‘ ¢ £210.... J .v
flu 1..... z...
p . . . . . .
S ...1
w. awake.
95.34
0‘.

 

. ¢
. _ .. .1...1..m....
1.1.9 1... .

.‘a f.
. u.

 

2.. .
9 .39.;
. .mv.

 

.17.... . . . . . .

3»wa . . . . . . . , , . . , 31 1..

. Vi): . . . . .. rig»: .
xvfit.‘ h. . . . . . . . . , . . . . «W Waite».
3n . , \ .

£131.: 7
In

 

r
,3“...

. 1c..."
A. . ~tau

.1.
.\I.5I¢.H.n!

ARGENM” u 1
mm... 8.11...
v... 43 v... 1.331%“.
\..c . :4 Vltp‘.’ .A i
. stun-4:... . run... .31.

.r “3..

.. . 1 )3;
$114

 

     
     

 
 

 

. : 1,...1 Knuc.\-1v.. . ........ .161. .5"... J»)n.l...¥m.+i:1.;muwwwu1uwwfl as?) «tkvmv. , . .. i .Jfi. an...“ .5. twain .1 .
.. . . 3 . .. .. 1...... . . ...r. ........... 3.»! 1.. E :1... . H . .
l! '1! _ . 1 . .. ... .1 ..,.. . , . .. . . 1, q..u_.....11..»iu§_.5.......4.
| ' 1 1 0
‘
I ll

 

02
1001
This is to certify that the

dissertation entitled

ECOLOGY AND CONSERVATION OF WILD BEES ASSOCIATED
WITH HIGHBUSH BLUEBERRY FARMS IN MICHIGAN

mi

presented by

Julianna Kristen Tuell

LIBRARY
ichigan Sta
University

 

L111

has been accepted towards fulfillment
of the requirements for the

Entomology and the Program in
Ecology, Evolutionary Biology,

 

 

 

 

 

PhD. degree in and Behavior
MajorProfes rs ignature
s\n\o‘7
Date

MSU is an afiinnative-action, equal—opportunity employer

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:lC|RC/DateDue.indd-p.1

ECOLOGY AND CONSERVATION OF WILD BEES ASSOCIATED WITH
HIGHBUSH BLUEBERRY FARMS IN MICHIGAN

By

Julianna Kristen Tuell

A DISSERTATION

Submitted to
Michigan State University
in partial fillfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Entomology
and the Program in Ecology, Evolutionary Biology, and Behavior

2007

ABSTRACT

ECOLOGY AND CONSERVATION OF WILD BEES ASSOCIATED WITH
HIGHBUSH BLUEBERRY FARMS IN MICHIGAN

By

Julianna Kristen Tuell

The objectives of these studies were: 1) to develop a method for pan trapping bees
in highbush blueberry (Ericaceae: Vaccim'um corymbosum L.); 2) to describe the
structure of the endemic bee community associated with blueberry and determine to what
extent wild bees contribute to its pollination; 3) to examine the relationship between wild
bee community structure and both local habitat features and pest management practices;
4) to evaluate native perennial plants for attracting bees; and 5) to examine blueberry bee
community structure in the context of landscape spatial scales. Pan traps elevated in the
canopy of flowering blueberry collected more bees than traps placed on the ground or
above the canopy. This method was used at 15 blueberry farms in southwest Michigan to
characterize the bee community present prior, during, and after blueberry bloom for three
years (2004-6). Honey bees, primarily from rented hives, comprised 50-66% of all the
bees captured in pan traps each year during bloom. The remainder were wild bees, mame
soil nesting bees in the families Andrenidae and Halictidae. Andrena carolina, a solitary
bee that is oligolectic on Ericaceae, was the most abundant species captured in pan traps
during bloom (14%). Bees collected while foraging (n = 22 species) and pollen carried by
bees in pan traps (n = 126 specimens) were used to confirm which bees were foraging on
blueberry. In pan trap samples during bloom, there was high species turnover from year
to year with 79 species recorded on average each year out of a total of 120 species over

the three years. Wild bees were more often captured in traps at the field edge than in the

interior. In simple linear regressions (SLR) of bees and field characteristics, bee species
richness increased with flowering plant species in 2005, and declined with the local
frequency of adjacent blueberry fields. Bee diversity (H’) was also lower in fields with
more nearby blueberry fields. Native bee abundance and richness in 2004 along with bee
diversity in 2005 declined with increasing insecticide program toxicity (IPT). In a
redundancy analysis of the same characters, IPT explained the greatest amount of
variation in the bee species data in 2004, and vegetation attributes explained variation in
the species in 2005 and 2006. Of the 43 native perennial plants that were evaluated, all
but 2 were visited at least once by non-Apis bees (n = 1393), but there were 9 that were
visited most frequently. Floral area explained 33% of the bee diversity at particular plant
species. The response of wild bees to the proportion of different landscape types around
blueberry fields was evaluated at 5 nested scales using 250, 500, 750, 1000, and 1500 m
radius circles. Forest margins (<10 m from the forest edge), human settlement, annual
cropland, and blueberry plantations were the most abundant landscape types at each of
the spatial scales. Total bee abundance, richness, and diversity did not vary significantly
with any of 6 categories of land use at any of the spatial scales, however, A. carolina
responded to the proportion of human settlement at the 1500 m scale, and Ceratina
calcarata/dupla responded to the proportion of blueberry plantations and semi-natural
habitat at the 250 m scale. Three of 4 blueberry fi'uit yield attributes increased with the
proportion of blueberry habitat and decreased with the proportion of semi-natural habitat
within 500 m of the focal field. Implications for the conservation of native bees and their

importance in blueberry pollination are discussed.

Copyrighted by
JULIANNA KRISTEN TUELL
2007

DEDICATION

To Matthew

ACKNOWLEDGEMENTS

A great many people assisted me along the way in the preparation and writing of
this dissertation. Ideas for bee sampling methods and for what bees I might expect to see
in a blueberry agroecosystem initially came from Kenna MacKenzie of the Atlantic Food
and Horticulture Research Centre in Nova Scotia and Frank Drummond of the University
of Maine. Frank also suggested ways to quantify floral abundance. Long productive days
in the field would not have been possible without the able assistance of Chelsea McLean
(2003), Jesse Sieman (2004), Matthew Tuell (2004-5), Christina McEmber (2006), Kat
Roltsch (2007), and especially Jack Langdon (2005-7), who helped collect, process and
sort samples, plant native plants during grueling hot weather, and enter data Jack
Langdon also helped with preliminary identifications of bees to the genus and/or species
level, particularly for bees in the Halictidae family, and he did much of the sorting of the
bees in the collection. Sam Droege’s online key to bees of the eastern US made this
process a lot faster. John Ascher of the American Museum of Natural History identified
most of the bees in the family Andrenidae and verified identifications we made for bees
in other families. Colleagues I met at The Bee Course run by the American Museum of
Natural History at the Southwest Research Station in Portal, Arizona — especially Annette
Meredith and Berry Brosi — helped me become and stay connected to the world of bee
biologists. Paul Jenkins helped me navigate SAS 9.1 and provided me links to excellent
programming shortcuts for doing mixed model analyses; Berry Brosi helped me with
species accumulation curves and spatial autocorrelation analyses. Anna Fiedler was
generous with her bee samples and floral attribute data from the native plant plot on

campus. Members of the Isaacs lab, especially Keith Mason, helped with field work

logistics and with edits on some earlier portions of this document. The staff at the Trevor
Nichols Research Complex, especially John Wise, provided me access to several semi-
abandoned sites in southwest Michigan and allowed me to manipulate managed bees in
the blueberry blocks at the station. None of the landscape level or habitat comparisons
would have been possible without the eleven generous commercial blueberry growers
who allowed me to trespass on their properties: Rant, Tiles, Carini, Stansby, Wassink,
Bowerman, DeJonge, Earl (Wood), Bodtke, DeGrandchamp, and Galens. Funding was
provided in part by a predoctoral fellowship in sustainable agriculture fi'om the CS. Mott
Foundation, grants from the USDA-SARE program, USDA Sustainable Agriculture and
Food Systems Program at MSU, MSU Project GREEEN, and Michigan Blueberry
Growers Association. Moral support was provided along the way by the fabulous Jill
Kolp, my family and fiiends, and especially friends from my graduate school cohort: with
special reference to Paul Jenkins and the “girls”: Kirsten Pelz-Stelinski, J areé Johnson,
Kristi Zurawski, Julianne Heinlein, and Mollie McIntosh My advisory committee
members, Douglas Landis, James Miller, Zachary Huang, and L. Alan Prather, offered
their sage advice and kept me on my toes. Finally, Rufus Isaacs faced his fear of bees and
allowed me to fill up his lab with them, and then helped me design ways to study them,

write about them, and stay focused.

Thank you!

Julianna Tuell

August 16, 2007

vii

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... x
LIST OF FIGURES ...................................................................................................... xiii
CHAPTER 1: THE IMPORTANCE OF BEES FOR CROP POLLINATION WITH AN
EMPHASIS ON BLUEBERRY ...................................................................................... 1
Introduction ......................................................................................................... 2
History and Current Status of Crop Pollination in North America ........................ 3
An Overview of the Biology of Non-Apis Bees .................................................... 8
Agricultural Practices Likely to Affect Endemic Bees ........................................ 10
Habitat Quality and Land Use ............................................................................ l3
Pollination Requirements of Blueberry .............................................................. 17
Focus of This Project ......................................................................................... 22
CHAPTER 2: ELEVATED PAN TRAPS FOR MONITORING BEES IN CROP
CANOPIES: RESPONSE OF BLUEBERRY POLLINATORS ..................................... 24
Introduction ....................................................................................................... 25
Materials and Methods ....................................................................................... 27
Results ............................................................................................................... 30
Discussion ......................................................................................................... 39

CHAPTER 3: WILD BEES ASSOCIATED WITH THE HIGHBUSH BLUEBERRY

AGROECOSYSTEM IN MICHIGAN .......................................................................... 42
Introduction ....................................................................................................... 43
Materials and Methods ....................................................................................... 45
Results ............................................................................................................... 49
Discussion ......................................................................................................... 69

CHAPTER 4: RESPONSE OF NATIVE BEES TO HABITAT QUALITY AND

PRODUCTION PRACTICES IN HIGHBUSH BLUEBERRY ..................................... 74
Introduction ....................................................................................................... 75
Materials and Methods ....................................................................................... 77
Results ............................................................................................................... 85
Discussion ....................................................................................................... 100

CHAPTER 5: COMPARISON OF NATIVE PLANTS FOR USE IN AGRICULTURAL

BEE CONSERVATION PROGRAMS IN MID-WESTERN U.S ................................ 104
Introduction ..................................................................................................... 105
Materials and Methods ..................................................................................... 108
Results ............................................................................................................. 111
Discussion ....................................................................................................... 123

viii

CHAPTER 6: RESPONSE OF NATIVE BEES TO LAND USE PATTERNS IN

BLUEBERRY AGROECOSYSTEMS ........................................................................ 128
Introduction ..................................................................................................... 129
Materials and Methods ..................................................................................... 131
Results ............................................................................................................. 137
Discussion ....................................................................................................... 147

LITERATURE CITED ................................................................................................ 151

APPENDIX A: GPS COORDINATES FOR FIELD SITES ........................................ 166

APPENDIX B: LIST OF VOUCHER SPECIMENS ................................................... 168

APPENDIX C: A RECORD OF BEE SPECIES ASSOCIATED WITH VACCINI UM
SPP. IN MICHIGAN ................................................................................................... 173

ix

Table

2.1

2.2

2.3

3.1

3.2

3.3

3.4

3.5

3.6

3.7

4.1

LIST OF TABLES

Page

Bee species recovered from pan traps placed either on the ground, in mid-canopy
(0.46 to 1.2 m), or above the canopy (1.5 to 1.8 m) in a highbush blueberry field
during bloom near F ennville, Michigan in 2004-05. ........................................... 34

Number and species diversity of bees recovered fi'om blue, white, or yellow pan
traps elevated 1.2 m in the canopy of a mature highbush blueberry stand during
bloom near Fennville, Michigan in 2004. ........................................................... 37

Comparison of bees captured in pan traps placed in the blueberry canopy with
bees observed foraging on blueberry during timed observations, in one blueberry
field over two days of sampling in southwest Michigan in 2005. ........................ 38

Dates between which pan trapping was conducted in Michigan blueberry fields
prior to blueberry bloom (Pre-bloom), during blueberry bloom (Bloom), and after
blueberry bloom (Post-bloom), and the number of times traps were deployed in
each field during each time period (n). ............................................................... 47

Bee species collected in pan traps in blueberry fields in southwest Michigan over
a period of three years beginning in 2004. .......................................................... 51

Species of native bees (n = 120 species and 3228 specimens) listed in order of
most to least abundant over three years of collecting during bloom in 15
southwest Michigan blueberry farms fiom 2004-O6. .......................................... 60

Proportion of Vaccim'um pollen on the bodies of the most commonly collected
native bees found in Michigan blueberry fields in 2004 and 2005 (n = 126). ...... 63

Non-Apis bee species collected while foraging on blueberry blooms in Michigan
in 2004-06. ........................................................................................................ 64

New species range extensions for bees captured in pan traps in Michigan
blueberry fields in 2004-06. ............................................................................... 65

AN OVA results of the response to pan trap position (edge vs. field interior) and
color (white vs. yellow) of the eight most abundant native bees known to forage
on Vaccinium. .................................................................................................... 69

Categories of habitat features used to characterize the habitat immediately
surrounding highbush blueberry fields where bees were sampled in southwest
Michigan fiom 2004-O6. .................................................................................... 80

4.2

4.3

4.4

4.5

4.6

5.1

5.2

5.1

6.1

6.2

6.3

6.4

Areas in which non-crop vegetation was managed in and around focal blueberry
fields in southwest Michigan and how they were scored for intensity. ............... 82

Insecticides used by blueberry growers in southwest Michigan whose farms were
sampled for bees in 2004-06. ............................................................................. 83

Plant species that were found in the perimeter of focal blueberry fields in
southwest Michigan. .......................................................................................... 87

Regression coefficients for simple linear regressions of the 5 most abundant
native bee blueberry foragers and (a) kgAl/l-Ia/LDso, (b) kgAI/Ha/ToxRate, (c)
vegetation management intensity, ((1) adjacent deciduous woods, (e) adjacent
ditches, (1) adjacent blueberry fields, and (g) flowering plant species richness, at
15 blueberry farms in southwest Michigan. ........................................................ 94

Summary of the RDA analyses of bee community abundance in blueberry fields
by year. .............................................................................................................. 96

Bees collected during 3 vacuum sampling periods during peak bloom at a research
plot of 43 native prairie forb species in Ingham County, Michigan. .................. 113

Floral attributes, bloom period, and the average number of wild bees collected
during vacuum sampling for floral visitors during peak bloom of each plant in
2005. ............................................................................................................... 119

Results of multiple linear regressions of the abundance and diversity of bees
collected at native flowering plants during peak bloom against three floral
characters ......................................................................................................... 123

Categories of landscape types used in the digitization of aerial photos . ......... 136

Composition and quantification of the 1500 m radius landscape sectors in
southwestern Michigan. .................................................................................. 138

