersed long distances by wind, and may ultimately spread into most
soybean production areas of the U.S. (Venette and Ragsdale 2004). Aphis glycines exhibits a
heteroecious, holocyclic life cycle, overwintering on Rhamnus sp., and emigrating to soybean in
early spring (Ragsdale et al. 2004), where it causes yield losses of up to 40% in the North Central
US (DiFonzo and Hines 2002). Heavy infestations may result in early defoliation, shortened
stems and leaves, underdeveloped roots, distorted foliage, reduced branch number, lower seed
and pod counts, reduced seed weight, and stunting (Wu et al. 2004). In the North Central U.S.
soybean was rarely sprayed with insecticide before A. glycines arrival; however, soybean aphid
presence has led to increased pesticide use, with 2.8 million ha sprayed in the North Central U.S.
in 2003 (Landis et al. 2003). Soybean aphid can be a vector of crop viruses, including soybean
mosaic virus, soybean dwarf virus, soybean stunt virus, and several others (Heimpel et al. 2004,
Wu et al. 2004). Aphis glycines could negatively affect native aphids and other herbivores
through direct or apparent competition (Heimpel et al. 2004). Soybean aphid abundances may
have driven increases in populations of exotic coccinellids, leading to the displacement of native
species (Heimpel et al. 2004).
Factors such as natural enemy abundance, timing of soybean field invasion, temperature,
and moisture may impact soybean aphid outbreaks. For example, wetter and cooler weather
favors soybean aphid population growth (Ragsdale et al. 2004). In the North Central U.S.

13

maximum daily temperatures in June are often 22-25 °C, temperatures which favor rapid
soybean aphid development (Ragsdale et al. 2004). Outbreaks of soybean aphids are mediated by
the number of overwintering eggs, synchronization of aphid and soybean development, and crop
management practices including sowing time and crop rotation (Wu et al. 2004).
Coccinellids and biological control of soybean aphid. Effective biocontrol of A.
glycines could reduce economic and environmental costs associated with insecticide application
if pests are kept below economic injury thresholds (Heimpel et al. 2004). Ragsdale et al. (2007)
estimated an economic threshold (ET) of 273 ± 38 soybean aphids per plant, with a seven day
lead time before populations would be expected to exceed the Economic Injury Level (EIL) of
674 ± 95 aphids per plant. In order to effectively suppress soybean aphid, natural enemies must
attack early in the season when aphid populations exhibit low density and slow growth (Wu et al.
2004), or aggregate in areas of high aphid density later in the season, and have high attack rates
(Rutledge et al. 2004). Generalist predators have been shown to suppress A. glycines during both
early season aphid establishment and mid-season aphid population increase (Costamagna and
Landis 2006). In particular, coccinellids occur early in the season and in sufficient numbers to
potentially prevent outbreaks (Rutledge et al. 2004). Coccinellid beetles are thought to contribute
the greatest amount to control of this pest in China (Wu et al. 2004). In an observational study of
the soybean aphid predator guild, adults and larvae of large coccinellid species such as
Coccinella septempunctata and Harmonia axyridis accounted for a majority of soybean aphid
predation (Costamagna and Landis 2007). These species had the highest per capita consumption
rate of A. glycines, consumed aphids rapidly, and increased in abundance and predation rate in
response to increased soybean aphid density. Coccinellids in general have been shown to
efficiently control soybean aphid populations, leading to significant increases in plant biomass
14

and soybean yield (Costamagna and Landis 2007). Harmonia axyridis acts to reduce aphid
densities, arriving early in the season before aphids, and responding numerically to aphid
abundances (Rutledge et al. 2004). Additionally, intraguild predation by large coccinellids has
been shown not to release soybean aphid from biological control, as H. axyridis supplies
sufficient control to make up for lost intermediate predators (Costamagna and Landis 2007,
Gardiner and Landis 2007). Furthermore, Costamagna and Landis (2007) did not observe
intraguild predation during the course of their observational study and recorded very few
predation events on alternate prey, indicating that soybean aphid is a primary prey item for
coccinellids in soybean fields.
Coccinellid responses to landscape and habitat management. An abundance of
evidence suggests that coccinellids respond to both local and landscape conditions. Coccinellids
have been found to be more abundant within habitat management strips themselves (Dong et al.
2012), and also more abundant in vineyards with adjacent noncrop habitat than in vineyards
without such adjacent habitats (Thomson and Hoffmann 2010). Additionally, the increased
abundances of coccinellids and other predators lead to higher predation rates of pest eggs in the
vineyards with adjacent noncrop vegetation (Thomson and Hoffmann 2010), suggesting that
habitat management can be a useful tool for improving biocontrol services from coccinellids. At
larger scales, coccinellids are highly responsive to landscape composition in agroecoystems
(Colunga-Garcia et al. 1997, Elliott et al. 2002, Gardiner et al. 2009a), although specific
responses may differ by species (Colunga-Garcia et al. 1997, Gardiner et al. 2009b).
Furthermore, the response of coccinellids to landscape structure has been related to greater
suppression of A. glycines in landscapes with more diverse non-crop habitat (Gardiner et al.
2009a). Such evidence of coccinellid response to both habitat management and landscape

15

structure, and the related implications for pest suppression at both scales, suggests a need to
compare effects at these two scales simultaneously and to consider possible interactions between
them. However, because studies have shown contrasting responses of coccinellids to overall
abundance of crop and non-crop habitat in a landscape (Gardiner et al. 2009a, Caballero-Lopez
et al. 2012), it is necessary to begin examining coccinellid’s relationship to landscape structure in
more detail. Additionally, coccinellid immigration and resulting densities at a patch level are
influenced by small-scale fragmentation, as well as patch isolation (Grez et al. 2004, Grez et al.
2005, Zaviezo et al. 2006). These influences on the aggregation behavior of coccinellids could
influence aphid outbreaks (Kareiva 1987), suggesting a need to examine the relationship between
coccinellid immigration and soybean aphid suppression.
Conclusions
The occurrence of natural enemy communities that provide biocontrol services within
agroecosystems is dependent upon landscape characteristics. Biocontrol services can also be
augmented with habitat management practices that provide natural enemies with additional food
and habitat resources to natural enemies. It is unknown, however, whether landscape structure
and habitat management interact, such that habitat management practices improve biocontrol
services more in certain types of landscapes than in others. Highly diverse landscapes may
already provide a sufficient supply of natural enemies for biocontrol, while extremely simplified
landscapes likely support few natural enemies to respond to additional resources (Tscharntke et
al. 2005). Furthermore, multiple aspects of landscape structure are relevant for natural enemy
communities, while prior studies have typically focused on a single aspect of landscape
composition. Finally, because landscape structure may influence natural enemy dispersal as well

16

as abundance, it is important to understand how changes in natural enemy dispersal could
influence pest suppression.
The objectives of this dissertation are to: 1) examine the effectiveness of habitat
management for soybean aphid biological control in landscapes of varying complexity, 2) to
examine in detail specific aspects of landscape structure that influence coccinellid abundances,
3) to examine the influence of coccinellid immigration on soybean aphid suppression, and 4) to
compare diel patterns in the soybean aphid natural enemy community.

17

LITERATURE CITED

18

LITERATURE CITED

Altieri, M. A. 1999. The ecological role of biodiversity in agroecosystems. Agriculture
Ecosystems & Environment 74:19-31.
Andow, D. 1983. The extent of monoculture and its effects on insect pest populations with
particular reference to wheat and cotton. Agriculture Ecosystems & Environment 9:2535.
Aviron, S., F. Burel, J. Baudry, and N. Schermann. 2005. Carabid assemblages in agricultural
landscapes: impacts of habitat features, landscape context at different spatial scales and
farming intensity. Agriculture Ecosystems & Environment 108:205-217.
Batáry, P., A. Baldi, G. Szel, A. Podlussany, I. Rozner, and S. Erdos. 2007. Responses of
grassland specialist and generalist beetles to management and landscape complexity.
Diversity and Distributions 13:196-202.
Benjamin, R., G. Cedric, and I. Pablo. 2008. Modeling spatially explicit population dynamics of
Pterostichus melanarius I11. (Coleoptera: Carabidae) in response to changes in the
composition and configuration of agricultural landscapes. Landscape and Urban Planning
84:191-199.
Bianchi, F., N. A. Schellhorn, and W. van der Werf. 2009. Predicting the time to colonization of
the parasitoid Diadegma semiclausum: The importance of the shape of spatial dispersal
kernels for biological control. Biological Control 50:267-274.
Bianchi, F. J. J. A., C. J. H. Booij, and T. Tscharntke. 2006. Sustainable pest regulation in
agricultural landscapes: a review on landscape composition, biodiversity and natural pest
control. Proceedings of the Royal Society B-Biological Sciences 273:1715-1727.
Bianchi, F. J. J. A., P. W. Goedhart, and J. M. Baveco. 2008. Enhanced pest control in cabbage
crops near forest in The Netherlands. Landscape Ecology 23:595-602.
Bianchi, F. J. J. A. and W. van der Werf. 2003. The effect of the area and configuration of
hibernation sites on the control of aphids by Coccinella septempunctata (Coleoptera:
Coccinellidae) in agricultural landscapes: A simulation study. Environmental
Entomology 32:1290-1304.
Bianchi, F. J. J. A., W. K. R. E. van Wingerden, A. J. Griffioen, M. van der Veen, M. J. J. van
der Straten, R. M. A. Wegman, and H. A. M. Meeuwsen. 2005. Landscape factors
affecting the control of Mamestra brassicae by natural enemies in Brussels sprout.
Agriculture Ecosystems & Environment 107:145-150.
Bigger, D. S. and W. E. Chaney. 1998. Effects of Iberis umbellata (Brassicaceae) on insect pests
19

of cabbage and on potential biological control agents. Environmental Entomology
27:161-167.
Bommarco, R. 1998. Reproduction and energy reserves of a predatory carabid beetle relative to
agroecosystem complexity. Ecological Applications 8:846-853.
Burel, F., C. Lavigne, E. J. P. Marshall, A. C. Moonen, A. Ouin, and S. L. Poggio. 2013.
Landscape ecology and biodiversity in agricultural landscapes. Agriculture, Ecosystems,
and Environment 166:2.
Caballero-Lopez, B., R. Bommarco, J. M. Blanco-Moreno, F. X. Sans, J. Pujade-Villar, M.
Rundlof, and H. G. Smith. 2012. Aphids and their natural enemies are differently affected
by habitat features at local and landscape scales. Biological Control 63:222-229.
Chang, G. C. and P. Kareiva. 1999. The case for indigenous generalists in biological control.in
B. A. Hawkins and H. V. Cornell, editors. Theoretical Approaches to Biological Control.
Cambridge University Press, Cambridge.
Clough, Y., A. Kruess, D. Kleijn, and T. Tscharntke. 2005. Spider diversity in cereal fields:
comparing factors at local, landscape and regional scales. Journal of Biogeography
32:2007-2014.
Clough, Y., A. Kruess, and T. Tscharntke. 2007. Local and landscape factors in differently
managed arable fields affect the insect herbivore community of a non-crop plant species.
Journal of Applied Ecology 44:22-28.
Colunga-Garcia, M., S.H. Gage, and D.A. Landis. 1997. Response of an assemblage of
Coccinellidae (Coleoptera) to a diverse agricultural landscape. Environmental
Entomology 26:8.
Costamagna, A. C. and D. A. Landis. 2004. Effect of food resources on adult Glyptapanteles
militaris and Meteorus communis (Hymenoptera: Braconidae), parasitoids of Pseudaletia
unipuncta (Lepidoptera: Noctuidae). Environmental Entomology 33:128-137.
Costamagna, A. C. and D. A. Landis. 2006. Predators exert top-down control of soybean aphid
across a gradient of agricultural management systems. Ecological Applications 16:16191628.
Costamagna, A. C. and D. A. Landis. 2007. Quantifying predation on soybean aphid through
direct field observations. Biological Control 42:16-24.
Cumming, G. S. and B. J. Spiesman. 2006. Regional problems need integrated solutions: Pest
management and conservation biology in agroecosystems. Biological Conservation
131:533-543.
deVries, H. H., P. J. denBoer, and T. S. vanDijk. 1996. Ground beetle species in heathland
20

fragments in relation to survival, dispersal, and habitat preference. Oecologia 107:332342.
DiFonzo, C. D. and R. Hines. 2002. Soybean aphid in Michigan: Update from 2001 season.
Michigan State University Extension Bulletin E-2746.
Ditner, N., O. Balmer, J. Beck, T. Blick, P. Nagel, and H. Luka. 2013. Effects of experimentally
planting non-crop flowers into cabbage fields on the abundance and diversity of
predators. Biodiversity and Conservation 22:1049-1061.
Dixon. 1985. Aphid Ecology. Blackie & Son Ltd., Glasgow.
Dixon, A. F. G., J. L. Hemptinne, and P. Kindlmann. 1997. Effectiveness of ladybirds as
biological control agents: Patterns and processes. Entomophaga 42:71-83.
Dong, Z. K., F. J. Gao, and R. Z. Zhang. 2012. Use of ryegrass strips to enhance biological
control of aphids by ladybirds in wheat fields. Insect Science 19:529-534.
Drapela, T., D. Moser, J. G. Zaller, and T. Frank. 2008. Spider assemblages in winter oilseed
rape affected by landscape and site factors. Ecography 31:254-262.
Duelli, P., M. K. Obrist, and D. R. Schmatz. 1999. Biodiversity evaluation in agricultural
landscapes: above-ground insects. Agriculture Ecosystems & Environment 74:33-64.
Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect
populations in complex landscapes. Oikos 65:169-175.
Elliott, N. C., R. W. Kieckhefer, and D. A. Beck. 2002. Effect of aphids and the surrounding
landscape on the abundance of Coccinellidae in cornfields. Biological Control 24:214220.
Elliott, N. C., R. W. Kieckhefer, B. W. French, and J. H. Lee. 1998. A comparison of size, sex
ratio, and dispersal in three coccinellids. Southwestern Entomologist 23:137-145.
Elliott, N. C., R. W. Kieckhefer, J. H. Lee, and B. W. French. 1999. Influence of within-field and
landscape factors on aphid predator populations in wheat. Landscape Ecology 14:239252.
Fahrig, L., J. Baudry, L. Brotons, F. G. Burel, T. O. Crist, R. J. Fuller, C. Sirami, G. M.
Siriwardena, and J. L. Martin. 2011. Functional landscape heterogeneity and animal
biodiversity in agricultural landscapes. Ecology Letters 14:101-112.
Frank, S. D. and P. M. Shrewsbury. 2004. Effect of conservation strips on the abundance and
distribution of natural enemies and predation of Agrotis ipsilon (Lepidoptera: Noctuidae)
on golf course fairways. Environmental Entomology 33:1662-1672.

21

Gardiner, M. M. and D. A. Landis. 2007. Impact of intraguild predation by adult Harmonia
axyridis (Coleoptera: Coccinellidae) on Aphis glycines (Hemiptera: Aphididae) biological
control in cage studies. Biological Control 40:386-395.
Gardiner, M. M., D. A. Landis, C. Gratton, C. D. DiFonzo, M. O'Neal, J. M. Chacon, M. T.
Wayo, N. P. Schmidt, E. E. Mueller, and G. E. Heimpel. 2009a. Landscape diversity
enhances biological control of an introduced crop pest in the north-central USA.
Ecological Applications 19:143-154.
Gardiner, M. M., D. A. Landis, C. Gratton, N. Schmidt, M. O'Neal, E. Mueller, J. Chacon, and
G. E. Heimpel. 2010. Landscape composition influences the activity density of Carabidae
and Arachnida in soybean fields. Biological Control.
Gardiner, M. M., D. A. Landis, C. Gratton, N. Schmidt, M. O'Neal, E. Mueller, J. Chacon, G. E.
Heimpel, and C. D. DiFonzo. 2009b. Landscape composition influences patterns of
native and exotic lady beetle abundance. Diversity and Distributions 15:554-564.
Geiger, F., F. Wackers, and F. J. J. A. Bianchi. 2009. Hibernation of predatory arthropods in
semi-natural habitats. BioControl 54:529-535.
Grez, A., T. Zaviezo, L. Tischendorf, and L. Fahrig. 2004. A transient, positive effect of habitat
fragmentation on insect population densities. Oecologia 141:444-451.
Grez, A. A., T. Zaviezo, and M. Rios. 2005. Ladybird (Coleoptera: Coccinellidae) dispersal in
experimental fragmented alfalfa landscapes. European Journal of Entomology 102:209216.
Gurr, G. M., S. D. Wratten, and J. M. Luna. 2003. Multi-function agricultural biodiversity: pest
management and other benefits. Basic and Applied Ecology 4:107-116.
Haaland, C., R. E. Naisbit, and L. F. Bersier. 2011. Sown wildflower strips for insect
conservation: a review. Insect Conservation and Diversity 4:60-80.
Haenke, S., B. Scheid, M. Schaefer, T. Tscharntke, and C. Thies. 2009. Increasing syrphid fly
diversity and density in sown flower strips within simple vs. complex landscapes. Journal
of Applied Ecology 46:1106-1114.
Harmon, J. P., A. R. Ives, J. E. Losey, A. C. Olson, and K. S. Rauwald. 2000. Colemegilla
maculata (Coleoptera: Coccinellidae) predation on pea aphids promoted by proximity to
dandelions. Oecologia 125:543-548.
Heimpel, G. E., D. W. Ragsdale, R. Venette, K. R. Hopper, R. J. O'Neil, C. E. Rutledge, and Z.
S. Wu. 2004. Prospects for importation biological control of the soybean aphid:
Anticipating potential costs and benefits. Annals of the Entomological Society of
America 97:249-258.

22

Hickman, J. M. and S. D. Wratten. 1996. Use of Phacelia tanacetifolia strips to enhance
biological control of aphids by hoverfly larvae in cereal fields. Journal of Economic
Entomology 89:832-840.
Hogg, B. N. and K. M. Daane. 2010. The role of dispersal from natural habitat in determining
spider abundance and diversity in California vineyards. Agriculture Ecosystems &
Environment 135:260-267.
Idris, A. B. and E. Grafius. 1995. Wildflowers as nectar sources for Diadegma insulare
(Hymenoptera: Ichneumonidae), a parasitoid of diamondback moth (Lepidoptera:
Yponomeutidae). Environmental Entomology 24:1726-1735.
Isaia, M., F. Bona, and G. Badino. 2006. Influence of landscape diversity and agricultural
practices on spider assemblage in Italian vineyards of Langa Astigiana (northwest Italy).
Environmental Entomology 35:297-307.
Jonsson, M., S. D. Wratten, D. A. Landis, and G. M. Gurr. 2008. Recent advances in
conservation biological control of arthropods by arthropods. Biological Control 45:172175.
Kaiser, M. E., T. Noma, M. J. Brewer, K. S. Pike, J. R. Vockeroth, and S. D. Gaimari. 2007.
Hymenopteran parasitoids and dipteran predators found using soybean aphid after its
midwestern United States invasion. Annals of the Entomological Society of America
100:196-205.
Kareiva, P. 1987. Habitat fragmentation and the stability of predator prey interactions. Nature
326:388-390.
Kindlmann, P. and A. F. G. Dixon. 1999a. Generation time ratios - Determinants of prey
abundance in insect predator-prey interactions. Biological Control 16:133-138.
Kindlmann, P. and A. F. G. Dixon. 1999b. Strategies of aphidophagous predators: lessons for
modelling insect predator-prey dynamics. Journal of Applied Entomology-Zeitschrift Fur
Angewandte Entomologie 123:397-399.
Kindlmann, P. and A. F. G. Dixon. 2001. When and why top-down regulation fails in arthropod
predator-prey systems. Basic and Applied Ecology 2:333-340.
Koji, S., Z. R. Khan, and C. A. O. Midega. 2007. Field boundaries of Panicum maximum as a
reservoir for predators and a sink for Chilo partellus. Journal of Applied Entomology
131:186-196.
Krauss, J., I. Steffan-Dewenter, and T. Tscharntke. 2003. How does landscape context contribute
to effects of habitat fragmentation on diversity and population density of butterflies?
Journal of Biogeography 30:889-900.

23

Kremen, C. 2005. Managing ecosystem services: what do we need to know about their ecology?
Ecology Letters 8:468-479.
Kruess, A. and T. Tscharntke. 1994. Habitat fragmentation, species loss, and biological-control.
Science 264:1581-1584.
Landis, D., M. J. Brewer, and G. E. Heimpel. 2003. Soybean aphid parasitoid questionnaire
2003. East Lansing, MI, USA.
Landis, D. A. and W. van der Werf. 1997. Early-season predation impacts the establishment of
aphids and spread of beet yellows virus in sugar beet. Entomophaga 42:499-516.
Landis, D. A., S. D. Wratten, and G. M. Gurr. 2000. Habitat management to conserve natural
enemies of arthropod pests in agriculture. Annual Review of Entomology 45:175-201.
Langer, A. and T. Hance. 2004. Enhancing parasitism of wheat aphids through apparent
competition: a tool for biological control. Agriculture Ecosystems & Environment
102:205-212.
Lee, J. C., F. B. Menalled, and D. A. Landis. 2001. Refuge habitats modify impact of insecticide
disturbance on carabid beetle communities. Journal of Applied Ecology 38:472-483.
Losey, J. E. and M. Vaughan. 2006. The economic value of ecological services provided by
insects. Bioscience 56:311-323.
Marino, P. C. and D. A. Landis. 1996. Effect of landscape structure on parasitoid diversity and
parasitism in agroecosystems. Ecological Applications 6:276-284.
McGarigal, K., S. A. Cushman, and E. Ene. 2012. Fragstats v4: Spatial pattern analysis program
for categorical and continuous maps. Computer software program produced by the
authors at University of Massachusetts, Amherst. Available at the following website:
http://umass.edu/landeco/research/fragstats/fragstats.html.
Meek, B., D. Loxton, T. Sparks, R. Pywell, H. Pickett, and M. Nowakowski. 2002. The effect of
arable field margin composition on invertebrate biodiversity. Biological Conservation
106:259-271.
Millennium Ecosystem Assessment. 2005. Ecosystems and Human Well-being: Synthesis. Island
Press, Washington, D.C.
Murdoch, W. W., S. L. Swarbrick, and C. J. Briggs. 2006. Biological control: lessons from a
study of California red scale. Population Ecology 48:297-305.
Nielsen, C. and A. E. Hajek. 2005. Control of invasive soybean aphid, Aphis glycines
(Hemiptera: Aphididae), populations by existing natural enemies in New York State, with
emphasis on entomopathogenic fungi. Environmental Entomology 34:1036-1047.
24

Oberg, S., B. Ekbom, and R. Bommarco. 2007. Influence of habitat type and surrounding
landscape on spider diversity in Swedish agroecosystems. Agriculture Ecosystems &
Environment 122:211-219.
Östman, O., B. Ekbom, and J. Bengtsson. 2001. Landscape heterogeneity and farming practice
influence biological control. Basic and Applied Ecology 2:365-371.
Pfiffner, L., H. Luka, C. Schlatter, A. Juen, and M. Traugott. 2009. Impact of wildflower strips
on biological control of cabbage lepidopterans. Agriculture Ecosystems & Environment
129:310-314.
Pluess, T., I. Opatovsky, E. Gavish-Regev, Y. Lubin, and M. H. Schmidt-Entling. 2010. Noncrop habitats in the landscape enhance spider diversity in wheat fields of a desert
agroecosystem. Agriculture Ecosystems & Environment 137:68-74.
Polis, G. A., W. B. Anderson, and R. D. Holt. 1997. Toward an integration of landscape and food
web ecology: The dynamics of spatially subsidized food webs. Annual Review of
Ecology and Systematics 28:289-316.
Power, A. G. 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philosophical
Transactions of the Royal Society B-Biological Sciences 365:2959-2971.
Prasifka, J. R., K. M. Heinz, and R. R. Minzenmayer. 2004. Relationships of landscape, prey and
agronomic variables to the abundance of generalist predators in cotton (Gossypium
hirsutum) fields. Landscape Ecology 19:709-717.
Purtauf, T., I. Roschewitz, J. Dauber, C. Thies, T. Tscharntke, and V. Wolters. 2005. Landscape
context of organic and conventional farms: Influences on carabid beetle diversity.
Agriculture Ecosystems & Environment 108:165-174.
Ragsdale, D. W., B. P. McCornack, R. C. Venette, B. D. Potter, I. V. Macrae, E. W. Hodgson,
M. E. O'Neal, K. D. Johnson, R. J. O'Neil, C. D. DiFonzo, T. E. Hunt, P. A. Glogoza, and
E. M. Cullen. 2007. Economic threshold for soybean aphid (Hemiptera: Aphididae).
Journal of Economic Entomology 100:1258-1267.
Ragsdale, D. W., D. J. Voegtlin, and R. J. O'Neil. 2004. Soybean aphid biology in North
America. Annals of the Entomological Society of America 97:204-208.
Ricklefs, R. E. and I. J. Lovette. 1999. The roles of island area per se and habitat diversity in the
species-area relationships of four Lesser Antillean faunal groups. Journal of Animal
Ecology 68:1142-1160.
Roland, J. and P. D. Taylor. 1997. Insect parasitoid species respond to forest structure at
different spatial scales. Nature 386:710-713.

25

Roschewitz, I., M. Hucker, T. Tscharntke, and C. Thies. 2005. The influence of landscape
context and farming practices on parasitism of cereal aphids. Agriculture Ecosystems &
Environment 108:218-227.
Rutledge, C. E., R. J. O'Neil, T. B. Fox, and D. A. Landis. 2004. Soybean aphid predators and
their use in integrated pest management. Annals of the Entomological Society of America
97:240-248.
Schlapfer, F., B. Schmid, and L. Seidl. 1999. Expert estimates about effects of biodiversity on
ecosystem processes and services. Oikos 84:346-352.
Schmidt, M. H., I. Roschewitz, C. Thies, and T. Tscharntke. 2005. Differential effects of
landscape and management on diversity and density of ground-dwelling farmland
spiders. Journal of Applied Ecology 42:281-287.
Schmidt, M. H., C. Thies, W. Nentwig, and T. Tscharntke. 2008. Contrasting responses of arable
spiders to the landscape matrix at different spatial scales. Journal of Biogeography
35:157-166.Schmidt, M. H. and T. Tscharntke. 2005. Landscape context of sheetweb
spider (Araneae: Linyphiidae) abundance in cereal fields. Journal of Biogeography
32:467-473.
Settle, W. H., H. Ariawan, E. T. Astuti, W. Cahyana, A. L. Hakim, D. Hindayana, A. S. Lestari,
Pajarningsih, and Sartanto. 1996. Managing tropical rice pests through conservation of
generalist natural enemies and alternative prey. Ecology 77:1975-1988.
Smith, J., S. G. Potts, B. A. Woodcock, and P. Eggleton. 2008. Can arable field margins be
managed to enhance their biodiversity, conservation and functional value for soil
macrofauna? Journal of Applied Ecology 45:269-278.
Snyder, W. E. and D. H. Wise. 1999. Predator interference and the establishment of generalist
predator populations for biocontrol. Biological Control 15:283-292.
Steffan-Dewenter, I. 2003. Importance of habitat area and landscape context for species richness
of bees and wasps in fragmented orchard meadows. Conservation Biology 17:1036-1044.
Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tscharntke. 2002. Scaledependent effects of landscape context on three pollinator guilds. Ecology 83:1421-1432.
Steffan-Dewenter, I., U. Munzenberg, and T. Tscharntke. 2001. Pollination, seed set and seed
predation on a landscape scale. Proceedings of the Royal Society of London Series BBiological Sciences 268:1685-1690.
Swinton, S. M., F. Lupi, G. P. Robertson, and S. K. Hamilton. 2007. Ecosystem services and
agriculture: Cultivating agricultural ecosystems for diverse benefits. Ecological
Economics 64:245-252.

26

Symondson, W. O. C., K. D. Sunderland, and M. H. Greenstone. 2002. Can generalist predators
be effective biocontrol agents? Annual Review of Entomology 47:561-594.
Thies, C., I. Roschewitz, and T. Tscharntke. 2005. The landscape context of cereal aphidparasitoid interactions. Proceedings of the Royal Society B-Biological Sciences 272:203210.
Thies, C., I. Steffan-Dewenter, and T. Tscharntke. 2003. Effects of landscape context on
herbivory and parasitism at different spatial scales. Oikos 101:18-25.
Thies, C. and T. Tscharntke. 1999. Landscape structure and biological control in
agroecosystems. Science 285:893-895.
Thomas, M. B., S. D. Wratten, and N. W. Sotherton. 1992. Creation of island habitats in
farmland to manipulate populations of beneficial arthropods - predator densities and
species composition. Journal of Applied Ecology 29:524-531.
Thomson, L. J. and A. A. Hoffmann. 2010. Natural enemy responses and pest control:
Importance of local vegetation. Biological Control 52:160-166.
Tilman, D., K. G. Cassman, P. A. Matson, R. Naylor, and S. Polasky. 2002. Agricultural
sustainability and intensive production practices. Nature 418:671-677.
Treacy, M. F., J. H. Benedict, M. H. Walmsley, J. D. Lopez, and R. K. Morrison. 1987.
Parasitism of bollworm (Lepidoptera: Noctuidae) eggs on nectaried and nectariless
cotton. Environmental Entomology 16:420-423.
Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan-Dewenter, and C. Thies. 2005. Landscape
perspectives on agricultural intensification and biodiversity - ecosystem service
management. Ecology Letters 8:857-874.
Tscharntke, T., I. Steffan-Dewenter, A. Kruess, and C. Thies. 2002. Characteristics of insect
populations on habitat fragments: A mini review. Ecological Research 17:229-239.
Turner, M. G., R. H. Gardner, and R. V. O'Neill. 2001. Landscape Ecology in Theory and
Practice: Pattern and Process. Springer-Verlag, New York.
Vasseur, C., A. Joannon, S. Aviron, F. Burel, J. M. Meynard, and J. Baudry. 2013. The cropping
systems mosaic: how does the hidden heterogeneity of agricultural landscapes drive
arthropod populations? Agriculture, Ecosystems and Environment 166:12.
Venette, R. C. and D. W. Ragsdale. 2004. Assessing the invasion by soybean aphid (Homoptera:
Aphididae): Where will it end? Annals of the Entomological Society of America 97:219226.
Walton, N. J. and R. Isaacs. 2011. Influence of native flowering plant strips on natural enemies
27

and herbivores in adjacent blueberry fields. Environmental Entomology 40:697-705.
Weibull, A. C., O. Ostman, and A. Granqvist. 2003. Species richness in agroecosystems: the
effect of landscape, habitat and farm management. Biodiversity and Conservation
12:1335-1355.
Werling, B. P. and C. Gratton. 2008. Influence of field margins and landscape context on ground
beetle diversity in Wisconsin (USA) potato fields. Agriculture Ecosystems &
Environment 128:104-108.
Werling, B. P. and C. Gratton. 2010. Local and broadscale landscape structure differentially
impact predation of two potato pests. Ecological Applications 20:1114-1125.
Westphal, C., I. Steffan-Dewenter, and T. Tscharntke. 2003. Mass flowering crops enhance
pollinator densities at a landscape scale. Ecology Letters 6:961-965.
Wilby, A., L. P. Lan, K. L. Heong, N. P. D. Huyen, N. H. Quang, N. V. Minh, and M. B.
Thomas. 2006. Arthropod diversity and community structure in relation to land use in the
Mekong Delta, Vietnam. Ecosystems 9:538-549.
With, K. A., D. M. Pavuk, J. L. Worchuck, R. K. Oates, and J. L. Fisher. 2002. Threshold effects
of landscape structure on biological control in agroecosystems. Ecological Applications
12:52-65.
Woltz, J. M., B. P. Werling, and D. Landis. 2012. Natural enemies and insect outbreaks in
agriculture: a landscape perspective. in P. Barbosa, D. Letourneau, and A. Agrawal,
editors. Insect Outbreaks Revisited. Wiley-Blackwell.
Wu, Z. S., D. Schenk-Hamlin, W. Y. Zhan, D. W. Ragsdale, and G. E. Heimpel. 2004. The
soybean aphid in China: A historical review. Annals of the Entomological Society of
America 97:209-218.
Zaviezo, T., A. A. Grez, C. F. Estades, and A. Perez. 2006. Effects of habitat loss, habitat
fragmentation, and isolation on the density, species richness, and distribution of
ladybeetles in manipulated alfalfa landscapes. Ecological Entomology 31:646-656.