Regression coefficients for wild bee abundance, species richness, diversity, and
the 5 most abundant Vaccim'um-foragers averaged over three years and the
proportion of (a) forest margin, (b) settlement, (c) annual cropland, (d) blueberry
plantations, (e) semi-natural habitat, and (f) semi-natural and forest margins
together at 5 spatial scales. ............................................................................... 139

Regression coefficients for the proportion of fi'uit set at harvest, fi'uit weight per
berry at harvest, diameter of the largest berry, and the number of seeds per largest
berry averaged over three years and the proportion of (a) forest margin, (b)
settlement, (c) annual cropland, (d) blueberry plantations, (e) semi-natural habitat,
and (f) semi-natural and forest margins together at 5 spatial scales ................... 143

xi

GPS coordinates for the 15 blueberry fields in southwest Michigan in which bees
were sampled between 2004-6. ........................................................................ 167

List of voucher specimens ................................................................................ 169

Bee species for which there are either pollen, nectar, or floral visitation records on
Vaccinium spp. ................................................................................................ 174

xii

LIST OF FIGURES

Figure Page

2.1 Average number of bees recovered from pan traps placed on the ground or
elevated 1B“, 2B“, or above the canopy within highbush blueberry stands m
2004 and 2005. .................................................................................................. 33

2.2 Average number of bees 1n the families Andrenidae and Halictidae recovered
fi‘om pan traps placed on the ground or elevated 18”, 2B“, or above the canopy
within highbush blueberry stands m 2005 .......................................................... 33

3.1 Location of 15 bee collection sites in relation to the top five blueberry producing
counties in Michigan. ......................................................................................... 46

3.2 Species accumulation curve generated from 100 permutations of the 2004 pan
trap sampling data ............................................................................................. 50

3.3 Proportion of W1ld (non-Apis) bees by family trapped during blueberry bloom at
15 farms 1n southwest Michigan from 2004- O6. ................................................. 58

3.4 Proportion of wild (non-Apis) bees in each nesting guild trapped throughout the
season across 15 blueberry farms in southwest Michigan over three years .......... 59

3.5 Incidence throughout the season of the most abundant native bee species that are
known to forage on Vaccinium spp. plus all the Bombus spp. that were trapped
throughout the study in 13 commercial and 2 semi-abandoned highbush blueberry
fields in southwest Michigan from 2004-06. ...................................................... 61

3.6 Proportion of honey bees and wild bees observed during blueberry bloom at semi-
abandoned and commercial blueberry fields in southwest Michigan, 2004-O6. 63

3.7 Native bee response to trap placement (edge vs. interior of the field) and color
(white vs. yellow) across three years during bloom in highbush blueberry fields in
southwest Michigan. .......................................................................................... 67

3.8 Honey bee response to trap placement (edge vs. interior of the field) and color
(white vs. yellow) across three years during bloom in highbush blueberry fields in
southwest Michigan. .......................................................................................... 68

4.1 Diagram of habitat feature sampling method around the perimeter of blueberry
fields, depicted here as the shaded square ........................................................... 79

4.2 Regression analyses of wild (A, B, C) bee abundance, (D, E, F) bee species

richness, and (G, H, I) bee diversity with the number of flowering plant species
found in the field margin at 15 farms in southwest Michigan. ............................ 89

xiii

4.3

4.4

4.5

4.6

4.7

4.8

5.1

5.2

5.3

5.4

Regression analyses of wild (A, B, C) bee abundance, (D, E, F) bee species
richness, and (G, H, I) bee diversity with the abundance of blueberry fields
bordering the field margin in eight 45 degree directions (see Materials and
Methods for details). .......................................................................................... 9O

Regression analyses of wild (A, B, C) bee abundance, (D, E, F) bee species
richness, and (G, H, I) bee diversity with the insecticide program toxicity score
based on LDso for honey bees from the year previous to bee sampling (see
Materials and Methods for details). .................................................................... 91

Regression analyses of wild (A, B, C) bee abundance, (D, E, F) bee species
richness, and (G, H, I) bee diversity with the insecticide program toxicity score
based on ratings in the 2007 Michigan Fruit Management Guide from the year
previous to bee sampling (see Materials and Methods for details). ..................... 92

Redundancy analysis of the abundance of 38 bee species known to forage on
Vaccim‘um and 13 enviromnental characters, including two measurements of crop
management intensity at 15 blueberry farms in southwest Michigan in 2004 ...... 97

Redundancy analysis of 30 bee species known to forage on Vaccinium and 13
enviromnental characters, including two measurements of crop management
intensity at 15 blueberry farms in southwest Michigan in 2005. ......................... 98

Redundancy analysis of 28 bee species known to forage on Vaccinium and 13
environmental characters, including two measurements of crop management
intensity at 15 blueberry farms in southwest Michigan in 2006. ......................... 99

Average abundance (number of bees per plant species) and richness (number of
bee taxa per plant species) of all wild (non-Apis) bees collected at native plants in
2005 in Ingham Co., Michigan via vacuum sampling during peak bloom. ........ 112

Total number of (A) wild bees, and (B) honey bees collected during 15, 30 sec
vacuum samples for floral visitors at 43 native plants in Ingham Co., Michigan in
2005. .............................................................................................................. 118

Average number of (A) wild bees, and (B) honey bees noted during 5 timed
observations ..................................................................................................... 121

Comparison of bee sampling methods using simple linear regression of (A) the
proportion of wild bees and (B) the proportion of honey bees caught or recorded
using each method at 38 of the plant species tested in Ingham County, Michigan
in 2005. ........................................................................................................... 122

xiv

6.1

6.2

6.3

6.4

6.5

6.6

Examples of the aerial photographs used to digitize landscape features; (A) is a
site with a high proportion of annual and nursery crops and depicts the 5 different
radius (meters) sectors used in the analyses, (B) is a site near Lake Michigan (to
the west) with a high proportion of settlement area, and (C) is a site with a high
proportion of blueberry plantations. ................................................................. 135

Scale dependent effect of landscape structure on the number of Andrena carolz’na
bees collected in pan traps at 15 highbush blueberry fields in southwestern
Michigan. ....................................................................................................... 1 40

Scale dependent effects of landscape structure on the number of Ceratina
calcarata or dupla (they are morphologically indistinguishable) female bees
collected in pan traps at 15 highbush blueberry fields in southwestern Michigan.
........................................................................................................................ 1 41

Scale-dependent effects of the proportion of the landscape in blueberry production
on three different fi'uit attributes at harvest at 15 highbush blueberry fields in
southwestern Michigan. ................................................................................... 144

Scale-dependent effects of the proportion of open uncultivated land on two
different fruit attributes at harvest at 15 highbush blueberry fields in southwestern
Michigan. ............................................... 145

Scale-dependent effects of the proportion of semi-natural land, including

woodland habitat, on two different fruit attributes at harvest at 15 highbush
blueberry fields in southwestern Michigan ...................................................... 146

XV

CHAPTER 1:
THE IMPORTANCE OF BEES FOR CROP POLLINATION,

WITH AN EMPHASIS ON HIGHBUSH BLUEBERRY

INTRODUCTION

Arthropod-mediated pollination is an evolved mutualism in most angiosperms
(Willemstein 1987) and is an essential ecosystem service that directly contributes to plant
productivity in natural and agricultural landscapes (Kevan 1991, Kearns and Inouye
1997, Nabhan and Buchmann 1997, Kevan and Phillips 2001, Potts et a1. 2003, Fontaine
et a1. 2006). Ecosystem services are “the conditions and processes through which natural
ecosystems, and the species that make them up, sustain and fulfill human life” (Daily
1997). Many agricultural crops require insect-mediated pollination for economical yields,
and this is usually provided by bees (Free 1993, Delaplane and Mayer 2000). It has been
estimated that 35% of global crop yields of fruit, vegetables, and seeds are dependent
upon bee-mediated pollination (Klein et a1. 2006).

For crops that require pollination, large commercial farms typically rely on
thousands of honey bee (Apis mellifera L.) colonies that are trucked in to the fields by
itinerate beekeepers. However, there are more than 4000 other bee species that are native
to North America (Michener 2000) and contrrbute to crop pollination at an estimated
value of $3.07 billion of mu and vegetables in the United States annually (Losey and
Vaughan 2006). There is growing concern about the health of both honey bee and wild
bee populations around the world (Allen-Wardell et a1. 1998), and the conservation of
wild bees may be viewed as a strategy towards sustainable crop pollination (Southwick

and Southwick 1992, Kevan and Phillips 2001, Klein et a1. 2006). Bee conservation

efforts have been underway for more than a decade in Europe (Edwards 1996), but there
has been relatively little awareness of the need for pollinator conservation in the US. A
recent report fi'om the National Academy of Sciences (2006) has highlighted that we
know very little about the status of pollinators in North America.

This review will include: 1) a brief history and current status of crop pollination
by bees in North America; 2) the biology of non-Apis bees; 3) agricultural practices that
are likely to affect endemic bee communities; 4) the influence of habitat quality and
surrounding land use on endemic insect communities in agricultural ecosystems; and 5)

the pollination requirements of highbush blueberry (Vaccim'um corymbosum L.).

HISTORY AND CURRENT STATUS OF CROP POLLINATION
IN NORTH AMERICA

Honey bees — the first managed pollinator. Honey bees are native to the Eurasian
continent and have been managed for pollination for the past 50 years (DeGrandi-
Hoffman 2003). Honey and beeswax production have been important to human
civilization for thousands of years, as seen in Neolithic rock paintings, ancient Egyptian
hieroglyphics, and in the writing of Aristotle, Virgil and Pliny (Martin and McGregor
1973). Recognition of the role of bees in plant pollination was not reported until 1682,
and formally documented by Koelreuter in 1761 (as cited in Martin and McGregor 1973).
The first record of honey bees (Apis mellifera L.) being introduced into North America
from Europe is from 1607 when early settlers brought them to Jamestown for their honey

and wax production (DeGrandi-Ho ffman 2003). Thomas Jefferson wrote in 1788 that

native Americans called honey bees the “white man’s flies” as they were often used as
indicators of nearby European settlements (as cited in DeGrandi-Hoffman 2003).

Part of the evolution in thinking about honey bees as crop pollinators was in
response to evolving agricultural practices. Transition fiom the small polycultural family
farms in relatively complex landscape matrices prior to the Great Depression and World
War II to the large monocultural and highly mechanized corporate farms in relatively
simple landscape matrices resulted in greater dependence on managed pollinators. In a
polycultural setting, native bees living in the vicinity were often plentiful enough to
provide crop pollination services and honey bees were kept for the honey and wax they
produced, contributing incidentally to crop pollination. In the current monocultural model
of farming, blocks of flowering crops have become too large for endemic bees to
pollinate, and so the practice of transporting thousandsof honey bee hives to pollinate
these vast plantations is now common (see DeGrandi-Hoffman 2003 for a succinct
review of honey bee history in North America). Indeed, the annual economic value of
honey bees in North America is estimated to be $14.6 billion (Morse and Calderone
2000) ($17.1 billion when adjusted for inflation to represent 2006 dollars).

The majority of honey bee workers are nectar foragers and are typically not the
most efficient pollinators of crop plants (Westerkarnp 1991). However, they make up for
their relatively low pollination efficiency in many crops by being available in great
numbers (Dogterom and Winston 1999). Interest in pollination by native and/or non-
managed pollinators is increasing as evidence mounts that many crops benefit fiom visits
by native bees (Torchio 1994). For instance, some crops have floral morphologies that

make them better suited to native bees that can sonicate, or vibrate, the anthers to release

pollen (e. g. members of the Solanaceae and Ericaceae families) (Buchmann 1983,
Morandin et al. 2001b). Other crops may have sufficient open floral morphology to allow
easy nectar access and pollen transfer, but they bloom in early spring when weather
conditions are highly variable and often colder than when honey bees are apt to forage
(e. g. cherry, apple, and early varieties of blueberry) (Boyle and Philogene 1983). Many
native solitary bee species that emerge in synchrony with the bloom of particular native
cr0p plants are well-adapted to these conditions (MacKenzie and Averill 1995,
MacKenzie and Eickwort 1996, Javorek et aL 2002). A foraging behavior, called trap-
lining, or foraging along a row instead of across rows, can also limit honey bee
pollination potential in crops that provide the greatest yield when cross-pollinated, such
as those that are planted with male pollenizers (apples) or in alternating rows of a
different cultivar (e. g. rabbiteye blueberries, V. ashei). There is some recent evidence that
the presence of native bees has a synergistic effect on honey bee pollination of
sunflowers for the production of hybrid seed, by causing trap-lining honey bees to be
disrupted (Greenleaf and Kremen 2006).

Challenges to traditional beekeeping. Over the past 20-30 years, a combination
of parasitic mites and diseases of honey bees have rapidly reduced the number of
beekeepers and the number of hives available for pollination. Winter mortality in honey
bee colonies in the United States has dramatically increased since the accidental
introduction of two parasitic mites: tracheal mites (Acarapis woodi) in 1984 and Varroa
mites (Varroa destructor (listed as V. jacobsoni in Frazier et a1. 2000)) in 1987 (Frazier et
a1. 2000). There are few effective treatments against the mites, and this has led to

concerns about chemical resistance (Frazier et aL 2000). Also, a number of debilitating

diseases, that affect mainly the brood in a hive, also pose tremendous challenges to
beekeepers (Free 1993). More recently, a mysterious ailment called “Colony Collapse
Disorder,” in which the bees disappear fi'om hives, has caused many commercial
beekeepers to lose large portions of their overwintering stock (Gill 2007).

In addition to disease and mite problems, “Africanized” bees are another
significant concern for the already beleaguered beekeeping industry. Afi'icanized bees
result from the hybridization of a European subspecies with an African subspecies (Apis
mellifera scutellata). The African subspecies is more easily irritated and aggressive and
swarms with greater frequency than the European subspecies, making Africanized
colonies particularly dangerous to nearby workers and livestock (V isscher et a1 1997).
Since 1956 when 24 swarms of imported A. m. scutellata escaped in Brazil, it was
predicted that they would eventually arrive in the United States (Martin and McGregor
1973). Hybridized or “Africanized” bees entered Texas in 1990, and have since spread to
all the states bordering Mexico (Visscher et a1. 1997). Although further encroachment
into the US has been much slower than predicted, Africanized bees have hybridized with
both wild and domesticated honey bee populations throughout the agricultural region of
southern California, causing many beekeepers to go out of business there. As a result hive
shortages have occurred in some areas (Visscher et al. 1997). Since a large proportion of
available hives in the US. pass through the California almond crop, it was predicted that
eventually the Africanized bee would spread further (DeGrandi-Hoffman 2003), and
recently, Afi'icanized bees have been detected in southern Oklahoma, southwestern

Arkansas, and Florida (DeGrandi-Hoffman et al. 2006).

Other managed bees. While honey bees are by far the most important and
abundant managed pollinators, there are several other bee species that are managed and
used for crop pollination. Commercially reared bumble bee colonies have been used in
greenhouses with much success (Morandin et a1. 2001 a, Morandin et al. 2001b), but they
have also been evaluated and used to pollinate crops in outdoor settings (Bohart 1972,
Stubbs and Drummond 2001a). In natural systems, bumble bee queens produced at the
end of the summer overwinter, emerging in early spring to initiate a colony. Most
commercial colonies were originally started by collecting wild Bombus queens and then
manipulating them into beginning their colony much earlier than they would in nature, so
that workers are available to. pollinate early blooming crops (Kearns and Thomson 2001).