28

CHAPTER I: Landscape structure and habitat management differentially influence coccinellids
in an agricultural landscape
Abstract
Woltz, J.M., R. Isaacs, and D.A. Landis, 2012. Landscape structure and habitat management
differentially influence insect natural enemies in an agricultural landscape. Agriculture,
Ecosystems, and Environment 152: 40-49.

Increasing evidence suggests that landscape composition is an important driver of
beneficial insect populations and resulting ecosystem services. Additionally, local-scale
manipulations such as planting floral strips are used at the field-level to provide resources for
beneficial insects to increase their services. It has been proposed that the benefits of local
manipulations will depend on the landscape context, with greater benefits in simplified
landscapes and smaller benefits in landscapes with an abundance of non-crop resources. To test
this, we used soybean aphid, Aphis glycines, and its coccinellid predators as a model system to
elucidate the effects of habitat management and landscape on biocontrol services in soybean. We
selected pairs of soybean fields in landscapes of varying composition and planted buckwheat,
Fagopyrum esculentum, strips adjacent to one field in each pair. We measured coccinellid
abundance and biocontrol in each field. Coccinellid abundance was higher in buckwheat than in
control field margins in all landscapes, and coccinellid abundance in soybean was positively
related to amount of semi-natural vegetation in the landscape. We found no evidence of an
interaction between landscape and local variables, and biocontrol services were high in all
contexts. For soybean aphid suppression, landscape factors are the key drivers of predator
abundance.

29

Introduction
Landscape complexity and local management of agricultural lands can influence the
diversity of plants, animals, and microorganisms (Altieri 1999, Tscharntke et al. 2005) which in
turn affect the provision of ecosystem services such as nutrient cycling, water regulation, and
pest suppression provided by beneficial insects (Altieri, 1999; Schlapfer et al., 1999; Tilman et
al., 2002; Kremen, 2005). Biological control of crop pests is an ecosystem service valued at US$
4.5 billion/yr in the United States alone (Losey and Vaughan, 2006), and it can reduce the need
for costly chemical pesticides and associated environmental and human health concerns (Meehan
et al., 2011). In order to support this valuable ecosystem service, various on-farm management
practices have been developed to increase natural enemy diversity and abundance. For example,
planting non-crop habitats adjacent to crop fields can provide natural enemies with additional
food resources and refuge from disturbance (Landis et al. 2000), and may decrease the negative
impacts of farming on local biodiversity (Concepción et al. 2008). However, recent studies
demonstrating the influence of landscape-scale factors on biocontrol services have led some to
hypothesize that the landscape context may constrain the ability of local management actions to
influence community dynamics and biocontrol services (Concepción et al. 2008, Tscharntke et
al. 2012).
Landscape composition influences natural enemy communities through the provision of
food and shelter resources. Complex landscapes with large amounts of semi-natural habitat may
benefit natural enemies by providing undisturbed areas that offer shelter from crop disturbances,
overwintering refuges, alternative hosts and prey, and additional nectar resources (Thies et al.,
2003; Bianchi et al., 2006). As such, the amount of non-crop or semi-natural habitat in a
landscape has been positively related to the abundance of a variety of predators, including

30

spiders (Drapela et al., 2008; Schmidt et al., 2008), carabid beetles (Purtauf et al., 2005),
coccinellid beetles (Gardiner et al., 2009), and opiliones (Gardiner et al., 2010). A meta-analysis
of 46 landscape-scale studies in agroecosystems showed positive relationships between the
amount of natural or non-crop habitat in the landscape and natural enemy abundance and
diversity, predation and parasitism (Chaplin-Kremer et al. 2011).
Landscape-mediated increases in natural enemy abundance and diversity can in turn lead
to higher predation and parasitism rates, and thus greater pest control. Thies and Tscharntke
(1999) found that high rates of plant damage by the rape pollen beetle, Meligethes aeneus
(Fabricius), and low rates of parasitism and larval mortality were correlated with landscape
simplicity, while parasitism by two key parasitoid species increased with landscape
heterogeneity. In a study by Gardiner et al. (2009), biocontrol of soybean aphid, Aphis glycines
Matsumara, was greater in diverse compared to agriculturally dominated landscapes, and in
landscapes with greater amounts of non-crop habitat. Similarly, the percentage of non-crop area
in the landscape has been positively related to parasitism-induced mortality of true armyworm
Pseudaleta unipuncta (Haworth) (Marino and Landis, 1996) and negatively related to pest
damage by rape pollen beetles (Thies et al., 2003).
At the field or farm scale, local habitat can also be manipulated to provide natural
enemies with resources to optimize their performance, a practice known as habitat management
(Landis et al., 2000). Floral resource strips are commonly used to increase natural enemy
efficacy by providing them with resources such as nectar, pollen, or alternative prey (Landis et
al., 2000). Efforts to increase biocontrol with floral strips have met with mixed success, with
some studies showing no effect on predator abundance within fields (Bigger and Chaney, 1998;
Koji et al., 2007) and others leading to increased predator or parasitoid abundance and reduced

31

pest pressure in crop fields (Hickman and Wratten, 1996; Langer and Hance, 2004; Lee and
Heimpel, 2005). Because the surrounding landscape supplies the natural enemies that may
respond to floral resource strips, it is possible that the ability of habitat management to influence
pest control depends on the landscape context in which a particular crop field is situated.
Tscharntke et al. (2005) suggested that local management practices may only increase biocontrol
services in partially simplified landscapes, presumably because complex landscapes already
supply natural enemies with necessary resources, while landscapes almost entirely cleared for
agriculture may not support sufficient natural enemy populations to allow significant effects of
habitat management. Thus, the increase in biocontrol services due to local management is
expected to be the greatest in landscapes of an intermediate complexity. Recent studies to test
this hypothesis have shown conflicting patterns, with landscape simplicity having positive (e.g.
Haenke et al., 2009) or negative (e.g. Winqvist et al., 2011) effects on the benefits of local
management practices.
Our goal was to examine the potential interaction between landscape traits and habitat
management in a typical field crop agroecosystem of the U.S. Midwest. To do so, we used the
soybean aphid and its coccinellid (lady beetle) predators as a model system to elucidate the
effects of habitat management and landscape on biocontrol services in soybean fields.
Coccinellids are a primary predator of soybean aphid in the north central U.S., with large species
like H. axyridis and C. septempunctata having the highest per capita consumption and
accounting for the majority of soybean aphid predation (Costamagna and Landis, 2007).
Coccinellid abundance has been correlated with landscape features such as diversity and
proportion of non-crop habitat (Gardiner et al. 2009) and may respond to floral resources
(Harmon et al., 2000). Many coccinellids are known to feed on pollen of a variety of flowering

32

plant species, including some aphidophagous species commonly found in soybean fields in north
central U.S., like C. septempunctata, P. quatourdecimpunctata, and C. maculata (Hodek and
Honek 1996). Pollen is an important protein source for carnivorous coccinellids when insect prey
are scarce (Hodek and Honek, 1996). Buckwheat is a commonly used flowering resource in
habitat management studies because of its demonstrated attractiveness to natural enemies
(Fiedler et al., 2008). Additionally, buckwheat strips can serve as a source of alternative prey,
like non-pest aphids and the eggs and larvae of lepidopterans and herbivorous beetles.
In this study, we hypothesized that: 1) coccinellid abundance will increase in response to
habitat management; 2) predators will significantly suppress soybean aphid populations; 3)
biocontrol services will be positively related to habitat diversity within landscapes; and 4) the
degree to which habitat management improves biocontrol services will depend on characteristics
of the surrounding landscape. To test these hypotheses, we manipulated floral resources adjacent
to soybean fields situated in landscapes of varying complexity and composition and measured
resultant changes in levels of biocontrol services.
Methods
Site design. This study was replicated in 17 sites across a gradient of landscape diversity
in southern Michigan (Figure 1.1; Tables 1.1, 1.2) during the summers of 2008 and 2009. Within
a year, each site consisted of two soybean fields, located from 0.5 – 1.8 km apart. In each site, a
control field was bordered by a fencerow or grassy field margin, and a treatment field was
adjacent to a floral resource strip of annual buckwheat (Fagopyrum esculentum cv. Mancan).
Hereafter, the control field margin and the buckwheat strip will collectively be referred to as
field edges. Within each field, sampling was replicated in four 30 m x 20 m plots (Figure 1.2). In
2008, plots in control fields were set up at least 40 m away from all edges and any non-soybean

33

vegetation. In the buckwheat fields in 2008 and in both fields in 2009, the plots were set up 15 m
away from the buckwheat strip or field margin and at least 40 m away from any non-soybean
vegetation on all other sides.
Buckwheat establishment and control margin characteristics. Buckwheat strips were
5 m wide x 200 m in length (Figure 1.2). The strips were planted using a 19 cm row spacing at a
rate of 70kg/ha using a Great Plains Seed Drill (Salina, Kansas) along field edges chosen by land
owners. In each year, the majority of strips (12 in 2008, 11 in 2009) were planted along
fencerows, with the remaining strips planted along woodlots (2 in 2008), driveways or roads (2
in 2008, 5 in 2009), or directly between two crop fields (1 in each year). Strips were planted
between 19 May and 6 June in 2008, and between 3 June and 7 July in 2009 (Tables 1.3, 1.4).
The strips were planted later in 2009 due to heavy rains in May. After reaching full bloom,
buckwheat height and density were measured at 10 random locations within each strip.
Control field margins were chosen to match the cardinal direction of the paired
buckwheat strip at each site, except in a few cases where field dimensions made this infeasible.
Control field margins were similar to the edges next to which buckwheat strips had been planted
(10 at fencerows, 1 next to a woodlot, 3 next to driveways or roads, 1 between crop fields and 2
next to larger patches of mown grass). Vegetation of control field margins consisted of grasses
and weedy herbs generally between half a meter and a meter in height.
Aphid counts. Soybean aphid abundance was estimated at weekly intervals using
destructive whole plant counts. Five randomly selected soybean plants in each plot were
removed and the number of alate and apterous aphids counted. Vegetative and reproductive
growth stages for each plant were recorded (Higley and Boethel, 1994).

34

Coccinellid counts. Aerially dispersing coccinellids were measured in soybean fields and
field edges with 23 x 28 cm unbaited yellow sticky cards (PHEROCON AM, Great Lakes IPM,
Vestaburg, MI). All coccinellids on the cards were counted and identified to species in the field.
Coccinellid voucher specimens were deposited in the Albert J. Cook Arthropod Research
Collection at Michigan State University (Table 1.5). In 2008, traps were placed in the center of
each plot in all study sites and at two locations within each buckwheat strip. Traps were
suspended at 120 cm height from plastic step-in fence posts (Zareba Systems, Lititz, PA) and
replaced every week from 1 June to 15 Aug. In 2009, four traps were placed in each field, one in
the center of each plot, and three sticky traps were placed in the field edge of both treatment and
control fields. The traps in the center of each plot were located 25 m away from the field edge,
and were suspended from step-in plastic fence posts at plant canopy height and were replaced
every week from 8 June to 26 Aug.
Biocontrol study. The effect of biocontrol services on soybean aphid populations was
determined by contrasting soybean aphid population growth in the presence and absence of
generalist predators, using methods modified from Gardiner et al. (2009). Predators were
excluded using cages erected from 7 July to 19 July in 2008 (Table 1.6) and from 9 July to 12
Aug in 2009 (Table 1.7). In each plot a single soybean plant was enclosed in either a 0.84 m tall,
0.30 m diameter (2008) or a 1.07 m tall, 0.36 m diameter (2009) wire tomato support cage. The
cages were covered with fine-mesh white netting (openings of 0.65 x 0.17 mm; Skeeta,
Bradenton, FL) sewn into sleeves of 137 cm circumference. The bottom of the mesh sleeve was
buried in the soil, and the top was tied to a metal “T” post. Each caged plant was paired with an
uncaged plant of the same vegetative and reproductive growth stage, located 1 m away in the
same row.

35

Each caged and uncaged plant was infested with 5 aphids in 2008 and 10 aphids in 2009.
In both years the aphids were of mixed age classes representing natural colonies in the field. The
aphids were transferred using a fine paintbrush from soybean plants in predator exclusion cages
established in the soybean fields earlier in the season to rear aphids. Fields in each site were
infested with aphids reared in that particular site, except where a field contained insufficient
populations. The number of aphids on each caged and uncaged plant was counted at 7 d and 14 d
after infestation using non-destructive sampling in the field. Apterous and alate aphids were
counted separately, and cages containing more than 10 alates were excluded from analyses to
limit the influence of alates that would have normally dispersed to other plants remaining and
reproducing within the cages.
Biocontrol services were measured as an index describing the difference in aphid
population size with and without naturally occurring predators as a proportion of the exclusion
treatment in each site (Gardiner et al., 2009). The Biocontrol Services Index (BSI) was
calculated as the difference between the number of aphids on the open plant and on the caged
plant, divided by the number of aphids on the caged plant, all at fourteen days after initial
infestation.
BSI = (Ac-Ao)/Ac
Ac = number of aphids on caged plant at 14 d
Ao = number of aphids on open plant at 14 d
BSI values range from 0-1, with 1 representing total aphid suppression and 0 representing no
aphid suppression. Negative BSI values created in cases where Ao > Ac were reset to 0. Values
of BSI were calculated for each plot in each year, and a field average was calculated.

36

All cage study results in 2008 are based on 11 sites. Three sites were excluded from the
cage study due to the unsuccessful establishment of their buckwheat strips. Three additional sites
were excluded from the analysis of cage data due to an inability to successfully establish rearing
and experimental cages at the site. All cage study results in 2009 are based on 12 sites. Two sites
were not included in the cage study due to the very late blooming of the buckwheat strips in
those sites. Two sites were excluded from the analysis because large numbers of alate aphids in
the fields during the week of the cage study increased the numbers on the open plants. Two
additional sites were excluded due to difficulties in establishing aphid populations due to
competing pest populations.
Landscape analysis. Landscapes were assessed by considering all landcover types
surrounding each field using the methods of Gardiner (2009). GPS coordinates were collected
from the center of each field using a handheld GPS receiver. The GPS coordinates were used to
obtain ortho-rectified digital aerial photos for each field from the Michigan Spatial Data Library.
Habitats around each field were digitized to a 2.0 km radius using ARC GIS 9.2 (2008) or 9.3
(2009; ESRI Redlands, CA). This range represents a reasonable foraging area of large species of
adult ladybeetles. Flights of approximately 1.8 – 2.0 km or greater are considered migratory (i.e.
entering or leaving overwintering habitat) for Hippodamia convergens (Hodek et al., 1993), a
lady beetle species of similar size and diet preferences as the dominant coccinellid predators of
soybean aphid. Shorter distances can be considered within the realm of “trivial flights,” or shortdistance flights for foraging, which is the type of dispersal relevant for biocontrol in crop fields.
Habitat types within field sites were ground-verified each summer and landscape changes were
corrected during digitization. During ground-truthing, the specific land cover within all habitat
polygons in each landscape was determined by assigning landscape categories to each polygon.

37

Landscape categories included field crops (primarily corn, soy, wheat, and forage crops) and
non-crop habitat types (primarily woodlots, old field, and residential areas) (Table 1.8).
Categories were added for any additional land cover types discovered during ground-truthing.
Total area of each habitat category for each site was calculated in ARC GIS 9.2 or 9.3 at four
scales: 0.5, 1.0, 1.5, and 2.0 km from the center of the focal field.
Proportion of semi-natural habitat and Simpson’s Diversity Index (D) (Simpson, 1949)
were used independently as metrics of landscape heterogeneity. Semi-natural habitat included
wooded habitats such as forests, woodlots, and fencerows, as well as grassy habitats such as old
field and Conservation Reserve Program lands. Simpson’s index -- typically used as a measure
of species diversity -- is used here as a measure of habitat diversity within each landscape,
according to the methods of Gardiner et al. (2009).
D=

 pi2,

where pi = proportion of landscape occupied by patch type i. Simpson’s index is less sensitive
than other diversity measures to category richness, in this case habitat richness (Magurran,
2004). Therefore, Simpson’s index effectively describes the variance in relative abundance of
habitat types without being skewed by the presence of rare habitats (Magurran, 2004).
Diversity decreases as D increases, therefore we calculated –lnD so that values intuitively
increase with increasing diversity. The negative logarithm is preferable to other transformations
such as the reciprocal of D, which can create variance problems (Magurran, 2004).
Because seminatural habitat and metrics of habitat diversity have been found to be highly
correlated in other studies (e.g. Thies and Tscharntke, 1999; Steffan-Dewenter et al., 2002;
Roschewitz et al., 2005b), we calculated Pearson correlation coefficients (PROC CORR, SAS

38

9.2) for the two landscape variables percentage of seminatural habitat and –lnD at each
landscape extent to determine the degree to which these two metrics measured different
components of landscape structure.
Statistical analysis. Coccinellid abundance: To test the hypothesis that coccinellid
abundance increases in response to habitat management, total coccinellid captures were
compared between treatment and control fields, and between buckwheat strips and control field
margins. Additionally, the response of coccinellid populations to landscape composition and the
interaction between landscape and habitat management were determined. Weekly coccinellid
counts were averaged for each field or field margin and analyzed separately for each year using
analyses of variance (ANOVA) (PROC GLIMMIX, SAS 9.2). For each response variable
(coccinellid abundance in soybean fields in 2008, coccinellid abundance in soybean fields in
2009, and coccinellid abundance in field edges in 2009), eight separate models were run
considering either the effect of landscape diversity (-lnD) or proportion of semi-natural habitat
measured at 0.5, 1.0, 1.5, or 2.0 km. Fixed effects were treatment, landscape (either –lnD or
semi-natural habitat), and the treatment by landscape interaction. Coccinellid abundance in field
edges was log transformed prior to analysis to meet the assumptions of ANOVA. It was
unnecessary to transform coccinellid abundance in soybean fields.
Additionally, individual coccinellid species abundance was compared between
buckwheat strips and control field margins, as well as between the soybean fields adjacent to
buckwheat fields and those without strips. Because of difficulty in meeting distributional
assumptions with the data for any individual coccinellid species, nonparametric tests were used.
Average summer abundance in each field or field edge were first ranked (PROC RANK, SAS
9.2), such that the field or edge with the highest abundance was given the highest rank. Then,

39

ranks were compared using mixed model ANOVA (PROC GLIMMIX, SAS 9.2), with treatment
as a fixed effect and site as a random effect. ANOVA on rank-transformed data does not depend
on the distribution of the data, is resistant to outliers, and is robust to non-normality of errors.
However, assumptions regarding homoscedasticity of variance still hold, thus we were only able
to compare the relatively common species H. axyridis, C. septempunctata, P.
quatourdecimpunctata, H. variegata, and C. maculata using this method.
Biocontrol study: To test the hypothesis that the presence of predators significantly
depresses soybean aphid populations, numbers of aphids on caged and uncaged soybeans were
compared. Aphid counts were log(x+1) transformed prior to analysis to meet the assumptions of
homogeneity of variance, and were analyzed using a repeated measures mixed model ANOVA
(PROC GLIMMIX, SAS 9.1). Fixed effects were treatment, cage status (i.e., plant caged or
uncaged), week and their two- and three-way interactions. Random effects were site and all
interaction terms involving site.
Results
Buckwheat establishment. In 2008, buckwheat strips were successfully established in
14 out of 17 sites (Figure 1.3). Three sites were not included in the analysis because of low
germination due to flooding or loss to hail. The average height of buckwheat in the 14 strips used
in the 2008 study was 89 cm, ranging from 50 cm to 109 cm (Table 1.3). The average density of
2

2

buckwheat in the 14 strips was 12.2 stems/100cm , ranging from 7.2 to 20.5 stems/100cm . In
2009, 17 strips were successfully established. The average height of all buckwheat strips in 2009
was 57.4 cm, ranging from 20.5 to 93.1 cm (Table 1.4). The average density of all buckwheat
2

2

strips in 2009 was 7.4 stems/100cm , ranging from 1.4 to 16.8 stems/100cm . Final buckwheat
height did not change significantly with planting date in either year (2008: F1,126=1.85, p=0.176;
40

2009: F1,153=0.09, p =0.762), nor did density (2008: F1,126=0.26, p=0.613; 2009: F1,153=0.28, p
=0.600).
Landscape analysis. The studied landscapes represented a gradient of landscape
diversity, with landscapes ranging from forest-dominated to agriculturally-dominated in both
years. All landscapes were predominantly made up of corn, soy, wheat, forest, grassland, and
residential areas, but proportions of each of these varied between landscapes (Table 1.9, 1.10). In
2008, diversity values (–lnD) ranged from 1.19 to 1.91, while the percentage of semi-natural
habitat ranged from 3 to 79% at the 2.0 km scale. In 2009, diversity values ranged from 0.83 to
1.95, with the percentage of semi-natural habitat ranging from 5 to 76 % at the 2.0 km scale.
Values were similar for smaller scales.
In 2008, the landscape variables -lnD and the proportion of semi-natural habitat were not
2

2

2

2

correlated at any scale (r =-0.11, p=0.59; r = -0.05, p=0.81; r = 0.08, p=0.69; r =0.05,
p=0.80; in descending order of landscape extent). In 2009, these landscape variables were
2

2

correlated at 2.0, 1.5, and 0.5 km. However, the correlations are low (r = -0.44, p=0.009; r = 2

0.33, p=0.05; r =0.34, p=0.05 respectively) and they explain little variation in the two variables.
2

Furthermore, the changing sign of the correlations and the lack of correlation at 1.0 km (r = 0.24, p=0.18), suggest that these two variables are not closely related within this set of
landscapes.
Aphid counts. Soybean aphids naturally colonized all fields used in the study, but
populations generally remained below the economic threshold of 250 aphids/ plant (Ragsdale et
al., 2007). In 2008, aphids were first detected on 2 June and aphids were discovered in all fields
by 17 July (Table 1.6). In 2009, aphids were initially detected during the first week of sampling
41

(9-12 June) in 6 fields, and all fields were infested by 23 July (Table 1.7). Only one field reached
the economic threshold during the last week of sampling in 2009.
Coccinellid abundance. In 2008, 1412 coccinellids were collected and identified in
soybean fields. Of these, 90% were exotic species, with 47 % H. axyridis, 31% C.
septempunctata, 10% P. quatourdecimpunctata, and 2% H. variegata. The most common native
species were C. maculata and H. parenthesis, which made up 6% and 2% of the total capture,
respectively. All other species made up 1% or less of captures. In 2009, 4276 coccinellids were
collected and identified in soybean fields. The majority were exotic species, with 33% C.
septempunctata, 28% H. axyridis, 18% P. quatourdecimpunctata, and 9% H. variegata. The
most common native species were C. maculata and H. parenthesis, which made up 5% and 3%
of the total capture, respectively. All other species made up 1% or less of captures.
Habitat manipulation altered predator abundance, with significantly more coccinellids
found in buckwheat strips than in control field edges in 2009, the only year for which the
comparison was made (Table 1.11, Figure 1.4). An average of 2.69 ± 0.42 lady beetles per card
were found in the buckwheat strips, as compared to an average of 1.32 ± 0.21 per card in the
control field margins.
Slightly more coccinellids were captured in fields with buckwheat strips than in control
fields in both years, although the effect was not significant in either year (Table 1.12; Figure
1.4). Coccinellid abundance in fields was not affected by landscape diversity at any scale in
either year, nor by the diversity by treatment interaction (Table 1.11). However, in 2009, lady
beetles were more abundant in soybean fields surrounded by landscapes with greater proportions
of semi-natural habitat measured at 1.5 and 2.0 km (Table 1.11, Figure 1.4). The treatment by
landscape interaction for this effect was not significant at any scale. . Statistical results for all

42

effects were identical at the 1.5 and 2.0 km scales, while no effect of landscape was shown at
smaller scales.
Within buckwheat strips, 81 % of lady beetles captured were exoticspecies (C.
septempunctata 28%, H. variegata 25%, P. quatourdecimpunctata 20%, H. axyridis 8%). The
native pink lady beetle C. maculata made up an additional 9% and the orange spotted lady
beetles Brachiacanthus sp. made up an additional 4% of the lady beetles found in buckwheat
strips. In contrast, there were more H. axyridis (17%) and fewer H. variegata (3%) in the control
field margins. Proportions of other lady beetles were similar (C. septempunctata 24%, P.
quatourdecimpunctata 34%, C. maculata 4%, Brachiacanthus sp.8%). However, the abundance
of lady beetles in field margins did not vary with landscape diversity or amount of semi-natural
habitat in the landscape in either year (Table 1.11).
While rank comparisons showed no difference in species abundance between treatment
and control fields in 2008 (Table 1.12), several coccinellid species were significantly more
abundant in buckwheat strips than in control field edges in 2009 (Table 1.13). Using rank
abundance, three common exotic species C. septempunctata, H. variegata, and P.
quatordecimpunctata, as well as the common native species C. maculata, were more commonly
found in buckwheat strips. C. maculata was also found to be more abundant in fields next to
buckwheat strips than in fields next to control field margins. In contrast, no species was found to
be more abundant in control field margins, and only H. axyridis was more common in control
soybean fields.
Biocontrol study. Exclusion of natural enemies resulted in dramatically increased
soybean aphid populations in both years (Figure 1.5). Across fields and sites, the number of
aphids on caged plants from which predators were excluded was significantly higher than on

43

uncaged plants which were exposed to predators (2008: F1,10=251.33, p<0.0001; 2009:
F1,10=114.03, p<0.0001). After 14 days, there were 38 fold more aphids per plant on caged than
uncaged plants in 2008, and 53 fold more aphids per plant on caged than uncaged plants 2009.
Buckwheat treatment did not significantly affect the size of the difference in aphid populations
on caged and uncaged plants in either year (2008: F1,10=1.62, p=0.2313; 2009: F1,10=0.62,
p=0.4509), nor were any of the interactions involving treatment significant in either year. The
differences in aphid populations between caged and uncaged plants were greater at 14 days than
at 7 days in both years (Figure 1.5), as indicated by a significant week by cage status interaction
(2008: F1,10=61.06, p<0.0001; 2009: F1,10=30.03, p=0.0003).
We found very high suppression of aphids by the existing predator populations in both
years in almost all fields. BSI values in 2008 ranged from 0.77 to 1, with an average value of
0.954 ± 0.01 and 22 of 23 fields had BSI values greater than 0.90 (Table 1.14). BSI values in
2009 ranged from 0.66 to 0.99, with an average value of 0.954 ± 0.02 and 20 of 23 fields had
BSI values of greater than 0.90 (Table 1.14). We did not conduct landscape analyses on BSI
because of a lack of variation available to be explained by this or any other variable.
Discussion
Landscape effects on local management to improve services. In this study, predators
had a significant negative impact on soybean aphid populations and lady beetle abundance
increased both in habitat management strips and with the proportion of semi-natural habitat in
the landscape, but we found no interaction between landscape context and local habitat
management on coccinellid abundance. In contrast, other studies have found that the degree to
which local management can influence communities is dependent upon the degree of
44

heterogeneity in the surrounding landscape. Local management was more effective for increasing
richness and abundance in simplified, crop-dominated landscapes than in diverse landscapes for
a variety of taxa (e.g. pollinators, (Carvell et al., 2011); bees, (Holzschuh et al., 2007; Rundlof et
al., 2008); butterflies, (Rundlof and Smith, 2006); birds, (Geiger et al., 2010; Smith et al., 2010);
aphidophagous syrphids, (Haenke et al., 2009); and plants, (Roschewitz et al., 2005a)). In
contrast, Winqvist et al. (2011) demonstrated the opposite effect, with organic farming methods
increasing biocontrol over conventional methods only in the most diverse landscapes.
Additionally, other studies have shown little impact of local management at all, with taxon
diversity and abundance responding primarily to landscape composition alone (Purtauf et al.,
2005; Schmidt et al., 2005). These contrasting results may indicate that these relationships are
highly taxon- and context-dependent. Indeed, even within a single study, different taxa exhibited
opposite responses to landscape and local management (Concepcion et al., 2008), while a recent
meta-analysis showed that the arthropod response to local management was influenced by
landscape context for populations in cropland habitats but not within grassland habitats (Batáry
et al., 2011).
Effects of local management on coccinellid beetles. The habitat management treatment
successfully attracted coccinellid beetles, with significantly more lady beetles in buckwheat
strips than in grassy field margins or fencerows. However, this increase in coccinellid abundance
at the field edges was not reflected in higher abundance in adjacent soybean fields. Other studies
have similarly demonstrated predators taking advantage of managed habitats without clear
effects on adjacent fields. For example, a guinea grass border did not affect the abundance of
predators within maize or the density of spotted stem borer, Chilo partellus, in spite of attracting
abundant predators to the strip (Koji et al., 2007). In contrast, Walton and Isaacs (2011) found

45

significantly higher abundance of both syrphid flies and predatory wasps in blueberry fields
adjacent to native wildflower plantings than next to mown grass strips. In other cases, increased
natural enemy abundance from floral resource strips have resulted in higher parasitism rates
(Langer and Hance, 2004) and lower pest abundance in adjacent fields (Hickman and Wratten,
1996). Inconsistencies in these patterns could be a result of varying predator biology or
landscape context.
Landscape effects on coccinellid beetles. In 2009, abundance of coccinellids in soybean
fields increased with proportion of semi-natural habitat in the surrounding landscape. Similar
positive effects of semi-natural habitat on natural enemies have been found for a variety of
predatory and parasitic taxa (Chaplin-Kramer et al., 2011). Lady beetles require undisturbed
habitat for overwintering, and H. axyridis, one of the two most abundant species collected in our
study, is known to overwinter in wooded areas. The abundance of potential overwintering sites in
landscapes with large proportions of semi-natural habitat could have led to higher overwinter
survival of coccinellids in those landscapes, leading to larger populations emerging in the spring.
Additionally, lady beetles in the north central U.S. emerge from overwintering in March and
April, before field crops such as corn and soy are planted in May or June. Thus, lady beetles are
dependent upon other food sources in the spring before foraging on soybean aphid in the
summer. Aphids in semi-natural habitat, such as those feeding on grasses or legumes in old
fields, could provide an abundant food source in the spring, allowing lady beetles in these
landscapes to build up larger populations by the time field crops are planted in summer.
In other studies, seminatural habitat has been used as a metric of landscape heterogeneity, as
in European landscapes it has been repeatedly found to be highly correlated with Shannon habitat
diversity (e.g. Thies and Tscharntke, 1999; Steffan-Dewenter et al., 2002; Roschewitz et al.,

46

2005b). In our study, however, Simpson’s Diversity, a similar metric, was not strongly correlated
with seminatural habitat at any scale. This could be because European landscapes are
functionally different than landscapes in the north central U.S. Alternatively, the landscapes
examined in this study may represent a greater maximum amount of seminatural habitat. In the
European studies cited, maximum semi-natural habitat was usually around 30 – 50%, whereas in
this study semi-natural habitat was as high as 79%. It is possible that the relationship between
habitat diversity and seminatural habitat breaks down at high levels of seminatural habitat
because of dominance of one or two habitat types. A landscape may have a large proportion of
seminatural habitat and low diversity if, for example, it is dominated by a single seminatural
landcover type such as forest. In our study, semi-natural habitat was a better predictor of
coccinellid abundance than habitat diversity. This may be because a single type of seminatural
habitat can still provide multiple resources, such as nesting and overwintering habitats,
temporary refuges, alternative prey and floral resources (Bianchi et al., 2006).
In contrast to the effect on lady beetles in crop fields, there was no effect of semi-natural
habitat on coccinellid abundance within the buckwheat strips. Haenke et al. (2009) found that the
abundance and species richness of syrphid flies in flower strips was positively related to the
amount of cropland in the surrounding landscape, presumably because the addition of floral
resources was more rewarding in crop-dominated landscapes than in diverse landscapes with
abundant resources. This suggests that coccinellids used the flowering buckwheat strips
regardless of the availability of other resources in the surrounding landscapes.
Effects of predators on soybean aphid density. Predators provided high levels of
biological control of soybean aphid in all sites. Regardless of landscape conditions or habitat
management, soybean aphid populations exposed to ambient predator densities were greatly

47

suppressed compared to the predator exclusion treatment. In light of the fact that the abundance
of the major predators of soybean aphid in this system, coccinellid beetles, was positively related
to the proportion of semi-natural habitat in the landscape surrounding study fields, it is somewhat
surprising that biocontrol services were not similarly lower in crop dominated landscapes and
higher in landscapes with more semi-natural habitats. The high levels of biocontrol in all
landscapes suggests that even the lowest abundance of coccinellids was sufficient to control
soybean aphid. This result stands in stark contrast to those of Gardiner et al. (2009) who found
that both coccinellid abundance and soybean aphid suppression were correlated with habitat
diversity and abundance of non-crop habitat in surrounding landscapes. This may indicate a shift
in this system from 2005-06 when Gardiner and others (2009) collected their data to 2008-09
when the data for this paper were collected. Elton (1958) postulated that changes within
biological communities that lead to increased competition and predation or parasitism of exotics
may be responsible for the transition of some exotic species from highly invasive to merely
naturalized members of the community. In a similar fashion, the predominant predators of the
soybean aphid, H. axyridis and C. septempunctata, may have become a greater proportion of the
existing natural enemy community in response to the abundant food source provided by soybean
aphids (Heimpel et al., 2010), thereby helping to suppress further aphid outbreaks. As further
evidence, soybean aphid and H. axyridis populations were initially cyclic, with aphid outbreak
years (2001, 2003, and 2005), followed by high populations of H. axyridis in 2002, 2004, and
2006 (Knapp et al. 2012). However, no soybean aphid outbreaks have occurred in Michigan
since 2005 and aphids and H. axyridis have not cycled since then, possibly indicating that a new
equilibrium has been reached in this system. Alternatively, management changes within soybean
fields could have limited the ability of soybean aphid populations to increase unchecked

48

throughout the growing season. Modeling studies indicate that in order to prevent soybean aphid
outbreaks, it is important for natural enemies to consume insipient aphid colonies before they
become well established and produce alates (Costamagna and Landis, 2011). In the past, initial
soybean aphid colonization of Midwestern soy fields occurred early in the season when
coccinellids have had limited time to increase from their overwintering populations. Increased
use of soybean seed treated with systemic insecticide in recent years (Magalhaes et al., 2008)
may delay aphid population growth in soybean fields, giving natural enemy populations time to
increase on alternative hosts and prey before soybean aphids become well-established.
Conclusions
In this study, landscape characteristics and local habitat affected soybean aphid natural
enemies in different ways. At a local scale, coccinellid abundance increased within the managed
habitats. Thus, within the same landscapes they were more commonly captured in flowering
buckwheat strips than in typical field edge habitat such as grassy margins or fencerows. In
contrast, overall coccinellid population levels in equivalent habitats (i.e. paired soybean fields)
were influenced by the abundance of semi-natural habitat at the landscape scale. This makes
sense in light of the large distances over which coccinellid species such as H. axyridis and C.
septempunctata disperse and forage. Furthermore, we found no evidence of an interaction
between local management and landscape diversity or semi-natural habitat in this study.
Coccinellid response in soybean fields to landscape-scale semi-natural habitatwas not affected by
the presence of adjacent buckwheat strips, indicating that for this pest, landscape characteristics
override the effect of enhanced local resources. This suggests that to manage for increased
biocontrol services of soybean aphid will require a focus on manipulating overall landscape
structure rather than local resources.