Several solitary nesting bees are managed or encouraged to nest near agricultural
fields as well. The alkali bee, Nomia melanderi, is encouraged to nest near alfalfa fields
by providing the right kind of soil in which it will nest (Bohart 1972). Various mason and
leafcutting bees have been encouraged to nest near agricultural fields by providing
manmade straws or blocks of wood drilled with holes. Some growers will incubate the
bees during the winter and then place them in ambient conditions in time for emergence
with a particular crop bloom Mason bee species that have been managed for pollination
of fruit crops include the hom-faced bee (Osmia cornifrons) (Bohart 1972) and the blue
orchard mason bee (0. lignaria) in cherry orchards (Bosch and Kemp 2001), 0. lignarz’a
in almond orchards (Bosch et al. 2000), and 0. atriventris in lowbush blueberry heaths
(Drummond and Stubbs 1997a). Osmia lignaria propinqua has been evaluated for the
pollination of meadowfoam (Limnanthes alba) (Jahns and Jolliff 1991), while the alfalfa

leafcutter bee (Megachile rotundata) has been managed for pollination of alfalfa (Bohart

1972) and lowbush blueberry (MacKenzie and J avorek 1997, Stubbs and Drummond
1997a). Overall, these studies indicate that there are several non-Apis managed bees that
could serve as effective alternatives or supplements to honey bee pollination for some

crops.

AN OVERVIEW OF THE BIOLOGY OF NON-APIS BEES

Taxonomy and morphology. Bees are Hyrnenoptera in the superfamily Apoidea,
and are closely related to sphecid wasps (Michener 2000). The main difference between
bees and their close relatives, aside fiom morphological differences, is that they provision
their nests with pollen mixed with nectar, as opposed to prey items. This behavior, along
with morphological characters such as plumose hairs that attract electrostatically charged
pollen (V aknin et a1. 2000) and dense pollen-carrying hairs on their hind legs, thorax, or
abdomen called scopa (Michener et a1. 1994) make them an important group of
pollinators. Apis mellifera is just one of more than 4000 different species of bees found in
North America (Michener 2000). In Michigan, 398 species have been recorded,
consisting of 48 genera in 6 families (Mitchell 1960, Hurd, Jr. 1979, Michener et a1.
1994)

Bees can be broadly grouped by family according to tongue length, and those with
the shortest tongues are thought to be most closely related to sphecid wasps. There are 5
main families of bees in Michigan. In order fi'om shortest to longest tongues, they are:
Colletidae (2 genera, 32 species), Andrenidae (5 genera, 91 species), Halictidae (13
genera, 104 species), Megachilidae (11 genera, 72 species), and Apidae (l6 genera, 98

species) (Mitchell 1960, Hurd, Jr. 1979, Michener 2000). The details of bee biology

described below are paraphrased fiom Michener’s The Bees of the World (2000), unless
otherwise noted.

Social guilds. Bees can be divided into ecological guilds in several ways. One
way is to consider their sociality. Bees can be solitary with each female provisioning a
nest containing her own offspring, communal with several females sharing the same nest
entrance but provisioning for their own offspring, semi-social with one to a few queens
and a number of female workers that may or may not lay some of their own eggs, social
with one queen and many female workers who are unlikely to lay their own eggs, or
cleptoparasitic or cuckoo bees with females laying their eggs in the nests of other bees.

Most species of bees are solitary, producing one generation per year. Many of
them nest in large aggregations. Males and females of solitary bees usually emerge at
about the same time fiom overwintering as adults or pupa, though there is often some
degree of protandry. Upon emergence they mate, and females either excavate or locate a
suitable nest in which to lay their eggs. The eggs develop into adults that emerge the
following season. Communal nesting is somewhat similar except for the shared nest
entrance and because the females will often take turns guarding the entrance. Many
solitary bees emerge in synchrony with particular plants and are considered to be
oligolectic, i.e. they collect pollen fi‘om plants within a single genus or family.

Semi-social and social bees produce multiple generations per season, and
probably as a consequence, they visit plants fiom multiple families across a broad
temporal range. There is a continuum of behaviors fiom some of the more loosely-social
groups of halictid bees, to the very highly eusocial behavior of honey bees. Division of

labor in semi-social bees can be very flexible. Except for the special case of honey bees,

who maintain a perennial colony, in general, female social bees emerge from
overwintering already mated to begin the establishment of a new nest. Queens must
forage until the first set of workers emerge. Towards the end of the season, drones and
new queens are produced, they mate, and then the new queens find an overwintering site.

Nesting guilds. Another way to group bees is by nesting behavior. Most bee
species excavate nests in soil. Soil nesting bees can be found in every bee family except
for the Megachilidae. Most megachild bees nest in preexisting cavities such as beetle
galleries in logs and use mud, pieces of leaves, or plant fibers to create nest cells. They
can be easily encouraged to nest in manmade straws or blocks of wood drilled with holes
(Shepherd et a1. 2003). Other megachilids use pebbles and resin to construct nests that
they attach to a substrate. Bumble bees (Bombus spp.) in the family Apidae nest in well-
insulated, preexisting cavities and many prefer abandoned rodent burrows or grassy
tussocks (Goulson 2003). Carpenter bees, also in the family Apidae, excavate nests in
solid wood (Xylocopa spp.) or in pithy plant stems (Ceratina spp.).

Both social and nesting guilds are important to consider in bee conservation
efforts. For instance, bees that produce multiple generations throughout the season are
likely to be constrained by the availability of season-long foraging resources and intensity
of agricultural practices, such as pesticide use. Likewise, bees that nest in preexisting

cavities are likely to be constrained by the availability of nesting sites.
AGRICULTURAL PRACTICES LIKELY TO AFFECT ENDEMIC BEES

Agricultural intensification. Mechanized agriculture and the introduction of

inexpensive pesticides developed during World War 11 significantly impacted crop

10

production methods, with the effect that more land was put into production, planted in
large monocultures, and maintained with heavy machinery and broad-spectrum pesticides
(Martin and McGregor 1973). While opportunities for pollination services increased with
these changes in crop production, beekeeping as an occupation became less profitable in
the US and some other countries in the 1950s and 605 due in part to these different land
use patterns that resulted in reduced year-long foraging habitats for bees (Martin and
McGregor 1973). By the 19705, it was recognized that intensive land use was a major
limiting factor to commercial beekeeping and to the conservation of valuable native bee
resources (Martin and McGregor 1973). Effects of agricultural intensification include
direct and indirect risks to bees fi'om pesticide use and the destruction and fi'agmentation
of natural habitats.

Pesticide poisoning to bees is an important problem facing beekeepers (Martin
and McGregor 1973) and the dangers of pesticides, in particular insecticides, to
pollinators are well documented (Johansen 1977, Johansen and Mayer 1990, Stevenson
2003, Chauzat et a1. 2006). Some herbicides are also poisonous to bees, but use of them is
more a concern with regard to the destruction of potential foraging resources (Kevan
1999). Most studies have dealt with toxicity and hazards of direct exposure of honey bees
to pesticides, but these results do not necessarily transfer well to other bees (Johansen and
Mayer 1990, Riedl et a1. 2006). Less known and often overlooked are the sublethal
effects that reduce longevity and adversely affect foraging, memory and navigational
abilities of some bees (Kevan 1999, Stevenson 2003, Chauzat et a1. 2006, Desneux et a1.
2007). One hypothesis about “Colony Collapse Disorder” is that a new neonicotinoid

insecticide, irnidacloprid, is causing the sublethal effect of memory loss, with the result

11

that upon exposure, honey bees are unable to find their way back to the hive (Gill 2007).
Evidence for sublethal effects of neonicotinoids has been recently presented for both
honey bees (Suchail et al. 2001) and bumble bees (Morandin and Winston 2003).

Pressure to reduce the use of organophosphate and other broad—spectrum
insecticides in US agriculture has increased with the implementation of the Food Quality
Protection Act of 1996. With the adoption of integrated pest management programs
designed to protect and enhance populations of non-target, beneficial insects and reduce
environmental impact, grower awareness about pesticide risks has increased. In crops that
rely on insect pollination, it is recommended that growers generally not apply insecticides
during bloom (McGregor 1976, Johansen and Mayer 1990). However, there are relatively
few studies on the response of native pollinators to the use of broad-spectrum insecticides
within typical pest management programs.

Although the risk of bee kill from insecticides in agroecosystems has been
reduced, there have been instances in Canada and in the US in which major losses of bees
have occurred due to pest management activities in urban and arboreal landscapes. These
have typically been associated with widespread mosquito control programs (Kevan
1999). Although not measured, the effects to populations of native bees have been
expected to be severe (Kevan 1999), but there is little information relating specifically to
the effects on non-Apis pollinators. In one documented case, application of the
organophosphate fenitrothion to reduce defoliation of spruce by budworms in eastern
Canada also devastated the native bee pollinators of adjacent lowbush blueberry and
other native flora (Kevan and Plowright 1989). The reduction in fruit set in these plants

reduced the amount of resources available to other animals and created broad disruption

12

to the ecosystem. These ecosystem effects support the idea that pollinators can be used as

sensitive bioindicators of habitat degredation (Kevan 1999).

HABITAT QUALITY AND LAND USE

Declines in abundance of pollinators and other beneficial insects are thought to be
the result of habitat loss and fi'agmentation due to anthropogenic land use and agricultural
practices associated with pest management (Kremen et a1. 2002, Tschamtke et a1. 2005,
Biesmeijer et a1. 2006). Many studies have examined the relationship between local
habitat resources and the abundance and diversity of pollinators (Kells et a1. 2001, Klein
et a1. 2004, Ricketts 2004, Forup and Memmott 2005, waell et a1. 2005, Shuler et a1.
2005, Hegland and Boeke 2006, Pollard and Holland 2006, Marshall et al 2006). Other
studies have looked at pollinator abundance and diversity in relation to landscape
characteristics on larger spatial scales (Steffan—Dewenter et a1. 2002, Westphal et a1.
2003, Westphal et a1. 2006, Winfree et a1. 2006, Chacoff and Aizen 2006). In general,
reduced natural communities of pollinators have been found to limit pollination and
reduce crop yields (i. e. the desired ecosystem service) (Kremen et a1. 2002). This has also
been documented in lowbush blueberry heaths in eastern Canada and Maine (Kevan et al.
1997) and recently in highland coffee production in Indonesia (Klein et a1. 2003).

Habitat quality. Most beekeepers move their hives of A. mellifera several times
throughout a growing season to follow the crops that need pollination services and to
provide their bees with alternative foraging when those crops are finished blooming. This
is not possible for native bees, yet alternative foraging resources are often required to

complete their life cycle. Generally regarded as weeds in crop production systems, the

13

value of non-crop forage to pollinators and other anthophiles is high. Flowering plants
have been considered for use in agricultural settings to help conserve populations of
beneficial insects, including insect natural enemies (Bugg et al. 1989, Maingay et a1.
1991, Bug and Waddington 1994, Pontin et al. 2006) and pollinators (Patten et a1. 1993,
Kearns and Inouye 1997, Carreck and Williams 1997). Flowering plants frequently have
been recommended for attracting beneficial insects in agricultural settings to reduce pest
populations (Baggen and Gurr 1998, Baggen et a1. 1999, Begum et a1. 2006). A few
studies in North America have evaluated native plants for their attraction to bees (Patten
et a1. 1993, Frankie et a1. 2005), and some studies in the United Kingdom have evaluated
pollinator attraction to cultivated (Comba et al. 1999a) and to native or naturalized
(Comba et al 1999b) flowering plants.

The effects of intensive agricultural practices on potential nesting, mating and
foraging habitats for bees (and other beneficial fauna) have been recognized everywhere
agriculture is practiced (Kevan 1999, Tschamtke et al. 2005, Klein et a1. 2006, Winfree et
al. 2006). A number of studies in agricultural systems have suggested that uncropped,
flower-rich habitats directly adjacent to crop fields will increase diversity and abundance
of beneficial insects in the field (Long et a1. 1998, Kells et al. 2001, Croxton et a1. 2002,
waell et a1. 2005, Marshall et a1. 2006), and that hedges adjacent to agricultural fields in
particular can hold high arthropod diversity (Pollard and Holland 2006). The creation of
flower-rich field borders to provide refirge habitats that might stimulate populations of
beneficial insects such as adophagous hoverflies (Syrphidae), ladybird beetles
(Coccinellidae), pollinators and parasitoids (Hyrnentoptera) has been examined in a few

cropping systems (Bugg et al. 1989, Maingay et a1. 1991 , Bugg and Waddington 1994,

14

Pontin et a1. 2006). In California, farmers are planting borders of flowering plants and
perennial grasses around their crops in order to attract beneficial insects (Long et al.
1998). In England, resource strips and set-asides have been studied for their ability to
provide resources in agricultural landscapes for bumble bees (Corbet et al. 1994, Croxton
et a1. 2002, Carvell et al 2006, Carvell et a1. 2007), and in Sweden one study found that
grassland adjacent to agricultural fields provides a source of butterflies and bumble bees
in adjacent agricultural landscapes (Ockinger and Smith 2007).

The encroachment of development and intensive agricultural methods has lead to
fragmentation and scarcity of natural habitats (Edwards 1996, Westrich 1996). This
disturbance is at odds with the needs of stem and ground-nesting bees and of bumble bees
that prefer abandoned field-mice dens and tall grass for their nest sites (Goulson 2003).
Availability of habitats for foraging and nest building is the key limiting factor for native
pollinator populations (Matheson 1996).

Land use effects and proximity to natural habitat. Kremen et a1. (2002)
compared bee abundance and diversity of native bees among melon farms that were
classified as organic-near (organic management surrounded by over 30% natural habitat
within a 1km radius of the field), organic-far (organic management surrounded by <1 %
natural habitat within a 1km radius), or conventional-far (conventional management
surrounded by <l% natural habitat within a 1km radius). These authors found that
proximity to natural habitat and diversity of native bees were key to sustainable
pollination of the melon crops. Because bee abundance and community assemblage

naturally fluctuate from year to year, they found that overall bee diversity was more

15

important than any particular bee species for ensuring that the melon flowers were
pollinated (Kremen et al. 2002).

Landscape context has been examined for many insects of relevance to
agriculture, and several studies have examined the effects of landscape matrix quality in
relation to biological control and biodiversity in and around crops. In Sweden, increasing
landscape diversity around cereal fields corresponded with a decrease in the
establishment of the bird cherry-oat aphid, Rhopalosiphum padi, due to increased
biological control agents inhabiting field margins (Ostman et al. 2001). In South Dakota
alfalfa fields, the quality of the landscape surrounding the field explained a greater
proportion of the variation in predator abundance than did the aphid populations within
the field (Elliott et al. 2002). In southern Mexico coffee plantations, ground-foraging ant
diversity was greater in more complex landscape matrices (Perfecto and Vandermeer
2002). These studies support the concept of landscape complexity as a driver of
biodiversity and improved biologically-based food production.