49

However, as the studies reviewed above suggest, the influences of landscape
characteristics on the relationship between local habitats and natural enemy populations are
likely system-specific and dependent upon the biology of the natural enemies in question.
Soybean aphid is not the only pest of soybean in these landscapes and other crops have
additional pest-natural enemy associations that may be successfully managed at local scales. A
greater understanding of these complex relationships will enable growers and researchers to
develop more effective management systems suited to specific landscapes, prevailing pests, and
their natural enemy communities. Thus, we may anticipate that in the future a combination of
local and landscape management practices may be required to maximize overall pest suppression
in the larger agroecosystem.

50

APPENDIX

51

Chapter I Tables and Figures

52

Table 1.1: Township, county and coordinates from soybean treatment and control fields used in 2008.

Site

County

Cassopolis
Leroy
Vicksburg
Springport
Stockbridge
Mason
Fowlerville
Westphalia
Vermontville
Charlotte
Ithaca
Breckenridge
Saginaw
Lapeer
Richfield
Fostoria
Silverwood

Cass
Calhoun
Kalamazoo
Jackson
Ingham
Ingham
Livingston
Clinton
Eaton
Eaton
Gratiot
Gratiot
Saginaw
Lapeer
Genesee
Tuscola
Lapeer

Treatment field
Longitude
Latitude
85° 58' 26.97"W
85° 12' 57.38"W
85° 31' 54.45"W
84° 38' 30.91"W
84° 11' 57.12"W
84° 25' 46.29"W
84° 5' 58.49"W
84° 47' 46.38"W
84° 58' 45.88"W
84° 53' 49.79"W
84° 37' 10.78"W
84° 30' 11.98"W
84° 6' 55.80"W
83° 18' 50.30"W
83° 29' 38.31"W
83° 23' 9.50"W
83° 13' 3.61"W

41° 59' 58.15"N
42° 10' 2.46"N
42° 6' 46.86"N
42° 25' 5.02"N
42° 32' 21.76"N
42° 32' 31.71"N
42° 44' 56.09"N
42° 52' 21.92"N
42° 38' 29.62"N
42° 32' 12.39"N
43° 11' 32.22"N
43° 21' 13.22"N
43° 22' 50.30"N
42° 58' 54.60"N
43° 5' 29.61"N
43° 14' 59.16"N
43° 17' 16.77"N

53

Control field
Longitude
Latitude
85° 59' 5.05"W
85° 11' 55.4"W
85° 31' 12.52"W
84° 37' 48.15"W
84° 12' 3.01"W
84° 26' 4.51"W
84° 6' 2.66"W
84° 47' 52.92"W
84° 57' 55.92"W
85° 0' 19.9"W
84° 36' 55.46"W
84° 31' 2.51"W
84° 6' 44.83"W
83° 18' 11.23"W
83° 29' 7.29"W
83° 23' 16.99"W
83° 13' 8.32"W

41° 59' 49.89"N
42° 9' 54.72"N
42° 6' 16.43"N
42° 25' 15.87"N
42° 31' 46.35"N
42° 32' 37.2"N
42° 44' 27.05"N
42° 52' 27.37"N
42° 38' 31.12"N
42° 32' 11.78"N
43° 11' 26.03"N
43° 20' 55.71"N
43° 22' 28.23"N
42° 58' 57.74"N
43° 5' 24.35"N
43° 15' 1.01"N
43° 17' 16.34"N

Table 1.2: Township, county and coordinates from soybean treatment and control fields used in 2009.

Site

County

Cassopolis
Leroy
Vicksburg
Springport
Charlotte
Vermontville
Westphalia
Stockbridge
Mason
Fowlerville
Grand Blanc
Oxford 1
Oxford 2
Metamora
Ithaca
Breckenridge
Silverwood

Cass
Calhoun
Kalamazoo
Jackson
Eaton
Eaton
Clinton
Ingham
Ingham
Livingston
Genesee
Oakland
Oakland
Lapeer
Gratiot
Gratiot
Lapeer

Treatment field
Longitude
Latitude
85° 59' 21.84" W
85° 11' 38.15" W
85° 30' 16.07" W
84° 37' 55.12" W
85° 1' 14.52" W
85° 0' 2.16" W
85° 47' 50.40" W
84° 12' 13.94" W
84° 26' 8.93" W
84° 6' 13.69" W
83° 40' 55.56" W
83° 14' 21.84" W
83° 18' 55.80" W
83° 15' 10.80" W
84° 33' 57.23" W
84° 30' 9.72" W
83° 12' 48.54" W

41° 59' 51.72" N
42° 9' 57.91" N
42° 05' 08.15" N
42° 24' 52.63"N
42° 29' 39.82" N
42° 41' 46.31" N
42° 51' 46.02" N
42° 32' 23.25" N
42° 32' 35.13" N
42° 45' 1.93" N
42° 53' 13.55" N
42° 51' 37.80" N
42° 51' 19.07" N
42° 56' 35.16" N
43° 12' 57.23" N
43° 21' 17.64" N
43° 17' 18.21" N

54

Control field
Longitude
Latitude
85° 59' 4.96" W
85° 11' 59.34" W
85° 29' 42.06" W
84° 37' 59.70" W
85° 1' 14.36" W
85° 0' 10.43" W
85° 47' 44.23" W
84° 12' 8.12" W
84° 25' 56.34" W
84° 6' 46.02" W
83° 40' 57.35" W
83° 14' 20.04" W
83° 18' 46.43" W
83° 15' 21.23" W
84° 34' 37.19" W
84° 30' 17.64" W
83° 13' 8.28" W

41° 59' 49.85" N
42° 10' 17.43" N
42° 6' 3.15" N
42° 25' 11.78" N
42° 29' 16.15" N
42° 41' 47.75" N
42° 52' 10.13" N
42° 31' 49.22" N
42° 32' 40.31" N
42° 44' 14.14" N
42° 53' 28.31" N
42° 51' 52.91" N
42° 51' 24.47" N
42° 56' 34.79" N
43° 13' 6.59" N
43° 20' 25.07" N
43° 17' 16.28" N

Table 1.3: Buckwheat strip data for 2008, including date each strip was planted and subsequently reached full bloom. Density
2
(stems/10 cm ) and height (cm) measurements were taken at full bloom.
Site
Cassopolis
Leroy
Vicksburg
Springport
Stockbridge
Mason
Fowlerville
Westphalia
Vermontville
Charlotte
Ithaca
Breckenridge
Saginaw
Lapeer
Richfield
Fostoria
Silverwood

Planting date
2 June
2 June
30 May
30 May
19 May
19 May
19 May
6 June
6 June
28 May
28 May
28 May
28 May
23 May
23 May
23 May
23 May

Full bloom
30 June
7 July
7 July
30 June
1 July
1 July
1 July
no record
no record
no record
3 July
3 July
3 July
26 June
26 June
26 June
8 July

Density
20.5 ± 1.62
19.6 ± 1.58
8.3 ± 1.02
9.4 ± 0.69
16.6 ± 18.7
16 ± 1.58
7.2 ± 0.51
2 ± 0.71
2 ± 0.76
n/a
8.5 ± 0.96
8.9 ± 0.6
8.1 ± 0.55
10.4 ± 0.64
9.4 ± 0.52
14.2 ± 0.99
13.5 ± 1.0

55

Height
95.7 ± 4.57
50.4 ± 8.81
69.6 ± 6.09
86.4 ± 5.51
89.9 ± 6.55
68.4 ± 4.87
128.7 ± 7.05
43.5 ± 13.36
36.7 ± 10.4
n/a
82.9 ± 4.04
108.5 ± 5.58
87.5 ± 5.21
74.6 ± 6.1
107.9 ± 6.38
113.1 ± 4.26
80.9 ± 6.54

Table 1.4: Buckwheat strip data for 2009, including date each strip was planted and subsequently reached full bloom. Density
2
(stems/10 cm ) and height (cm) measurements were taken at full bloom.
Site
Cassopolis
Leroy
Vicksburg
Springport
Charlotte
Vermontville
Westphalia
Stockbridge
Mason
Fowlerville
Grand Blanc
Oxford 1
Oxford 2
Metamora
Ithaca
Breckenridge
Silverwood

Planting Date
16 June
16 June
16 June
10 June
10 June
16 June
3 June
7 July
10 June
7 July
24 June
24 June
24 June
24 June
3 June
3 June
3 June

Full Bloom
20 July
20 July
20 July
15 July
16 July
23 July
9 July
5 Aug
22 July
19 Aug
29 July
10 Aug
28 July
28 July
17 July
6 July
14 July

Density
3.90 ± 0.75
3.60 ± 0.82
3.00 ± 1.13
4.70 ± 0.52
6.40 ± 1.07
5.50 ± 0.81
16.80 ± 1.28
11.00 ± 1.15
15.90 ± 0.84
11.50 ± 0.91
7.10 ± 1.03
6.00 ± 1.12
1.40 ± 0.37
4.30 ± 1.58
6.60 ± 1.10
6.50 ± 0.85
11.30 ± 1.64

56

Height
22.10 ± 2.52
45.30 ± 8.37
57.20 ± 14.37
55.30 ± 6.06
64.80 ± 4.23
76.30 ± 6.80
58.05 ± 5.42
66.30 ± 6.58
67.00 ± 2.27
89.20 ± 2.74
39.20 ± 2.80
71.10 ± 3.06
20.50 ± 6.03
31.50 ± 6.79
81.75 ± 3.46
93.10 ± 3.24
37.40 ± 5.37

Table 1.5: Record of deposition of voucher specimens for Chapter I. The specimens listed below have been deposited in the Albert J.
Cook Arthropod Research Collection, Michigan State University (MSU) as samples of those species or other taxa, which were used in
this research. Voucher recognition labels bearing the voucher number 2013-08 have been attached or included in fluid preserved
specimens.
Family
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Coccinellidae
Aphididae

Genus-Species
Coccinella septempunctata
Harmonia axyridis
Hippodamia variegata
Propylea quatuordecimpunctata
Coleomegilla maculata
Cycloneda munda
Hippodamia parenthesis
Hippodamia glacialis
Hippodamia convergens
Hippodamia tredecimpunctata
Brachiacanthus ursina
Brachiacanthus undulata
Psyllobora vigintimaculata
Aphis glycines

Life Stage
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adult
Adults, nymphs

57

Quantity
3
3
3
3
3
2
2
1
1
1
2
1
1

Preservation
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
70% EtOH

Table 1.6: Dates soybean aphids were first detected in each treatment and control field, and dates cage study was conducted in each
site in 2008.

Site
Cassopolis
Leroy
Vicksburg
Springport
Stockbridge
Mason
Fowlerville
Westphalia
Vermontville
Charlotte
Ithaca
Breckenridge
Saginaw
Lapeer
Richfield
Fostoria
Silverwood

Aphid detection
treatment
16 June
7 July
16 June
3 June
3 June
3 June
11 June
4 June
4 June
12 June
4 June
18 June
11 June
10 June
2 June
9 June
17 July

Aphid detection
control
30 June
30 June
3 June
3 June
3 June
16 June
10 July
4 June
4 June
6 June
4 June
18 June
11 June
10 June
2 June
9 June
8 August

Cage study
14-28 July
16-30 July
11-25 July
n/a
18 July-1 Aug
10-24 July
23 July - 6 Aug
n/a
11-25 July
n/a
13-27 July
7-21 July
8-22 July
15-29 July
9-23 July
17-31 July
17-31 July

58

Table 1.7: Dates soybean aphids were first detected in each treatment and control field, and date cage study was conducted in each site
in 2009.

Site
Cassopolis
Leroy
Vicksburg
Springport
Charlotte
Vermontville
Westphalia
Stockbridge
Mason
Fowlerville
Grand Blanc
Oxford 1
Oxford 2
Metamora
Ithaca
Breckenridge
Silverwood

Aphid detection
treatment
30 June
16 June
23 June
8 July
30 July
12 June
16 July
11 June
1 July
11 June
8 July
22 June
15 June
11 June
17 July
24 July
14 July

Aphid detection
control
30 June
16 June
6 July
1 July
23 July
24 June
22 June
11 June
16 June
25 June
29 June
11 June
22 June
14 July
9 June
17 July
7 July

Cage study
20 July - 3 Aug
20 July - 3 Aug
20 July - 3 Aug
15-29 July
16-30 July
23 July-6 Aug
9-23 July
n/a
15-29 July
n/a
12-26 Aug
28 July -10 Aug
10-24 Aug
10-24 Aug
10-24 July
10-24 July
14-28 July

59

Table 1.8: A) Crop and B) non-crop land covers found in landscapes surrounding study fields. All land covers found were used to
calculated Simpson’s diversity metric (-lnD). The categories of grassland and forest land covers were subsequently combined into
“semi-natural habitat” for landscape analyses.
A) Crop land covers

B) Non-crop land covers

Major annual crops
Soy
Corn
Wheat

Grasslands
Remnant prairie
Old Field
CRP lands

Forage crops
Alfalfa
Pasture
Orchard grass

Forest
Forest
Woodlots
Hedgerows/ fencerows

Less common annual crops
Dry beans
Sugar beats
Fruit
Unidentified small grains

Residential
Urban
Residential/ suburban
Mowed turf (golf course, cemetery)
Less common non-crop covers
Unplanted, tilled field
Gravel quarry
Open Water
Wetland

60

Table 1.9: Simpson’s diversity metric and proportion of each major land cover type surrounding study fields at the 2 km scale in 2008.
Site
Breckenridge
Cassopolis
Leroy
Lapeer
Westphalia
Vermontville
Fostoria
Springport
Saginaw
Vicksburg
Ithaca
Stockbridge
Charlotte
Richfield
Fowlerville
Silverwood
Mason

-lnD

Soy

1.189
1.272
1.376
1.403
1.536
1.563
1.609
1.615
1.623
1.695
1.719
1.722
1.740
1.769
1.816
1.818
1.911

0.24
0.15
0.07
0.05
0.28
0.25
0.28
0.16
0.21
0.05
0.20
0.22
0.22
0.22
0.16
0.26
0.15

Corn
Wheat
Forest
Grasslands
Landscape surrounding treatment field
0.47
0.13
0.02
0.02
0.45
0.00
0.22
0.13
0.08
0.07
0.46
0.16
0.02
0.05
0.43
0.30
0.33
0.08
0.10
0.17
0.16
0.00
0.31
0.22
0.10
0.05
0.30
0.13
0.24
0.05
0.32
0.13
0.14
0.04
0.34
0.09
0.10
0.02
0.29
0.11
0.30
0.14
0.12
0.19
0.22
0.13
0.23
0.12
0.19
0.08
0.27
0.15
0.07
0.02
0.26
0.18
0.26
0.02
0.21
0.23
0.13
0.21
0.13
0.21
0.18
0.04
0.23
0.19

61

Residential

Semi-natural

0.02
0.05
0.13
0.12
0.05
0.05
0.11
0.05
0.09
0.32
0.06
0.05
0.06
0.20
0.09
0.04
0.17

0.04
0.35
0.62
0.73
0.26
0.53
0.43
0.45
0.42
0.40
0.31
0.35
0.42
0.44
0.44
0.34
0.42

Table 1.9 (cont’d).
Site
Breckenridge
Leroy
Lapeer
Cassopolis
Westphalia
Vermontville
Springport
Richfield
Ithaca
Fostoria
Saginaw
Charlotte
Vicksburg
Stockbridge
Fowlerville
Silverwood
Mason

-lnD

Soy

1.142
1.170
1.210
1.308
1.515
1.569
1.597
1.609
1.614
1.673
1.674
1.726
1.765
1.772
1.794
1.889
1.889

0.27
0.06
0.06
0.17
0.27
0.19
0.21
0.16
0.19
0.26
0.21
0.23
0.08
0.20
0.16
0.23
0.14

Corn
Wheat
Forest
Grasslands
Landscape surrounding control field
0.47
0.08
0.00
0.03
0.09
0.06
0.52
0.08
0.02
0.01
0.49
0.29
0.43
0.00
0.21
0.14
0.34
0.06
0.09
0.18
0.10
0.01
0.36
0.25
0.23
0.06
0.31
0.08
0.07
0.03
0.36
0.17
0.36
0.10
0.09
0.17
0.10
0.06
0.29
0.16
0.23
0.04
0.26
0.11
0.18
0.08
0.28
0.14
0.18
0.03
0.27
0.10
0.25
0.12
0.20
0.15
0.28
0.04
0.18
0.25
0.17
0.18
0.14
0.23
0.17
0.04
0.24
0.20

62

Residential

Semi-natural

0.02
0.16
0.11
0.05
0.05
0.05
0.04
0.16
0.06
0.11
0.08
0.05
0.25
0.05
0.08
0.04
0.20

0.03
0.60
0.79
0.35
0.27
0.61
0.39
0.54
0.27
0.45
0.37
0.42
0.37
0.35
0.42
0.38
0.43

Table 1.10: Simpson’s diversity metric and proportion of each major land cover type surrounding study fields at the 2 km scale in
2009.
Site

-lnD

Soy

Metamora
Cassopolis
Leroy
Breckenridge
Grand Blanc
Vermontville
Silverwood
Vicksburg
Charlotte
Stockbridge
Oxford Liptons
Ithaca
Springport
Westphalia
Oxford
Mason
Fowlerville

0.828
1.196
1.240
1.311
1.391
1.481
1.499
1.574
1.575
1.610
1.675
1.725
1.753
1.765
1.783
1.920
1.951

0.04
0.14
0.16
0.38
0.05
0.22
0.40
0.21
0.20
0.22
0.05
0.22
0.14
0.19
0.21
0.15
0.19

Corn
Wheat
Forest
Grasslands
Landscape surrounding treatment fields
0.01
0.00
0.65
0.11
0.47
0.02
0.24
0.07
0.02
0.03
0.49
0.10
0.28
0.07
0.01
0.04
0.07
0.00
0.38
0.17
0.33
0.01
0.25
0.09
0.17
0.06
0.14
0.13
0.26
0.03
0.29
0.05
0.13
0.08
0.36
0.16
0.28
0.18
0.18
0.06
0.02
0.01
0.35
0.21
0.30
0.12
0.12
0.10
0.19
0.06
0.30
0.05
0.29
0.09
0.17
0.17
0.02
0.00
0.30
0.29
0.16
0.06
0.25
0.13
0.13
0.05
0.24
0.24

63

Residential

Semi-natural

0.08
0.05
0.15
0.02
0.27
0.06
0.04
0.05
0.04
0.05
0.20
0.10
0.05
0.05
0.09
0.24
0.10

0.76
0.32
0.59
0.05
0.56
0.34
0.27
0.35
0.52
0.24
0.56
0.22
0.35
0.34
0.59
0.38
0.49

Table 1.10 (cont’d).
Site

-lnD

Soy

Metamora
Cassopolis
Grand Blanc
Leroy
Breckenridge
Vermontville
Silverwood
Vicksburg
Westphalia
Charlotte
Oxford Liptons
Oxford
Springport
Stockbridge
Ithaca
Fowlerville
Mason

0.835
1.204
1.292
1.377
1.396
1.486
1.524
1.532
1.647
1.655
1.661
1.668
1.707
1.711
1.802
1.855
1.923

0.06
0.13
0.05
0.18
0.31
0.24
0.40
0.27
0.21
0.19
0.05
0.20
0.17
0.20
0.24
0.15
0.15

Corn
Wheat
Forest
Grasslands
Landscape surrounding control fields
0.00
0.00
0.65
0.11
0.48
0.02
0.22
0.10
0.05
0.00
0.40
0.18
0.03
0.03
0.43
0.13
0.35
0.13
0.02
0.05
0.31
0.01
0.26
0.09
0.15
0.05
0.15
0.13
0.22
0.00
0.30
0.06
0.34
0.10
0.11
0.16
0.12
0.12
0.34
0.15
0.02
0.01
0.35
0.17
0.02
0.00
0.34
0.24
0.20
0.05
0.31
0.05
0.27
0.16
0.19
0.11
0.24
0.13
0.14
0.13
0.21
0.06
0.26
0.22
0.14
0.07
0.26
0.13

64

Residential

Semi-natural

0.08
0.05
0.29
0.15
0.02
0.06
0.04
0.08
0.06
0.04
0.19
0.10
0.05
0.05
0.10
0.08
0.22

0.76
0.32
0.58
0.57
0.07
0.34
0.28
0.36
0.28
0.49
0.53
0.59
0.36
0.30
0.26
0.48
0.38

Table 1.11: Effects of landscape and local habitat management on coccinellid abundance in 2008 soybean fields, 2009 soybean fields,
and 2009 field margins. Separate analyses were run with A) Simpson’s diversity index –lnD, and with B) proportion of semi-natural
habitat in the landscape as the landscape descriptor variables. Results are shown for landscape values at the 2.0 km scale, but were
consistent at all scales
A
2008 Soy Fields

B
F value

pr > F

F value

pr > F

Buckwheat treatment

F1,24=2.72

0.112

Buckwheat treatment

F1,24=0.59

0.451

Diversity (-lnD)

F1,24=1.15

0.295

semi-natural habitat

F1,24=0.07

0.798

treatment*-lnD

F1,24=2.18

0.153

treatment*semi-natural

F1,24=1.81

0.191

F value

pr > F

F value

pr > F

Buckwheat treatment

F1,30=0.19

0.662

Buckwheat treatment

F1,30=0.47

0.496

Diversity (-lnD)

F1,30=0.14

0.713

semi-natural habitat

F1,30=6.11

0.019

treatment*-lnD

F1,30=0.31

0.580

treatment*semi-natural

F1,30=0.19

0.667

F value

pr > F

F value

pr > F

Buckwheat treatment

F1,30=10.58

0.003

Buckwheat treatment

F1,30=10.78

0.003

Diversity (-lnD)

F1,30=0.00

0.889

semi-natural habitat

F1,30=0.25

0.618

treatment*-lnD

F1,30=0.03

0.834

treatment*semi-natural

F1,30=0.40

0.532

2009 Soy Fields

2009 Field Margins

65

Table 1.12: Mean ± S.E. of coccinellid beetles of each species found in soybean fields or field margins during 2008. No comparisons
between treatment and control fields were significant in this year.

Coccinellids

Buckwheat
strip

Treatment
field

Control field

0.32 ±0.04
0.87 ±0.10
0.007 ±0.005

0.64 ±0.05
0.41 ±0.03
0.02 ±0.01

0.52 ±0.04
0.35 ±0.03
0.02 ±0.01

0.24 ±0.03

0.15 ±0.02

0.10 ±0.01

0.11 ±0.03

0.07 ±0.01

0.07 ±0.01

EXOTIC SPECIES
Harmonia axyridis
Coccinella septempunctata
Hippodamia variegata
Propylea
quatuordecimpunctata
NATIVE SPECIES
Coleomegilla maculata

Additionally, the native species Hippodamia convergens, Hi. parenthesis,
Brachiacantha ursina, Cycloneda munda, Hyperaspis binotata, and Adalia
bipunctata were all collected at mean abundances of less than 0.02 in each
habitat in 2008.

66

Table 1.13: Mean ± SE of coccinellid beetle species found in each field or field margin during 2009. Individual counts were rank
transformed and compared using ANOVA. Asterisks (*) indicate a significant difference in abundance between buckwheat strips and
control field margins, or between treatment and control fields (*** p < 0.01, ** p < 0.05, * p < 0.1).
Coccinellids

Buckwheat
strip

Control field
margin

Treatment
field

Control
field

0.31 ±0.03
1.07 ±0.09
0.96 ±0.13
0.78 ±0.08

0.27 ±0.03
0.38 ±0.05
0.04 ±0.01
0.54 ±0.05

***
***
*

1.07 ±0.08
1.36 ±0.08
0.50 ±0.05
0.71 ±0.06

1.09 ±0.07
1.18 ±0.07
0.21 ±0.02
0.71 ±0.05

0.35 ±0.06

0.06 ±0.01

***

0.24 ±0.03

0.16 ±0.02

EXOTIC SPECIES
Harmonia axyridis
Coccinella septempunctata
Hippodamia variegata
Propylea quatuordecimpunctata
NATIVE SPECIES
Coleomegilla maculata

*

Additionally, the native species Hippodamia parenthesis, Hi. convergens, Hi. tredecimpunctata, Hi. glacialis,
Brachiacantha ursina, Cycloneda munda, Hyperaspis undulata, Hy. binotata, Hy. proba, Hy. bigeminata,
and Psyllobora vigintimaculata were all collected at mean abundances less than 0.16 in each habitat 2009.

67

Table 1.14: Biocontrol Services Index values for aphid suppression by natural enemies for each soybean field in 2008 and 2009.
2008 BSI
Site
Breckenridge
Cassopolis
Fowlerville
Ithaca
Lapeer
Leroy
Mason
Richfield
Saginaw
Stockbridge
Vermontville
Vicksburg

Average

Treatment
0.993
0.901
0.948
1.000
0.990
0.972
0.932
0.992
0.976
0.776

2009 BSI

0.922

Control
0.973
0.943
0.995
1.000
0.959
0.983
0.965
0.958
0.941
0.979
0.924
0.918

0.942 ±0.002

Site
Breckenridge
Cassopolis
Charlotte
Grand Blanc
Ithaca
Mason
Oxford
Silverwood
Springport
Vermontville
Vicksburg
Westphalia

0.959 ±0.008

0.663
0.974
0.985
0.991
0.990
0.995
0.992
0.943
0.997

Control
0.983
0.902
0.968
0.830
0.927
0.970
0.998
0.997
0.993
0.983
0.992
0.968

0.937 ±0.031

68

Treatment
0.846
0.991

0.968 ±0.008

Figure 1.1: Site locations of paired soybean fields across southern Michigan sampled for aphids
and natural enemies in 2008 and 2009. Sites on the same or nearby farms from year to year
appear to overlap at this scale (crossed circles), but samples were collected in a different set of
fields in each year. For interpretation of the references to color in this and all other figures, the
reader is referred to the electronic version of this dissertation.

69

Figure 1.2: Schematic of sampling design showing relative locations of plots and buckwheat strip
or control field margin. Inset: a yellow sticky trap and a pair of caged and uncaged soybean
plants for the biocontrol experiment were located in the center of each plot.

70

Figure 1.3: An example buckwheat strip in full bloom in July 2009.

71

Figure 1.4: A) Coccinellid abundance in field margins was consistently and significantly higher in buckwheat strips than in control
field margins across a gradient of semi-natural habitat within landscapes. B) Coccinellid abundance in soybean fields in 2009
increased significantly with the proportion of seminatural habitat in the surrounding landscape at 2.0 km. The effect did not differ
between fields adjacent to or without buckwheat strips.