The spatial scale of plant-insect interactions is based on the biology and behavior
of the organism(s) in question (Thies and Tschamtke 1999, Steffan-Dewenter 2002,
Thies et al. 2002). The foraging range of most native bees is unknown. However, for
those that are, the size of the bee and the relative fi'agmentation of the habitats used for
nesting, mating and foraging appear to determine their range (Westrich 1996, Osborne et
al. 1999, Svensson et al. 2000, Goulson 2003). Recently, Tschamtke and Brandl (2004)
reviewed the current literature on population dynamics theory and trophic interactions

from a landscape perspective in relation to plant-insect interactions. They concluded that

16

the fragmentation of landscapes is important to consider when studying any plant-insect
interactions.

Bumble bees as a case study. Although bumble bee (Bombus spp.) nests can be
difficult to locate in the wild, the habitat requirements and foraging ranges of some
bumble bee species have been well studied compared to most other non-Apis bees. In a
landscape scale microsatellite study in England, B. terrestris and B. pascuorum were
found to have different nest densities and foraging ranges; B. pascuorum had more nests
per area and a smaller foraging range of less than 312 m versus B. terrestris, which will
forage up to 625 m from its nest (Darvill et al. 2004). Another study found that marked B.
terrestris workers could be recaptured while foraging on highly abundant resources, i.e.
crop plants, at distances up to 1750 m fi'om their nest compared to B. muscuorum that
tended to forage in patches that were closer to their nest (Walther-Hellwig and Frankl
2000). A study in Sweden found that bumble bees were most likely found in forest and
field boundaries and in uncultivated open landscapes (Svensson et a1. 2000), and that
while different species preferred different habitats, the preferred nesting sites were

withered grass and tussocks for all species.

POLLINATION REQUIREMENTS OF BLUEBERRY
Blueberry production in Michigan. Highbush blueberry (V. corymbosum) is one
of several cultivated species of Ericaceae plants grown in the United States. Its grth
habit is upright and bushes can reach more than 2 m high. It is planted in rows, and a
field can remain in production for more than 50 years when care is taken in annual

pruning. Large block plantings of single varieties are common to facilitate cultivation and

17

harvesting (Morrow 1943, Free 1993), although this limits opportunities for cross-
pollination Blueberries are dark bluish-black berries that are both hand- and
mechanically-harvested, with the first few pickings typically done by hand. In Michigan,
early cultivars of highbush blueberry begin to bloom in early May, and the latest
blooming cultivars finish by mid-June.

Michigan is the top producer of highbush blueberry in the U.S.; it has 18,500
acres in production in 2003 (2003-2004 Michigan Fruit Inventory) with a farm gate value
of ~$90 million per year. Most of the acreage is located in southwest Michigan in five
counties that border Lake Michigan and benefit fiom weather moderated by the large
body of fresh water. The top three cultivars in terms of acreage are ‘Jersey,’ ‘Bluecrop,’
and ‘Rubel,’ although much of the ‘Jersey’ and ‘Rubel’ acreage was planted prior to the
19705 (2003-2004 Michigan Fruit Inventory).

Pollination requirements of highbush blueberry. It is well-established that
highbush blueberry flowers require the transfer of pollen by insects, but that they do not
require cross-pollination. Merrill (1936) was the first to examine the pollination
requirements of the highbush blueberry and concluded that yields from self-pollination
were as satisfactory as cross-pollination. Later studies found that depending on the
cultivar, cross-pollination may produce slightly larger berries (Schaub and Baver 1942,
Morrow 1943, Meader and Darrow 1947) and faster ripening fi'uit than those that are self-
pollinated (Dorr and Martin 1966). As proof of its need for pollen transfer by bees, Dorr
and Martin (1966) and Brewer et al. (1969) found that blueberry bushes that were caged

to exclude bees produced an eighth of the yield produced by bushes caged with bees.

18

Up until the early 19503, most blueberry growers apparently obtained adequate
pollination fiom native bees, because it was not common for them to rent honey bee
hives. However, by 1965 growers felt that ‘Jersey,’ the predominant commercial cultivar
in production at the time, had been decreasing in yield and that this reduction seemed to
coincide with increasing pesticide use (Dorr and Martin 1966). Anecdotal evidence
suggests that, prior to cultural changes that included clean cultivation and the widespread
use of pesticides, native bees had been abundant pollinators of blueberry in Michigan.
Because of this perceived decline in native bees, honey bee pollination activity on
blueberry was examined and found to be adequate when supplied in sufficient numbers
(Dorr and Martin 1966, and more recently in rabbiteye blueberry by Dedej and Delaplane
2003). Placement of hives singly and evenly within a blueberry plantation was
recommended, because it was observed that pollination by honey bees was best close to
the hive (Martin and McGregor 1973).

In the meantime, new cultivars have been introduced with new studies to
determine whether there is a benefit of cross-pollination. Dogterom et al. (2000)
documented that yields and development of outcrossed ‘Bluecrop’ bushes versus self-
pollinated bushes were not significantly different. This corresponds with an earlier study
by MacKenzie (1997) in which cross-pollination in ‘Bluecrop’ was of no benefit over self
pollination, while two other cultivars (‘Patriot’ and “Northland’) were differentially
benefited, again indicating that any effect of cross-pollination is cultivar-dependent.

Dogterom et al. (2000) also found that 125 pollen tetrads were sufficient to
produce maximum fi'uit set and minimize the time to ripen in a for ‘Bluecrop.’ Based on

the known pollen deposition levels of honey bees (range of 5-20 tetrads/visit) and many

19

native bees (range of 6-1 07 tetrads/visit) in lowbush blueberry (V. angustifoilum) it is
clear that multiple visits from pollinators would be required to provide the optimum
number of pollen tetrads (Javorek et al. 2002).

Efficiency of honey bees vs. native bees. Although most growers rely on the
pollination services of honey bees in Michigan blueberry production, a variety of
commercial pollination alternatives are slowly becoming available and affordable (Free
1993, Torchio 1994, Roubik 1996, Winston 1998). Honey bees are not the most efficient
pollinators for every crop (Bohart 1972). Depending on their handling, some A. mellzfera
colonies can be induced to be better pollen foragers, but there is some question as to
whether that would really make them better pollinators. For example, Dogterom and
Winston (1999) found that honey bee colonies deprived of pollen increased pollen
foraging, but not in adjacent V. corymbosum (cultivar ‘Bluecrop’) fields. On the contrary,
nectar foraging honey bees were more likely to visit V. corymbosum and were probably
providing pollination services as a secondary effect of nectar foraging (Dogterom and
Winston 1999).

Other bees that do not produce harvestable honey have been managed
commercially or experimentally as alternative pollinators to honey bees, and several are
particularly important and proven to be consistently effective pollinators of some crops
(Martin and McGregor 1973, Torchio 1990). Flora] morphology can sometimes prevent
direct access to nectar (Delaplane and Mayer 2000) and highbush blueberry and other
blueberry species fall into this category with their long corolla tubes. Bees with longer
tongues such as bumble bees, or bees that are small enough to climb inside the flower

such as many halictids, may have an easier time manipulating flowers with long corolla

20

tubes (Delaplane and Mayer 2000). In addition, pollen is more easily accessed from the
porous anthers of the blueberry blossom via sonication (or buzz-pollination) achieved by
some native bees such as bumble bees through vibration of their wing muscles upon
landing on the flower (Buchmann 1983, Morandin et al. 2001b). Another advantage of
non-Apis pollinators, even those not able to sonicate flowers, is that some, such as the
alfalfa leaf-cutting bee (Megachile rotundata and other species), will begin foraging
earlier in the day during cooler temperatures when A. mellifera are not yet active
(Delaplane and Mayer 2000, Goulson 2003).

Other managed blueberry pollinators. There are other pollinators with a high
degree of potential for pollination of blueberry, and studies have already been conducted
to evaluate a few non-Apis bees for management in these blueberry crops. An orchard
mason bee (Osmia ribifloris, family Megachilidae) was found to be an effective managed
pollinator for rabbiteye blueberry (Vaccinium ashei) in the southeastern US. (Sampson et
al. 2006). Osmia ribifloris and two other megachilids, Osmia atriventris and the alfalfa
leafcutting bee, Megachile rotundata, were found to be effective managed pollinators for
lowbush blueberry (Vaccinium angustifolium) pollination in Maine (Drummond and
Stubbs 1997a, Stubbs and Drummond 1997a, Stubbs and Drummond 1997b).
Anthophora pilipes villosula and commercially reared Bombus impatiens, both in the
family Apidae, were also shown to be promising alternative managed bees for the
pollination of lowbush blueberry (Stubbs and Drummond 2000, Stubbs and Drummond
2001a)

Community studies of blueberry pollinators. A handfirl of native bee community

studies have been conducted in cultivated blueberry fields. One study identified the three

21

main bee taxa visiting rabbiteye blueberry in southeastern US, which were honey bees,
several species of bumble bees, and a solitary bee, Habropoda laboriosa (Cane and
Payne 1993). The bee community of lowbush blueberry in Maine was sampled by
Drummond and Stubbs (1997b), but the community composition was not reported in
detail. Two studies were conducted in highbush blueberry, one as part of a survey of
native bees in multiple berry cr0ps in the Lower Fraser Valley of British Columbia,
Canada (MacKenzie and Winston 1984), and another in upstate New York (MacKenzie
and Eickwort 1996). A detailed listing of bee species collected in New York was
provided, and a total of 42 species were observed, with the three most abundant species
being honey bees, and two solitary andrenid bees, Andrena carlini and Andrena carolina
(MacKenzie and Eickwort 1996). These studies demonstrate the variation that is typically

observed fiom one blueberry production region to another.

FOCUS OF THIS PROJECT

The rich natural habitats that often surround Michigan croplands may provide the
appropriate habitat for nesting, mating and alternate nectar and pollen resources for non-
Apis bees. But to date, there has been no systematic survey of bees associated with
Michigan blueberry. There is significant potential for conservation of native bees in and
around V. corymbosum fields, to provide a pollinator community that is more reliable
during early spring weather conditions and that can provide sustainable pollination
services to the blueberry crOp during the relatively short window of opportunity for
pollination in Michigan. Bee-toxic pesticides are generally avoided during blueberry

bloom, but insecticide use later in the season may be suppressing the long-term grth of

22

native bee populations in and around fruit crops. Habitat quality and non-crop habitat
surrounding fields may buffer against intensive management practices within fields.

My objectives were: (1) to determine optimum pan-trap placement for capturing
bees in highbush blueberry; (2) to characterize the native bee community and to
determine the degree to which native species contribute to pollination of blueberries in
Michigan; (3) to determine which commercial production practices and local habitat
attributes surrounding V. corymbosum fields are related to the abundance and diversity of
native bees within the fields during bloom; (4) to evaluate native perennial plants for their
relative attractiveness to Michigan bee species; and (5) to determine at what spatial scale

landscape complexity affects bee abundance and diversity in highbush blueberry fields.

23

CHAPTER 2:
ELEVATED PAN TRAPS FOR MONITORING BEES IN CROP CANOPIES:

RESPONSE OF BLUEBERRY POLLINATORS

24

INTRODUCTION

Bees are highly visual insects that respond to visible and UV color patterns on
flowers (Chittka and Menzel 1992, Kevan et al. 1996, Gumbert 2000). This attraction to
color can be eXploited for monitoring bee populations by the use of passive colored pan
traps that provide an inexpensive method to capture bees (Aizen and Feinsinger 1994,
Leong and Thorp 1999, Baum et al. 2006). Pan traps are made from colored receptacles
holding a dilute, aqueous soap solution to trap bees. These traps are typically plastic party
bowls (15 cm diameter) or condiment cups (6 cm diameter) that are already colored. Or
they may be painted with fluorescent paint (LeBuhn et a1. 2002).

The pan trapping approach has important advantages compared to more
traditional bee collection methods that use nets or aspirators. Pan traps eliminate collector
bias, which is particularly important when comparing data across different studies or
when using multiple collectors in the same study. They also provide an inexpensive
method for sampling bee populations using consistent, easily replicated sampling
intensity. Finally, collectors with minimal training can sample over a longer period of
time and at multiple sites simultaneously. As with any sampling method, there are some
drawbacks: pan traps may be biased toward attracting some bee taxa over others, and
they do not allow determination of direct relationships between bee and flower species. In

combination with other methods, however, pan trapping can be very effective. Williams

25

et al. (2001) recommend their use in combination with netting and trap nesting methods
for monitoring bee communities.

The use of pan traps to monitor bee communities is a recent development, but it is
increasingly used as a sampling tool by researchers in natural landscapes and in studies
comparing bee abundance in urban versus natural landscapes. For example, Leong and
Thorp (1999) used blue, white, and yellow bowl traps to sample the bee community
associated with a natural vernal pool plant community in California. McIntyre and
Hostetler (2001) used blue and yellow pans to compare the bee communities associated
with urban land use in southwest U.S. desert ecosystems and surrounding natural
habitats. Russell et a1 (2005) used light blue, dark blue, yellow, and white bowls to
compare bee communities associated with powerline easements and grasslands in
Massachusetts. Brosi et a1. (2007) used pan traps to monitor the bee community
associated with pastureland among forest fragments in Costa Rica. In all of these studies,
the traps were placed directly on the ground among open, or low-growing vegetation.
This deployment of pan traps is appropriate for many natural habitats and so is generally
recommended (see: http://online.sfsu.edu/~beeplot/).

The majority of monitoring for native bees has been in natural habitats, often with
little to no overhead plant canopy (e. g. extensive sampling for bees with pan traps has
been conducted in the Grand Staircase-Escalante National Monument by T. Griswold et
al., unpublished data). However, if this technique is applied in agricultural habitats, it will
be important to consider that many crops have a vertical structure. Canopy structure has
been found to be important when monitoring for pest insects, including several

Rhagoletis flies (Diptera: Tephritidae). Vertical placement of traps has significant

26

implications for monitoring apple maggot fly (Drummond et a1. 1984), blueberry maggot
fly (Teixeira and Polavarapu 2001), and the eastern cherry fi'uit fly (Pelz-Stelinski et al.
2006). Mature highbush blueberry (Vaccinium corymbosum) plants can reach over 2 m,
with the majority of flowers produced in the upper half of the plant, so it is expected that
pan traps placed on the ground would be less visible to bees than traps placed in the
canopy where most bee foraging occurs.

The goal of this study was to determine the response of bees to vertical placement
of traps relative to the canopy of highbush blueberry plants, with the expectation that pan
traps in the highbush blueberry canopy would capture more bees than traps placed on the
ground or above the canopy. I also tested whether three colors of commercially-available
plastic bowls varied in the abundance and diversity of bees collected. Specifically, I was
interested in determining whether bee species foraging on Vaccinium responded
differently to height and color than species not known to forage on Vaccinium. I also
compared timed observations, a method used in previous studies of bees foraging in

blueberry fields, with pan trapping on the same day.

MATERIALS AND METHODS
Response of bees to trap height. This study was conducted in a highbush
blueberry planting (V. corymbosum) at the Trevor Nichols Research Complex near
Fennville, Michigan in May of 2004 and 2005. The height of yellow pan traps was varied
within the canopy of the blueberry plants by placing traps on the ground or on different
lengths of PVC pipe stabilized with rebar. Traps were placed on the ground, mounted

1/3rd of the way up in the canopy (between 0.46-0.6 m), 2/3“‘15 of the way up in the canopy

27

(between 0.9-1.2 m), or above the canopy (between 1.5-1.8 m). Traps were spaced 5 m
apart and arranged in a 4 x 6 Latin square design, with six replications.

In 2004 the study was conducted on 19 May in a mature stand (cv. ‘Rubel’) with
an average height of 1.5 Hi. In 2005 the study was repeated on 25 May and conducted in
two blocks in the mature planting described above, and two blocks in a younger, adjacent
plantation (cv. ‘Bluecrop’), that had an average height of l m. Traps were deployed
during full bloom from 10:00-19:00 11, when weather conditions met the following
criteria: minimum temperature of 13°C with clear or partly cloudy skies or 17°C with any
sky condition other than rain (waell et al. 2005).

Pan traps were constructed fiom 355 ml yellow plastic bowls (sunshine yellow,
Amscan, Inc., Ehnsford, NY) with a PVC coupler (female, double-ended, accepting 2.7
cm PVC) glued to the bottom of each bowl using plumber’s cement so that the bowls
could be mounted onto PVC poles of 2.7 cm diameter. The PVC poles were stabilized by
slipping them over 0.6 m lengths of rebar that were pounded 0.3 m into the ground. Each
pan was half filled (approx. 150 ml) with a 2% unscented soap solution (Dawn® dish
soap, Procter & Gamble, Cincinnati, OH). Insects collected in the pans were strained into
plastic bags, which were then stored in a cooler for transport to a -12°C freezer for later
processing. Specimens were thawed at room temperature, washed in 70% ethanol, and
then placed in a mesh bag through which they were fluffed and dried with a hairdryer
before pinning and identification (LeBuhn et al. 2002).

Preliminary identifications of bees were made using two published dichotomous
keys (Mitchell 1960, Michener et a1 1994) and the online key available through

www.Discoverlife.org. Verification and further identifications were made by Dr. J .S.

28

Ascher of the American Museum of Natural History, Division of Invertebrate Zoology.
Voucher specimens are held in the Albert J. Cook Arthropod Research Collection at
Michigan State University.

Assumptions for equal variance (Levene’s test) and normality (Shapiro-Wilk test)
were met without transformation of the 2004 bee abundance data, which were tested in
relation to trap placement using a mixed model analysis of variance (PROC MIXED,
SAS V9.1). Trap height was the fixed effect and replicate (n = 6) was the random effect.
In 2005, bee abundance was summed across the six traps of each treatment (ground, 1/3rd
campy, 2/3rd canopy, and above canopy) within each block, and then data were square-
root transformed to meet assumptions of normality (Shapiro-Wilk test) and equal
variance (Levene’s test). A mixed model AN OVA with trap height as the fixed effect and
block (n = 4) as the random effect was used to determine whether bee abundance varied
significantly with trap placement. Also in 2005, andrenid and halictid bee abundance
were examined separately using the same methods as for the complete bee fauna. The
Tukey-Kramer means separation test was conducted for all analyses.

Response of bees to trap color. Blue, white, and yellow bowls (marine blue, snow
white, and sunshine yellow, respectively; Amscan, Inc., Ehnsford, NY) were compared
during bloom on 19 May 2004 at the same highbush blueberry (cv. ‘Rube1’) planting
described above. Traps were mounted with PVC couplers on 1.2 m lengths of PVC pipe
as described above, were spaced 5 m apart, and arranged in a 3 x 6 Latin square design
(six replications), from 10:00-19:00 h when minimum weather conditions for bee activity
were met (see above). Specimen handling and identification methods used were the same

as in the height study. Bee abundance was normally distributed without transformation

29

and was analyzed to determine whether bee captures varied among trap colors, using a
mixed model AN OVA with trap color as the fixed effect and replicate as the random
effect. A Pearson chi-square analysis was conducted to determine 1) whether Vaccinium
foragers were more likely to be collected in one color over the other, and 2) whether
andrenids and halictids preferred one color over the others.

Comparison of pan trap sampling to timed observations. On 2 and 7 June 2005,
20 pan traps were deployed between 8:00 and 16:00 hrs at a semi-abandoned blueberry
field near Douglas, Michigan. Pan traps were elevated on PVC poles as described
previously, to within 2/3rd of the highbush blueberry canopy. Meanwhile, 15 l-min
observations were made four times (8:00-9:00, 10:00-11:00, 12:00-13:00, and 14:00-
15:00 hrs) on each day at blueberry bushes in the same field. Bees collected in pan traps
were identified to genus or species using the methods described above. Bees observed
foraging on blueberry were identified to genus (with the exception of genera in the Tribe
Augochlorini) or species when only one species was known to be present in the area (e. g.
Apis mellifera and Xylocopa virginica virginica). Abundance and generic composition of

the samples were compared.

RESULTS
Response of bees to trap height. In 2004, bee captures varied significantly with
trap height (Fug = 9.17, P = 0.001) (Figure 2.1), and the number of bees recovered from
traps 1/3rd of the way up the blueberry canopy was significantly greater than traps placed

at the other three levels (Tukey-Cramer P<0.05). Andrenid and halictid bees were the

30

predominant groups collected, comprising 41 and 32% of the total number of bees
collected, respectively.

When the study was repeated in 2005, including the adjustment of traps relative to
a shorter blueberry stand in two of the replicates, the trend for traps in the canopy to
collect more bees than those on the ground or above the canopy remained (Figure 2.1),
but the number of bees captured was not significantly different among trap positions in
the canopy (F39 = 3.16, P = 0.08). Andrenid and halictid bees were trapped in different
proportions from the year before (28 and 66% of the total number of bees collected,
respectively), and I examined how andrenids and halictids responded to trap height.
Andrenid bees were found to vary significantly with trap height (F 3,9 = 6.93, P = 0.01)
(Figure 2.2), and were significantly more abundant among the bees captured in the
canopy level traps compared to those on the ground or above the canopy (Figure 2.2). In
contrast, halictid bee captures did not vary significantly with trap placement (F39 = 2.36,
P = 0.14) (Figure 2.2).

Bees in the family Andrenidae, of which Andrena carolina (cited as A. longifacies
in LaBerge 1980) is a specialist on Vaccinium, were always recovered from mid-canopy
traps, but less often from ground level or above canopy traps (Table 2.1). Both Apis
mellifera (honey bees) and the bumble bee Bombus impatiens were recovered only fiom
traps elevated above the ground (Table 2.1). Smaller bees such as Ceratina spp. and
Hylaeus affinis were found primarily in ground and mid-canopy traps. However, many of
the small Lasioglossum (Dialictis) spp. were found more often in mid-canopy and above
canopy traps. Eighteen bee species were present in both years of this study, suggesting

they are common in this location during blueberry bloom. There were 6 unique bee

31

species captured in 2004 and 20 unique species captured in 2005, although the total
increase in the number of species captured between years was likely the result of

increased sampling effort in 2005 (Table 2.1).

32

 

 

2004 2005

 

   

 

 

0 2 4 6 8 0 10 20 30 40 50

No. of bees 1 SE.

Figure 2.1. Average number of bees recovered fiom pan traps placed on the ground or
elevated 1/3'd, 2/3'd, or above the canopy within highbush blueberry stands in 2004 and
2005. Bars with the same letter are not significantly different fi'om one another (Tukey-
Kramer means separation, P<0.05).

above h b Andrenidae
2m _.
1,... ba ——

 

Halictidae

 

4
ground _-—4 ab

0 5 1o 15 0 5 10 15 20 25 30
No. of bees :1: SE.

 

 

Trap placement In
relation to the canopy

 

 

 

Figure 2.2. Average number of bees in the families Andrenidae and Halictidae recovered
from pan traps placed on the ground or elevated l/3'd, 2/3'd, or above the canopy within
highbush blueberry stands in 2005. Means with different letters are significantly different
fiom one another (Tukey—Kramer means separation, P<0.05).

33

Table 2.1. Bee species recovered from pan traps placed either on the ground, in mid-
canopy (0.46 to 1.2 m), or above the canopy (1.5 to 1.8 m) in a highbush blueberry field
during bloom near F ennville, Michigan in 2004-05. A (J) indicates the year in which
each species was collected. Presence of female (9) or male (6) bees is indicated by an x.
Family 2004 2005 ground level mid-canopy 3:3),
Species 9 6 9 <3 9 6
Andrenidae
Andrena algida
Andrena arabis
Andrena carlini'l'I J
Andrena carolina'l‘I J
Andrena commoda
Andrena cressonii J
Andrenaforbesii'l' J
Andrena hippotesT
Andrena imitatrix'l' or morrisonella“
Andrena miserabilis
Andrena morrisonella
Andrena nasonii
Andrena perplexa
Andrena vicinaTI
Andrena sp.
Apidae
Apis mellrfera'l'
Bombus impatiensT
Ceratina calcarata
Ceratina calcarata or dupla‘l‘* J
Ceratina dupla‘l’
Ceratina strenua‘l' J x x
Eucera hamata
Colletidae
Hylaeus aflinis
Halictidae
Agapostemon virescens
Augochlorella aurata (=striata)TI
Dufourea marginata
Halictus confususI
Halictus Iigatus
Halictus rubicundusi
Lasioglossum admirandum
Lasioglossum anomalum
Lasioglossum coriaceumI J
Lasioglossum cressoniiI
Lasioglossum imitatumI
Lasioglossum leucozonium J
Lasioglossum nymphaearum
Lasioglossum pectorale
Lasioglossum pilosum J
Lasioglossum quebecensel

'\\\'\\'\
><

\\
\\'\\\
><

'\ \'\\'\
\\
xxxxxxxxxxxxxxx

\\'\'\'\
X
X

\
'\ \
><
>4
><

\\\\
\\
xxxx
xx
xx

\\\'\\
xxxx
xxxxxx

\'\\\\'\\
><><><><><><><><

34

 

ground level mid-canopy above

 

 

 

 

 

Family 2004 2005 canopy
Species 9 (3‘ S? 6 92 6‘
Lasioglossum rohweri J J x x
Lasioglossum tegulare J x x
Lasioglossum spp. J x
Sphecodes dichrous J x
Sphecodes spp. J J x x

Total # of species: 24 38
No. of unique species: 6 20
Total # bees: 72 278

 

TVaccinium floral record noted by Hurd, Jr. (1979). IVaccim‘um floral record based on collections made by
MacKenzie and Eickwort (1996) on highbush blueberry in the Finger Lakes area of central New York.
*These specimens could not be distinguished fiom one another: Andrena imitatrix and A. mom'sonella are
difficult to distinguish when hair on the thorax has been matted down after being wet; Ceratina calcarata
and C. dupla females are morphologically indistinct.

35

Response of bees to trap color. Bee captures, both overall abundance and species
richness, did not vary significantly with trap color (F2,1o = 0.31, P = 0.74) (Table 2.2).
The lowest number of bee species associated with Vaccinium was trapped in the blue
traps, with only two species of Andrena, compared to six and seven species in the white
and yellow traps, respectively (Table 2.2). Andrenid bees were more likely to be found in
white bowls, whereas halictid bees were more likely to be found in blue bowls (df = 2, x2
= 8.75, P = 0.01). However, there was no significant relationship between trap color and
whether bees trapped were known to forage on Vaccinium (df = 2, x2 = 2.22, P = 0.33).

Comparison of pan trap sampling to timed observations. A total of 320 pan-trap
hours and 2 hours of observations were made over the two days of sampling. Pan trap
samples contained more species but fewer specimens compared with the number of bees
observed while foraging on blueberry (Table 2.3). The number of honey bees observed
foraging was 13 times greater than the number collected in traps (Table 2.3). X ylocopa
virginica virginica was only observed foraging and was not captured in pan traps, as was
a species of Megachile. Parasitic bee species were captured in pan traps, but not observed
foraging on blueberry blooms (e. g. Nomada and Sphecodes spp., Table 2.3). A single
bumble bee was collected in a pan trap, but no bumble bees were observed foraging on

Vaccinium during timed observations.

36

Table 2.2. Number and species diversity of bees recovered fiom blue, white, or yellow
pan traps elevated 1.2 m in the canopy of a mature highbush blueberry stand during
bloom near Fennville, Michigan in 2004. All specimens were female unless otherwise
noted.

 

 

 

Family
Species blue white yellow total
Andrenidae
Andrena carliniTI 8 10 3 21
Andrena carolina'l'I l l
Andrena commoda l l
Andrena hippotes‘l' l l
Andrena imitatrixT“ or morrisonella * l 1
Andrena mandibularis l 1
Andrena miserabilis 6 6
Andrena nasonii 2 2 4
Andrena perplexa 1 1
Andrena vicinaTI 2 l 3
Andrena sp. 1 1
Apidae
Apis melliferafi 5 8 8 21
Ceratina calcarataI (6‘ only) 1 1
Ceratina calcarataT"I or dupla'l‘I“ 1 1
Eucera hamata (8 only) 3 3
Colletidae
Colletes thoracicus‘l l l
Halictidae
A gapostemon texanus l 1
Agapostemon virescens l 1 2
Augochlorella striata‘l'I l 1
Hal ictus confususi 2
Halictus ligatus 3 l l 5
Halictus rubicundusi 1 l
Lasioglossum admirandum 2 l 3
Lasioglossum cressoniiI l l
Lasioglossum imitatumi l l 2
Lasioglossum leucozonium 4 l 5
Lasioglossum pilosumI 3 l 4 8
Lasioglossum rohweri 1 l 2
Lasioglossum sp. 3 3
Total abundance: 35 40 29 104
Total no. of species: 13 16 15 29
No. of species previously recorded on Vaccinium: 6 8 9 15

 

TVaccinium floral record noted by Hurd, Jr. (1979). IVaccinium floral record based on collections made by
MacKenzie and Eickwort (1996) on highbush blueberry in the Finger Lakes area of central New York.
*These specimens could not be distinguished fi'om one another: Andrena imitatrix and A. morrisonella are
difficult to distinguish when hair on the thorax has been matted down after being wet; Ceratina calcarata
and C. dupla females are morphologically indistinct.

37

Table 2.3. Comparison of bees captured in pan traps placed in the blueberry canopy with
bees observed foraging on blueberry during timed observations, over two days of
sampling in a southwest Michigan blueberry field in 2005.

 

 

 

 

Observation Bees observed while
Pan-trapped bees Pan Trap total total foraging on blueberry
Apis mellifera 15 201 Apis mellifera
Bombus impatiens l 0
Xylocopa virginica 0 8 Xylocopa virginica
Ceratina calcarata/dupla 15 1 Ceratina sp.
Ceratina strenua 2 -
Nomada sp. 1 0
Andrena sp. 1 l Andrena sp.
Andrena carlini 7 29 Andrena carlini/vicina
Andrena carolina 4 14 Andrena carolina
Andrena hippotes 1 -
Andrena nasonii l -
Andrena vicina l -
Colletes sp. 1 11 Colletes sp.
Hylaeus sp. 3 l Hylaeus sp.
Augochlora para 6 5 Augochlorini
Augochlorella sp. 1 -
Halictus confitsus 4 6 Halictus sp.
Halictus parallelus 1 -
Halictus rubicundus 2 -
Lasioglossum bruneri 2 26 Lasioglossum sp.
Lasioglossum coriaceum 4 -
Lasioglossum leucozonium 2 -
Lasioglossum pilosum 6 -
Sphecodes sp. 1 0
Megachile sp. 0 l Megachile sp.
Osmia spg 2 l Osmia sg
Total: 84 305
Minimum no. of species: 24 13*

 

*T his is a conservative estimate based on the generic level listed here. A more liberal and perhaps
more accurate estimate would be 18 species, based on what was collected in pan traps and what is
found in the floral record literature (see Appendix C).