B

Buckwheat Strip
Control Field Margin
14

Treatment soybean field
Control soybean field
10

Mean Coccinellid Abundance in Soy Fields

Mean Coccinellid Abundance in Field Margins

A

12
10
8
6
4
2
0

8

6

4

2

0
0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.0

Proportion of Semi-Natural Habitat

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Proportion of Semi-Natural Habitat

72

0.8

Figure 1.5: Average number of aphids per soybean plant exposed to and excluding predators in
treatment and control fields, at 7 and 14 days after infestation in A) 2008 and B) 2009. The
horizontal dashed line indicates the economic threshold for soybean aphid of 250 aphids/ plant.
In both years, natural enemies dramatically suppressed soybean aphid populations, and the
differences in aphid populations between caged and uncaged plants were greater at 14 days than
at 7 days. However, buckwheat treatment did not significantly affect the size of the difference in
aphid populations on caged and uncaged plants in either year.
A

B
Treatment caged
Treatment open
Control caged
Control open

800

800

600
Aphids/ Plant

Aphids/ Plant

600

400

400

200

200

0

0
0 days

7 days

14 days

73

0 days

7 days

14 days

LITERATURE CITED

74

LITERATURE CITED

Altieri, M.A., 1999. The ecological role of biodiversity in agroecosystems. Agric. Ecosyst.
Environ. 74, 19-31.
Batáry, P., Andras, B., Kleijn, D., Tscharntke, T., 2011. Landscape-moderated biodiversity
effects of agri-environmental management: a meta-analysis. P. R. Soc. B. 278, 18941902.
Bianchi, F.J.J.A., Booij, C.J.H., Tscharntke, T., 2006. Sustainable pest regulation in agricultural
landscapes: a review on landscape composition, biodiversity and natural pest control. P.
R. Soc. B. 273, 1715-1727.
Bigger, D.S., Chaney, W.E., 1998. Effects of Iberis umbellata (Brassicaceae) on insect pests of
cabbage and on potential biological control agents. Environ. Entomol. 27, 161-167.
Carvell, C., Osborne, J.L., Bourke, A.F.G., Freeman, S.N., Pywell, R.F., Heard, M.S., 2011.
Bumble bee species' responses to a targeted conservation measure depend on landscape
context and habitat quality. Ecol. Appl. 21, 1760-1771.
Chaplin-Kramer, R., O'Rourke, M.E., Blitzer, E.J., Kremen, C., 2011. A meta-analysis of crop
pest and natural enemy response to landscape complexity. Ecol. Lett. 14, 922-932.
Concepcion, E.D., Diaz, M., Baquero, R.A., 2008. Effects of landscape complexity on the
ecological effectiveness of agri-environment schemes. Landsc. Ecol. 23, 135-148.
Costamagna, A.C., Landis, D.A., 2007. Quantifying predation on soybean aphid through direct
field observations. Biol. Control 42, 16-24.
Costamagna, A.C., Landis, D.A., 2011. Lack of strong refuges allows top-down control of
soybean aphid by generalist natural enemies. Biol. Control 57, 184-192.
Drapela, T., Moser, D., Zaller, J.G., Frank, T., 2008. Spider assemblages in winter oilseed rape
affected by landscape and site factors. Ecography 31, 254-262.
Elton, C.S., 1958. Ecology of Invasions by Animals and Plants. Chapman & Hall, London.
Fiedler, A.K., Landis, D.A., Wratten, S.D., 2008. Maximizing ecosystem services from
conservation biological control: The role of habitat management. Biol. Control 45, 254271.
Gardiner, M.M., Landis, D.A., Gratton, C., DiFonzo, C.D., O'Neal, M., Chacon, J.M., Wayo,

75

M.T., Schmidt, N.P., Mueller, E.E., Heimpel, G.E., 2009. Landscape diversity enhances
biological control of an introduced crop pest in the north-central USA. Ecol. Appl. 19,
143-154.
Gardiner, M.M., Landis, D.A., Gratton, C., Schmidt, N., O'Neal, M., Mueller, E., Chacon, J.,
Heimpel, G.E., 2010. Landscape composition influences the activity density of Carabidae
and Arachnida in soybean fields. Biol. Control 55, 11-19.
Geiger, F., de Snoo, G.R., Berendse, F., Guerrero, I., Morales, M.B., Onate, J.J., Eggers, S., Part,
T., Bommarco, R., Bengtsson, J., Clement, L.W., Weisser, W.W., Olszewski, A.,
Ceryngier, P., Hawro, V., Inchausti, P., Fischer, C., Flohre, A., Thies, C., Tscharntke, T.,
2010. Landscape composition influences farm management effects on farmland birds in
winter: A pan-European approach. Agric. Ecosyst. Environ. 139, 571-577.
Haenke, S., Scheid, B., Schaefer, M., Tscharntke, T., Thies, C., 2009. Increasing syrphid fly
diversity and density in sown flower strips within simple vs. complex landscapes. JAppl.
Ecol. 46, 1106-1114.
Harmon, J.P., Ives, A.R., Losey, J.E., Olson, A.C., Rauwald, K.S., 2000. Colemegilla maculata
(Coleoptera: Coccinellidae) predation on pea aphids promoted by proximity to
dandelions. Oecologia 125, 543-548.
Heimpel, G.E., Frelich, L.E., Landis, D.A., Hopper, K.R., Hoelmer, K.A., Sezen, Z., Asplen,
M.K., Wu, K.M., 2010. European buckthorn and Asian soybean aphid as components of
an extensive invasional meltdown in North America. Biol. Invasions 12, 2913-2931.
Hickman, J.M., Wratten, S.D., 1996. Use of Phacelia tanacetifolia strips to enhance biological
control of aphids by hoverfly larvae in cereal fields. J. Econ. Entomol. 89, 832-840.
Higley, L.G., Boethel, D.J. (Eds.), 1994. Handbook of Soybean Insect Pests. Entomological
Society of America, Lanham, MD.
Hodek, I., Honek, A., 1996. Ecology of Coccinellidae. Kluwer Academic Publishers, Dordrecht,
The Netherlands.
Hodek, I., Iperti, G., Hodkova, M., 1993. Long distance flights in Coccinellidae (Coleoptera).
Eur. J. Entomol. 90, 403-414.
Holzschuh, A., Steffan-Dewenter, I., Kleijn, D., Tscharntke, T., 2007. Diversity of flowervisiting bees in cereal fields: effects of farming system, landscape composition and
regional context. J. Appl. Ecol. 44, 41-49.
Knapp, A.K., Smith, M.D., Hobbie, S.E., Collins, S.L., Fahey, T.J., Hansen, G., Landis, D.A.,
Melillo, J., Seastadt, T.R., Shaver, G.R., Webster, J.R., 2012. Long-term experiments in
the LTER network: past, present and future roles. Bioscience, in press.

76

Koji, S., Khan, Z.R., Midega, C.A.O., 2007. Field boundaries of Panicum maximum as a
reservoir for predators and a sink for Chilo partellus. J. Appl. Entomol. 131, 186-196.
Kremen, C., 2005. Managing ecosystem services: what do we need to know about their ecology?
Ecol. Lett. 8, 468-479.
Landis, D.A., Wratten, S.D., Gurr, G.M., 2000. Habitat management to conserve natural enemies
of arthropod pests in agriculture. Annu. Rev. Entomol. 45, 175-201.
Langer, A., Hance, T., 2004. Enhancing parasitism of wheat aphids through apparent
competition: a tool for biological control. Agric. Ecosyst. Environ. 102, 205-212.
Lee, J.C., Heimpel, G.E., 2005. Impact of flowering buckwheat on Lepidopteran cabbage pests
and their parasitoids at two spatial scales. Biol. Control 34, 290-301.
Losey, J.E., Vaughan, M., 2006. The economic value of ecological services provided by insects.
Bioscience 56, 311-323.
Magalhaes, L.C., Hunt, T.E., Siegfried, B.D., 2008. Development of methods to evaluate
susceptibility of soybean aphid to imiacloprid and thiamethoxam at lethal and sublethal
concentrations. Entomol. Exp. Appl. 128, 330-336.
Magurran, A.E., 2004. Measuring Biological Diversity. Blackwell Science Ltd, Malden, MA.
Marino, P.C., Landis, D.A., 1996. Effect of landscape structure on parasitoid diversity and
parasitism in agroecosystems. Ecol. Appl. 6, 276-284.
Meehan, T.D., Werling, B.P., Landis, D.A., Gratton, C., 2011. Agricultural landscape
simplification and insecticide use in the Midwestern U.S. Proc. Natl. Acad. Sci.
www.pnas.org/cgi/doi/10.1073/pnas.1100751108.
Purtauf, T., Roschewitz, I., Dauber, J., Thies, C., Tscharntke, T., Wolters, V., 2005. Landscape
context of organic and conventional farms: Influences on carabid beetle diversity. Agric.
Ecosyst. Environ. 108, 165-174.
Ragsdale, D.W., McCornack, B.P., Venette, R.C., Potter, B.D., Macrae, I.V., Hodgson, E.W.,
O'Neal, M.E., Johnson, K.D., O'Neil, R.J., DiFonzo, C.D., Hunt, T.E., Glogoza, P.A.,
Cullen, E.M., 2007. Economic threshold for soybean aphid (Hemiptera : Aphididae). J.
Econ. Entomol. 100, 1258-1267.
Roschewitz, I., Gabriel, D., Tscharntke, T., Thies, C., 2005a. The effects of landscape
complexity on arable weed species diversity in organic and conventional farming. J.
Appl. Ecol. 42, 873-882.
Roschewitz, I., Hucker, M., Tscharntke, T., Thies, C., 2005b. The influence of landscape context

77

and farming practices on parasitism of cereal aphids. Agric. Ecosyst. Environ. 108, 218227.
Rundlof, M., Nilsson, H., Smith, H.G., 2008. Interacting effects of farming practice and
landscape context on bumblebees. Biol. Conserv. 141, 417-426.
Rundlof, M., Smith, H.G., 2006. The effect of organic farming on butterfly diversity depends on
landscape context. J. Appl. Ecol. 43, 1121-1127.
Schlapfer, F., Schmid, B., Seidl, L., 1999. Expert estimates about effects of biodiversity on
ecosystem processes and services. Oikos 84, 346-352.
Schmidt, M.H., Roschewitz, I., Thies, C., Tscharntke, T., 2005. Differential effects of landscape
and management on diversity and density of ground-dwelling farmland spiders. J. Appl.
Ecol. 42, 281-287.
Schmidt, M.H., Thies, C., Nentwig, W., Tscharntke, T., 2008. Contrasting responses of arable
spiders to the landscape matrix at different spatial scales. J. Biogeogr. 35, 157-166.
Simpson, E.H., 1949. Measurement of diversity. Nature 163, 688-688.
Smith, H.G., Danhardt, J., Lindstrom, A., Rundlof, M., 2010. Consequences of organic farming
and landscape heterogeneity for species richness and abundance of farmland birds.
Oecologia 162, 1071-1079.
Steffan-Dewenter, I., Munzenberg, U., Burger, C., Thies, C., Tscharntke, T., 2002. Scale
dependent effects of landscape context on three pollinator guilds. Ecology 83, 1421-1432.
Thies, C., Steffan-Dewenter, I., Tscharntke, T., 2003. Effects of landscape context on herbivory
and parasitism at different spatial scales. Oikos 101, 18-25.
Thies, C., Tscharntke, T., 1999. Landscape structure and biological control in agroecosystems.
Science 285, 893-895.
Tilman, D., Cassman, K.G., Matson, P.A., Naylor, R., Polasky, S., 2002. Agricultural
sustainability and intensive production practices. Nature 418, 671-677.
Tscharntke, T., Klein, A.M., Kruess, A., Steffan-Dewenter, I., Thies, C., 2005. Landscape
perspectives on agricultural intensification and biodiversity - ecosystem service
management. Ecol. Lett. 8, 857-874.
Tscharntke, T., Tylianaskis, J., Rand, T., Didham, R., Fahrig, L., Batary, P., Bengtsson, J.,
Clough, Y., Crist, T., Dormann, C., Ewers, R., Holt, R., Holzschuh, A., Klein, A., Kleijn,
D., Kremen, C., Landis, D., Laurance, B., Lindenmayer, D., Scherber, C., Sodhi, N.,
Steffan-Dewenter, I., Thies, C., van der Putten, W., Westphal, C. 2012. Landscape

78

moderation of biodiversity patterns and processes – eight hypotheses. Biol. Reviews, 87,
661-685.
Walton, N.J., Isaacs, R., 2011. Influence of native flowering plant strips on natural enemies and
herbivores in adjacent blueberry fields. Environ. Entomol. 40, 697-705.
Winqvist, C., Bengtsson, J., Aavik, T., Berendse, F., Clement, L.W., Eggers, S., Fischer, C.,
Flohre, A., Geiger, F., Liira, J., Part, T., Thies, C., Tscharntke, T., Weisser, W.W.,
Bommarco, R., 2011. Mixed effects of organic farming and landscape complexity on
farmland biodiversity and biological control potential across Europe. J. Appl. Ecol. 48,
570-579.

79

CHAPTER II: Coccinellid response to landscape structure across an agricultural intensification
gradient.
Abstract
Arthropod natural enemies provide valuable pest suppression in agricultural landscapes,
and numerous studies have demonstrated the effects of landscape structure on natural enemy
abundance and diversity. In these studies, the amount of semi-natural habitat present in a
landscape is frequently used as a proxy for all other aspects of landscape structure. Coccinellid
beetles are important generalist predators in a variety of agricultural landscapes. While
coccinellids have also been found to be related to the amount of semi-natural habitat in the
surrounding landscape, other aspects of landscape structure may be more amenable to
management. My goal is to simultaneously assess the effects of landscape composition,
including measures of land cover and habitat diversity, and landscape configuration on
coccinellid abundances in soybean fields. Here I use Partial Least Squares regression to examine
coccinellid response to aspects of landscape structure in more detail. My analyses show that
coccinellid abundances were lowest in landscapes dominated by annual crops grown in large
fields and highest in landscapes dominated by semi-natural habitat, with a higher diversity of
crops grown in smaller fields. My findings suggest a need for studies to separate out potential
contributions of these different aspects of landscape structure, especially as some (field size and
crop diversity) are more easily managed by growers than semi-natural habitat.
Introduction
Agricultural landscapes cover at least one quarter of terrestrial surfaces on our planet
(Millennium Ecosystem Assessment, 2005) and are critical for providing food and fiber for our
growing global population. In addition to the production of food and fiber, agricultural
landscapes can be managed to produce a variety of ecosystem services, including support of
80

biodiversity and the regulation of water quality, soil fertility, and greenhouse gases, (Power,
2010; Swinton et al., 2007) and to minimize disservices like soil and nutrient runoff and
exposure to pesticides (Power, 2010). Changes in landscape structure influence the provision of
these services and disservices. For example, increases in proportional cover of cropland in
agricultural landscapes are related to increased pest pressure and insecticide use (Meehan et al.,
2011).
Agricultural landscapes also benefit from ecosystem services like pollination and natural
pest control (Power, 2010). In the U.S., native and naturalized natural enemies provide pest
suppression services to agricultural landscapes worth an estimated $4.6 billion annually (Losey
and Vaughan, 2006). The ultimate success of pest suppression is dependent upon the provision of
natural enemies to crop fields by the surrounding landscape. Landscape structure affects the
composition of insect communities and the abundance of individual species, which can in turn
lead to higher biocontrol rates and less crop damage in more complex landscapes (Thies and
Tscharntke, 1999) and in landscapes with greater proportions of non-crop habitat (Thies et al.,
2003).
A number of different metrics have been proposed to quantify landscape structure,
including composition metrics that describe which habitats are present and their relative
abundances, and configuration metrics that indicate how those habitats are spatially arranged
(Turner et al., 2001). Numerous studies have examined the influence of landscape structure on
natural enemy communities, yet most of these have focused strictly on landscape composition –
specifically, the amount of non-crop or semi-natural habitat (Burel et al., 2013). Semi-natural
habitats can provide natural enemies with food resources and shelter during disturbances in crop
fields (Bianchi et al., 2006), leading to higher natural enemy abundance (Colunga-Garcia et al.,

81

1997; Gardiner et al., 2009), species richness (Schmidt et al., 2005) and fecundity (Bommarco,
1998). However, other types of land covers could also influence natural enemy communities. For
example, perennial crops may provide many of the same benefits as semi-natural habitat,
including reduced disturbance and alternative prey (Landis et al., 2000). Although studies
examining perennial crops have been few in number, landscapes which contain more perennial
crops have been found to have higher natural enemy activity density (Oberg et al., 2007) and
diversity (Elliott et al., 1999), and lower pest rates (Östman et al., 2001).
Habitat diversity, another metric of landscape composition (Turner et al. 2001,
McGarigal et al. 2012) may also be important for naturally occurring biological control. Multiple
land covers in a local area may provide different resources needed by a single organism
(Dunning et al., 1992), which can in turn influence communities in agricultural landscapes
(Fahrig et al., 2011). For example, a predatory insect may require one type of habitat for
overwintering and another for foraging and egg-laying. As such, habitat diversity within the
landscape has also been positively related to natural enemy abundance (Isaia et al., 2006) and to
herbivore suppression (Gardiner et al., 2009; Kruess, 2003). Similarly, crop diversity may
provide many of the same benefits as overall habitat diversity, as different crops with different
phenologies may provide complementary resources at different times of the year (Vasseur et al.,
2013).
The influence of landscape configuration, the spatial arrangement of habitats within a
landscape, on natural enemies has been largely overlooked in this body of work (Burel et al.
2013). Yet the configuration of habitats can affect the rate of arrival of natural enemies into crop
fields (Bianchi et al., 2006), which can in turn influence natural enemy efficacy (Hogg and
Daane, 2010). In particular, modeling has shown that an even distribution of non-crop habitats

82

within a landscape should facilitate colonization of fields by natural enemies (Bianchi and van
der Werf, 2003), suggesting that landscapes with small crop fields and greater dispersion of noncrop habitat could promote higher natural enemy abundances in crop fields. Because farmers
may have more ability to manipulate some aspects of landscape structure than others,
understanding the relationship between these different elements of landscape structure and
natural enemy communities has important management implications (Fahrig et al., 2011).
Ladybird beetles (Coleoptera: Coccinellidae) are generalist predators in a variety of
agricultural landscapes (Colunga-Garcia et. al, 1997; Obrycki and Kring, 1998). They make a
good model taxon to study landscape structure effects because they are known to respond to
landscape composition (Gardiner et al., 2009; Woltz et al., 2012), and theoretical models suggest
that they should also respond to configuration (Bianchi and van der Werf, 2003). I examined the
coccinellid community of soybean fields as part of a separate study examining their efficacy in
suppressing soybean aphid, Aphis glycines Matsumura, in agricultural landscapes across a
gradient of semi-natural habitat (Woltz et al., 2012). Previous work indicates that coccinellids are
of primary importance for suppression of soybean aphid (Costamagna and Landis, 2007;
Costamagna et al., 2007), the most important insect pest of soybean in the North Central US
(Ragsdale et al., 2011). In this study, I assessed the effects of both landscape composition and
configuration on coccinellid abundances in soybean fields. I hypothesize that coccinellids will be
positively correlated with a) non-crop habitat, b) perennial crops, c) crop diversity, and d) degree
of intermixing of crop and non-crop habitat, and negatively correlated with e) annual crop
habitats and f) field size. Understanding the relationship between these measures of landscape
structure and generalist predators has implications for management of agricultural landscapes
globally.

83

Methods
Site design. This study was replicated in 17 sites across a gradient of landscape diversity
in southern Michigan, USA during the summer of 2009 (see Table A.2 in Woltz et al. 2012 for
field coordinates). Sites ranged from 6 to 46 km apart, with an average nearest neighbor distance
of 23 km. Agriculture in this region consists of a mix of annual field crops (maize, soybean, and
wheat) and perennial forage crops (pasture and perennial mixes for hay). Deciduous forests and
grasslands or old fields make up the primary non-crop land cover in the area.
Coccinellid counts. Aerially dispersing coccinellids were collected in soybean fields
using 22.9 x 27.9 cm unbaited yellow sticky traps (PHEROCON AM, Great Lakes IPM,
Vestaburg, MI). Four traps were used in each field, placed 30 m apart from one another and 25 m
from a field edge. Field edges were chosen based on ease of access from roads and were
distributed among cardinal directions (North n=4; East n=4; South n=7; West n=3 fields). Traps
were suspended from step-in plastic fence posts (Woodstream Corporate, Lititz, PA) at plant
canopy height and were replaced every week for 8-11 weeks, from 8 June to 26 Aug. All
coccinellids captured were identified to species in the field. The number of coccinellids captured
per trap per week were used to calculate mean number of coccinellid captures per field.
Landscape data. Ortho-rectified digital aerial photos for each field were obtained from
the Michigan Spatial Data Library based on GPS coordinates collected from the center of each
field using a handheld GPS receiver. Using these photographs, habitats within a 2.0 km radius
around each field were digitized as polygons using ARC GIS 9.3 (ESRI Redlands, CA). For
linear patches (e.g. fencerows, roadside vegetation), patches 20 m wide or greater were digitized
as distinct habitats. Land cover types and habitat boundaries within field sites were groundverified, and changes from the aerial photos were noted and corrected during digitization. The

84

specific land cover within all habitat polygons in each landscape was determined by manually
assigning land cover categories to each polygon. The digitized polygon land cover layers for
each site were converted to raster format in ARC GIS. The exported rasters were input into
Fragstats 4.1 (McGarigal et al., 2012) to calculate landscape metrics (hereafter referred to in
italics).
Measures of landscape composition. Abundance of primary land-covers: The total area
of the three most common crop (Maize, Soybean, Wheat) and non-crop (Grasslands, Forests,
Residential areas) land covers were calculated for each landscape. Grasslands included old
fields, prairie plantings and Conservation Reserve Program lands (areas planted to “long-term,
resource-conserving covers” to provide a variety of ecosystem services (USDA 2013)).
Residential areas included private residences and public mowed areas like cemeteries and
municipal parks. Small woodlots were included in the Forest total. Additionally, the abundance
of pasture, alfalfa, orchard grass, and other forage crops were calculated and combined into a
single variable, Forage crops, representing perennial agricultural crops.
Cover diversity: Simpson’s Diversity Index (SDI) (Simpson, 1949) was used as a metric
of landscape heterogeneity (McGarigal et al., 2012). This index is less sensitive than other
diversity measures to the number of land cover types because SDI is based on the proportion of
area that each habitat type constitutes rather than the absolute area, and rare habitats will
contribute smaller proportions (Magurran, 2004). Therefore, it effectively describes the variance
in relative abundances of habitat types without being skewed by the presence of rare habitats
(Magurran, 2004). Overall habitat diversity within the landscape, SDIhabitat, was calculated
using the proportional cover of every land cover type identified in ground-truthing using the
formula:
85

SDI = 1 -

 pi2,

where pi = proportion of landscape occupied by patch type i. This formulation of the index varies
from 0 to 1 and can be intuitively interpreted as the probability that any two randomly chosen
pixels are of different composition.
In addition to overall habitat diversity (SDIhabitat), I also calculated crop diversity
(SDIcrop) using the proportion of every crop type identified. These two metrics were not highly
2

correlated with one another (r = 0.01) and could represent two different pools of resources for
natural enemies.
Measures of landscape configuration. Crop mesh size: I calculated an effective mesh
size (MESH) using only crop land covers, to represent average field size in each landscape,
according the formula:
MESH =

  aij2/ A ,
2

2

where aij is the area of patch ij (m ), and A is the total landscape area (m ) (McGarigal et al.,
2012). MESH is a metric of average size (ha) of landscape patches. Thus, landscapes with larger
crop MESH values have fewer, larger fields and landscapes with smaller MESH values have
more, smaller fields.
Landscape shape index: I re-coded the landscape maps identifying each patch as either
crop or non-crop and calculated a Landscape Shape Index (LSI),
LSI = 0.25E*/ A ,

86

where 0.25 is an adjustment for raster data, E* is the total length of edge in the landscape in
meters, including the boarders of all patches identified during digitization. A is the total area of
the landscape in square meters (McGarigal et al., 2012). LSI is greater than or equal to 1, with
one indicating a landscape comprised of one large patch. LSI increases as the amount of edge in
the landscape increases. LSI is a measure of both the total amount of edge existing between land
covers of different types and of habitat aggregation. As a particular land cover becomes more
dispersed throughout a landscape, the amount of edge increases. Thus, higher LSI values
represent landscapes with greater dispersion of crop and non-crop habitat, while lower LSI values
represent landscapes in which crop and non-crop habitats are aggregated into larger patches.
Statistical analyses. I used a Partial Least Squares (PLS) regression approach (PROC
PLS SAS 9.2) to model variation between the landscape variables and mean coccinellid
abundance. The goal of PLS is to extract underlying factors, consisting of linear combinations of
predictor variables that account for variation in predictor and response variables. PLS analyses
can handle multiple, highly correlated predictor variables (Wold et al., 2001) making it more
flexible than the more commonly used multiple linear regression, in which correlated predictor
variables lead to unstable parameter estimates. Using PLS regression allows us to consider the
relationship between multiple correlated landscape variables, instead of using semi-natural
habitat as a proxy for all the others, as has been done in the past (Burel et al., 2013).
PLS regression algorithms extract factors that account for the greatest amount of
variation in predictor and response variables simultaneously. While PLS is closely related to
Principal Components Analysis (PCA), in PCA factors are chosen that explain the most variation
in predictor variables independently of response variables, regardless of whether those factors
are relevant to the predictors (Abdi, 2010). Additionally, PLS can handle data sets in which there

87

are more predictor variables than observations (Carrascal et al., 2009). While PLS has been most
commonly used in econometrics and chemistry, the ability to analyze datasets in which there are
a large number of correlated predictor variables and a relatively small sample size makes this
method quite useful for ecological studies (Carrascal et al., 2009). For example, Rusch and
colleagues (2012) used PLS to examine landscape influences on the abundance of pollen beetles
(Meligethes aeneus F.).
It is common practice to transform predictor and response variables to achieve symmetry
prior to conducting PLS regression (Wold et al., 2001). I used a Shapiro-Wilk statistic to
determine if variables were symmetrical. I square-root transformed proportions of Wheat,
Forage crops, Residential areas and Grasslands, as well as MESH, and cubed SDIhabitat and
repeated Shapiro-Wilk tests to confirm that symmetry (normality) had been achieved. Results
using the transformed and untransformed datasets were almost identical, and I report the results
from the transformed data set to conform with standards in the published literature (Wold et al.
2001).
Data were centered to a mean of zero and scaled to a standard deviation of 1 to give each
variable equal weight in the analysis (Wold et al., 2001). I used one-at-a-time cross-validation to
determine the appropriate number of factors to extract from the data. In this technique, one
observation is removed and the model is fit with the remaining observations and used to predict
the missing observation. The difference between the actual and predicted value for the missing
observation is calculated. This procedure is repeated for each observation in the data set and the
sum of squares for the error from each model is used to calculate the predicted residual sum of
squares (PRESS). PRESS is analogous to the residual sum of squares in regular regression and is
an indicator of the model’s predictive power. To select the appropriate number of factors for a
88

given model, first multiple orthogonal factors are estimated that are each linear combinations of
the predictors variables. Each subsequent factor explains less of the overall variation in the
response and predictor variables. In the cross-validation procedure, the PRESS is calculated
separately for models with different numbers of factors, and the model that minimizes PRESS is
chosen as the best. Because in some cases a model with more factors may have only slightly
lower PRESS, I used Van der Voet’s test to determine if models with fewer extracted factors had
insignificantly higher PRESS at a 10% level (SAS Institute Inc., 2011).
After the appropriate number of factors was selected, I assessed the model’s goodness of
2

2

2

2

fit by examining R X, R Y, and Q . R X is the amount of variation in the matrix of predictor
2

variables that the model explains, R Y is the amount of variation in the matrix of response
2

variables that the model explains, and Q is the amount of variation in the response variable that
the model predicts through the cross-validation procedure. I then plotted the loadings of predictor
and response variables for each factor to look for curvature or outliers, and examined the
coefficients and variable influence on projection (VIP) values to for each variable to determine
the contribution of each variable to the model (Wold et al., 2001). Variables with low VIP
values (below 0.80) and regression coefficients near zero have little influence on the response
variable and make little contribution to the structure of the overall PLS model (SAS Institute
Inc., 2011).
Results
Coccinellid captures. In 11 weeks of sampling, I captured a total of 1,989 coccinellids.
Average captures per trap per week ranged from 1.9 to 6.4 coccinellids/trap. The coccinellid
community within soybean fields was dominated by exotic coccinellids, with Coccinella

89

septempunctata making up 33%, Harmonia axyridis 31%, Propylea quatourdecimpunctata 20%,
and Hippodamia variegata 6%, of all coccinellids captured. The most common native
coccinellids captured were Coleomegilla maculata and Hippodamia parenthesis, making up 5
and 3% of total captures. Other species made up 1% or less of captures.
Landscape variables. Landscapes represented a gradient from agriculturally-dominated
to forest-dominated. Soybean, Maize, Forest, and Residential areas had the biggest ranges across
landscapes, while Grasslands, Forage crops, and Wheat had smaller gradients (Table 2.1). These
7 main composition variables combined made up 79.6 to 99.4% of all landscapes studied.
SDIhabitat and SDIcrop had similar ranges, although they were not correlated with one another.
The range in MESH indicated a gradient of landscapes with average crop fields of roughly 1 to
84 ha (Table 2.1).
In addition to the three most common annual crops and forage crops, several other crops
were identified in the studied landscapes in lower abundances, which also contribute to crop
diversity, SDIcrop. These included: dry beans (comprising a maximum of 8.1% of all landscapes
studied), non-wheat small grains (e.g. barley, oats; 2.2%), sugar beets (1.9%), clover cover crop
(1.5%), vegetables (0.3%) and fruit trees (0.1%). Additionally, although alfalfa, orchard grass,
and pastures were grouped as forage crops, they were included separately in crop diversity
calculations, each contributing a maximum of 8.1, 6.1, and 6.0 % of landscapes, respectively.
PLSR model. The model with one extracted factor had the lowest PRESS, and the
PRESS was significantly lower than the model with zero factors (p<0.04). Factor 1 explained
2

51.7% of variation in the landscape predictor variables (R X) and 63.8% of variation in mean

90

2

coccinellid abundances (R Y). Furthermore, the model with one factor explained 51.2% of the
2

variation in mean coccinellid abundance (Q ).
The plot of response and predictor scores from factor 1 shows a linear positive
relationship between response and predictor scores and indicates no outliers, grouping of data, or
curvature in this relationship (Figure 2.1). Residual plots for response and predictor variables
indicate no heterogeneities.
Factor 1 primarily explains a gradient from landscapes dominated by a few annual crops
grown in large individual fields to landscapes with higher crop diversity, smaller fields, and more
abundant non-crop habitats. Mean coccinellid abundance, SDIcrop, and the abundance of
Grasslands, Forests, and Residential areas load positively with factor 1, while MESH (average
field size) and the abundance of Maize, Soybean and Wheat load negatively (Figure 2.2). The
variables Grassland and Residential area have the highest correlation with mean coccinellid
abundance. The variables Forage and SDIhabitat had low Variable Importance for Projection
(VIP; Table 2.1) and regression coefficients near zero, indicating that they were not important in
structuring factor 1. All other variables had greater influence on the structure of the model (e.g.
VIP variables > 0.8; Table 2.1). The model explained moderate amounts of variation in the
proportion of wheat in the landscape and LSI, and 60-80% of variation in the other variables.
Because variables with low VIP and coefficients close to zero have little influence on the
model, it is common to delete them and rerun the PLS to obtain a better PLS model for
predictive purposes (SAS Institute Inc., 2011). Rerunning the model without SDIhabitat and

91

2

2

2

Forage increased the R X to 62.5%, but other results (i.e. R Y, Q , VIP values, residual plots)
were nearly unchanged.
Discussion
Naturally occurring pest suppression from native or naturalized natural enemies is an
important ecosystem service in crop production (Losey and Vaughan, 2006; Swinton et al.,
2006). This study showed that both landscape composition and configuration may influence the
abundance of a group of important generalist predators. Landscapes in this study represented a
range of agricultural intensification typical of the upper Midwestern U.S., with landscapes
dominated by semi-natural habitat and having high crop diversity and small crop field size at one
end of the gradient and landscapes dominated by a few annual crops grown in large fields at the
other.
Mean summer coccinellid abundance in soybean fields increased with increasing noncrop cover in the landscape. Similar positive effects of non-crop habitat on predator and
parasitoid abundances have been found in a variety of systems, including canola, wheat and
cotton (Prasifka et al., 2004; Schmidt and Tscharntke, 2005; Thies and Tscharntke, 1999). Noncrop habitat could have contributed to increased coccinellid abundances by providing resources
such as overwintering habitat, nectar and pollen, and alternative hosts and prey (Bianchi et al.,
2006).
However, because it is unlikely that many farmers will remove substantial amounts of
land from production to increase non-crop area, it is also useful to consider the benefits of
aspects of landscape structure more under farmer control, such as crop type, overall crop
diversity, and field size. In this study, the diversity of crops grown in the landscape and mean
coccinellid abundance both loaded strongly positively with factor 1, which explained a gradient
92

from landscapes dominated by a few annual crops to those with more natural habitat, higher
crops diversity, and smaller field sizes. Crop diversity could be important for natural enemies
because resources in one crop may become available as resources in another crop are
diminishing, due to differing crop phenologies (Vasseur et al., 2013). Surprisingly, I did not find
a relationship between perennial forage crops and coccinellid abundance. This may be because
coccinellids correlated strongly along an agricultural intensification gradient, while perennial
forage crops were grown in both agriculturally dominated and non-crop dominated landscapes.
The negative relationship between average crop field size and coccinellid abundance may
suggest that coccinellids are more likely to colonize small fields. In landscapes with smaller
fields, more soybean area will be adjacent to edges shared with other cover types (Bianchi et al.,
2006), which could provide food and refuge resources to coccinellids. Furthermore, the increased
area of cover type interfaces allows increased arrival rates of dispersing coccinellids from other
cover types into soybean (Bianchi and van der Werf, 2003). Alternatively, the relationship
between field size and coccinellids could be an artifact of landscape change and the association
between coccinellids and crop cover. In this region, the landscapes in which the most maize and
soybean are grown tend to have fertile soils and flat, well-drained land, allowing these crops to
be grown in larger fields.
Contrary to my hypothesis, I found an inverse relationship between LSI and coccinellids.
This may be the result of correlations between LSI and other variables that are stronger drivers of
coccinellids abundance. For example, coccinellids are highly correlated with the amount of noncrop habitats. In this set of landscapes, greater amounts of non-crop habitat tends to be
aggregated into larger patches of grassland or forest, which would lead to having lower LSI
values. Therefore, the strong effect of non-crop habitat on coccinellids could be driving this

93

counter-intuitive relationship. Alternatively, the fact that coccinellids are strong fliers might
negate the influences of landscape configuration in this system. It would be informative to
conduct a similar study measuring a suite of natural enemies with different dispersal abilities, to
determine if the strength of configuration effects increase as dispersal abilities decline.
I have shown several landscape factors that are strongly related to both coccinellid
captures in soybean fields and to each other. Such correlations among landscape variables are
common, making it difficult for researchers to assert causal relationships. For example, many
studies of landscape influences on natural enemies and biological control use semi-natural
habitat as their primary measure of landscape structure, using it as a proxy for habitat diversity
and citing the strong correlation between these two variables in their studies (Roschewitz et al.,
2005; Steffan-Dewenter et al., 2002; Thies and Tscharntke, 1999). Notably, in this study, overall
habitat diversity was not strongly correlated with any of the non-crop land covers. Instead, I
found overall habitat diversity was highest at intermediate ranges of non-crop habitat, and had
low explanatory value in the model. This example highlights the need for caution when drawing
conclusions about the influences of a particular landscape variable on natural enemies,
particularly when there are legitimate hypotheses for why multiple landscape factors may be
involved. Studies which carefully select landscapes that vary independently in key landscape
variables may be particularly informative; however, that may be easier in theory than practice
(Fahrig et al., 2011).
Conclusions
There is growing evidence that crop pest populations cannot be effectively managed
without consideration of landscape influences (Cumming and Spiesman, 2006; Schellhorn et al.,
2008). Understanding the influence of landscape composition and configuration on natural

94

enemy abundance and movement in agroecosystems is one prerequisite to such region-wide pest
management. Given that farmers may have more control over their crop selection and field size
than over semi-natural habitat surrounding their farms, future studies to isolate these aspects of
landscape structure will be necessary to move forward on landscape-scale pest management.
Additionally, given that landscapes dominated by annual crop covers are likely to experience not
only greater pest pressure but also increased insecticide use (Meehan et al., 2011; O'Rourke and
Jones, 2011), a broad understanding of the relationship between landscape structure and natural
pest suppression is critical for mitigating potential ecosystem disservices in agricultural
landscapes.