38

DISCUSSION

Response of bees to trap height. Bee captures were greatest in traps elevated in
the blueberry canopy versus those placed on the ground or above the canopy. Members of
the bee family Andrenidae, many of whom tend to be oligolectic (Michener 2000), were
more likely to be found in the canopy traps nearer to the blooms on which they were
presumably foraging before they were captured. In contrast, halictid bees, which tend to
be polylectic (Michener 2000), were not confined to the canopy traps, presumably
because they were more likely to be searching a broader area for forage. Future studies
that use pan traps to monitor bee communities associated with plants that have a vertical
structure, in which other sampling methods may be difficult (e. g. net sampling) or time
consuming (e. g. observations), should consider placement of pan traps in the canopy. The
optimum height to obtain samples with the highest bee abundance and diversity should be
tested for each plant community.

Response of bees to trap color. Known Vaccinium-foraging bees did not exhibit
marked preferences for trap color, but andrenid bees preferred white traps and halictid
bees preferred blue traps. Leong and Thorp (1999) found that male and female Andrena
limnanthis, an oligolectic bee of white-flowering Limnanthes douglasii rosea (Benth.)
Mason, were most attracted to white pan traps over blue or yellow. Additionally they
found that non-A. limnanthis bees, consisting of generalist and specialist species of
yellow-colored flowers, were most attracted to yellow pan traps. In contrast, I found that
known Vaccinium-foragers, not necessarily all specialists on this plant genus, were found
in similar abundance in white, blue, and yellow traps. However, andrenid species found

in this study were more likely to be found in white traps (although, curiously, the one

39

Vaccinium specialist, A. carolina, was only found in yellow traps — see also Chapter 3). I
also found that halictids, a group of bees that tend to be generalists, favored blue traps
over white and yellow, whereas honey bees showed no significant preference (Table 2.2,
but see Chapter 3, Figure 3.8).

Comparison of pan trapping and timed observations. More bees were observed
foraging on blueberry flowers during the same time period in which pan trap sampling
was being conducted, but species richness was greater in the pan traps. Apis mellifera was
13 times more abundant at blueberry flowers than in pan traps. Species composition was
similar between the methods, although parasitic bees were only caught in pan traps and
X ylocopa virginica virginica was only observed while foraging. That more species were
collected in pan traps than on blueberry, could be a result of sampling at a single plant
species as opposed to sampling at any other plants that may have been blooming in the
landscape. However, when this crop is in bloom, it is the most abundant floral resource in
this landscape (personal observation).

Pan trapping and honey bees. Apis mellifera (honey bees) are often rarely caught
in studies that use pan traps, so it has been generally assumed that pan traps are not a
good method for monitoring honey bees (Cane et al. 2001). Recently in southern Costa
Rica, honey bees were rarely caught in pan traps whereas they were collected in great
abundance in netting samples (Brosi et al. 2007). When intensive net sampling was
compared with pan trapping for bees in northern Virginia, only a single honey bee was
captured in pan traps compared with 204 honey bees netted or observed foraging in the
same area (Roulston et al. 2007). Both of these studies placed pans directly on the ground

and the lack of honey bees in ground level traps in this study follows that pattern.

40

However, honey bees were captured with similar frequency to other bee species in the
height study (Table 2.1), and in greater fiequency than other bee species in the pan trap
color experiment (Table 2.2), when traps were elevated in the canopy. This emphasizes
the need to place pan traps in the appropriate niche used by bees for foraging when
deploying this method for monitoring bees.

Conclusion. The use of pan traps has important advantages compared to more
traditional bee collection methods. Pan traps eliminate collector bias, are relatively
inexpensive, are easily replicated, and can be used over a longer period of time at
multiple sites simultaneously. Although pan traps may be biased toward some bee taxa
over others, in combination with other methods, pan trapping can be very effective for
monitoring bee communities (Kearns and Inouye 1993, pg 269). This study aimed to
optimize pan trapping methods for monitoring the bee community in a crop that has a
vertical structure. From the results presented here, I suggest that attention should be given
to vertical plant structure and that elevated pan traps may ensure the greatest abundance

and diversity of bees in pans.

41

CHAPTER 3:

NATIVE BEES ASSOCIATED WITH THE HIGHBUSH BLUEBERRY

AGROECOSYSTEM IN MICHIGAN

42

INTRODUCTION

Highbush blueberry (Vaccinium corymbosum L.) is a native North American crop
that is dependent upon pollination for optimum yields (McGregor 1976, Free 1993,
Delaplane and Mayer 2000). A number of native bee species, such as several Andrena,
Osmia, and especially Bombus spp., are efficient pollinators of Vaccinium. Some are able
to sonicate the porous anthers of Vaccinium flowers or will forage under cooler weather
conditions than honey bees (Buchmann 1983, Heinrich 2004). Some visit more flowers
per minute and deposit more tetrads per visit than honey bees (Dogterom 1999, Sampson
and Cane 2000, J avorek et al. 2002, Sampson et a1. 2006). Together, these traits make
native bees efficient pollinators of blueberry.

Prior to the current large-scale production of highbush blueberry, endemic native
bees and feral honey bees were largely responsible for its pollination (Marucci and
Moulter 1977, DeGrandi-Hoffman 1987). It became necessary for commercial growers to
supplement wild pollinators with managed honey bee (Apis mellifera) hives when
commercial acreage increased and pest management practices grew more intensive,
including the use of herbicides to clear vegetation surrounding fields that would have
supported wild pollinators when the crop was not in bloom. Although honey bees are not
the most efficient at pollinating Vaccinium, when there are enough of them, adequate
pollination can be achieved (Dogterom and Winston 1999, Dedej and Delaplane 2003).

Therefore, honey bees have become indispensable for most crops that require pollination

43

to produce profitable yields (Southwick and Southwick 1992, Roubik 1996) including
blueberry (Dorr and Martin 1966). In the mid-sixties, Wood (1965) reported that bumble
bees and solitary bees were common in North America, but usually not in adequate
numbers to pollinate commercial crops.

About the time that honey bees began to be managed extensively and transported
across the US. for pollination, beekeepers began to face annual hive losses due to several
illnesses and parasitic mite infestations, such that by the late 1980S and early 1990s,.there
was concern about the state of the honey bee industry (Torchio 1990, Watanabe 1994,
DeGrandi-Ho ffrnan 2003). As of this writing, honey bees are threatened with a
mysterious ailment called “colony collapse disorder,” in which beekeepers find hives full
of honey but with no bees (New York Times, February 2007). In the winter of 2006-7,
some beekeepers have reported losses of up to 90% of their colonies. It remains to be
seen how this will impact crop pollination, and whether these beekeepers will recover
quickly from these losses.

At the same time, there has been increasing concern over the perceived loss of
pollinator biodiversity around the world, with a call for studies to better understand the
current extent of pollinator populations, including native bees (Kearns 1998, Allen-
Wardell et al. 1998, Cane and Tepedino 2001, National Academy of Sciences 2006).
Surveys of native bees associated with lowbush blueberry production in Maine
(Drummond and Stubbs 1997), rabbiteye blueberry in South Carolina (Cane and Payne
1993, Sampson and Cane 2000), and highbush blueberry in upstate New York
(MacKenzie and Eickwort 1996) have been conducted previously. These studies have

focused on bees foraging during blueberry bloom, but many of these species are likely to

be present prior to and/or after blueberry bloom, which means that in order to help
conserve and eventually increase their abundance, they require floral resources, nesting
habitat, and protection fiom production practices aimed at pest insects outside of the
bloom period of the crop. To my knowledge, no native bee survey on Michigan blueberry
has been conducted, even though Michigan is the leading producer of blueberries in the
US, with 18,500 acres in production (USDA 2004) valued at ~$90 million per year.

Pan trapping, direct observation, and pollen analysis from bee specimens were
used to determine the relative abundance and diversity of wild bees associated with
highbush blueberry agroecosystems in southwest Michigan before, during, and after
bloom.

MATERIALS AND METHODS

A three-year study was conducted to characterize the bee community active
during blueberry bloom at 13 commercial blueberry farms and 2 semi-abandoned
blueberry fields located in the highbush blueberry production region of southwest
Michigan (Figure 3.1). Six sites were located in Ottawa County, north of Holland,
Michigan, five sites were located in Allegan County, and the remaining four sites were
located in Van Buren County (see Appendix A). Each sampled field was at least 3 km
away fiom any others in this study.

Passive collections of bees were made using pan traps (Chapter 2) and direct
collections of bees were made during timed observations. To determine which bees were
pollinating blueberry, the proportion of Vaccinium pollen was analyzed fiom bees
collected in pan traps and while foraging on blueberry flowers. Bee sampling was

conducted when weather conditions met the following criteria: minimum temperature of

45

13°C with clear or partly cloudy skies or 17°C with any sky condition other than rain

(waell et al. 2005).

 

 

 

 

 

 

 

 

 

 

 

 

i
= - 3 _L_‘_
a, -
1 l
J l /

_L3J

 

 

 

 

 

 

 

Figure 3.1. Location of 15 bee collection sites in relation to the top five blueberry
producing counties in Michigan. Van Buren, Ottawa, and Allegan Counties are 15‘, 2““,
and 3rd in blueberry acreage in Michigan with 7550, 5300, and 2750 acres, respectively
(USDA 2004). This figure is presented in color.

46

Pan trapping. Each of the fifteen farms was sampled during bloom (2004-O6)
using pan traps (Table 3.1). Due to varying weather conditions from year to year,
trapping was conducted two (2004, 2006) or three (2005) times during bloom in each
field. Pre-bloom pan trapping was conducted in 2005-6 and an additional post-bloom pan

trapping sample was conducted in 2005-6.

Table 3.1. Dates between which pan trapping was conducted in Michigan blueberry
fields prior to blueberry bloom (Pre-bloom), during blueberry bloom (Bloom), and after
blueberry bloom (Post-bloom), and the number of times traps were deployed in each field
during each time period (11).

Year Pre-bloom 11 Bloom Post-bloom

 

n
2004 n/a 16 May — 3 June 15 June - 4 Sept 2
3
3

n
2
2005 15 —21 April 1 16—25 May 3 22 June— 15 Sept
2006 19—26 Amril 1 17— 31 May 2 12 June— 10 August

 

Five pairs of white and yellow pan traps mounted on 1.2 in PVC poles were
placed 5 m apart along each of two transects running perpendicular to the orientation of
the rows. One transect was established within 1 m of the field edge and the other was
established 25 m into the field. Traps were set out between 8:00-12:00 h and were
collected between 16:00—20:00 h for a minimum trapping period of 6 h on days when
suitable weather conditions, as described above, were met.

Pan traps half filled with a 2% unscented soap solution (Dawn® dish soap,
Procter & Gamble, Cincinnati, OH), were constructed from 355 ml white and yellow
plastic bowls (Amscan, Inc., Ehnsford, NY) mounted onto 2.7 diameter PVC poles
stabilized with rebar (see Chapter 2, page 37). After the sampling period, pan trap

contents were strained into plastic bags and stored in a -12°C freezer for later processing.

47

Specimens were thawed at room temperature prior to washing in a 70% ethanol solution.
Honey bees were separated out and counted, then stored in 70% ethanol solution. Pollen
samples were taken (when present) from bees collected during bloom, then wild bees
were placed in a mesh bag through which they were fluffed and dried with a hairdryer
before pinning and identification.

Species accumulation curves were based on randomized re-sampling of bee
trapping observations with 100 permutations in R 3.2.1 (“vegan” package, specaccum
fimction). A 2-way analysis of variance (PROC GLM, SAS 9.1) was conducted to
examine the response of bees to trap position in the field (edge vs. interior) and trap color
(white vs. yellow) by year with Tukey means separation. This model was used to test the
response of native (non-Apis) bee abundance (log n+1), native bee species richness,
native bee diversity (Shannon-Wiener H’), and honey bee abundance. The abundance of
eight of the most common species that have been recorded foraging on blueberry (Hurd,
Jr. 1979, MacKenzie and Eickwort 1996, and data from this study) was pooled across
years and also tested for response to trap position and color.

Pollen analysis. Pollen samples were brushed from corbiculae on honey bees and
scopa on all other bees collected during timed observations and in pan traps, using a fine
paint brush Each pollen sample was stained by mixing with melted basic fiischin gel on
glass microscope slides (Kearns and Inouye 1993). Pollen slides were examined under a
400x light microscope and the number of tetrad pollen grains (i.e. Vaccinium) out of 100
was recorded. The proportion of Vaccinium pollen was calculated per bee and averaged

over each bee species from which pollen was collected.

48

Direct bee observations. Timed observations of bees visiting blueberry flowers
were conducted at three of the commercial sites and at the two semi-abandoned blueberry
farms in 2004-06. Fifteen randomly selected bushes were observed for one minute each,
on three occasions during bloom in each field. Observations were conducted during times
when conditions were suitable for bee activity. Bees were identified as honey bees,
bumble bees, or “other” bees. “Other” bees were collected for identification (n = 62).

Species identifications. Preliminary identifications of bees were made using two
published dichotomous keys (Mitchell 1960, Michener et a1. 1994) and the online key
available through www.discoverlife.org. Further identifications and verifications were
made by J .S. Ascher of the American Museum of Natural History, Division of
Invertebrate Zoology. Voucher specimens are held in the Albert J. Cook Arthropod

Research Collection at Michigan State University (See Appendix B).

RESULTS
Over three years across the 15 farms, 7929 bees were collected in pan traps,
representing at least 174 species, in five families and 30 genera (Table 3.2). Each site was
sampled 17 times, representing more than 1300 pan-days of trapping effort. Species
accumulation curves created using pan trapping data collected during bloom in 2004
approached an asymptote (Figure 3.2), indicating that pan trapping effort was sufficient

to represent the community of bees likely to be captured in pan traps in this habitat.

49

 

80
I

60

20
l

 

 

 

 

I I I I l l I
0 5 10 15 20 25 30
Index

Figure 3.2. Species accumulation curve generated from 100 permutations of the 2004
pan trap sampling data Species curves generated fiom 2005-6 data were similar.

50

Table 3.2. Bee species collected in pan traps in blueberry fields in southwest Michigan
over a period of three years beginning in 2004. Samples were taken prior to blueberry
bloom (pre), during bloom (bloom), and after bloom (post).