95

APPENDIX

96

Chapter II Tables and Figures

97

Table 2.1: Landscape variables used in PLS analysis organized by variable type. Variables
analyzed are shown in italics. Minimum and maximum values for each variable within the
2
studied landscapes are shown, along with variable influence for project (VIP) values. R x
indicates the percent of variation in each predictor variable explained by Factor 1.
2

min.

max.

VIP

R X

4.8
0.4
0.0

39.7
47.2
15.6

1.15
0.85
1.05

65.0
60.5
45.4

0.4

14.7

0.10

0.1

2.4
1.9
3.5

64.7
29.5
14.8

0.85
1.35
1.35

63.1
73.3
59.2

0.57

0.85

0.35

7.8

SDIcrop

0.76

1.00

1.10

78.4

MESH
Landscape Shape Index

0.94
3.25

83.98
7.55

1.15
0.85

80.1
35.8

COMPOSITION
Primary Land Covers (%):
Annual crops:
Soybean
Maize
Wheat
Forage crops

1

Non-crop:
Forest
Residential
Grasslands
Cover diversity
SDIhabitat
2

CONFIGURATION

1

Forage crops consist of pasture, alfalfa, orchard grass and other crops grown for hay

2

In addition to the crops listed above, the amount of dry beans, small grains, sugar
beets, clover, vegetables and fruit trees were included in crop diversity calculations.

98

Figure 2.1: Relationship between coccinellid abundance in soybean fields in 2009 and landscape
variable scores from PLS analysis. Predictor and explanatory variables show a positive
correlation within factor 1, suggesting that this factor captures a real relationship between these
variables.

3

Y Scores
Mean coccinellid counts

2

1

0

-1

-2
-4

-2

0

2

X Scores
Landscape Variables

99

4

6

Figure 2.2: Loadings of response (coccinellids, in box) and predictor (landscape) variables from
PLS regression. The abundance of coccinellids loads positively along factor one, along with noncrop habitats like grasslands, residential areas, and forests, and with crop diversity (SDIcrop). In
contrast, annual crops (maize, soybean, wheat) and crop field size (MESH) load negatively. The
numbers 1 through 17, representing the landscapes, are generally spread-out and show no
obvious outliers.

100

LITERATURE CITED

101

LITERATURE CITED

Abdi, H., 2010. Partial least squares regression and projection on latent structure regression (PLS
regression). Wiley Interdisciplinary Reviews: Computational Statistics. 2, 97-106.
Bianchi, F. J. J. A., Booij, C. J. H., Tscharntke, T., 2006. Sustainable pest regulation in
agricultural landscapes: a review on landscape composition, biodiversity and natural pest
control. Proceedings of the Royal Society B-Biological Sciences. 273, 1715-1727.
Bianchi, F. J. J. A., van der Werf, W., 2003. The effect of the area and configuration of
hibernation sites on the control of aphids by Coccinella septempunctata (Coleoptera :
Coccinellidae) in agricultural landscapes: A simulation study. Environmental
Entomology. 32, 1290-1304.
Bommarco, R., 1998. Reproduction and energy reserves of a predatory carabid beetle relative to
agroecosystem complexity. Ecological Applications. 8, 846-853.
Burel, F., Lavigne, C., Marshall, E. J. P., Moonen, A. C., Ouin, A., Poggio, S. L., 2013.
Landscape ecology and biodiversity in agricultural landscapes. Agriculture, Ecosystems,
and Environment. 166, 2.
Carrascal, L. M., Galvan, I., Gordo, O., 2009. Partial least squares regression as an alternative to
current regression methods used in ecology. Oikos. 118, 681-690.
Colunga-Garcia, M., S.H. Gage, and Landis, D.A, 1997. Response of an assemblage of
Coccinellidae (Coleoptera) to a diverse agricultural landscape. Environmental
Entomology. 26, 8.
Costamagna, A. C., Landis, D. A., 2007. Quantifying predation on soybean aphid through direct
field observations. Biological Control. 42, 16-24.
Costamagna, A. C., Landis, D. A., DiFonzo, C. D., 2007. Suppression of soybean aphid by
generalist predators results in a trophic cascade in soybeans. Ecological Applications. 17,
441-451.
Cumming, G. S., Spiesman, B. J., 2006. Regional problems need integrated solutions: Pest
management and conservation biology in agroecosystems. Biological Conservation. 131,
533-543.
Dunning, J. B., Danielson, B. J., Pulliam, H. R., 1992. Ecological processes that affect
populations in complex landscapes. Oikos. 65, 169-175.

102

Elliott, N. C., Kieckhefer, R. W., Lee, J. H., French, B. W., 1999. Influence of within-field and
landscape factors on aphid predator populations in wheat. Landscape Ecology. 14, 239252.
Fahrig, L., Baudry, J., Brotons, L., Burel, F. G., Crist, T. O., Fuller, R. J., Sirami, C.,
Siriwardena, G. M., Martin, J. L., 2011. Functional landscape heterogeneity and animal
biodiversity in agricultural landscapes. Ecology Letters. 14, 101-112.
Gardiner, M. M., Landis, D. A., Gratton, C., DiFonzo, C. D., O'Neal, M., Chacon, J. M., Wayo,
M. T., Schmidt, N. P., Mueller, E. E., Heimpel, G. E., 2009. Landscape diversity
enhances biological control of an introduced crop pest in the north-central USA.
Ecological Applications. 19, 143-154.
Hogg, B. N., Daane, K. M., 2010. The role of dispersal from natural habitat in determining
spider abundance and diversity in California vineyards. Agriculture Ecosystems &
Environment. 135, 260-267.
Isaia, M., Bona, F., Badino, G., 2006. Influence of landscape diversity and agricultural practices
on spider assemblage in Italian vineyards of Langa Astigiana (northwest Italy).
Environmental Entomology. 35, 297-307.
Kruess, A., 2003. Effects of landscape structure and habitat type on a plant-herbivore-parasitoid
community. Ecography. 26, 283-290.
Landis, D. A., Wratten, S. D., Gurr, G. M., 2000. Habitat management to conserve natural
enemies of arthropod pests in agriculture. Annual Review of Entomology. 45, 175-201.
Losey, J. E., Vaughan, M., 2006. The economic value of ecological services provided by insects.
Bioscience. 56, 311-323.
Magurran, A. E., 2004. Measuring Biological Diversity. Blackwell Science Ltd, Malden, MA.
McGarigal, K., Cushman, S. A., Ene, E., 2012. Fragstats v4: Spatial pattern analysis program for
categorical and continuous maps. Computer software program produced by the authors at
University of Massachusetts, Amherst. Available at the following website:
http://umass.edu/landeco/research/fragstats/fragstats.html.
Meehan, T. D., Werling, B. P., Landis, D. A., Gratton, C., 2011. Agricultural landscape
simplification and insecticide use in the Midwestern United States. Proceedings of the
National Academy of Sciences of the United States of America. 108, 11500-11505.
Millennium Ecosystem Assessment, Ecosystems and Human Well-being: Synthesis. 2005.
Island Press, Washington, D.C.
O'Rourke, M. E., Jones, L. E., 2011. Analysis of landscape-scale insect pest dynamics and
pesticide use: an empirical and modeling study. Ecological Applications. 21, 3199-3210.
103

Oberg, S., Ekbom, B., Bommarco, R., 2007. Influence of habitat type and surrounding landscape
on spider diversity in Swedish agroecosystems. Agriculture Ecosystems & Environment.
122, 211-219.
Obrycki, J. J., Kring, T. J., 1998. Predaceous Coccinellidae in biological control. Annual Review
of Entomology. 43, 295-321.
Östman, O., Ekbom, B., Bengtsson, J., 2001. Landscape heterogeneity and farming practice
influence biological control. Basic and Applied Ecology. 2, 365-371.
Power, A. G., 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philosophical
Transactions of the Royal Society B-Biological Sciences. 365, 2959-2971.
Prasifka, J. R., Heinz, K. M., Minzenmayer, R. R., 2004. Relationships of landscape, prey and
agronomic variables to the abundance of generalist predators in cotton (Gossypium
hirsutum) fields. Landscape Ecology. 19, 709-717.
Ragsdale, D. W., Landis, D. A., Broduer, J., Heimpel, G. E., Desneux, N., 2011. Ecology and
management of the soybean aphid in North America. Annual Review of Entomology. 56,
375-99.
Roschewitz, I., Hucker, M., Tscharntke, T., Thies, C., 2005. The influence of landscape context
and farming practices on parasitism of cereal aphids. Agriculture Ecosystems &
Environment. 108, 218-227.
Rusch, A., Valantin-Morison, M., Roger-Estrade, J., Sarthou, J. P., 2012. Local and landscape
determinants of pollen beetle abundance in overwintering habitats. Agricultural and
Forest Entomology. 14, 37-47.
SAS Institute Inc., 2011. The PLS Procedure. SAS/STAT 9.3 User's Guide. SAS Institute Inc.,
Cary, NC.
Schellhorn, N. A., Macfadyen, S., Bianchi, F., Williams, D. G., Zalucki, M. P., 2008. Managing
ecosystem services in broadacre landscapes: what are the appropriate spatial scales?
Australian Journal of Experimental Agriculture. 48, 1549-1559.
Schmidt, M. H., Roschewitz, I., Thies, C., Tscharntke, T., 2005. Differential effects of landscape
and management on diversity and density of ground-dwelling farmland spiders. Journal
of Applied Ecology. 42, 281-287.
Schmidt, M. H., Tscharntke, T., 2005. Landscape context of sheetweb spider (Araneae:
Linyphiidae) abundance in cereal fields. Journal of Biogeography. 32, 467-473.
Simpson, E. H., 1949. Measurement of diversity. Nature. 163, 688-688.

104

Steffan-Dewenter, I., Munzenberg, U., Burger, C., Thies, C., Tscharntke, T., 2002. Scaledependent effects of landscape context on three pollinator guilds. Ecology. 83, 14211432.
Swinton, S. M., Lupi, F., Robertson, G. P., Hamilton, S. K., 2007. Ecosystem services and
agriculture: Cultivating agricultural ecosystems for diverse benefits. Ecological
Economics. 64, 245-252.
Swinton, S. M., Lupi, F., Robertson, G. P., Landis, D. A., 2006. Ecosystem services from
agriculture: Looking beyond the usual suspects. American Journal of Agricultural
Economics. 88, 1160-1166.
Thies, C., Steffan-Dewenter, I., Tscharntke, T., 2003. Effects of landscape context on herbivory
and parasitism at different spatial scales. Oikos. 101, 18-25.
Thies, C., Tscharntke, T., 1999. Landscape structure and biological control in agroecosystems.
Science. 285, 893-895.
Turner, M. G., Gardner, R. H., O'Neill, R. V., 2001. Landscape Ecology in Theory and Practice:
Pattern and Process. Springer-Verlag, New York.
USDA Farm Services Agency. Website “Conservation Programs”
https://www.fsa.usda.gov/FSA/webapp?area=home&subject=copr&topic=crp
accessed 26 Apr 2013
Vasseur, C., Joannon, A., Aviron, S., Burel, F., Meynard, J. M., Baudry, J., 2013. The cropping
systems mosaic: how does the hidden heterogeneity of agricultural landscapes drive
arthropod populations? Agriculture, Ecosystems and Environment. 166, 12.
Wold, S., Sjostrom, M., Eriksson, L., 2001. PLS-regression: a basic tool of chemometrics.
Chemometrics and Intelligent Laboratory Systems. 58, 109-130.
Woltz, J. M., Isaacs, R., Landis, D. A., 2012. Landscape structure and habitat management
differentially influence insect natural enemies in an agricultural landscape. Agriculture
Ecosystems & Environment. 152, 40-49.

105

CHAPTER III: Coccinellid immigration to infested host patches influences suppression of Aphis
glycines in soybean.

Woltz, J.M. and D.A. Landis, 2013. Coccinellid immigration rate influences suppression of
Aphis glycines in soybean. Biological Control. 64: 330-337.
Abstract
Generalist natural enemies may be well adapted to annual crop systems in which pests
and natural enemies re-colonize fields each year. In addition, for patchily-distributed pests, a
natural enemy must also disperse within a given crop field to arrive at infested host patches. As
they typically have long generation times compared to their prey, theory suggests that generalist
natural enemies need high immigration rates both to and within fields to effectively suppress pest
populations. The soybean aphid, Aphis glycines Matsumura, is a pest of an annual crop and is
predominantly controlled by coccinellids. To test if rates of coccinellid arrival at aphid-infested
patches are crucial for soybean aphid control, we experimentally varied coccinellid immigration
to 1m2 soybean patches using selective barriers and measured effects on A. glycines population
dynamics. In a year with low ambient aphid pressure, naturally-occurring levels of coccinellid
immigration to host patches were sufficient to suppress aphid population growth. In contrast,
decreasing coccinellid immigration rates resulted in large increases in soybean aphid populations
within infested patches. Activity of other predatory taxa was very low in this year, suggesting
that most of the differences in aphid population growth were due to changes in coccinellid
immigration. Alternatively, in a year in which alate aphids continually colonized plots, aphid
suppression was incomplete and increased activity of other predatory taxa contributed to adult
coccinellid predation of A. glycines. Our results suggest that in a system in which natural enemy

106

populations cannot track pest populations through reproduction alone, immigration of natural
enemies to infested patches can compensate for reproduction rates and result in pest control.
Introduction
Natural enemies of crop pests exhibit a wide range of life history traits that influence
their effectiveness as biological control agents. Modeling studies suggest that effective pest
suppression may be obtained from either specialist natural enemies inflicting high mortality
rates, or from generalist natural enemies with high immigration rates into an infested area
(Chang and Kareiva, 1999). Additionally, each of these strategies may be better adapted to
particular crop types. For example, specialist natural enemies may be well-adapted for perennial
crops in which temporal stability allows equilibrium to be achieved between natural enemy and
pest populations, while generalists may be better suited for annual crops which both pests and
natural enemies have to recolonize each growing season (Murdoch et al., 2006). Furthermore,
because generalist natural enemies like coccinellids (Coleoptera: Coccinellidae) tend to have
long generation times relative to those of their prey and thus cannot rapidly respond numerically
to pest population dynamics (Dixon et al., 1997; Kindlmann and Dixon, 1999a; Kindlmann and
Dixon, 1999b; Kindlmann and Dixon, 2001), generalists must exhibit high immigration rates to
be effective (Murdoch et al., 2006). In fact, higher immigration rates have been demonstrated to
decrease the predator/ prey ratios necessary for effective pest suppression (van der Werf, 1995),
and manipulating immigration rates of generalist predators has resulted in lower pest abundance
and higher yields (Snyder and Wise, 1999).
Natural enemy immigration occurs at different temporal and spatial scales. In annual
agriculture, many natural enemies overwinter outside of the crop and must first immigrate into
crop fields after their establishment. Once in the crop habitat, they must then disperse and locate

107

prey-infested patches within the field. In this manuscript, we focus on the process of patch-level
immigration and its impact on populations of the soybean aphid, Aphis glycines Matsumura
(Hemiptera: Aphididae), an invasive and economically damaging pest of soybean (Glycines max
(L.) Merrill).
In the north central U.S., soybean is typically grown in annual rotation with other crops
and both A. glycines and many of its natural enemies must colonize the crop each season
(Ragsdale et al., 2011). Furthermore, because A. glycines is patchily distributed in soybean fields
(Costamagna and Landis 2011), it is necessary for predators to locate and immigrate to aphidinfested patches within soybean fields. In North America, generalist predators provide the
majority of natural control of A. glycines, and multiple studies indicate that if predators are
reduced or excluded entirely, A. glycines quickly reach outbreak levels (Brosius et al., 2007;
Costamagna and Landis, 2006; Costamagna et al., 2007; Desneux et al., 2006; Donaldson et al.,
2007; Fox et al., 2005; Gardiner et al., 2009; Rhainds et al., 2007; Schmidt et al., 2007). Large,
mobile predators play a particularly important role in A. glycines suppression. Both Gardiner and
Landis (2007) and Costamagna and others (2008) showed that small, less mobile predators like
Orius insidiosus (Say) and various predatory larvae were unable to control A. glycines, while
aphid populations were significantly lower when they were accessible by large predators like
coccinellids. Furthermore, the large coccinellid beetles Coccinella septempunctata L. and
Harmonia axyridis Pallas have been shown to be highly efficient A. glycines predators, having
the highest per capita consumption rate of the A. glycines predators studied to date, and
contribute to a majority of A. glycines predation in the North Central U.S. (Costamagna and
Landis, 2007). In addition, these two species are dominant in soybean fields in Michigan. H.

108

axyridis and C. septempunctata together accounted for 75 and 62 % of total lady beetle captures
in Michigan soybean fields in 2008 and 2009 (Woltz, unpub. data).
Most coccinellid beetles have long generation times relative to aphids, suggesting an
inability to suppress aphid prey through reproduction alone (Dixon et al., 1997; Kindlmann and
Dixon, 1999a; Kindlmann and Dixon, 1999b; Kindlmann and Dixon, 2001). Furthermore,
although H. axyridis and C. septempunctata over-wintered adults are active in spring foraging in
crops like alfalfa and wheat (Colunga-Garcia and Gage, 1998; LaMana and Miller, 1996), there
is very little production of larvae in soybean fields before mid-summer when A. glycines
populations are rapidly increasing. This is because coccinellids require minimum prey densities
to oviposit (Seagraves, 2009), tend to oviposit near aphid colonies, and lay more eggs when more
prey are available (Hodek and Honek, 1996). For example, in surveys of 34 soybean fields
throughout the soybean-producing region of Michigan in 2009, lady beetle larvae were not
detected in soybean fields until 20 July even though aphids were present by the first week of
June (Woltz unpub. data). Thus, we predict that early-summer predation of A. glycines on small
soybean plants due to adult coccinellids regularly arriving at and predating aphid-infested
patches within soybean fields will be crucial to overall A. glycines suppression.
Here, we measure the importance of patch-level coccinellid immigration for suppressing
2

aphid populations by experimentally varying coccinellid immigration to 1 m soybean patches
by using selective barriers. We hypothesize that: 1) our manipulations would create a gradient of
coccinellid immigration within the plots and that 2) decreases in early summer growing season
lady beetle immigration would result in increases in aphid populations within plots.

109

Methods
Site. This study was conducted during the summers of 2010 and 2011 on and around the
Kellogg Biological Station, Long Term Ecological Research Site (KBS LTER; 42.405714 N,
85.377658 W). Agriculture in this region consists of a mix of annual field crops (corn, soy, and
wheat) and perennial forage crops (pasture and perennial mixes for hay). Deciduous forests and
old field successions make up the primary non-crop land cover in the area. Study fields were
planted between 10 and 24 May in 2010 and between 9 May and 2 June in 2011. Sampling took
place weekly from 16 June to 9 July in 2010, and 5 to 26 July in 2011. Sampling was later in
2011 because a rainy spring delayed soybean herbicide applications in this region and thus our
access to fields. During the first week of the 2010 experiment, soybean plants ranged from the
two to five trifoliate growth stage (V2-5) (Higley and Boethel, 1994), while during the first week
of the 2011 experiment, soybean plants ranged from the four to six trifoliate growth stage (V46).
Plot establishment. This study was replicated in n=8 soybean fields in each year. Each
field contained three treatment types manipulating patch-level immigration, crossed with two
sampling types for a total of 6 plots per field and 48 plots per year. Predators and A. glycines
were monitored in 1 x 1 m plots in which the immigration of coccinellids to host patches was
differentially manipulated by use of selective barriers. Selective cages and barriers can be used to
manipulate the movement of key coccinellids through plots without altering levels of other
predators or abiotic conditions (Costamagna et al., 2007; Gardiner and Landis, 2007; Gardiner et
al., 2009). In a previous experiment, we observed that 1 m x 1 m plots surrounded by four-sided
40 cm tall polyethylene plastic fences (i.e. fully open tops) reduced the number of adult
coccinellids entering the plot. This occurred even though prey density was significantly higher

110

on plants inside of plots (Landis and van der Werf, 1997). Upon encountering such vertical
barriers, adult coccinellids frequently climb upward to the top edge, and typically initiate a short
flight (D. Landis pers. obs.). While some subsequently land inside the enclosed plot, many do
not, and thus even simple barriers can be used to manipulate the density of transient adult
coccinellids. Plots were placed at least 10 m from the edge of each field and a minimum of 20 m
from other plots. Coccinellids are strong fliers capable of dispersing long distances (Hodek et al.,
1993). In our previous studies, we have found that coccinellids in soybean fields are captured on
sticky traps at similar rates at distances of 25, 50, and 100 m away from field edges. Therefore,
the distance of our treatment plots from the field edge or from each other should not influence
the potential numbers of coccinellids entering each plot. The same design was used in each field,
and barrier treatments and sampling types were randomly assigned to each location. Plots were
centered on a soybean row such that each contained 3 rows of soybean plants grown 38.1 cm
apart. All fields received commercial levels of fertilizer and weed control but no insecticides
were applied.
Three treatments were used to manipulate coccinellid immigration within the 1 m x 1 m
plots (Figure 3.1). The “open” treatment had no barriers to immigration of coccinellids, and plot
corners were simply demarcated with stakes. In a second treatment, “2-sided” barriers were
erected on opposite sides of the plot, blocking immigration from those two sides while leaving
3

the top and other sides open. A 1m PVC frame was erected around each plot and secured into
3

the soil. The east and west facing sides of the 1m frame were wrapped in coarse mesh netting (2
mm openings) that allowed smaller predators to pass through, but prevented coccinellids from
entering (Costamagna et al., 2007). The mesh extended from the top of the frame sides to 2 cm
above the ground, allowing ground-dwelling predators (Carabidae, Aranae, etc.) to freely enter
111

and exit the plots (Figure 3.1). The effect of wind on coccinellid flight is debated (Hodek and
Honek, 1996). At our site the dominant wind direction is west to east, thus our barriers
consistently blocked coccinellids traveling with or against the wind irrespective of row
3

orientation. Finally, in a third treatment, “5-sided,” the 1 m PVC frame was covered on all four
sides and the top with coarse mesh netting (2 mm openings) also raised 2 cm above the ground,
blocking coccinellid immigration from all directions. Previous research has shown little or no
effect of predator exclusion cages on microclimate within cages (Fox et al., 2004; Hopper et al.,
1995; Schmidt et al., 2003), little or no effect of cages on soybean growth (Fox et al., 2004),
biomass, or yield (Costamagna et al., 2008), and little or no effects on soybean aphid growth
separate from changes in predator density (Costamagna et al., 2008; Fox et al., 2004).
Each field contained two sets of each treatment. One set used as “sampling plots” in
which predators were sampled without replacement. The other set was used as “observation
plots” in which aphid populations were monitored. This allowed us to eliminate any effects of
predator removal on aphid abundance. At the initiation of the experiment, every soybean plant in
each plot was visually inspected and all insects found were removed.
Predator monitoring. To determine the efficacy of our barrier treatments for
manipulating immigration to plots, coccinellids and other aerially-dispersing predators
(chrysopids (Neuroptera: Chrysopidae), hemerobeids (Neuroptera: Hemerobediae), lampyrids
(Coleoptera: Lampyridae), cantharids (Coleoptera: Cantharidae), and dolichopodids (Diptera:
Dolichopodidae)) were measured within sampling plots with 23 x 28 cm unbaited yellow sticky
cards (PHEROCON AM, Great Lakes IPM, Vestaburg, MI). All coccinellids on the cards were
counted and identified to species in the field. The traps were placed in the center of each
sampling plot suspended from step-in plastic fence posts at plant canopy height and were
112

replaced weekly (Figure 3.1 A-C). We also measured activity-density of ground-dwelling
predators to test our assumption that they would not be affected by the barrier treatments.
Ground-dwelling predators including carabids (Coleoptera: Carbidae), staphylinids (Coleoptera:
Staphylinidae), elaterids (Coleoptera: Elateridae), formicids (Hymenoptera: Formicidae), spiders
(Aranae) and harvestmen (Opiliones), were collected in 11 cm diameter pitfall traps filled with
1:1 propylene glycol: water solution. Pitfall traps were collected and reset each week. Foliardispersing predators, those residing on the foliage and likely to disperse short distances from
plant to plant, were assessed with weekly visual plant surveys. Each week the 5 plants surveyed
for aphids within each plot were also inspected for syrphid larvae (Diptera: Syrphidae),
coccinellid larvae, chrysopid eggs and larvae, and anthocorid (Hemiptera: Anthocoridae) adults
and nymphs.
Aphid monitoring. Naturally occurring aphid populations within each study field were
monitored weekly using whole plant counts. During the first week of sampling, the centermost
plant in all plots was infested with 10 greenhouse-reared (2010) or field-collected (2011) aphids
of mixed age classes representing natural incipient colonies. Aphid abundances were counted
weekly on the infested plant and on 4 additional randomly selected plants within each plot.
Apterous and alate aphids were counted separately to account for the different modes in which
aphid populations could be increasing in the plots. Briefly, aphid densities can increase through
reproduction, triggering the production of large numbers of winged alate dispersal morphs.
Alternatively, the appearance of alate aphids at low aphid densities suggests the arrival of new
colonizers from other infested patches, fields, or hosts. Plot counts were stopped after 4 (2010) or
5 (2011) weeks due to a build-up of alates in the 5-sided treatment.

113

Aphid populations within fields were monitored weekly by randomly placing a 1m

2

quadrat within the soybean field. Five quadrat samples were taken each week in each field. Five
random plants within each quadrat were inspected for apterous and alate A. glycines, coccinellid
adults and larvae, lacewing larvae, syrphid larvae, and anthocorid adults and nymphs. The
vegetative growth stage of each surveyed plant was determined according to Higley and Boethel
(1994) and recorded. Finally, the number of soybean plants within each quadrat was recorded.
Statistical analysis. For all of the insect taxa, neither data nor residuals were normal and
variance was heterogenous and increased with the mean. Therefore, insect data were analyzed
using generalized linear models, which estimate parameters with less bias than modeling
transformed data (O'Hara and Kotze, 2010) and avoid complications of making inferences in the
untransformed scale (Venables and Dichmont, 2004). We analyze data for each taxon separately
for each year. Soybean aphid dynamics are known to vary dramatically from year to year (e.g.
Rhainds et al., 2010) and we did not wish to average over this important variation. We report
results for all tests for which levels are different at <0.1, using Tukey’s adjustment for multiple
comparisons.
Predator counts: The activity-density of coccinellids, lampyridae (2011 only), six taxa of
ground-dwelling predators (formicids, carabids, staphylinids, elaterids, spiders, and harvestmen),
and anthocorids (2011 only), were analyzed using separate generalized linear mixed models
(PROC GLIMMIX; SAS 9.2). With the exception of anthocorids, individuals per trap for each
taxon were modeled with a negative binomial distribution, a common distribution for overdispersed count data. Data from each year were analyzed separately. The lognormal distribution
was the best fit for the average number of anthocorids (combined adults and nymphs) per plant in
each plot in 2011, the only foliar-dispersing predator abundant enough to analyze statistically.
114

Other predator taxa were not captured frequently enough to allow for statistical comparisons
between treatments. Captures rates for all taxa (untransformed mean  s.e.) are summarized in
the results.
For each taxon we analyzed statistically, we treated treatment, week and their interaction
as fixed effects and field as a random block. Laplace estimation was used to estimate means and
Wald F- and t-tests were used for hypothesis testing (Bolker et al., 2009). Least square means
comparisons were constructed on the linked scale. Because of difficulty in model convergence
for coccinellid data, non-significant effects were dropped from the model and least square means
comparisons were run on models containing only significant effects.
Aphid abundance: Aphid counts from observation and sampling plots were modeled with
a negative binomial distribution using a repeated measures generalized linear model (PROC
GENMOD; SAS 9.2). Plot was the subject of the repeated measures across weeks. Repeated
measures were used in the case of aphid abundance because we followed the same population of
aphids within a plot over time. In contrast, predator data were drawn from random samples of the
broader population and are not expected to be correlated from week to week. Data from each
year were analyzed separately.
Results
Adult coccinellid captures. A total of 231 and 355 adult coccinellids were collected on
sticky traps in 2010 and 2011. Four exotic and five native species of adult coccinellids were
captured in both years (Tables 3.1, 3.2). Three exotic species accounted for 84 and 85 % of all
coccinellids collected in 2010 and 2011, respectively. H. axyridis accounted for 38 and 47 %, P.
quatuordecimpunctata accounted for 34 and 20 %, and C. septempunctata accounted for 11 and
18 %. No other coccinellid species accounted for more than 6 % of total captures in either year.

115

We successfully manipulated adult coccinellid immigration into plots in both years. In 2010,
adult coccinellid immigration into plots differed by treatment (F2,103 = 2.82, p=0.06; Figure 3.2),
but not by week (F3,103=1.02, p=0.39), and the treatment by week interaction was not significant
(F6,103=0.65, p=0.69). Trap catches averaged 3.1 ± 0.41 coccinellids per week in the open
treatment, 2.1 ± 0.40 coccinellids per week in the 2-sided treatment, and 0.42 ± 0.24 in the 5sided treatment. There were significantly more adult coccinellids in the open treatment than in 5sided treatments (t112=6.79, p<0.0001), and significantly more in 2-sided than in 5-sided
(t112=5.77, p<0.000l). There was a trend toward higher adult coccinellid captures in open than in
2-sided plots, but this difference was not significant at Tukey-adjusted levels. The three most
common species followed these patterns of having captures in the open, intermediate in the 2sided, and lowest in the 5-sided plots (Tables 3.1, 3.2).
In 2011 adult coccinellid immigration into plots differed by treatment (F2,84 = 8.01,
p=0.0007; Figure 3.2), but not by week (F3,84=1.30, p=0.28), and the treatment by week
interaction was not significant (F6,84=0.57, p=0.75). There were 6.03 ± 1.04 adult coccinellids in
the open treatment, 2.38 ± 0.33 in the 2-sided treatment, and 0.26 ± 0.10 in the 5-sided treatment.
There were significantly more adult coccinellids in the open treatment than in 2-sided (t86=4.16,
p<0.0001) or in 5-sided (t86=7.43, p<0.0001), and significantly more in 2-sided than in 5-sided
(t86=5.50, p<0.0001).
Ground-dwelling predators. Average captures for each ground-dwelling taxon are
shown in Table 3.3. Of the six taxa examined, only carabidae varied between treatments in 2010
116

(F2,99=3.13, p=0.05). Carabids were captured more frequently in 5-sided than in two-sided plots
(t99=2.48, p=0.04). Captures in open plots were intermediate to the 5- and 2-sided plots and not
statistically different from either. Several taxa varied between weeks (Formicidae: F3,99=14.13,
p<0.0001; Carabidae: F3,99=3.36, p=0.02; Staphylinidae: F3,99=20.27, p<0.0001; Elateridae:
F3,99=10.01, p<0.001; Opiliones: F3,99= 2.53, p=0.06), generally increasing over time. No taxon
showed a significant week by treatment interaction in 2010.
Only Carabidae and Opiliones varied between treatments in 2011 (F2,98=3.92, p=0.02;
F2,98=6.60, p=0.002, respectively). Carabids were captured more frequently in 5-sided than open
plots (t98=2.76, p=0.02). Captures in 2-sided plots were intermediate to the 5-sided and open
plots and not statistically different from either. Opiliones were captured more frequently in both
5- and 2-sided plots than in open plots (t98=3.06, p=0.008; t98=3.39, p=0.003, respectively), but
did not differ between 5- and 2-sided plots. Spiders, opiliones, formicids and staphylinids varied
between weeks (F4,98=3.89, p=0.006; F4,98=14.02, p<0.0001; F4,98=10.96, p<0.0001;
F4,98=5.61, p=0.0004, respectively). Captures of all five of these taxa increased and decreased
erratically between weeks, with no consistent patterns. Additionally, there was a significant week
X treatment interaction for formicids in 2011 (F8,98=2.21, p=0.03), although no differences in
treatments were consistent over time, and no individual comparisons were significant using
adjusted p-values.

117

Foliar- and aerially-dispersing predators. Most foliar-dispersing predators were found
in low abundances with no systematic differences between treatments (Table 3.4). Anthocorids
in 2011 did not vary by week (F3,122=2.16, p=0.10) or by treatments (F2,122=2.11, p=0.13).
However, there were significant interaction effects (F6,122=2.67, p=0.02). There were more
anthocorids per plant in two-sided than in the open treatment in week 2 (t122=2.17, p=0.03). In
week three, there were more anthocorids per plant in both open and two-sided than in the 5-sided
treatment (t122=2.03, p=0.04; t122=3.68, p=0.0003, respectively).
Lampyrids in 2011 varied by week (F4,98=6.53, p=0.0001) and treatment (F2,98=24.25,
p<0.0001) but there was no interactive effect (F8,98=1.43, p=0.19). There were more lampyrids
captured in the open treatment than in either the 2- or 5-sided treatment (t98=5.10, p<0.0001;
t98=5.54, p<0.0001, respectively). Other aerially-dispersing predators were collected in fairly
low numbers relative to coccinellid captures, and in general, were highest in the open treatment
and lowest in the 5-sided treatment (Table 3.5).
Aphid abundance. In 2010, natural aphid pressure in the study fields was very low,
never reaching more than 1.72 aphids / plant in any field in any week. Within infested plots,
2

2

aphid numbers varied between treatment (χ 2=109.02, p<0.0001), week (χ 3=98.23, p<0.0001),
2

and their interaction (χ 6=142.47, p<0.0001). All plots had the same number of aphids when
infested, with treatment differences developing over time (Figure 3.3). In the first week there
were no differences between aphid counts. In week 2 there were significantly more aphids in
2

2

both 5-sided and 2-sided treatments than in the open treatment (χ 1=7.65, p=0.006; χ 1=4.22,
118

p=0.04, respectively), but not between 5- and 2-sided. In week 3, all treatments had significantly
different numbers of aphids. In week 4, there were significantly more aphids in the 5-sided
treatment than in both the open and 2-sided treatments, but no difference between open and 2sided treatments. By the fourth week, there were 105 ± 38 aphids/ plant in the 5-sided treatment,
but only 7 ± 3 aphids/ plant in 2-sided treatments and 3 ± 1 aphids/ plant in the open treatment.
2

In 2011, aphids within the treatment plots increased over each week (χ 3=1319.91,
2

p<0.0001), but there were no differences by treatment (χ 2=2.31, p=0.31) and no interactive
2

effects (χ 6=6.50, p=0.37). However, in 2011 arrival of alate aphids in fields was higher than in
2010 (Figure 3.3). In 2010, no naturally colonizing alate aphids were observed in any soybean
fields, although there was some build-up of alatae in 2- and 5-sided treatments during weeks 3
and 4 (Table 3.6). In 2011, however, naturally-colonizing alatae were observed in every week,
and were found in all treatments from week 2 on (Table 3.6).
Discussion
We set out to determine if the patch-level immigration of adult coccinellids influenced
aphid suppression in soybean fields. Our treatments were effective at creating a range of
coccinellid immigration to local host patches. In both years there were significantly fewer
coccinellids captured in 5-sided than open treatments, with captures in 2-sided treatments
intermediate to these. Lampyrids were also captured at higher rates in open than 2- or 5-sided
plots. However, lampyrids are expected to make limited contributions to soybean aphid
suppression because their capture rates in soybean fields tend to be lower than coccinellids, they
are considered to be somewhat specialized predators of earthworms and snails (Debach and
Rosen, 1991), and they have never been confirmed to feed on soybean aphid despite numerous

119

surveys of predators in North American soybean fields (Ragsdale et al., 2011). In contrast to
coccinellids and lampyrids, activity-density of other predatory taxa was not consistently
influenced by treatment (e.g. anthocorids), was highest in the 5-sided treatment (e.g. opiliones
and carabids), or was low in comparison to coccinellid activity-density.
Changing the patch-level immigration rate of coccinellids altered A. glycines dynamics,
as populations showed trends toward higher numbers in the plots with the lowest coccinellid
captures in both years. Aphid populations were highest in the 5-sided treatment in spite of
increased activity-density of some ground-dwelling predators in these plots. Increased opiliones
and carabid activity-densities in 5-sided plots may be a result of aggregation to higher prey
resources as the bottoms of treatment barriers were open and allowed free access to these
predators.
There are several possible explanations for the different aphid dynamics in 2010 and
2011 despite similar coccinellid immigration rates into the different treatment plots for the first
three weeks of each year of the study. In both years, aphid populations in the open and two-sided
plots remained similar to ambient levels in the surrounding soybean fields. A steady influx of
alate aphids from the surrounding field or landscape into the plots in 2011 could be responsible
for weekly population increases within the plots. Few alates were found in 2010, and only in
plots with high densities of aphids, suggesting that these alates were produced within those plots,
as aphids typically produce dispersal morphs in response to crowding (Dixon, 1985). In contrast,
in 2011 alates were found in our treatment plots and in the surrounding field in every week after
our initial infestation, even in plots with low densities of aphids, suggesting that these alates
represented additional migrants to the plots. The reproduction of the alates within our plots
would have boosted populations beyond what would be expected from reproduction from our

120

original 10 aphids alone. Additionally, while the abundance of foliar-dispersing predators was
low and variable in both years, abundances were generally higher in 2011. The increase of these
predators across all treatments in 2011 may have contributed to the lack of significant differences
in aphid populations in this year, while helping to hold populations in all treatments below
economic threshold. However, due to their low consumption rates (Costamagna and Landis,
2007), it is unlikely that these foliar-dispersing predators such as Orius insidiosus and coccinellid
and chrysopid larvae were primarily responsible for A. glycines suppression in either year.
Finally, the higher captures of opiliones and carabids in 5- and 2-sided plots in 2011 may have
offset reductions in coccinellid immigration into these two treatment plots and be at least
partially responsible for the lack of significant differences in aphid abundance between the
treatments in this year. One species of opiliones has been observed feeding on soybean aphid in
North America (Allard and Yeargan, 2005) and nine species of carabids captured in soybean
fields fed on soybean aphid in laboratory no-choice trials (Hajek et al., 2007; Hannam et al.,
2008; Rutledge et al., 2004).
Coccinellid captures on sticky traps were within the range typically seen in Michigan
soybean fields (e.g. Woltz et al., 2012). With low ambient aphid populations in the field in 2010,
a coccinellid immigration rate of approximately 2 – 3 individuals captured in a plot per week was
sufficient to prevent A. glycines population increase in both open and 2-sided treatments. This
seemingly low activity-density of coccinellids implies that high numbers of coccinellids are not
necessary to suppress soybean aphid populations if individuals are regularly moving through and
predating in aphid-infested patches. In a study of the spatial dynamics of soybean aphid
infestations in soybean fields, Costamagna and Landis (2011) found that areas of high Aphis
glycines density in one week typically became areas of low density in subsequent weeks,

121

suggesting that predators regularly located patches with increasing aphid density and drove
populations back to average field levels. In contrast, in a year like 2011 with higher external
aphid pressure from alates regularly colonizing soybean fields, coccinellid immigration alone
may not be sufficient to counter the influx of new aphids and aphid densities may rise. Under
such conditions, the densities of foliar-dispersing predators such as anthocorids and coccinellid
and chrysopid larvae are likely to increase, increasing their contribution to aphid suppression.
The changing roles of various predators under different levels of ambient aphid pressure is an
interesting question that warrants further study.
Previous studies have shown that adult coccinellids aggregate in patches of high prey
density (Hodek and Honek, 1996), a behavior resulting from decreased emigration from host
patches (Krivan, 2008). However, for patches with low aphid densities, we have shown that the
immigration of cocinellid beetles to aphid infested patches is important for A. glycines
suppression. Similar results have been found in other systems. For example, experimentally
decreasing carabid and lycosid immigration into squash plots resulted in higher pest densities
and lower squash yield (Snyder and Wise, 1999). Furthermore, although coccinellid emigration
from a patch is related to characteristics such as prey density, the presence of potential predators
or competitors, immigration into a patch is random with respect to the quality of that patch
(Krivan, 2008). Rather, rates of immigration into a particular patch are determined by coccinellid
decisions to leave nearby patches (Grez and Prado 2000) as well as their overall densities in the
landscape. Because coccinellid immigration rates to patches within soybean fields should be
higher when their overall density in the landscape is higher, management actions to increase
coccinellid immigration to aphid colonies must take place outside of the crop. Increasing
evidence suggests that landscape composition at scales of multiple kilometers influences the

122

activity-density of coccinellids in crop fields, likely because the presence and abundance of
particular habitats determines the provision of resources necessary for coccinellid survival and
reproduction (Gardiner et al., 2009; Woltz et al., 2012).

123

APPENDIX

124

Chapter III Tables and Figures

125

Table 3.1: Means ± standard errors for all adult coccinellid species captured on sticky traps in 2010.

0

2010
2

5

1.275 ± 0.277
0.350 ± 0.122
1.000 ± 0.238
0.050 ± 0.035

0.700 ± 0.249
0.275 ± 0.134
0.725 ± 0.175
0.050 ± 0.035

0.184 ± 0.184
0
0.211 ± 0.086
0

0.125 ± 0.053
0.200 ± 0.073
0.075 ± 0.042
0.050 ± 0.035
0

0.150 ± 0.067
0.125 ± 0.053
0.025 ± 0.025
0.050 ± 0.035
0

0
0.026 ± 0.026
0
0
0

EXOTIC SPECIES
Harmonia axyridis
Coccinella septempunctata
Propylea quatuordecimpunctata
Hippodamia variegata
NATIVE SPECIES
Cycloneda munda
Coleomegilla maculata
Brachiacantha ursina
Hippodamia parenthesis
Hi. convergens

126

Table 3.2: Means ± standard errors for all adult coccinellid species captured on sticky traps in 2011.

0

2011
2

5

3.450 ± 0.757
1.275 ± 0.293
0.800 ± 0.130
0.275 ± 0.107

0.675 ± 0.191
0.325 ± 0.083
0.900 ± 0.171
0.075 ± 0.042

0.026 ± 0.026
0.026 ± 0.026
0.077 ± 0.043
0.026 ± 0.026

0.225 ± 0.091
0.050 ± 0.035
0.050 ± 0.035
0.075 ± 0.042
0.050 ± 0.034

0.200 ± 0.073
0.100 ± 0.060
0.050 ± 0.035
0.025 ± 0.025
0.025 ± 0.025

0.051 ± 0.036
0.051 ± 0.051
0
0
0

EXOTIC SPECIES
Harmonia axyridis
Coccinella septempunctata
Propylea quatuordecimpunctata
Hippodamia variegata
NATIVE SPECIES
Cycloneda munda
Coleomegilla maculata
Brachiacantha ursina
Hippodamia parenthesis
Hi. convergens

127

Table 3.3: Means ± standard errors for the most common ground-dwelling predator species captured in pitfall traps in 2010 and 2011.
Carabidae

Elateridae

Staphylinidae

Formicidae

Spiders

Opiliones

open
2-sided
5-sided

4.925 ± 1.035
3.800 ± 1.016
8.275 ± 1.894

1.450 ± 0.716
1.200 ± 0.731
0.525 ± 0.193

7.725 ± 1.799
11.075 ± 3.094
10.200 ± 2.421

9.250 ± 1.812
7.000 ± 1.371
6.950 ± 1.757

15.325 ± 2.403
11.600 ± 2.079
9.000 ± 1.512

9.700 ± 1.239
10.575 ± 3.152
10.575 ± 3.152

2011
open
2-sided
5-sided

4.350 ± 0.602
5.500 ± 0.781
8.650 ± 1.673

0.525 ± 0.179
1.075 ± 0.466
0.875 ± 0.355

3.400 ± 0.724
3.675 ± 0.578
6.750 ± 1.860

18.300 ± 1.687
22.700 ± 4.389
16.225 ± 2.927

17.050 ± 2.471
18.775 ± 3.349
20.700 ± 3.984

4.750 ± 1.247
6.225 ± 1.037
6.475 ± 1.159

2010

128

Table 3.4: Means ± standard errors for foliar-dispersing predators observed on soybean plants in 2010 and 2011.
Syrphid larvae

Coccinellid
larvae

Anthocorid
adults

Anthocorid
nymphs

Chrysopid
larvae

2010
open
2-sided
5-sided

0.031 ± 0.011
0.088 ± 0.043
0.250 ± 0.105

0.006 ± 0.004
0.003 ± 0.003
0.003 ± 0.003

0.006 ± 0.004
0.006 ± 0.004
0.016 ± 0.007

0.028 ± 0.011
0.009 ± 0.005
0.019 ± 0.011

0.006 ± 0.004
0.016 ± 0.008
0.013 ± 0.006

2011
open
2-sided
5-sided

0.005 ± 0.004
0.003 ± 0.003
0

0.038 ± 0.010
0.038 ± 0.013
0.020 ± 0.007

0.103 ± 0.028
0.113 ± 0.022
0.110 ± 0.036

0.346 ± 0.040
0.533 ± 0.055
0.333 ± 0.037

0.035 ± 0.013
0.020 ± 0.009
0.058 ± 0.017

129

Table 3.5: Means ± standard errors for aerially-dispersing predator species captured on sticky traps in 2010 and 2011.
Coccinellidae

Chrysopidae

Hemerobeidae

Lampyridae

Cantharidae

Dolichopodidae

2010
open
2-sided
5-sided

3.125 ± 0.405
2.100 ± 0.398
0.421 ± 0.243

0.200 ± 0.089
0.300 ± 0.103
0.079 ± 0.044

0.100 ± 0.100
0.075 ± 0.042
0

1.150 ± 0.575
0.300 ± 0.114
0.132 ± 0.086

0.075 ± 0.042
0.050 ± 0.035
0

0.125 ± 0.064
0
0

2011
open
2-sided
5-sided

6.025 ± 1.041
2.375 ± 0.330
0.256 ± 0.095

0.075 ± 0.042
0.075 ± 0.042
0.026 ± 0.026

0.075 ± 0.042
0.025 ± 0.025
0.026 ± 0.026

3.800 ± 0.712
1.200 ± 0.249
1.051 ± 0.345

0.075 ± 0.055
0
0

1.000 ± 0.275
0.375 ± 0.106
0.128 ± 0.066

130

Table 3.6: Average number of alate aphids observed per plant in each week in each plot in 2010 and 2011.
2010
week
1
2
3
4
5
6
7
8

field
0
0
0
0
0
0
0

open
0
0
0
0
0
-----

2
0
0
0
0.050 ± 0.039
-------

5
0
0
0.025 ± 0.025
4.675 ± 4.178
-------

2
0
0.088 ± 0.044
0.025 ± 0.018
0.025 ± 0.018
0.063 ± 0.027
-------

5
0
0.088 ± 0.044
0.013 ± 0.013
0
0.200 ± 0.156
-------

2011
week
1
2
3
4
5
6
7
8

field
0.385 ± 0.112
0.060 ± 0.022
0.040 ± 0.016
0.060 ± 0.018
0.141 ± 0.036
0.135 ± 0.025
0.085 ± 0.026
0.217 ± 0.044

open
0
0.100 ± 0.042
0.063 ± 0.027
0.013 ± 0.013
0.225 ± 0.073
-------

131

2

Figure 3.1: Predator activity-density and aphid abundance were measured in 1 m plots. Barriers
to coccinellid immigration were constructed with 2 mm opening mesh netting and were raised 2
cm above the ground to allow full access by small and ground-dwelling predators. Open plots (A
& D) had no barriers to immigration. 2-sided plots (B & E) had barriers erected on the east and
west sides of plot. 5-sided plots (C & F) had barriers on all sides. Each field contained two sets
of plots. In sampling plots (A-C), predators were sampled without replacement. In observation
plots (D-F), aphid populations were monitored without disturbing predator populations.

132

Figure 3.2: Mean number of adult coccinellids captured in treatment plots in 2010 and 2011. In 2010, coccinellid immigration was
higher in open and 2-sided than in 5-sided treatment plots. In 2011, coccinellid immigration rates were different in all treatments.
2010

2011

16

16
Open
2-sided
5-sided

Mean adult coccinellids/ trap

14

14

12

12

10

10

8

8

6

6

4

4

2

2

0

0
1

2

3

4

1

Week

2

3
Week

133

4

5

Figure 3.3: Mean number of A. glycines observed per plant each week in soybean fields outside of plots and in open, 2- and 5-sided
treatment plots in 2010 and 2011. Symbols indicate significantly more aphids in (+) 5-sided and 2-sided than in open plots, (*)
between all treatments and (x) in 5-sided than in open or 2-sided plots.
2010

2011

160

160
Open
2-sided
5-sided
Field

140

Aphids/ plant

120

x
140
120

100

100

80

80

60

60

*

40
20

40

+

20

0

0
1

2

3

4

1

Week

2

3
Week

134

4

5

LITERATURE CITED

135

LITERATURE CITED

Allard, C. M., Yeargan, K. V., 2005. Effect of diet on development and reproduction of the
harvestman Phalangium opilio (Opiliones: Phalangiidae). Environmental Entomology.
34, 6-13.
Bolker, B. M., Brooks, M. E., Clark, C. J., Geange, S. W., Poulsen, J. R., Stevens, M. H. H.,
White, J. S. S., 2009. Generalized linear mixed models: a practical guide for ecology and
evolution. Trends in Ecology & Evolution. 24, 127-135.
Brosius, T. R., Higley, L. G., Hunt, T. E., 2007. Population dynamics of soybean aphid and
biotic mortality at the edge of its range. Journal of Economic Entomology. 100, 12681275.
Chang, G. C., Kareiva, P., The case for indigenous generalists in biological control. In: B. A.
Hawkins, H. V. Cornell, Eds.), Theoretical Approaches to Biological Control. Cambridge
University Press, Cambridge, 1999.
Colunga-Garcia, M., Gage, S. H., 1998. Arrival, establishment, and habitat use of the
Multicolored Asian Lady Beetle (Coleoptera: Coccinellidae) in a Michigan landscape.
Environmental Entomology. 27, 1574-1580.
Costamagna, A. C., Landis, D. A., 2006. Predators exert top-down control of soybean aphid
across a gradient of agricultural management systems. Ecological Applications. 16, 16191628.
Costamagna, A. C., Landis, D. A., 2007. Quantifying predation on soybean aphid through direct
field observations. Biological Control. 42, 16-24.
Costamagna, A. C., Landis, D. A., 2011. Lack of strong refuges allows top-down control of
soybean aphid by generalist natural enemies. Biological Control. 57, 184-192.
Costamagna, A. C., Landis, D. A., Brewer, M. J., 2008. The role of natural enemy guilds in
Aphis glycines suppression. Biological Control. 45, 368-379.
Costamagna, A. C., Landis, D. A., DiFonzo, C. D., 2007. Suppression of soybean aphid by
generalist predators results in a trophic cascade in soybeans. Ecological Applications. 17,
441-451.
Debach, P., Rosen, D., 1991. Biological Control by Natural Enemies. Cambridge University
Press, Cambridge.
Desneux, N., O'Neil, R. J., Yoo, H. J. S., 2006. Suppression of population growth of the soybean
aphid, Aphis glycines Matsumura, by predators: The identification of a key predator and
136

the effects of prey dispersion, predator abundance, and temperature. Environmental
Entomology. 35, 1342-1349.
Dixon, 1985. Aphid Ecology. Blackie & Son Ltd., Glasgow.
Dixon, A. F. G., Hemptinne, J. L., Kindlmann, P., 1997. Effectiveness of ladybirds as biological
control agents: Patterns and processes. Entomophaga. 42, 71-83.
Donaldson, J. R., Myers, S. W., Gratton, C., 2007. Density-dependent responses of soybean
aphid (Aphis glycines Matsumura) populations to generalist predators in mid to late
season soybean fields. Biological Control. 43, 111-118.
Fox, T. B., Landis, D. A., Cardoso, F. F., DiFonzo, C. D., 2004. Predators suppress Aphis
glycines Matsumura population growth in soybean. Environmental Entomology. 33, 608618.
Fox, T. B., Landis, D. A., Cardoso, F. F., DiFonzo, C. D., 2005. Impact of predation on
establishment of the soybean aphid, Aphis glycines in soybean, Glycine max. Biocontrol.
50, 545-563.
Gardiner, M. M., Landis, D. A., 2007. Impact of intraguild predation by adult Harmonia axyridis
(Coleoptera: Coccinellidae) on Aphis glycines (Hemiptera: Aphididae) biological control
in cage studies. Biological Control. 40, 386-395.
Gardiner, M. M., Landis, D. A., Gratton, C., DiFonzo, C. D., O'Neal, M., Chacon, J. M., Wayo,
M. T., Schmidt, N. P., Mueller, E. E., Heimpel, G. E., 2009. Landscape diversity
enhances biological control of an introduced crop pest in the north-central USA.
Ecological Applications. 19, 143-154.
Grez, A. A., Prado, E., 2000. Effect of plant patch shape and surrounding vegetation on the
dynamics of predatory coccinellids and their prey Brevicoryne brassicae (Hemiptera:
Aphididae). Environmental Entomology. 29, 1244-1250.
Hajek, A. E., Hannam, J. J., Nielsen, C., Bell, A. J., Liebherr, J. K., 2007. Distribution and
abundance of Carabidae (Coleoptera) associated with soybean aphid (Hemiptera:
Aphididae) populations in central New York. Annals of the Entomological Society of
America. 100, 876-886.
Hannam, J. J., Liebherr, J. K., Hajek, A. E., 2008. Climbing behaviour and aphid predation by
Agonum muelleri (Coleoptera : Carabidae). Canadian Entomologist. 140, 203-207.
Higley, L. G., Boethel, D. J. Eds.), 1994. Handbook of Soybean Insect Pests. Entomological
Society of America, Lanham, MD.
Hodek, I., Honek, A., 1996. Ecology of Coccinellidae. Kluwer Academic Publishers, Dordrecht.

137

Hodek, I., Iperti, G., Hodkova, M., 1993. Long-Distance Flights in Coccinellidae (Coleoptera).
European Journal of Entomology. 90, 403-414.
Hopper, K. R., Aidara, S., Agret, S., Cabal, J., Coutinot, D., Dabire, R., Lesieux, C., Kirk, G.,
Reichert, S., Tronchetti, F., Vidal, J., 1995. Natural Enemy Impact on the Abundance of
Diuraphis-Noxia (Homoptera: Aphididae) in Wheat in Southern France. Environmental
Entomology. 24, 402-408.
Kindlmann, P., Dixon, A. F. G., 1999a. Generation time ratios - Determinants of prey abundance
in insect predator-prey interactions. Biological Control. 16, 133-138.
Kindlmann, P., Dixon, A. F. G., 1999b. Strategies of aphidophagous predators: lessons for
modelling insect predator-prey dynamics. Journal of Applied Entomology-Zeitschrift Fur
Angewandte Entomologie. 123, 397-399.
Kindlmann, P., Dixon, A. F. G., 2001. When and why top-down regulation fails in arthropod
predator-prey systems. Basic and Applied Ecology. 2, 333-340.
Krivan, V., 2008. Dispersal dynamics: Distribution of lady beetles (Coleoptera: Coccinellidae).
European Journal of Entomology. 105, 405-409.
LaMana, M. L., Miller, J. C., 1996. Field observations on Harmonia axyridis Pallas (Coleoptera:
Coccinellidae) in Oregon. Biological Control. 6, 232-237.
Landis, D. A., van der Werf, W., 1997. Early-season predation impacts the establishment of
aphids and spread of beet yellows virus in sugar beet. Entomophaga. 42, 499-516.
Murdoch, W. W., Swarbrick, S. L., Briggs, C. J., 2006. Biological control: lessons from a study
of California red scale. Population Ecology. 48, 297-305.
O'Hara, R. B., Kotze, D. J., 2010. Do not log-transform count data. Methods in Ecology and
Evolution. 1, 118-122.
Ragsdale, D. W., Landis, D. A., Broduer, J., Heimpel, G. E., Desneux, N., 2011. Ecology and
management of the soybean aphid in North America. Annual Review of Entomology. 56,
375-99.
Rhainds, M., Roy, M., Daigle, G., Brodeur, J., 2007. Toward management guidelines for the
soybean aphid in Quebec. I. Feeding damage in relationship to seasonality of infestation
and incidence of native predators. Canadian Entomologist. 139, 728-741.
Rhainds, M., Yoo, H. J. S., Kindlmann, P., Voegtlin, D., Castillo, D., Rutledge, C., Sadof, C.,
Yaninek, S., O'Neil, R. J., 2010. Two-year oscillation cycle in abundance of soybean
aphid in Indiana. Agricultural and Forest Entomology. 12, 251-257.

138

Rutledge, C. E., O'Neil, R. J., Fox, T. B., Landis, D. A., 2004. Soybean aphid predators and their
use in integrated pest management. Annals of the Entomological Society of America. 97,
240-248.
Schmidt, M. H., Lauer, A., Purtauf, T., Thies, C., Schaefer, M., Tscharntke, T., 2003. Relative
importance of predators and parasitoids for cereal aphid control. Proceedings of the
Royal Society B-Biological Sciences. 270, 1905-1909.
Schmidt, N. P., O'Neal, M. E., Singer, J. W., 2007. Alfalfa living mulch advances biological
control of soybean aphid. Environmental Entomology. 36, 416-424.
Seagraves, M. P., 2009. Lady beetle oviposition behavior in response to the trophic environment.
Biological Control. 51, 313-322.
Snyder, W. E., Wise, D. H., 1999. Predator interference and the establishment of generalist
predator populations for biocontrol. Biological Control. 15, 283-292.
van der Werf, W. 1995. How do immigration rates affect predator/prey interactions in field
crops? Predictions from simple models and an example involving the spread of aphidborne viruses in sugar beet. In: S. Toft, W. Riedel, Eds., Arthropod Natural Enmies in
Arable Land I: Density, Spatial Heterogeneity and Dispersal. Aarhus University Press,
Aarhus.
Venables, W. N., Dichmont, C. M., 2004. GLMs, GAMs and GLMMs: an overview of theory for
applications in fisheries research. Fisheries Research. 70, 319-337.
Woltz, J. M., Isaacs, R., Landis, D. A., 2012. Landscape structure and habitat management
differentially influence insect natural enemies in an agricultural landscape. Agriculture
Ecosystems & Environment. 152, 40-49.

139

CHAPTER IV: Diel variation in natural enemy communities of Aphis glycine in soybean.