 

 

 

2004 2005 2006
Family bloom post pre bloom post pre bloom post
Species 11 = 2 2 l 3 3 l 2 3 total
Andrenidae

Andrena (Melandrena)

sp. - - - - - 1 - - l
Andrena algida 1 - 27 l - 9 l - 39
Andrena alleghaniensis 3 - - 8 - - 9 2 22
Andrena arabis - - l - - 1 - - 2
Andrena barbilabris - - l 8 - l 2 - 12
Andrena bisalicis — - - l l 5 5 4 l6
Andrena carlini 59 - 21 60 - 21 18 - 179
Andrena carolina l 14 - 19 237 - 14 101 4 489
Andrena ceanothi 4 - - 2 - - 8 l 15
Andrena clarkella - - 1 1 - l - - 3
Andrena commoda 2 - - 2 - — - - 4
Andrena crataegi 7 - - l - - 3 3 l4
Andrena cressom'i 9 2 27 22 l 9 3 77
Andrena dunningi l - - - - - - - 1
Andrena erigeniae - - - 2 - - - - 2
Andrena erythrogaster - - 1 l - l - - 3
Andrena erythronii - - 8 - - 1 1 - 10
Andrenaforbesii 5 - 6 - - 5 7 - 23
Andrenafrigida - - 2 - - 2 - - 4
Andrena geranii - - - - - - - 1 l
Andrena hippotes 5 - 2 4 - 4 - 17
Andrena hirticincta - 2 - - - - - 2
Andrena imitatrix or

morrisonella 8 - 5 27 - 3 19 - 62
Andrena integra 2 - - - - - - - 2
Andrena mandibularis - - - 2 - - - 2
Andrena mariae l - - - - - - - l
Andrena milwaukeensis l - - - - - - - l
Andrena miserabilis 9 - 19 45 - 56 37 l 167
Andrena morrisonella 3 - - 4 - - 3 l 11
Andrena nasonii 7 - 22 12 - l4 9 1 65
Andrena neonana l - - - - - - l 2
Andrena nigrae 3 - - - - - 5 - 8
Andrena nivalis 1 - - - - - - - l
Andrena nuda 15 - - 2 - - 6 9 32
Andrena perplexa 7 - - l - 2 1 l 12
Andrena persimulata - - - - - - 1 - 1
Andrena placata - 5 - - - - - - 5
Andrena platypaea - - - - - - - 1 l
Andrena pruni 2 - - l - - - 3
Andrena rehni 2 - - - - - - - 2
Andrena robertsonii 1 - - l - - 2 - 4

51

 

 

 

2005 2006
bloom post pre bloom post pre bloom post total

Andrena rugosa 10 - l 4 - 1 6 - 22
Andrena salictaria l - - 5 - - - - 6
Andrena Sigmundi - - - - - 2 - - 2
Andrena sp. (females

only) - - 3 - - - - - 3
Andrena spp. (males

only) 70 4 361 138 - 304 70 10 957
Andrena thespii - - - - - - 2 - 2
Andrena tridans - - 1 - - - - - 1
Andrena vicina 49 - 7 43 - 6 24 - 129
Andrena wellesleyana - - 4 - - 1 - - 5
Andrena wilkella - - - - - — - 3 3
Calliopsis

andreniformis - l - - l - - 1 3
Perdita octomaculata - - - - 1 - - - 1
Pseudopanurgus

nebrascensis - - - - 3 - - - 3

Apidae (except for Apis mellifera,which is shown at the end of the table)

(Anthophorini) spp. - - - 1 3 - - - 4
Anthophora terminalis - 8 - - - - - 2 10
Bombus bimaculatus 6 - - - — - 2 - 8
Bombus citrinus 16 l - - - - 1 1 19
Bombusfervidus - 1 - 3 - - 2 - 6
Bombus griseocollis 3 1 - - - - 4 3 1 1
Bombus impatiens 2 - - 2 10 - - 5 19
Bombus perplexus 4 1 - - 1 1 - 2 9
Bombus vagans - l - - - - - 1
Ceratina calcarata

(males only) 3 23 296 13 22 94 32 7 490
Ceratina calcarata or

dupla (females only) 56 162 123 54 89 17 13 160 674
Ceratina dupla (males

only) 1 34 7 1 l 3 15 62
Ceratina strenua 9 45 39 6 3 l l 1 - 39 180
Eucera atriventris l - - - - - - - 1
Eucera hamata - - - - - - - 3 3
Melissodes agilis - - - - 1 - - - 1
Melissodes apicata - 2 - - - - - - 2
Melissodes bimaculata - 5 - - 3 - - 1 9
Melissodes communis - 3 - - - - - - 3
Melissodes druriella - - - - 3 - - - 3
Melissodes spp. - 6 - - 3 - - 1 10
Melissodes tridonis - l - - - - - 2 3
Nomada cressonii - - - l - - - - 1
Nomada denticulata - - - 2 - - - - 2
Nomada depressa - - 3 - - - - - 3
Nomada luteoloides - - 7 - - - - - 7
Nomada maculata - - l7 1 - - - - 18
Nomada obliterata - - 2 - - - - - 2

52

 

 

 

2004 2005 2006
bloom post pre bloom post pre bloom post total
Nomada ovata - - - l - - - - 1
Nomada pygmaea - - 1 - - - - - 1
Nomada spp. 32 - 86 47 l 70 13 40 289
Triepeolus lunatus - l - - - - - - 1
X ylocopa virginica
virginica 2 - - 1 - - - - 3
Colletidae
Colletes inaequalis - - 39 15 1 - 3 - 58
Colletes sp. - l 2 - - - - - 3
Colletes thoracicus 24 - 4 4 - 14 1 - 47
Colletes validus - ~ - 2 - - - - 2
Hylaeus afiinis 2 37 - 3 36 - 1 5 84
Hylaeus rudbeckiae - - - 3 2 - 2 26 33
Hylaeus sp. - - - 1 - - - 1
Halictidae
Agapostemon sericeus 1 2 1 3 - - 4 3 14
Agapostemon splendens 2 4 - l 2 - - - 9
Agapostemon texanus — - - 2 l - 1 - 4
Agapostemon virescens 3 9 - 5 10 - 3 3 33
Augochlora para 42 3 10 13 l '1 - l l 1 91
Augochlorella aurata 58 56 16 32 36 15 64 47 324
Augochlorella gratiosa 10 2 - - - - - - 12
Augochloropsis
sumptuosa 1 - - - - - - _ 1
Dufourea marginata - - - - 1 - - - 1
Halictus confusus 2 6 5 27 6 - 4 4 54
Halictus ligatus 19 33 - 20 35 l 10 9 127
Halictus parallelus 2 3 - 1 - - 2 2 10
Halictus rubicundus 3 2 - 7 - 1 3 1 17
Lasioglossum
(Dialictus) spp. - ll - - - - - 6 17
Lasioglossum
(Evylaeus) sup. 8 4 - - - - - - 12
Lasioglossum
acuminatum 4 - - 3 l l 5 l 15
Lasioglossum
admirandum 27 6 5 13 14 6 19 16 106
Lasioglossum
anomalum 1 l - 4 3 - 1 1 1 1
Lasioglossum
athabascense - - - - - 1 - - 1
Lasioglossum boreale 2 - l - l - - - 4
Lasioglossum bruneri 3 2 - 1 1 1 - 1 9
Lasioglossum
coeruleum 8 - 2 5 3 l 4 1 24
Lasioglossum
coriaceum 61 2 3 17 3 1 12 7 106
Lasioglossum cressonii 79 3 - 17 6 3 l7 19 144
Lasioglossumfattigi l - - 1 l - - - 3
Lasioglossumforbesii - - - - 1 1 - _ 2

53

 

 

 

2004 2005 2006
bloom post pre bloom post pre bloom post total
Lasioglossum
firscipenne 5 - - l - 1 l 1 9
Lasioglossum illinoense - - - - — - - 1 1
Lasioglossum imitatum 25 12 23 72 10 13 8 21 42 343
Lasioglossum
leucozonium 68 136 - 27 171 - 56 174 632
Lasioglossum nelumonis l - - - - - - - 1
Lasioglossum
nigroviride 3 - - - - - 3 - 5
Lasioglossum
nymphaearum l 2 - l l - 2 - 7
Lasioglossum nymphale - - l - - - - - 1
Lasioglossum oblongum 2 - - - - - - - 2
Lasioglossum pectorale 3 l4 - 14 24 - 1 l 25 91
Lasioglossum pilosum 45 26 54 147 62 4O 67 46 487
Lasioglossum
quebecense 8 l 3 6 - 3 7 3 31
Lasioglossum rohweri 47 6 7 28 3 2 l4 5 l 12
Lasioglossum spp. - - l - 9 - - - 10
Lasioglossum suvianae - - - - l - - - 1
Lasioglossum tegulare 6 2 - 5 5 1 14 14 47
Lasioglossum versans l - - - - - 1 - 2
Lasioglossum vierecki 1 6 3 6 l3 4 - 17 50
Sphecodes confertus l 1 l - p - - - - 3
Sphecodes cressonii - - - - 4 - - - 4
Sphecodes mandibularis - - - - 3 - - - 3
Sphecodes ranunculi l - - - - - - - 1
Sphecodes spp. 4 6 5 l 14 4 2 6 42
Megachilidae
Anthidium manicatum - 5 - - - - - 3 8
Ashmeadiella sp. - - - l - - - - 1
Coelioxys sp. - - - - 1 - - - 1
Dianthidium simile - l4 - - 2 - - - 16
Heriades leavitti - - - - - - - 1 1
Heriades variolosus - l - - - - - - 1
Hoplitis producta - - - 2 1 - - - 3
Hoplitis spoliata - - - l - - - - 1
Megachile albatarsis - - - - - _ 7 8
Megachile brevis - - - - - - - 1 1
Megachile
centuncularis - - - - 1 - _ - 1
Megachile mendica - - - - l - - - 1
Megachile montivaga - l - - - - - - 1
Megachile pugnata - 5 - - - - - - 5
Megachile rotundata - - - - 1 - - - 1
Megachile spp. 2 1 - 2 l - 1 6 13
Osmia a/m/p 8 3 1 l6 2 - - 1 130
Osmia atriventris - - 6 - - 15 - 2 23
Osmia atriventris or
pumila - - - - - 25 - - 25

54

 

 

 

 

 

 

Osmia bucephala 3 - l6 2 - 8 l 3 33
Osmia conjuncta - - 5 2 - - - 1 8
Osmia distincta - - - - - 1 - l 2
Osmiafelti - - - - - - - l l
Osmia georgica - - - - - - - 1 l
Osmia lignaria - - 7 - - l - - 8
Osmia michiganensis or
illinoensis - - 3 - - l l - 5
Osmia pumila - - 3 3 - 23 - 4 33
Osmia simillima - - - - - l - - l
Osmia spp. - - 5 2 l - l - 9
Osmia subfasciata - - 7 - - - - - 7
Osmia virga - - - - - 4 — - 4
2004 2005 2006
bloom post pre bloom post pre bloom post “It!“
Abundance
(without A. mellifera): 1136 704 1501 1298 681 969 794 846 7929
Apis mellifera
abundance: 2382 40 16 1347 6 8 851 55 4705
Total abundance: 12634
No. of species: 86 58 61 84 62 59 70 73
Shannon-Wiener H'
(calculated without A.
mellifera): 3.54 2.86 2.78 3.25 2.91 2.63 3.41 3.05
Total no. species: 175

 

55

Bee community structure. Apis mellifera was the most abundant species trapped
during bloom, with all other bees comprising one third to one half captured during bloom
Most of the non-Apis bees in the family Apidae were Ceratina species. Although Bombus
spp. were caught in low numbers, 7 species are represented (Table 3.2). Bees in the
families Andrenidae (between 34-46%) and Halictidae (45-55%) were among the most
abundant native bees trapped during blueberry bloom in all three years (Figure 3.3). The
most speciose genus was Andrena at 49 species, followed by at least 28 Lasioglossum
spp. (Table 3.2). Bees in families Colletidae and Megachilidae were rare in the pan
trapping samples each year. The overall proportion of bees within each family remained
relatively stable from year to year, however, fewer bees were trapped during bloom in
2006 compared to the other years, likely due to cooler spring weather conditions in that
year.

Looking at bees across the entire season by nesting guild, soil nesters were the
most abundant bees collected each year (66-75%), followed by pithy stem nesters (15-
21%) (Figure 3.4). Cleptoparasitic bees and cavity nesting bees were in equal abundance
all three years (between 5-6%) (Figure 3.4). More rare were wood-boring bees (e. g.

X ylocopa virginica virginica), and pebble and resin nest builders (e. g. Dianthidium
simile) (<1-2%).

By far the most abundant non-Apis species captured during bloom over the three
years at 14% of the total bee abundance was the Vaccinium specialist, Andrena carolina
(Table 3.3). Lasioglossum pilosum comprised 8% of the total, followed by Augochlorella
aurata and L. leucozonium at 5% (Table 3.3). Eight species were abundant between 3-

4%, 18 species were abundant between 1-2%, and the rest (90 species) were present at

56

less than 1% of the total. A total of 120 bee species were trapped in bowl traps during
blueberry bloom over the three years (Table 3.3).

The eight most abundant native bee species present every year during bloom that
are known to forage on Vaccinium were Andrena carolina, An. carlini, An. vicina,
Ceratina calcarata (or dupla; the females are indistinguishable, however there were
many more male C. calcarata than there were C. dupla), Augochlorella aurata (which
includes the species formerly known as Au. striata), Lasioglossum (L. ) coriaceum, L.
(Dialictus) imitatum, and L. (Dialictus) pilosum. All except for C. calcarata are ground
nesting bees (Michener 2000). All of these species were also present prior to bloom
(2005 and 2006, Figure 3.5). Five of these species were also present after bloom: Au.
aurata, Bombus spp., C. calcarata/dupla, L. imitatum, and L. pilosum (2004-06, Figure

3.5).

57

.830 E Bianca m can Mat. 60-38
Sow SamEouz 5an38 E mg 2 as 883 @3053 wage page: >35.“ .3 83 35...:05 pm? me :oEomoE .m.m .95»:—

       

        

        

3 882V 3 88,2: 3 8&9:
«mun: guru: 0min:
ooou mean 3cm
$3
a 9.
gene acumen: cougar
882...“:
$3
$3. 3:.
$25.22 omv_uo._uc< cau.co._uc<
«on so .
o\om o\ F omua< as; 82?. xi
822 .\. F 82.2.".«82 . a.” 82:680.). .i mau___5m82

832.8 832.8 882.8

58

.830 E 36305 mm can flag. .883 88:22 Baa mm .3205 weumoz 6.39m ooh: B>o newEomE 803558
E madam zbonoan 2 $06.... season 05 30:95:: comman— EEM wagon Juno 5 moon QEVéQG 25 mo come—098m dam 95mm..—

aconu: 09%.: 32. u:
coon mocu VOON
{own

$8 $2
mcozm>wox0 MCOSG>8XO
:8

     

                

           

   

9.26598
.8
.\. N =8 $2 _.
wcozgmoxo mco=m>moxo
$3 Eofi >55 Eon Eta
mco=w>ooxo o
5.... an...
sew acuflxooa o mczflxoou
8:38 o o \n. v
mczflxooa \o v \o@ wco=m>moxo $N
o\ m cozmfioxo 02853520 woo;
0
86230320 took 9.30..

053.8

wcoam>8xo .x. v .0320

89.5 9.52 Eweh a $323

    

oer
Ewe. w $.38

.xh
wco=w>aoxo
poo; 9.32

59

Table 3.3. Species of native bees (n = 120 species and 3228 specimens) listed in order of
most to least abundant over three years of collecting during bloom in 15 southwest
Michigan blueberry farms from 2004-06.