Abstract
Nocturnally-active predators can make important contributions to insect pest suppression
in agroecosystems. In particular, nocturnal predators have been shown to be responsible for
significant predation of herbivores within soybean fields. In spite of this fact, much of the
existing research on A. glycines natural enemies has focused on members of the diurnal predator
community. Here, we assess diel variation of the predator community and A. glycines predation
events in soybean with vacuum samples, direct observations, and video observations. The
majority of predation events observed (95%) occurred during daylight hours and primarily by
anthocorids (81%), followed by chrysopid larvae (10%), and one each by a coccinellid, and a
predatory mite. The sole noctural predation event observed was a by a spider. We found that
anthocorids and nabids were most active in the afternoons and carabids were more active at
night. Spiders and anthocorids were the most abundant predatory arthropods captured in vacuum
samples. In contrast, on video samples, lacewing larvae were most active during the day and
lacewing larvae, spiders, opiliones and carabids were similarly active at night. The differences
between the two sample methods suggests that vacuum samples may represent predator
abundance more accurately, while video data can be used to better assess relative time spent
foraging.
Introduction
The soybean aphid, Aphis glycines, is an invasive insect pest from China with the
potential to cause significant yield losses in soybean (Difonzo and Hines 2002). Since A. glycines
was first discovered in the U.S. in 2000, it has been the subject of much study (Ragsdale et al.
140

2011). A number of studies have focused on identifying the natural enemies that attack A.
glycines (Rutledge et al. 2004, Allard and Yeargan 2005, Mignault et al. 2006, Hajek et al. 2007,
Kaiser et al. 2007, Pike et al. 2007, Hannam et al. 2008, Noma and Brewer 2008) and their
relative efficacy (Fox et al. 2005, Desneux et al. 2006, Costamagna and Landis 2007). However,
most of these studies have focused on the diurnal predator community within soybean fields. For
example, Costamagna and Landis (2007), conducted direct observations of soybean aphid
predation within soybean fields, between 10:00 and 21:00 h.
In contrast, little work has examined the effect of nocturnal natural enemies on soybean
aphid populations, despite the fact that these taxa have been demonstrated to make important
contributions to pest suppression in soybean fields. Nocturnally-active nabids were the most
abundant predator and were responsible for up to half of Helicoverpa zea egg predation observed
in soybean fields in Kentucky (Pfannenstiel and Yeargan 2002). Similarly, nocturnal spiders
were responsible for 98% of lepidopteran egg predation in soybean fields in Texas (Pfannenstiel
2008). Nocturnally-active spiders have been observed at similar abundances and foraging at
similar frequencies as diurnally-active coccinellid beetles in Michigan soybean fields (L.
Petersen and J.M. Woltz unpub. data). Spiders can be effective natural enemies and an important
component of the natural enemy guild (Riechert and Lockley 1984, Sunderland 1999) and are
capable of suppressing aphid populations (Gavish-Regev et al. 2009). Additionally, spiders can
cause trophic cascasdes. Higher spider density can lead to more pest consumption and lower leaf
damage in soybean (Carter and Rypstra 1995, Rypstra and Carter 1995).
Multiple sampling methods are available for assessing natural enemy communities. The
efficacy of different methods varies with the vagility and mode of dispersal of taxa sampled
(Schmidt et al. 2008). Additionally, various trade-offs exist between spatial and temporal

141

coverage of sampling methods and the type of information each can generate. Traditional
sampling methods such as pitfall traps, sweep netting, and sticky traps can be employed
repeatedly over large areas and can provide information about natural enemy community
composition, but provide no information about the relative efficacy of various taxa. Direct
human observation can elucidate specific interactions, but is typically limited in the length and
number of observations possible. In contrast to these temporal limitations, video observations
can be made continuously, although over limited spatial areas (Grieshop et al. 2012).
Given the potential for significant soybean aphid suppression from the nocturnal
community, our objectives were to: 1) assess diel variation in the predator community in
soybean; 2) contrast rates of soybean aphid predation during night and day. Additionally, our
third objective was to 3) compare the predator communities described by vacuum sampling,
direct observations, and video observations.
Methods: Direct observation study
Study locations. This study was conducted between 6 June and 13 July 2012 in three
conventionally-managed soybean fields in southern Michigan (Table 4.1). The study was
conducted early in the growing season (soybean plants between growth stages of V1-V5)
because early season dynamics are important for preventing outbreaks (Rutledge et al. 2004). No
fields planted with systemic insecticide-treated seed were used in this study. Agriculture in this
region consists primarily of annual field crops (corn, soy, wheat) and perennial forage crops
(pasture, alfalfa), while non-crop areas are primarily deciduous forest. Three separate 24 h
predation studies were conducted in each field for a total of n=9 surveys.
Plot establishment. In each field, four 10 x 10 m plots were established for soybean
aphid predation surveys (Figure 4.1). Each plot encompassed 14 rows of soybean planted on a

142

row spacing of 76.2 cm. Two “edge” plots were situated between 0 and 10 m into the field, while
two “interior” plots were situated between 50 and 60 m into the field. The plots were at least 50
m from other field edges in all other directions.
Aphid infestation. Aphis glycines used in this study were reared under conditions of
predator exclusion in Lumite field cages (Lumite Inc, Baldwin, GA) in soybean fields at the
Michigan State University Department of Entomology Research Farm. Soybean leaves
containing aphids were removed from rearing cages and transferred to experimental fields in
coolers.
Within each plot, ten plants were randomly selected for aphid infestation (Figure 4.1).
First, ten numbers between one and fourteen were randomly generated without duplication to
select rows within plots. Then, ten numbers between one and ten were randomly selected to
determine the position in meters of the plant within the row. Ten aphids of mixed age classes
were transferred to the selected plants using a fine-haired paint brush. Aphids were placed in clip
cages on the underside of the upper most fully-opened leaf and allowed to settle for at least 1 h
before cages were removed. Plants were searched for naturally-occurring soybean aphids at
infestation, and any found were removed.
During each monitoring period, one interior and one edge plot were infested between
18:00 and 20:00 h. Clip cages were removed at dusk (approx. 21:00-22:00 for each of our
surveys) to begin nocturnal surveys. The other interior and edge plots were infested between
3:00 and 5:00, and clip cages were removed at sunrise (approx. 6:00 for each of our surveys) to
begin diurnal surveys.
Additionally, ten “caged” plants from which predators were excluded were used to
account for aphid mortality not due to predation. Five predator exclusion cages were located

143

between the two interior plots and five between the two edge plots (Figure 4.1). Exclusion cages
consisted of two 50 cm flexible metal rods bent over the plant in crossed arches to create a
frame. A 19 L paint strainer bag was placed around the frame and buried in the soil to a depth of
5 cm. The tip of the bag was cut to allow aphid counts and tightly retied with a twist-tie in
between surveys. Five caged plants were infested prior to dusk and five additional plants were
infested prior to dawn, at the same time as the interior and edge plots.
Aphid surveys. The number of aphids on each plant was counted at clip cage removal to
account for reproduction or mortality that might have occurred within the clip cages. Aphids on
plants infested before dusk were counted every 3 h for 24 h at 21:00, 0:00, 3:00, 6:00, 9:00,
12:00, 15:00, 18:00, and 21:00. Aphids on plants infested before dawn were counted on each
plant at 6:00, 9:00, 12:00, 15:00, 18:00, and 21:00. With three replicate surveys within three
fields and two sampling locations (edge and interior) and two infestation periods (pre-dusk and
pre-dawn) and ten plants within each plot, 360 plants were surveyed at each time period (3 x 3 x
2 x 2 x 10=360), with the exception of the 0:00 and 3:00 time periods which were not replicated
in pre-dawn infested plots (3 x 3 x 2 x 10=180). The 6:00 samples approximately coincided with
dawn and the 21:00 samples coincided with dusk. At night, aphids were counted using Petzl
Tactikka Plus LED headlamps with red filters to minimize disturbance of nocturnally-active
insects.
Predator sampling. During the predation surveys, prior to counting aphids, plants were
visually inspected and any predatory insects or arachnids observed on the infested leaves were
recorded as well as whether or not they were observed feeding on A. glycines. Additionally,
predators were sampled on foliage three times/ d using a leaf blower (BG 56 C-E; Stihl,
Waiblingen, Germany) used in reverse as a vacuum to suck insects off of plants and into fine

144

mesh collection bags. At 10:00, 16:00, and 22:00, field workers walked along five 20 m
transects outside of the survey plots with the vacuum tube placed at the tops of the soybean
plants. Insects sucked into the tube were captured in 1-gallon mesh paint strainer bags placed
inside the mouth of the tube and secured with rubber bands. Transects outside of the edge plots
ran from 0-20 m into the field. Transects outside of interior plots ranged from 40-60 m into the
field. Each transect was at least 50 m from the plot to avoid removing predators that might have
otherwise migrated into the plots during subsequent sampling periods, and different transects
were used after each sampling period. A single mesh bag was used for all five transects during a
sampling period, unless excess leaf debris reduced suction. When multiple bags were used,
samples were combined prior to sorting. After each collection, the mesh bags were secured with
rubber bands and placed into a 18.9 L bucket with a lid, containing plaster of Paris soaked in
ethyl acetate to kill the insects. Insects and arachnids were identified to family or order in the
laboratory. Voucher specimens were deposited in the Albert J. Cook Arthropod Research
Collection at Michigan State University (Table 4.2).
Temperature. A thermometer (Extech 445713, Extech Instruments, Nashua, NH) was
placed at soybean canopy height in the center of each plot to record temperature. Readings were
taken after each predation survey was completed, so as not to disturb predator feeding.
Statistical analysis. Aphid surveys: The change in aphids over each 3 h time period was
analyzed using a general linear mixed model (PROC GLIMMIX; SAS 9.2, SAS Institute, Cary,
NC). Plant location (Interior, Exterior, or Caged) and time period were fixed effects and site X
survey was treated as a random block. Change in aphid numbers was not treated as repeated
measures. Direct population counts are typically treated as repeated measures because the
population in one time period is dependent upon the population in the previous time period. In

145

contrast, the change in aphid numbers during one time period is not expected to be dependent
upon the change in the previous one. Additionally, a plot of the model residuals by time period
showed no patterns, indicating that repeated measures were unnecessary to account for temporal
autocorrelation.
Vacuum samples: Analysis of Similarity (ANOSIM, PRIMER 6.0) was used to describe
differences in the predator community by time of day (morning, afternoon, evening) and site.
Preliminary analyses indicated no community differences between edge and interior samples.
First, a Bray-Curtis Similarity Index was calculated for each pair of samples, indicating the
degree of similarity in community composition between the samples. A value of zero indicates
complete dissimilarity (no common species) and a value of 100 represents completely similar
samples (both samples have the same species in the exact same abundances). The similarity
between a given sample and every other sample is ranked, and this is repeated for each sample.
ANOSIM is conducted on the rank similarities, with the null hypothesis that site and time of day
have no effect on the rank similarities between samples. An R statistic is calculated (analogous to
an F statistic in ANOVA). The significance of the R statistic is calculated by randomly
reassigning samples to different sites and times of day and recalculating the R statistic. This is
repeated 999 times, and the number of times that an R statistic greater than or equal to the one
calculated for the actual data is achieved by chance is divided by the number of permutations to
achieve a significance level.

146

Methods: Video observation study
Study location. This experiment was conducted in a 1 ha soybean field on the Michigan
State University Department of Crop and Soil Sciences Research Farm (Table 4.1). Eight
soybean plants within the field were selected for around-the-clock monitoring of aphid colonies
and predation events from 22 to 29 June 2012.
Aphids surveys. The uppermost trifoliate of each focal soybean plant was infested with
ten soybean aphids of mixed age class at the beginning of the experiment. After infestation,
aphids were enclosed in a clip cage for 1 h to allow them time to settle and begin feeding. Clip
cages consisted of a 1 cm section of 1.8 cm PVC pipe. One opening of the pipe was covered in
2

fine-mesh brass screen to allow air exchange. The other was enclosed with a 1 cm plastic
microscope slide. The two coverings were attached to the PVC with a metal hair barrett,
allowing the cage to be attached to plant leaves. The number of aphids on each soybean plant
was counted every 3 h for 7 d at 0:00, 3:00, 6:00, 9:00, 12:00, 15:00, 18:00, and 21:00 h.
Video observations. Digital surveillance equipment was used to record predators
foraging on soybean plants 24 hr /d, adapting the methods of Greishop et al. (2012). Eight waterproof outdoor video surveillance cameras (QOCDC36, Q-See.com, Anaheim, CA) were
deployed in a soybean field. The cameras were distributed in a semi-circle around the central
digital video recorded (DVR; Figure 4.2). Cameras were suspended 10 cm above the tops of
focal soybean plants from PVC frames placed perpendicular to soybean rows and anchored with
rebar (Figure 4.2). The cameras were connected by coaxial cables to a multi-channel H.264
Network DVR (Security Hardware Store, LLC, Boise, ID), where video was recorded onto a
500GB Seagate SV35.5 internal hard drive. The cameras and DVR were powered by a pair of
Die Hard RV/ marine batteries linked in series. The DVR and batteries were housed in a 75.7 L
147

plastic storage tub with lid to keep the equipment dry. Holes were cut in the sides of the tub to
feed power and coaxial cables through and to allow for ventilation to prevent the DVR from
over-heating.
Cameras were moved up the frames as needed to maintain ten cm height above soybean
canopy. Cameras were pointed at the uppermost trifoliate, and adjusted twice daily to account for
changes in the plant’s position. Video could be viewed as it was being recorded on a 18 cm
digital LCD TV (Haier) connected to the output channel of the DVR. This allowed us to confirm
that we were recorded the desired part of the plant.
Video processing. After 7 d of recording, the video files were transferred from the hard
drive in the DVR to an external hard drive (Fantom GreenDrive, MicroNet, Torrance, CA) and
converted to AVI files for viewing. Lab workers watched four videos at a time at 4x speed, to
identify frames with insect activity, after which, insect activity was viewed one video at a time at
1x speed. The taxonomic identity (family or order), arrival time, behavior, and departure time of
each insect was recorded. Arthropods of the same taxonomic group that left and re-entered the
screen within five minutes were considered a single individual. An arthropod of the same
taxonomic group that re-entered the screen after a period of greater than five minutes was
recorded as a separate individual.
Statistical analysis. Aphid surveys: The change in aphids over each 3 h time period was
compared using a general linear mixed model (PROC GLIMMIX; SAS 9.2, SAS Institute, Cary,
NC). Time period was treated as a fixed effect and day was used as a random block. As in the
direct observation study, change in aphid numbers was not treated as repeated measures and a
plot of the model residuals by time period showed no evidence of temporal autocorrelation.

148

Vacuum samples: ANOSIM (PRIMER 6.0) was used as described above to examine
differences in the predator community by time of day (morning, afternoon, evening).
Video data: ANOSIM (PRIMER 6.0) was also used to examine differences in the
predator community observed on the video at night (predators viewed between 21:00 and 6:00)
and day (predators viewed between 6:00 and 21:00).
Results: Direct observation study
Colony establishment. Establishment success varied among infestation periods (Figure
4.3). Establishment was more successful in the evening versus pre-dawn infestation, but in both
cases was similar between caged and uncaged plants, suggesting that initial establishment
success or failure was not directly related to predation. Six hours after pre-dusk infestation, aphid
populations on both caged and uncaged plants decreased in 32.5% and 35.6% of cases, remained
the same in 22.5% and 28.3% of cases, and increased in 45% and 35.6% of cases. However, for
pre-dawn infested plants, initial mortality was much greater. Aphid populations on caged and
uncaged plants decreased in 67.5% and 68.3% of cases, remained the same in 20% and 15%, and
increased in only 12.5% and 16.7% of cases.
Aphid surveys. On average, during the first 24 h after infestation, aphids either
decreased or changes were not significantly different from zero in all time periods. There was no
effect of location within field on change in aphid numbers (F2,15=1.22, p=0.32), suggesting that
aphid population changes did not differ between exterior and interior plots. Additionally,
population changes did not differ between caged and uncaged plants. There were marginal
differences in aphid decreases between time periods (F2,56=1.97, p=0.08). There were
marginally larger aphid decreases from 6:00-9:00 than from 18:00-21:00 and 21:00-0:00, and
significantly higher aphid decreases from 6:00-9:00 than from 3:00-6:00, 12:00-15:00, and
149

15:00-18:00 (Table 4.3). There were significant time period X location interactions
(F14,105=2.01, p=0.02); however, there were no consistencies for location effects across time
periods.
Temperature. Temperatures ranged from 11.8 to 41.3 C. Recorded temperatures were
within physiological tolerance limits of A. glycines (McCornack et al. 2004, McCornack et al.
2005) except for the periods from 12:00-15:00 and 15:00-18:00 in the second and third surveys
in Westphalia and all three surveys in Mason. Temperatures during these periods exceeded 35
C. However, given that aphid decreases during the time periods 12:00-15:00 and 15:00-18:00
were actually significantly lower than from 6:00-9:00, it is unlikely that high temperatures
explain aphid mortality. Additionally, removing data points from times in which temperatures
exceeded 35 C did not alter statistical outcomes.
Predator observations. Predators were observed infrequently during our direct
observations. During 2,520 observations, we directly observed 68 predators foraging on our
infested soybean plants during aphid surveys. Anthocorid nymphs and adults were the most
abundant predator observed (n=44 individuals), followed by spider adults and immatures (8
individuals) and coccinellid adults (6 individuals; Table 4.4). With the exception of anthocorids,
which were most frequently observed in the afternoon hours, other predators were not observed
at sufficient frequency to ascertain diel patterns.
Additionally, we directly observed 22 predation events on soybean aphid. Of these
events, seventeen were by anthocorids, two by chrysopid larvae, one by a coccinellid, one by a
spider and one by a predatory mite. All directly observed predation events occurred during the
daylight hours, except for one predation event by a spider which occurred at 0:00 hours.

150

Vacuum samples. Spiders and anthocorids were the most abundant predatory arthropods
captured in vacuum samples. In 49 vacuum samples, we collected 780 anthocorids and 371
spiders. Eight coccinellids were collected. The predator community sampled differed
significantly between all sites (R=0.492, p=0.001) and by time of day (R=0.182, p=0.002).
Predator communities were not different in the morning (10:00) and afternoon (16:00) samples
(R=0.080, p=0.16), but were significantly different between night samples (22:00) and morning
or afternoon samples (R=0.173, p=0.02; R=0.297, p=0.003, respectively). In general, more
anthocorids and nabids were collected in the afternoon (Table 4.5). Elaterids were found only at
night, and coccinellids and staphylinids were found only during the day (Table 4.6).
Results: Video observation study
Aphid surveys. There was a significant effect of time on the amount of change in aphid
numbers (F7,39=5.24, p=0.0003). Aphids increased by an average of 7.29 ± 1.37 individuals/
plant between 6:00 and 9:00, while aphid changes during all other time periods were not
significantly different from zero (Table 4.7).
Temperature. Temperatures ranged from 7.9 to 40.85 C. Recorded temperatures were
within physiological tolerance limits of A. glycines (McCornack et al. 2004, Mccornack et al.
2005) except for the periods from 12:00-15:00 and 15:00-18:00 on 26-28 June and 9:00-12:00 on
28 June. Temperatures during these periods exceeded 35 C. However, removing data points
from times in which temperatures exceeded 35 C did not alter statistical outcomes.
Vacuum samples. Spiders and anthocorids were the most abundant predatory arthropods
captured in vacuum samples (Table 4.8). In 18 vacuum samples we collected 143 anthocorids
and 247 spiders. Six coccinellids were captured. The predator community differed significantly
by time of day (R=0.206, p=0.025). Predator communities were not different in the morning
151

(9:00) and afternoon (15:00) samples (R=0.181, p=0.12), but were significantly different
between night (0:00) and morning or afternoon (R=0.259, p=0.073; R=0.206, p=0.031,
respectively). In general, there were more anthocorids and nabids in the afternoon (Table 4.5)
and coccinellids were only found during the day (Table 4.8). Unlike the vacuum samples from
the direct observation study, no elaterids, staphylinids, or lampyrids were captured (Table 4.8).
Video data. Due to intermittent technical difficulties with our video recording
equipment, we were not able to record continuously during the observation period. However, we
have three to six replicate days of video data following eight plants for each 3 h time period
(Tables 4.9, 4.10), providing a total of 792 hours of video watched. Chrysopid larvae were the
most common predator observed (n=41), followed by spiders (n=22), carabids (n=29), and
coccinellids (n=8). Additionally, 35 coleopterans that could not be identified to family level were
observed. The predator community observed on video did not differ between time periods
(R=0.032, p=0.31), or broadly between night and day (R=0.029, p=0.29), although there were
more carabids observed during the night (Table 4.5). However, the data showed similar patterns
to that of Petersen and Woltz (in prep.), with opiliones only present at night, and coccinellids
only present during the day (Tables 4.9, 4.10). Ants and lacewing larvae were the most abundant
predators observed during the day. Ants, carabids, spiders, opiliones, and lacewing larvae were
the most abundant predators observed at night.
Discussion
Aphid colony establishment. Given the demonstrated importance of nocturnal predators
on other pests in soybean (Pfannenstiel and Yeargan 2002, Pfannenstiel 2008), we expected to
observe significant soybean aphid reduction during night hours. Instead, most directly observed
predation events occurred during daylight hours. Unfortunately, the nature of our directly

152

observed aphid data prevents us from attributing aphid population changes to predation.
Specifically, the lack of difference between aphid population patterns on caged and open plants
in the direct observation study suggests that aphid population declines in this study were not
solely the result of predation events. While we have successfully transferred aphids for
experiments many times in the past (Fox et al. 2004, 2005, Costamagna and Landis 2006,
Costamagna et al. 2007, Costamagna et al. 2008, Gardiner et al. 2009, Woltz et al. 2012, Woltz
and Landis 2013), we have typically followed the resulting aphid populations over weekly
periods. As such, we have rarely monitored establishment success rate in the short-term. In one
instance in which aphid populations were monitored within 24 h of infestation, there was
approximately 25% mortality in exclusion cages from which aphids could not emigrate and to
which predators had no access (Fox 2002). The average 24 h declines in adult aphids on our predusk infested plants was about 40-45% regardless of cage status, indicating high initial mortality.
We suspect that the demands of our study, i.e. transferring large numbers of aphids in a short
time period, may have led to this increased mortality and masked any potential predator effects
in the 24 h observation period. In contrast, the video study, which was conducted over 7 d,
revealed a diel pattern in aphid population change with aphids increasing in the early daylight
hours and declining or remaining steady thereafter.
Predator sampling method comparison. Our study does allow us to assess diel patterns
in soybean predator communities and to compare methods of predator sampling. Video data
suggest that carabids are more active at night. Vacuum samples and direct observations suggest
that anthocorids are more active in the afternoons. Additionally, vacuum samples from the two
different studies showed similar statistical patterns and trends for several taxa. In both, spiders
and anthocorids were the most abundant predatory taxa collected. Similarly, in both, coccinellids

153

were only captured during the day, and anthocorids and nabids were more abundant in afternoon
samples. The major differences, including the absence of some taxa and differences in diel
patterns of others, could reflect differences in the broader predator community over space (i.e.
the samples were collected from sites spread across lower Michigan) or time (the samples were
collected during different periods of the summer growing season). Additionally, differences
occurring at the species level could influence overall diel patterns.
In contrast to the vacuum samples, spiders and anthocorids were not the most abundant
predatory arthropods observed on the video. Instead, ants and chrysopid larvae were the most
abundant taxa observed during the day and ants, chrysopid larvae, carabids, spiders and opiliones
were observed most frequently at night. These differences could be due to multiple factors. In the
case of anthocoridae, these tiny insects (1-5 mm) move relatively slowly and are very difficult to
identify on video. Therefore, they were likely not recorded in exact proportion to their actual
abundance. For the other, larger, easier to identify taxa, the differences could reflect differences
in relative abundance versus time spent foraging. Chrysopid larvae may have had lower overall
abundances within the field, but would be captured more frequently on video if they spent more
time foraging than more abundant taxa. While this may not reflect their true abundances as well
as a vacuum sample, this type of foraging-activity data may be more relevant for biological
control since it could better reflect the likelihood that a natural enemy will encounter a prey item.
Regardless of the sampling methods used, the relative efficacy of predatory taxa must be
considered when assessing the predator community. Single, particularly effective natural enemy
species can be responsible for the majority of pest suppression in simplified agricultural systems
(Hawkins et al. 1999, Straub and Snyder 2006, Frank et al. 2007). Therefore, without estimates
of the relative efficacy of different predators, observation methods assessing the numbers of

154

predators on plants could underestimate the importance of highly mobile predators to soybean
aphid suppression (Schmidt et al. 2008). For example, infrequent visits by mobile predators may
be sufficient to suppress soybean aphid populations in years with low aphid pressure (Woltz and
Landis 2013).
Coccinellids were sampled infrequently relative to resident predators like spiders and
anthocorids. The low coccinellid activity-density is not surprising given that aphid densities
within fields were low (Woltz pers. obs.). While at high aphid densities coccinellids will begin to
aggregate and to initiate ovipositon, at low aphid densities coccinellids disperse through soybean
fields consuming aphids they encounter. This predation from infrequently occurring coccinellids,
combined with predation from resident predators, is thought to be sufficient to suppress insipient
aphid colonies and prevent them from contributing to outbreaks at the field-level (Costamagna
and Landis 2011). Anthocorids, in contrast, were abundant by most sampling measures.
Anthocorids are present in soybean fields throughout the growing season and consume a variety
of soybean herbivores, with the result that anthocorid abundances within soybean fields are less
closely linked to soybean aphid density than those of coccinellids (Rutledge et al. 2004). In fact,
the ability of anthocorids to be present in soybean fields prior to soybean aphid arrival allows
them to help prevent early-season aphid population increases (Rutledge et al. 2004, Rutledge and
O'Neil 2005).
Use of video data in natural enemy surveys. For the aphid populations followed in the
video study, there are no paired caged plants to assess how well the populations were
established. Ideally, the video data should allow us to match up population changes on individual
plants with predators foraging on those plants during that period of time. However, due to
technical difficulties, video was not available for every period of time on every plant. For the

155

video that is available, large population decreases in aphids occur during some 3 h blocks in
which no predators were recorded. This does not mean that no predators were present on the
plant at those times. In some cases, arthropods were observed but were not recorded because
identification even to order was impossible. In other cases, aphids dispersed from the focal point
of the camera to other parts of the plant. In these cases, predation could have happened offscreen. Given these limitations, I conclude that video surveillance has limited usefulness for
recording predation of mobile prey on moving plants. In contrast, others (Merfield et al. 2004,
Frank et al. 2007, Grieshop et al. 2012) have had excellent success recording predation events of
immobilized prey on non-moving surfaces such as Petri dishes or tree limbs. Despite our
difficulties in observing aphid predation, video surveillance remains useful as a tool to survey the
predator community, even in a habitat with moving focal plants. Video data can give more
detailed information about frequency of foraging on plants, in comparison with more traditional
sampling methods (e.g. pitfall traps, sticky traps, vacuum sampling), which can indicate relative
abundance but do not necessarily provide information about time spent on plants. We found
differences in the relative abundances of predators indicated by the vacuum samples and their
relative foraging frequencies as observed on video. Similar differences have been found for
ground-foraging predators. In a study in which carabids and chilopodans represented the
majority of captures in pitfall traps, these taxa were infrequently recorded consuming sentinel
prey (Grieshop et al. 2012). Additionally, video may provide better evidence of true diel activity
patterns as vacuum samples may capture inactive individuals that remain in the habitat during
their inactive period.

156

APPENDIX

157

Chapter IV Tables and Figures

158

Table 4.1: Field locations and characteristics for direct and video observation of A. glycines predator communities and predation
events.

Study
Direct observation

Location
Frankenmuth, MI
Westphalia, MI
Mason, MI

Coordinates
43 24' 37.77" N 83 42' 06.90" W
42 51' 57.77" N 84 48' 04.87" W
42 37' 38.58" N 84 25'50.15" W

Video observation

East Lansing, MI

42 42' 51.86" N

84 27' 52.31" W

159

Size (ha)
30
14
32

Survey
dates
6-12 June
16-20 June
10-13 July

Vegetative
stage
V1-V2
V1-V2
V5

1

22-29 June

V3

Table 4.2: Record of deposition of voucher specimens for Chapter IV. The specimens listed
below have been deposited in the Albert J. Cook Arthropod Research Collection, Michigan State
University (MSU) museum as samples of those species or other taxa, which were used in this
research. Voucher recognition labels bearing the voucher number 2013-08 have been attached or
included in fluid preserved specimens.
Order
Coleoptera
Coleoptera
Coleoptera
Coleoptera
Hemiptera
Hemiptera
Neuroptera
Opiliones
Araneae

Family
Lampyridae
Elateridae
Carabidae
Staphylinid
Nabidae
Anthocoridae

Life Stage
Adult
Adult
Adult
Adult
Adult
Adult
Larvae
Adult/nymph
Adult/nymph

Quantity
3
3
3
3
3
3
5
5
5

160

Preservation
Pinned
Pinned
Pinned
Pinned
Pinned
Pinned
70% EtOH
70% EtOH
70% EtOH

Table 4.3: Means ± S.E. of aphid change during each time period of the direct observation study.
Bold indicates means that were significantly different from zero. Different letters indicate means
which are significantly different from one another.

Time period
6:00-9:00
9:00-12:00
0:00-3:00
18:00-21:00
21:00-0:00
15:00-18:00
3:00-6:00
12:00-15:00

Mean aphid change ±
SE
-1.16 ± 0.26
-0.67 ± 0.26
-0.52 ± 0.28
-0.51 ± 0.26
-0.43 ± 0.28
-0.32 ± 0.26
-0.23 ± 0.28
0.10 ± 0.26

A
AB
AB
AB
AB
B
B
B

161

Table 4.4: Predators directly observed during 24 h aphid surveys. Observations were made 180 times at 0:00 and 3:00 h, and 360
times at all other time periods.

Araneae
Opiliones
Cantharid
Carabidae
Coccinellid
Lampyrid
Anthocoridae
Chrysopidae larva
Syrphidae larva

6:00
1
1
0
2
0
1
5
0
1

9:00
0
0
1
0
1
0
5
0
0

12:00
1
0
0
0
1
0
9
0
0

15:00
1
0
0
0
0
0
11
0
0

18:00
0
0
0
0
2
0
7
2
0

162

Day total
3
1
1
2
4
1
37
2
1

21:00
0
0
0
0
0
0
1
0
0

0:00
4
0
0
0
1
0
3
0
0

3:00
1
0
1
1
1
0
3
0
0

Night total
5
0
1
1
2
0
7
0
0

Table 4.5: Diel comparisons in predatory taxa from vacuum and video samples.

Study
Vacuum samples, direct
observation study

Taxon
anthocoridae

Diel comparison
afternoon > morning

t46=2.09, p=0.04
t46=2.32, p=0.02

nabidae

afternoon > night

t46=2.08, p=0.04

anthocoridae

afternoon > morning

t15=1.88, p=0.08

afternoon > night

t15=2.48, p=0.03

afternoon > morning

t15=1.89, p=0.08

afternoon > night

Vacuum samples, video
observation study

afternoon > night

t15=2.42, p=0.03

night > day

t262=3.00, p=0.003

nabidae
Video samples

carabidae

163

Table 4.6: Mean ± S.E. number of predator individuals collected in vacuum samples of the direct observation study at morning
(10:00), afternoon (16:00), and night (22:00).

Araneae
Opiliones
Carabidae
Coccinellidae
Elateridae
Lampyridae
Staphylinidae
Anthocoridae
Nabidae
Chrysopidae larvae

Morning
6.73 ± 2.72
1.33 ± 0.60
0.07 ± 0.07
0.33 ± 0.16
0±0
0.13 ± 0.09
0.60 ± 0.40
11.4 ± 2.96
0.40 ± 0.16
0.47 ± 0.34

Sample times
Afternoon
7.06 ± 1.50
2.61 ± 1.14
0.33 ± 0.23
0.17 ± 0.12
0±0
0.06 ± 0.06
0.28 ± 0.13
24.78 ± 5.53
1.72 ± 0.88
0.67 ± 0.27

Night
8.94 ± 1.47
1.44 ± 0.68
0±0
0±0
0.69 ± 0.20
0.25 ± 0.25
0±0
10.19 ± 2.30
0.38 ± 0.26
0.69 ± 0.36

164

Table 4.7: Mean ± S.E. of aphid colony size change during each time period of the video observation study. Means shown in bold are
significantly different form zero. Different letters indicate means which are significantly different from one another.
Time period
6:00-9:00
9:00-12:00
18:00-21:00
3:00-6:00
21:00-0:00
12:00-15:00
15:00-18:00
0:00-3:00

Mean aphid change ± SE
7.29 ± 1.37
0.52 ± 1.37
0.09 ± 1.37
-1.17 ± 1.48
-1.3 ± 1.37
-1.52 ± 1.37
-2.05 ± 1.37
-2.67 ± 1.48

A
B
B
B
B
B
B
B

165

Table 4.8: Mean ± S.E. number of predator individuals collected in vacuum samples during the video observation study at morning
(9:00), afternoon (15:00) and night (0:00).

Araneae
Opiliones
Carabidae
Coccinellidae
Elateridae
Lampyridae
Staphylinidae
Anthocoridae
Nabidae
Chrysopidae larvae

Morning
24.75 ± 2.02
0.25 ± 0.25
0±0
0.5 ± 0.29
0±0
0±0
0±0
4.75 ± 3.47
1.00 ± 0.71
0.75 ± 0.75

Sample times
Afternoon
13.57 ± 5.19
0.43 ± 0.30
0.14 ± 0.14
0.57 ± 0.43
0±0
0±0
0±0
14.71 ± 5.65
8.71 ± 3.52
0.14 ± 0.14

Night
7.57 ± 1.62
0.14 ± 0.14
0.57 ± 0.43
0±0
0±0
0±0
0±0
3.00 ± 1.38
0.71 ± 0.42
1.00 ± 1.00

166

Table 4.9: Mean ± S.E. number of predator individuals observed in daylight hours during each three hour period of video surveillance.

# days video
Aranae
Opiliones
Carabidae
Coccinellidae
Coccinellidae larvae
Lampyridae
Anthocoridae
Nabidae
Chrysopidae
Chyrsopidae larvae

6-9
3

9-12
3

12-15
4

15-18
6

18-21
5

Day total

0±0
0±0
0.08 ± 0.06
0±0
0±0
0.04 ± 0.04
0.04 ± 0.04
0±0
0±0
0.04 ± 0.04

0.21 ± 0.10
0±0
0.08 ± 0.06
0.21 ± 0.10
0±0
0±0
0±0
0.08 ± 0.06
0.04 ± 0.04
0.38 ± 0.19

0.13 ± 0.13
0±0
0.06 ± 0.04
0±0
0±0
0.06 ± 0.04
0±0
0±0
0±0
0.13 ± 0.07

0.04 ± 0.03
0±0
0±0
0.02 ± 0.02
0±0
0.04 ± 0.03
0.06 ± 0.04
0.02 ± 0.02
0±0
0.13 ± 0.06

0.03 ± 0.03
0±0
0.05 ± 0.03
0.05 ± 0.03
0.03 ± 0.03
0±0
0.15 ± 0.09
0±0
0.03 ± 0.03
0.25 ± 0.10

0.07 0.03
0±0
0.05 ± 0.02
0.05 ± 0.02
0.01 ± 0.01
0.03 ± 0.01
0.06 ± 0.02
0.02 ± 0.01
0.01 ± 0.01
0.18 ± 0.04

167

Table 4.10: Mean ± S.E. number of predator individuals observed in night hours during each
three hour period of video surveillance.

# days video
Araneae
Opiliones
Carabidae
Coccinellidae
Coccinellidae larvae
Lampyridae
Anthocoridae
Nabidae
Chrysopidae
Chyrsopidae larvae

21-0
5

0-3
4

3-6
3

Night total

0.10 ± 0.05
0±0
0.20 ± 0.13
0±0
0±0
0.03 ± 0.03
0±0
0.10 ± 0.05
0.05 ± 0.03
0.08 ± 0.04

0.06 ± 0.04
0.19 ± 0.08
0.31 ± 0.13
0±0
0±0
0±0
0±0
0.03 ± 0.03
0±0
0.22 ± 0.12

0.17 ± 0.10
0.21 ± 0.10
0.13 ± 0.07
0±0
0±0
0±0
0.04 ± 0.04
0±0
0.04 ± 0.04
0.04 ± 0.04

0.10 ± 0.03
0.11 ± 0.04
0.22 ± 0.07
0±0
0±0
0.01 ± 0.01
0.01 ± 0.01
0.05 ± 0.02
0.03 ± 0.02
0.11 ± 0.04

168

Figure 4.1: Diagram of plot set-up for observation study. A) Two exterior plots were situated
from 0 – 10 m into the field. B) Two interior plots were situated from 50 – 60 m into the field.
All plots were 50 m away from all other field edges. C) Expanded view of plot demonstrated
random selection of plants for aphid infestation. D) Caged plants located in between plots were
used to account for mortality in the absence of predation.

169

Figure 4.2: Video observation set-up. A) Aphids and natural enemies were recorded with video
surveillance cameras suspended form PVC frames. B) 8 cameras were arranged radially around a
DVR. C) The DVR was powered by 2 marine batteries and connected to a removable hard drive
rack that allowed video files to be taken to the lab for uploading and viewing.

170

Figure 4.3: The percentage of plants on which aphid colonies decreased, remained steady, or
increased within 6 h of infestation. Plants were infested prior to dusk or dawn and were caged to
exclude predators or uncaged to be exposed to predators.

0.8
decreased
same
increased
% of infested plants

0.6

0.4

0.2

0.0
sk

re

B

o
ef

ed
ag
c

du

Be

re
fo

ed
ag
nc

d

u
sk
u

f
Be

e
or

wn
da

ed
ag
c

f

Be

171

e
or

d

n
aw

u

ed
ag
nc

LITERATURE CITED

172

LITERATURE CITED

Allard, C. M. and K. V. Yeargan. 2005. Effect of diet on development and reproduction of the
harvestman Phalangium opilio (Opiliones: Phalangiidae). Environmental Entomology
34:6-13.
Carter, P. E. and A. L. Rypstra. 1995. Top-down effects in soybean agroecosystems - spider
density affects herbivore damage. Oikos 72:433-439.
Costamagna, A. C. and D. A. Landis. 2006. Predators exert top-down control of soybean aphid
across a gradient of agricultural management systems. Ecological Applications 16:16191628.
Costamagna, A. C. and D. A. Landis. 2007. Quantifying predation on soybean aphid through
direct field observations. Biological Control 42:16-24.
Costamagna, A. C. and D. A. Landis. 2011. Lack of strong refuges allows top-down control of
soybean aphid by generalist natural enemies. Biological Control 57:184-192.
Costamagna, A. C., D. A. Landis, and M. J. Brewer. 2008. The role of natural enemy guilds in
Aphis glycines suppression. Biological Control 45:368-379.
Costamagna, A. C., D. A. Landis, and C. D. DiFonzo. 2007. Suppression of soybean aphid by
generalist predators results in a trophic cascade in soybeans. Ecological Applications
17:441-451.
Desneux, N., R. J. O'Neil, and H. J. S. Yoo. 2006. Suppression of population growth of the
soybean aphid, Aphis glycines Matsumura, by predators: The identification of a key
predator and the effects of prey dispersion, predator abundance, and temperature.
Environmental Entomology 35:1342-1349.
DiFonzo, C. D. and R. Hines. 2002. Soybean aphid in Michigan: Update from 2001 season.
Michigan State University Extension Bulletin E-2746.
Fox, T. B. 2002. Biological control of the soybean aphid Aphis glycines Matsumura
(Homoptera: Aphididae). Michigan State University, East Lansing.
Fox, T. B., D. A. Landis, F. F. Cardoso, and C. D. DiFonzo. 2004. Predators suppress Aphis
glycines Matsumura population growth in soybean. Environmental Entomology 33:608618.
Fox, T. B., D. A. Landis, F. F. Cardoso, and C. D. DiFonzo. 2005. Impact of predation on
establishment of the soybean aphid, Aphis glycines in soybean, Glycine max. BioControl
50:545-563.
173

Frank, S. D., S. D. Wratten, H. S. Sandhu, and P. M. Shrewsbury. 2007. Video analysis to
determine how habitat strata affects predator diversity and predation of Epiphyas
postvittana (Lepidoptera: Tortricidae) in a vineyard. Biological Control 41:230-236.
Gardiner, M. M., D. A. Landis, C. Gratton, C. D. DiFonzo, M. O'Neal, J. M. Chacon, M. T.
Wayo, N. P. Schmidt, E. E. Mueller, and G. E. Heimpel. 2009. Landscape diversity
enhances biological control of an introduced crop pest in the north-central USA.
Ecological Applications 19:143-154.
Gavish-Regev, E., R. Rotkopf, Y. Lubin, and M. Coll. 2009. Consumption of aphids by spiders
and the effect of additional prey: evidence from microcosm experiments. BioControl
54:341-350.
Grieshop, M. J., B. P. Werling, K. Buehrer, J. Perrone, R. Isaacs, and D. Landis. 2012. Big
brother is watching: studying insect predation in the age of digital surveillance. American
Entomologist 58:172-182.
Hajek, A. E., J. J. Hannam, C. Nielsen, A. J. Bell, and J. K. Liebherr. 2007. Distribution and
abundance of Carabidae (Coleoptera) associated with soybean aphid (Hemiptera:
Aphididae) populations in central New York. Annals of the Entomological Society of
America 100:876-886.
Hannam, J. J., J. K. Liebherr, and A. E. Hajek. 2008. Climbing behaviour and aphid predation by
Agonum muelleri (Coleoptera: Carabidae). Canadian Entomologist 140:203-207.
Hawkins, B. A., N. J. Mills, M. A. Jervis, and P. W. Price. 1999. Is the biological control of
insects a natural phenomenon? Oikos 86:493-506.
Kaiser, M. E., T. Noma, M. J. Brewer, K. S. Pike, J. R. Vockeroth, and S. D. Gaimari. 2007.
Hymenopteran parasitoids and dipteran predators found using soybean aphid after its
midwestern United States invasion. Annals of the Entomological Society of America
100:196-205.
Mccornack, B. P., M. A. Carrillo, R. C. Venette, and D. W. Ragsdale. 2005. Physiological
constraints on the overwintering potential of the soybean aphid (Homoptera: Aphididae).
Environmental Entomology 34:235-240.
McCornack, B. P., D. W. Ragsdale, and R. C. Venette. 2004. Demography of soybean aphid
(Homoptera: Aphididae) at summer temperatures. Journal of Economic Entomology
97:854-861.
Merfield, C. N., S. D. Wratten, and S. Navntoft. 2004. Video analysis of predation by
polyphagous invertebrate predators in the laboratory and field. Biological Control 29:513.

174

Mignault, M. P., M. Roy, and J. Brodeur. 2006. Soybean aphid predators in Quebec and the
suitability of Aphis glycines as prey for three Coccinellidae. BioControl 51:89-106.
Noma, T. and M. J. Brewer. 2008. Seasonal abundance of resident parasitoids and predatory flies
and corresponding soybean aphid densities, with comments on classical biological
control of soybean aphid in the Midwest. Journal of Economic Entomology 101:278-287.
Pfannenstiel, R. S. 2008. Spider predators of lepidopteran eggs in south Texas field crops.
Biological Control 46:202-208.
Pfannenstiel, R. S. and K. V. Yeargan. 2002. Identification and diel activity patterns of predators
attacking Helicoverpa zea (Lepidoptera: Noctuidae) eggs in soybean and sweet corn.
Environmental Entomology 31:232-241.
Pike, K. S., P. Stary, M. J. Brewer, T. Noma, S. Langley, and M. Kaiser. 2007. A new species of
Binodoxys (Hymenoptera: Braconidae: Aphidiinae), parasitoid of the soybean aphid,
Aphis glycines Matsumura, with comments on biocontrol. Proceedings of the
Entomological Society of Washington 109:359-365.
Ragsdale, D. W., D. A. Landis, J. Broduer, G. E. Heimpel, and N. Desneux. 2011. Ecology and
management of the soybean aphid in North America. Annual Review of Entomology
56:375-399.
Riechert, S. E. and T. Lockley. 1984. Spiders as biological-control agents. Annual Review of
Entomology 29:299-320.
Rutledge, C. E. and R. J. O'Neil. 2005. Orius insidiosus (Say) as a predator of the soybean aphid,
Aphis glycines Matsumura. Biological Control 33:56-64.
Rutledge, C. E., R. J. O'Neil, T. B. Fox, and D. A. Landis. 2004. Soybean aphid predators and
their use in integrated pest management. Annals of the Entomological Society of America
97:240-248.
Rypstra, A. L. and P. E. Carter. 1995. The web-spider community of soybean agroecosystems in
southwestern Ohio. Journal of Arachnology 23:135-144.
Schmidt, N. P., M. E. O'Neal, and P. M. Dixon. 2008. Aphidophagous predators in Iowa
soybean: A community comparison across multiple years and sampling methods. Annals
of the Entomological Society of America 101:341-350.
Straub, C. S. and W. E. Snyder. 2006. Species identity dominates the relationship between
predator biodiversity and herbivore suppression. Ecology 87:277-282.
Sunderland, K. 1999. Mechanisms underlying the effects of spiders on pest populations. Journal
of Arachnology 27:308-316.

175

Woltz, J. M., R. Isaacs, and D. A. Landis. 2012. Landscape structure and habitat management
differentially influence insect natural enemies in an agricultural landscape. Agriculture
Ecosystems & Environment 152:40-49.
Woltz, J. M. and D. A. Landis. 2013. Coccinellid immigration to infested host patches influences
suppression of Aphis glycines in soybean. Biological Control 64:330-337.

176

CHAPTER V: Conclusion
Implications for managing landscapes for ecosystem services
Due to the conflicting demands for ecosystem services and limits to the land that
produces them, there is increasing interest in multifunctional landscapes, i.e. those that are
managed to provide multiple ecosystem services simultaneously (Romstad et al. 2000).
Agricultural landscapes produce multiple services, including the provision of food, fuel and
fiber, the support of soil fertility and nutrient cycling, regulation of water quality, greenhouse gas
sequestration, and cultural services (Swinton et al. 2007, Power 2010). The provision of these
services in agricultural landscapes is dependent upon the composition and configuration of
agricultural landscapes (Goldman et al. 2007). Therefore, in order to simultaneously manage
agricultural landscapes for multiple services, we must understand the specific relationships
between the physical structure of the landscapes and the services they provide. The objectives of
this dissertation were to: 1) examine the effectiveness of habitat management for soybean aphid
biological control in landscapes of varying complexity, 2) to examine in detail specific aspects of
landscape structure that influence coccinellid abundances, 3) to examine the influence of
coccinellid immigration on soybean aphid suppression, and 4) assess diel variation in the
soybean aphid natural enemy community.
Insect-mediated services like pollination and biological control are dependent upon the
insect communities. In this work, I contributed to a growing body of evidence that important
natural enemy communities are influenced by landscape structure (Chaplin-Kramer et al. 2011). I
found that coccinellids increased with the proportion of semi-natural habitat in the surrounding
landscape. Coccinellids are dependent upon semi-natural habitats for over-wintering and for
foraging during periods in which prey is not available in agricultural habitats. Additionally, the

177

presence of alternative prey in semi-natural areas in the spring can allow coccinellids to achieve
the nutritional status necessary to initiate reproduction, allowing the population to increase and
contributing to greater pest control (Bianchi and van der Werf 2004).
Secondly, I have shown that coccinellid beetles utilize habitat management plantings,
whether for floral resources (Harmon et al. 2000) or for alternative prey (Dong et al. 2012).
However, coccinellid use of our habitat management plantings did not result in increased
coccinellid activity within adjacent fields. This suggests that, while coccinellids are able to locate
and use locally-provided resources, these resources may be insufficient to change population
dynamics. For natural enemies like coccinellids that disperse over large distances (Hodek et al.
1993), the availability of resources at the landscape scale is likely more important for
determining overall population levels, further emphasizing the need to manage for biological
control services at landscape scales. Despite this result, habitat management may still be useful
as a tool to increase biological control services in a given locale. Carefully timed management,
such as cutting the habitat management strip at a key time for aphid suppression, could force the
coccinellids using the strip to disperse into the adjacent crop (Dong et al. 2012).
Third, I demonstrated that the arrival rate of coccinellids at patches infested with soybean
aphids can be critical to prevent aphid increases. This suggests that understanding how local
conditions influence that rate could provide new tools for habitat management. In systems like
this, in which transient predators are important, it is not sufficient to manage for natural enemy
abundance. Rather, management must also consider dispersal and foraging behaviors of those
transient predators. For example, Grez and others (2010) found that different border types affect
the immigration rate of coccinellids into adjacent fields, and that effects were species specific.

178

Several studies have suggested that habitat configuration may influence coccinellid
dispersal and immigration to patches (Grez and Prado 2000, With et al. 2002, Grez et al. 2004,
Zaviezo et al. 2006). However, in my examination of specific aspects of landscape-level
composition and configuration that could influence coccinellids in chapter 2, I saw no strong
effects of landscape-level configuration in terms of habitat interspersion on the coccinellid
communities I studied. These differences are likely due to the scale at which configuration was
examined. The above studies cited examining the effects of habitat fragmentation and isolation
2

on coccinellid dispersal were primarily conducted within small patches (200-900 m ), suggesting
that these conditions influence coccinellid dispersal behavior at small scales. In contrast, overall
abundance of coccinellids is likely determined by resource availability at larger scales.
Finally, natural enemy communities within soybean fields vary throughout the diel cycle.
Additionally, the foraging rates of natural enemies, and their subsequent encounters with pest
individuals, may not be well-reflected by sampling methods that measure relative abundance.
The activity-density and species richness of several of the taxa most actively foraging at night
are well known to be dependent upon landscape structure (e.g. spiders (Clough et al. 2005, Isaia
et al. 2006, Oberg et al. 2007, Drapela et al. 2008); carabids (Bommarco 1998, Varchola and
Dunn 1999, Aviron et al. 2005, Purtauf et al. 2005, Werling and Gratton 2008); and Opiliones
(Gardiner et al. 2010)). This suggests that intentional landscape management to improve pest
suppression will likely have to consider the effects on multiple natural enemy taxa
simultaneously.
Context dependence of landscape effects
While this and other studies consistently emphasize the importance of landscape scale in
ecosystem service management, the contradictory nature of some studies indicates the context
179

dependence of these effects. For example, considering only coccinellid beetles, studies have
found surprisingly different patterns. In addition to the work presented here, others have found
increases in coccinellid abundance with increasing abundance of semi-natural habitat in the
landscape, specifically in field crops like soybean (Gardiner et al. 2009) and corn (Elliott et al.
2002). In contrast, several researchers have found increases in coccinellids with the amount of
crop habitat in the landscape, specifically in nettles (Rand and Tscharntke 2007) and organic
cereal fields (Caballero-Lopez et al. 2012). These seemingly contradictory results are possibly
the result of: 1) the nature of the habitat where beetles were collected, in comparison to
surrounding landscape; 2) the changing relationships between various landscape variables in
different sets of landscapes.
Regarding the first point, the abundance of a particular natural enemy captured in a given
habitat will depend both on the overall abundance of that natural enemy in the surrounding
landscape and the natural enemies preference for that habitat. Natural enemies may be found in
greater abundance because there are simply more of them available to colonize that habitat. On
the other hand, natural enemies may colonize a habitat disproportionately to their overall
abundance if they have a strong preference for that habitat. For example, syrphids locally show a
preference for floral habitats over grass habitats, but also show increased densities in the floral
habitats as cropland in the surrounding landscape increased (Haenke et al. 2009). This result is
not because syrphid densities are necessarily higher in agriculturally-dominated landscapes, but
rather that syrphids may concentrate in floral resources when there are fewer other resources to
utilize in the surrounding landscape. Others have found responses of ground-dwelling predators
to landscape structure on conventional but not organic farms (Weibull et al. 2003, Schmidt et al.
2005). This suggests that on conventional farms where resources are likely to be insufficient for

180

natural enemies, the occurrence of natural enemies is dependent upon how many are supplied by
the landscape, while organic farms may provide all of the resources necessary for natural
enemies to complete their life cycles, thus their abundance may be independent of surrounding
conditions. These examples indicate that on the one hand, the preference of a natural enemy for a
resource-rich habitat may depend upon the amount of resources available in the landscape, while
on the other hand, the abundance of natural enemies within a resource-poor habitat may depend
upon the ability of the landscape to supply those natural enemies. Future work examining these
context dependencies will be necessary to understand under what conditions landscape and local
management actions are expected to be most effective for increasing biocontrol services from
natural enemies.
A second mechanism for the context-dependence of natural enemy-landscape
relationships is the differing relationships between variables in different sets of landscapes. For
example, Werling et al. (2011) found decreases in the abundance of exotic coccinellids in corn,
switchgrass and prairie habitats with increases in herbaceous perennials like hayfield and
pastures. In contrast, a similar metric encompassing hayfields and pastures was not correlated
with coccinellid abundances in this work. Similarly, several studies have found relationships
between natural enemies and overall habitat diversity (Thies and Tscharntke 1999, SteffanDewenter et al. 2002, Roschewitz et al. 2005), while I found no relationship. In such cases,
natural enemies may be responding to a particular variable that is correlated with another
variable in one set of landscapes but not in others. In the case of habitat diversity, in studies in
which this metric was correlated with natural enemies, it was also correlated with amount of
semi-natural habitat, while in my work those two landscape variables were not related.
Therefore, selecting a single variable could lead to inappropriate interpretations of results. In this

181

work, I have shown that multiple landscape variables, all of which could reasonably affect
coccinellid biology, were all highly correlated with one another. This emphasizes the need to
more carefully examine all aspects of landscape, as opposed to using single metrics whose
relationships with other metrics may differ across landscapes and regions.
Additionally, a focus on single, broad metrics may over-look important “hidden
heterogeneity” within agricultural landscapes that could have important implications for natural
enemies (Vasseur et al. 2013). A given crop may be managed in a variety of different ways,
including nutrient regimes, use and choice of pesticides, and timing and choice of tillage
practices. These differences can drastically affect the quality of a crop habitat for natural enemies
and other taxa. For example, a reduction in fertilizer use resulting in decreased availability of
aphid prey can have greater impacts on cocccinellid populations than altering landscape
composition (Bianchi et al. 2007). Because crop management is more directly under farmer
control than other aspects of landscape heterogeneity, an increased research focus on this area
could yield more easily applicable management tools.
Expanding to underrepresented regions and agroecosystems
The context dependence of natural enemy relationships to landscape suggests an urgent
need to recreate these kinds of studies in other agroecosystem types and in under-represented
parts of the world. The majority of this type of research has been conducted in the temperate
climates in the northern hemisphere (i.e. in Europe and the United States), where most seminatural habitats consist of grasslands and forests. Even amongst these regions, differences exist.
Europe and the United States also have different agricultural histories. Agriculture spread slowly
across Europe thousands of years ago (Pinhasi et al. 2005). With agriculture being an integral
part of the landscape over such a long period of time, agroecosystems can actually support

182

substantial biodiversity and policies are in place to encourage agricultural practices that continue
to support biodiversity, e.g. Agriculture Environment Schemes, (Kleijn et al. 2011). In contrast,
the midwest and western U.S. was settled intentionally, systematically, and comparatively
rapidly. Agricultural intensification in these regions took place over periods of a few hundred
rather than thousands of years, giving species less time to adapt to agroecosystems. Thus, the
more intensified agricultural landscapes in the U.S. could result in lower natural enemy
abundance, as has been postulated for lower spider density in U.S. than European crop fields
(Nyffeler and Sunderland 2003).
Other agricultural regions around the world likely have even greater differences in the
degree and history of intensification, crops grown and agronomic practices utilized, and the
identity of both pest and beneficial species living within agroecosystems. We are aware of a
handful of landscape studies conducted outside of Europe and the U.S., including in arid
agroecosystems in the Middle East (Pluess et al. 2010) and in tropical agroecosystems in Asia
(Wilby et al. 2006). Many of the regions under-represented by this type of research also have
some of the fastest growing populations in the world and will likely have the greatest increases in
demands for food production in the near future (Tilman et al. 2011). As a consequence, these
areas will experience rapid changes in landscape structure as natural areas are converted for food
production. Research that would inform land management to preserve ecosystem services in
these regions in the face of these changes is critical.
Finally, agriculture is not limited to rural areas within a matrix of remaining semi-natural
habitat. Recently, much attention is being paid to urban agriculture as a means to address social
justice and food security issues in urban areas (Grewal and Grewal 2012). Until recently, there
has been little examination of arthropod biodiversity in urban areas, and much of that is limited

183

to surveys within different habitats (McIntyre 2000). Yet urban areas have the potential to
produce numerous ecosystem services, and there is a need to better understand the relationship
between urban and residential landscapes and the ecosystem services they provide (Cook et al.
2012, Jansson 2013). Research examining the ability of urban arthropod communities to provide
pollination and pest suppression services for urban food production is needed. This and other
urban ecosystem service work will be critical for urban sustainability planning, especially as
70% of the world’s population is expected to live in cities by 2050 (World Health Organization
2013).
Overall it is clear that landscape structure is integral to support natural enemy
communities and provision of pest suppression services. Continued study of these relationships,
including their context-dependence and amenability to manipulation, may yield information
critical to the future design of such landscapes for intentional provision of multiple ecosystem
services. Understanding how these relationships are, or are not, generalizable across ecosystems
and regions will be necessary to apply this knowledge on a global scale.

184

LITERATURE CITED

185

LITERATURE CITED

Aviron, S., F. Burel, J. Baudry, and N. Schermann. 2005. Carabid assemblages in agricultural
landscapes: impacts of habitat features, landscape context at different spatial scales and
farming intensity. Agriculture Ecosystems & Environment 108:205-217.
Bianchi, F. J. J. A., A. Honek, and W. van der Werf. 2007. Changes in agricultural land use can
explain population decline in a ladybeetle species in the Czech Republic: evidence from a
process-based spatially explicit model. Landscape Ecology 22:1541-1554.
Bianchi, F. J. J. A. and W. van der Werf. 2004. Model evaluation of the function of prey in noncrop habitats for biological control by ladybeetles in agricultural landscapes. Ecological
Modelling 171:177-193.
Bommarco, R. 1998. Reproduction and energy reserves of a predatory carabid beetle relative to
agroecosystem complexity. Ecological Applications 8:846-853.
Caballero-Lopez, B., R. Bommarco, J. M. Blanco-Moreno, F. X. Sans, J. Pujade-Villar, M.
Rundlof, and H. G. Smith. 2012. Aphids and their natural enemies are differently affected
by habitat features at local and landscape scales. Biological Control 63:222-229.
Chaplin-Kramer, R., M. E. O'Rourke, E. J. Blitzer, and C. Kremen. 2011. A meta-analysis of
crop pest and natural enemy response to landscape complexity. Ecology Letters 14:922932.
Clough, Y., A. Kruess, D. Kleijn, and T. Tscharntke. 2005. Spider diversity in cereal fields:
comparing factors at local, landscape and regional scales. Journal of Biogeography
32:2007-2014.
Cook, E. M., S. J. Hall, and K. L. Larson. 2012. Residential landscapes as social-ecological
systems: a synthesis of multi-scalar interactions between people and their home
environment. Urban Ecosystems 15:19-52.
Dong, Z. K., F. J. Gao, and R. Z. Zhang. 2012. Use of ryegrass strips to enhance biological
control of aphids by ladybirds in wheat fields. Insect Science 19:529-534.
Drapela, T., D. Moser, J. G. Zaller, and T. Frank. 2008. Spider assemblages in winter oilseed
rape affected by landscape and site factors. Ecography 31:254-262.
Elliott, N. C., R. W. Kieckhefer, and D. A. Beck. 2002. Effect of aphids and the surrounding
landscape on the abundance of Coccinellidae in cornfields. Biological Control 24:214220.
Gardiner, M. M., D. A. Landis, C. Gratton, C. D. DiFonzo, M. O'Neal, J. M. Chacon, M. T.
186

Wayo, N. P. Schmidt, E. E. Mueller, and G. E. Heimpel. 2009. Landscape diversity
enhances biological control of an introduced crop pest in the north-central USA.
Ecological Applications 19:143-154.
Gardiner, M. M., D. A. Landis, C. Gratton, N. Schmidt, M. O'Neal, E. Mueller, J. Chacon, and
G. E. Heimpel. 2010. Landscape composition influences the activity density of Carabidae
and Arachnida in soybean fields. Biological Control.
Goldman, R. L., B. H. Thompson, and G. C. Daily. 2007. Institutional incentives for managing
the landscape: Inducing cooperation for the production of ecosystem services. Ecological
Economics 64:333-343.
Grewal, S. S. and P. S. Grewal. 2012. Can cities become self-reliant in food? Cities 29:1-11.
Grez, A., T. Zaviezo, L. Tischendorf, and L. Fahrig. 2004. A transient, positive effect of habitat
fragmentation on insect population densities. Oecologia 141:444-451.
Grez, A. A. and E. Prado. 2000. Effect of plant patch shape and surrounding vegetation on the
dynamics of predatory coccinellids and their prey Brevicoryne brassicae (Hemiptera:
Aphididae). Environmental Entomology 29:1244-1250.
Grez, A. A., C. Torres, T. Zaviezo, B. Lavandero, and M. Ramirez. 2010. Migration of
coccinellids to alfalfa fields with varying adjacent vegetation in Central Chile. Ciencia E
Investigacion Agraria 37:111-121.
Haenke, S., B. Scheid, M. Schaefer, T. Tscharntke, and C. Thies. 2009. Increasing syrphid fly
diversity and density in sown flower strips within simple vs. complex landscapes. Journal
of Applied Ecology 46:1106-1114.
Harmon, J. P., A. R. Ives, J. E. Losey, A. C. Olson, and K. S. Rauwald. 2000. Colemegilla
maculata (Coleoptera: Coccinellidae) predation on pea aphids promoted by proximity to
dandelions. Oecologia 125:543-548.
Hodek, I., G. Iperti, and M. Hodkova. 1993. Long-Distance Flights in Coccinellidae
(Coleoptera). European Journal of Entomology 90:403-414.
Isaia, M., F. Bona, and G. Badino. 2006. Influence of landscape diversity and agricultural
practices on spider assemblage in Italian vineyards of Langa Astigiana (northwest Italy).
Environmental Entomology 35:297-307.
Jansson, A. 2013. Reaching for a sustainable, resilient urban future using the lens of ecosystem
services. Ecological Economics 86:285-291.
Kleijn, D., M. Rundlof, J. Scheper, H. G. Smith, and T. Tscharntke. 2011. Does conservation on
farmland contribute to halting the biodiversity decline? Trends in Ecology & Evolution
26:474-481.
187

McIntyre, N. E. 2000. Ecology of urban arthropods: A review and a call to action. Annals of the
Entomological Society of America 93:825-835.
Nyffeler, M. and K. D. Sunderland. 2003. Composition, abundance and pest control potential of
spider communities in agroecosystems: a comparison of European and US studies.
Agriculture Ecosystems & Environment 95:579-612.
Oberg, S., B. Ekbom, and R. Bommarco. 2007. Influence of habitat type and surrounding
landscape on spider diversity in Swedish agroecosystems. Agriculture Ecosystems &
Environment 122:211-219.
Pinhasi, R., J. Fort, and A. J. Ammerman. 2005. Tracing the origin and spread of agriculture in
Europe. PLoS Biology 3:2220-2228.
Pluess, T., I. Opatovsky, E. Gavish-Regev, Y. Lubin, and M. H. Schmidt-Entling. 2010. Noncrop habitats in the landscape enhance spider diversity in wheat fields of a desert
agroecosystem. Agriculture Ecosystems & Environment 137:68-74.
Power, A. G. 2010. Ecosystem services and agriculture: tradeoffs and synergies. Philosophical
Transactions of the Royal Society B-Biological Sciences 365:2959-2971.
Purtauf, T., I. Roschewitz, J. Dauber, C. Thies, T. Tscharntke, and V. Wolters. 2005. Landscape
context of organic and conventional farms: Influences on carabid beetle diversity.
Agriculture Ecosystems & Environment 108:165-174.
Rand, T. A. and T. Tscharntke. 2007. Contrasting effects of natural habitat loss on generalist and
specialist aphid natural enemies. Oikos 116:1353-1362.
Romstad, E., A. Vatn, P. K. Rorstad, and V. Soyland. 2000. Multifunctional Agriculture:
Implicatons for Policy Design. Agricultural University of Norway, Department of
Economics and Social Sciences.
Roschewitz, I., M. Hucker, T. Tscharntke, and C. Thies. 2005. The influence of landscape
context and farming practices on parasitism of cereal aphids. Agriculture Ecosystems &
Environment 108:218-227.
Schmidt, M. H., I. Roschewitz, C. Thies, and T. Tscharntke. 2005. Differential effects of
landscape and management on diversity and density of ground-dwelling farmland
spiders. Journal of Applied Ecology 42:281-287.
Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tscharntke. 2002. Scaledependent effects of landscape context on three pollinator guilds. Ecology 83:1421-1432.
Swinton, S. M., F. Lupi, G. P. Robertson, and S. K. Hamilton. 2007. Ecosystem services and

188

agriculture: Cultivating agricultural ecosystems for diverse benefits. Ecological
Economics 64:245-252.
Thies, C. and T. Tscharntke. 1999. Landscape structure and biological control in
agroecosystems. Science 285:893-895.
Tilman, D., C. Balzer, J. Hill, and B. L. Befort. 2011. Global food demand and the sustainable
intensification of agriculture. Proceedings of the National Academy of Sciences of the
United States of America 108:20260-20264.
Varchola, J. M. and J. P. Dunn. 1999. Changes in ground beetle (Coleoptera: Carabidae)
assemblages in farming systems bordered by complex or simple roadside vegetation.
Agriculture Ecosystems & Environment 73:41-49.
Vasseur, C., A. Joannon, S. Aviron, F. Burel, J. M. Meynard, and J. Baudry. 2013. The cropping
systems mosaic: how does the hidden heterogeneity of agricultural landscapes drive
arthropod populations? Agriculture, Ecosystems and Environment 166:12.
Weibull, A. C., O. Ostman, and A. Granqvist. 2003. Species richness in agroecosystems: the
effect of landscape, habitat and farm management. Biodiversity and Conservation
12:1335-1355.
Werling, B. P. and C. Gratton. 2008. Influence of field margins and landscape context on ground
beetle diversity in Wisconsin (USA) potato fields. Agriculture Ecosystems &
Environment 128:104-108.
Werling, B. P., T. D. Meehan, C. Gratton, and D. A. Landis. 2011. Influence of habitat and
landscape perenniality on insect natural enemies in three candidate biofuel crops.
Biological Control 59:304-312.
Wilby, A., L. P. Lan, K. L. Heong, N. P. D. Huyen, N. H. Quang, N. V. Minh, and M. B.
Thomas. 2006. Arthropod diversity and community structure in relation to land use in the
Mekong Delta, Vietnam. Ecosystems 9:538-549.
With, K. A., D. M. Pavuk, J. L. Worchuck, R. K. Oates, and J. L. Fisher. 2002. Threshold effects
of landscape structure on biological control in agroecosystems. Ecological Applications
12:52-65.
World Health Organization. 2013. Urban population growth. www.who.int/gho/urban_health.
Accessed 30 May 2013.
Zaviezo, T., A. A. Grez, C. F. Estades, and A. Perez. 2006. Effects of habitat loss, habitat
fragmentation, and isolation on the density, species richness, and distribution of
ladybeetles in manipulated alfalfa landscapes. Ecological Entomology 31:646-656.

189