 

 

proportion of

Family Species bees collected
Andrenidae Andrena carolinal‘z’ 3 0.14
Halictidae Lasioglossum pilosum2 0.08
Halictidae Augochlorella auratal’z' 3 0.05
Halictidae Lasioglossum leucozom'um 0.05
Andrenidae Andrena carlini 1’2’ 3 0.04
Apidae Ceratina calcarata or dupla (females only) 1’ 2’ 3 0.04
Halictidae Lasioglossum imitatumz’ 3 0.04
Andrenidae Andrena vicinal’ 2’ 3 0.04
Halictidae Lasioglossum cressoniiz' 3 0.04
Apidae Nomada spp. 3 0.03
Andrenidae Andrena miserabilis3 0.03
Halictidae Lasioglossum coriaceum2 0.03
Halictidae Lasioglossum rohweri 0.03
Halictidae Augochlora pural.2. 3 0.02
Halictidae Lasioglossum admirandum 0.02
Andrenidae Andrena imitatrix or morrisonella3 0.02
Halictidae Halictus ligatus 0.02
Apidae Ceratina calcarata (males only) 0.01
Andrenidae Andrena cressonii 0.01
Halictidae Halictus confususz‘ 3 0.01
Colletidae Colletes thoracicusl‘ 3 0.01
Andrenidae Andrena nasonii 0.01
Halictidae Lasioglossum pectorale 0.01
Halictidae Lasioglossum tegulare 0.01
Andrenidae Andrena nuda 0.01
Halictidae Lasioglossum quebecense2 0.01
Andrenidae Andrena alleghaniensis 0.01
Andrenidae Andrena rugosa2 0.01
Colletidae Colletes inaequalisz’ 3 0.01
Apidae Bombus citrinus 0.01
Halictidae Lasioglossum coeruleum 0.01

All other species (n = 90) present at <1% 0.19

 

I Vaccinium floral record in Hurd (1979).
2 Vaccinium floral record in MacKenzie and Eickwort (1996).

3 Vaccinium record in this study either from pollen samples or direct observations (Tables
3.4 and 3.5).

60

 

_ Andrena carlini

—— Andrena carolina
_ Andrena vicina

— Augochlorella aurata

' ' ' ' Bombus spp.

- ' ' ' Ceratina calcarata/dupla
' - - Lasioglossum coriaceum
- - Lasioglossum imitatum
— — Lasioglossum pilosum

No. bees rapped per da

 

 

No. bees trapped per day

    
 

No. bees trapped per day

 

 

 

. ’ 1‘“ I u
6%...- Mgkk
- .‘1L‘l";1E1,-§5 ‘hi‘-~i_

     

 

sample time In relation to
blueberry bloom

Figure 3.5. Incidence throughout the season of the most abundant native bee species that
are known to forage on Vaccinium spp. plus all the Bombus spp. that were trapped
throughout the study in fifieen highbush blueberry fields in southwest Michigan from
2004-06. The shaded area denotes blueberry bloom. This figure is presented in color.

61

Pollen analysis. Andrena carolina showed a high level of floral constancy for
Vaccinium; nearly all specimens were found to carry ~100% pure Vaccinium pollen
(n=37, Table 3.4). A. vicina (n=12) and A. carlini (n=12) carried pure loads of Vaccinium
pollen in about half of the specimens (0.40 and 0.53 respectively, Table 3.4). The species
group labeled Andrena sp. 2, which probably included mostly A. carolina (these were
specimens from which pollen was obtained before they were assigned an ID number and
could later be tracked to species identity), were more likely to carry 100% pure
Vaccinium pollen than other species of pollen (n = 15, on average 0.72, Table 3.4). Very
few samples of other species were analyzed for pollen composition, but the data agree
with previous records (Hurd, Jr. 1979) that several species of Colletes and several halictid
species also collect Vaccinium pollen (Table 3.4).

Direct observations of bees foraging on blueberry. From timed observations
made in commercial and semi-abandoned blueberry fields during bloom, honey bees far
outnumbered non-Apis bees in commercial fields by almost 33:1 (Figure 3.6). In semi-
abandoned fields, where no honey bee hives were installed, the ratio of honey bees to
wild bees was 3:1. During these observations, I collected 62 non-Apis bees visiting
blueberry for a total of 10 genera and 21 species (Table 3.5). The most abundant non-
Apis bees observed visiting bloom were Andrena carolina and An. carlini, followed by
Bombus bimaculatus and An. vicina (Table 3.5). Three species and two genera are
reported foraging on Vaccinium for the first time: An. miserabilis (pollen record, Table
3.4), An. morrisonella, Lasioglossum acuminatum, one Nomada and one Sphecodes

species (collected while foraging on blueberry, Table 3.5).

62

Table 3.4. Proportion of Vaccinium pollen on the bodies of the most commonly collected
native bees found in Micflan blueberry fields in 2004 and 2005 (n = 126).

prop Vaccinium

 

 

Species Family 11 pollen
Andrena carolina Andrenidae 37 0.99
Colletes inaequalis Colletidae 1 0.99
Augochlora pura Halictidae 1 0.90
Andrena sp.2 (medium) Andrenidae 15 0.72
Colletes validus Colletidae 2 0.65
Apis mellifera Apidae 5 0.62
Andrena vicina Andrenidae 12 0.53
Lasioglossum coriaceum Halictidae 6 0.53
Agapostemon sericeus Halictidae 1 0.50
Augochlorella aurata Halictidae 2 0.49
Andrena carlini Andrenidae 12 0.40
Colletes thoracicus Colletidae 3 0.36
Andrena sp.1 (large) Andrenidae 3 0.21
Lasioglossum (Evylaeus) sp. Halictidae 1 0.14
Andrena miserabilisT Andrenidae 7 0.13
Andrena sp.3 (small) Andrenidae 1 0.03
Andrena (Melandrena) sp. Andrenidae l 0.02
Halictus rubicundus Halictidae 1 0.02
Andrena imitatrix or morrisonella Andrenidae 6 0.01
Andrena (T rachandrena) sp. Andrenidae 3 <0.01
Andrena cressonii Andrenidae 3 0
Andrena morrisonella Andrenidae 1 0
Andrena nasonii Andrenidae 2 0
Andrena perplexa Andrenidae 1 0
1' New pollen record.

 

0.9 1 lhoneybees
0.8 - Dwild bees

.0 .o
O) V
l I

proportion of bees observed
during blueberry bloom
O O O O
'm 'o: is 'on

.0
.A
l

    

 

 

 

 

 

O
I

semi-abandoned commercial
type of blueberry field

Figure 3.6. Proportion of honey bees and wild bees observed during blueberry bloom at
semi-abandoned and commercial blueberry fields in southwest Michigan, 2004-06.

63

Table 3.5. Non-Apis bee species collected while foraging on blueberry blooms in
Michigan in 2004-06.
Family
Species No. bees
Andrenidae
Andrena carlini 11
Andrena carolina 13
Andrena morrisonella'l’ 1
Andrena nivalis
Andrena sp. (male)
Andrena vicina
Apidae
Bombus bimaculatus
Bombus griseocollis
Bombus impatiens
Bombus perplexus
Nomada sp.T
X ylocopa virginica virginica
Halictidae
Augochlora pura
Augochlorella aurata
Halictus confusus
Halictus rubicundus
Lasioglossum acuminatumT
Lasioglossum cressonii
Lasioglossum imitatum
Lasioglossum sp.
Sphecodes sp.T
Megachilidae
Osmia bucephala 1
Total: 62

Qp—tt—t

v—-—-.t>.N--\x

TNew floral records.

New species range extensions. Out of the three years of collecting, I obtained
seven new species that have never been recorded in southwest Michigan (Table 3.6). Six

are new state records, and are new northern extensions of their previously recorded

ranges. One specimen, Andrena nigrae, is new to southern Michigan, having been

reported previously in the northeastern portion of the lower peninsula of Michigan (Table

3.6).

Table 3.6. New species range extensions for bees captured in pan traps in Michigan

blueberry fields in 2004-06.
Family Species

Andrenidae Andrena neonana

Andrenidae Andrena nigrae

Andrenidae Andrena tridens

. Pseudo anurgus
Andrenidae p .
nebrascensw
Apidae E ucera atriventris
. Melissodes
Apidae .
apicata

Megachilidae Osmia virga

Notes

Van Buren County; new state record.

(USA: CT MI IL IN OH NY NJ DC TN NC GA FL
AR TX)

Allegan County; new site record; previously found
in the northeastern lower peninsula of MI.

(CAN: SKAB USA: ME CT MI IL IN OH NY NJ
PAMDVADCTNNCMSALGAFLMNIA
MO AR ND SD NE KS OK TX CO UT ID WA)

Van Buren County; new state record.

(CAN:? ON? USA: MA RI CT WI MI IL IN OH
NYNJPAMDWVVADCTNNCALGA
MN IA NE KS)

Allegan and Van Buren Counties; new state record.
(CAN: NS NB ON MB AB USA: NENG[—RI] WI
MIILINNYNJNCMSFLMNND SD
NE TX? CO)

Allegan County; new state record.
(USA: MA CT WI MI IL OH KY NY NJ PA MD
DE VA DC NC GA MN IA)

Allegan County; new state record, specialist on
Pontederia cordata [pickerelweed].

(USA: MENH MACT MI ILNYNJMDDCNC
GA FL)

Van Buren County; new state record, specialist on
Ericaceae.

(USA: ME MA CT WI MI IN NY NJ PA MD DE
WV VA NC)

 

Notes: In parentheses are the previous known state and province records for these species
per J .S. Ascher, personal communication.

65

Response of bees to trap position and color. Native bees were more likely to be
caught in edge traps than in interior traps in all three years (2004 F158 = 5.85, P = 0.02;
2005 F153 = 8.17, P = 0.006; 2006 F153 = 11.35, P = 0.001) (Figure 3.7). Native bee
richness followed a similar pattern with a greater number of species associated with traps
placed at the edge of the field (2004 F153 = 7.01, P = 0.01; 2005 F153 = 6.05, P = 0.02;
2006 F153 = 12.85, P = 0.0007). Native bee diversity (Shannon-Wiener H’) was also
greater at field perimeters than in the interiors in 2004 (F 1,53 = 6.17, P = 0.02) and 2006
(F153 = 8.06, P = 0.006), but there was no significant difference between positions in
2005 (F153 = 1.82, P = 0.18). Bee abundance, species richness, and diversity did not vary
with trap color in any of the three years (P>0.5). On the contrary, honey bees were more
likely to be caught in white traps (2004 F153 = 7.15, P = 0.01; 2005 F153 = 14.35, P =
0.0004; 2006 F356 = 12.18, P = 0.009) (Figure 3.8), and in 2006 they were more
frequently caught in interior traps than in edge traps (F153 = 5.77, P = 0.02), but
otherwise their capture did not vary with trap position (P>0.05).

Individual native species response to trap position and color varied across the
eight most abundant species known to forage on Vaccinium that were present in each year
(Table 3.7). Species responding significantly to color were trapped more frequently in
white over yellow traps, except for Andrena carolina, which was more often captured in
yellow traps (Table 3.7). In three species, more bees were trapped at the field edge than

the interior (Table 3.7).

66

N
O

 

D edge
I interior

1
1-

I

88888

No. of wild bees trapped
per farm during bloom
o 8

 

 

 

 

 

 

 

 

 

 

2005 . 2006

‘1
O

 

1:] white
I yellow

 

8888

-l N
O O O
1

No. of wild bees trapped
per farm during bloom

 

 

 

 

 

   

 

 

2004 2005 2006

Figure 3.7. Native bee response to trap placement (edge vs. interior of the field) and
color (white vs. yellow) across three years during bloom in highbush blueberry fields in
southwest Michigan. Stars indicate significantly different means within each year (P <
0.05). This figure is presented in color.

67

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

120
is Badge
5 B- 100 ‘ linterior
‘3. 80‘
is
:3 6°‘
SE 40-
‘8
15 u
008' 20"
z
0
2005 2006
180
i5160- * CIWhite
@8140. I Iyellow
"13
“1204
§1oo~ *
E3 80-
ig.§ 60' *
”5:2 40- .
:2 20‘
0

 

 

 

2004 2005 2006

Figure 3.8. Honey bee response to trap placement (edge vs. interior of the field) and
color (white vs. yellow) across three years during bloom in highbush blueberry fields in
southwest Michigan. Stars indicate significantly different means within each year (P <
0.05). This figure is presented in color.

68

Table 3.7. AN OVA results of the response to pan trap position (edge vs. field interior)
and color (white vs. yellow) of the eight most abundant native bees known to forage on
Vaccinium. Data across three years (2004—06) fiom 13 commercial and 2 semi-abandoned
blueberry fields in southwest Michigan are pooled in the analysis.

 

 

position color
Family Species F157 P F157 P (g :k; $1.5)

Andrenidae Andrena carolina 1.53 0.22 7.62 0.008 yellow
Andrenidae An. carliniI 0.50 0.48 2.33 0.13 ns
Andrenidae An. Vicinax 0.30 0.59 7.27 0.009 white
Apidae Ceratina calcarata or dupla 10.87 0.002 0.03 0.87 edge
Halictidae Augochlorella aurataI 8.84 0.004 0.99 0.32 edge
Halictidae Lasioglossum coriaceumI 0.41 0.52 7.03 0.01 white
Halictidae L. imitatum 0.93 0.34 0.03 0.86 ns
Halictidae L. pilosum 4.44 0.04 3.13 0.08 gge

 

TThe last column shows the trap attributes that were associated with significantly greater bee abundance for
each bee species in pairwise comparisons of edge vs. interior and white vs. yellow; (ns) means neither
comparison was significant.

ILog (n+1) transformed prior to analysis to fit assumptions of normality and equal variance.

DISCUSSION

Bee comm unity associated with blueberry. A pis mellifera was the most abundant
species captured during bloom at all sites, comprising one third to one half of all bees
captured in pan traps. During observations in commercial and semi-abandoned fields that
were in bloom, honey bees out numbered native bees by almost 20:1 in commercial
fields, whereas they were 3 times more abundant as native bees in semi-abandoned fields
in which managed honey bee hives were absent (Figure 3.6). This suggests that the
blueberry production region is abundant with honey bees during bloom, because even at
locations without managed hives, there were still honey bees present. Prior to bloom and
the addition of managed hives, honey bees were rarely found (Table 3.2) and this was
also the case after bloom when hives are removed from fields. The lack of feral honey

bee colonies remaining near the blueberry fields emphasizes the dependence of highbush

69

blueberry production on native bees or managed honey bees (National Academy of
Sciences 2007).

In total, at least 174 species of non-Apis bees were trapped throughout the
growing season in and around blueberry fields, with the majority of these species tending
to be rare (1-2 specimens) and not appearing in every year of the study (Table 3.2). Of
the native bees captured during bloom, soil nesting bees in the families Andrenidae and
Halictidae were the most abundant, with the genus Andrena being the most speciose (49
species, Table 3.2). This finding agrees with a previous study of bees associated with
highbush blueberry in upstate New York (MacKenzie and Eickwort 1996, see also
Appendix C). The most abundant known Vaccinium foragers were three Andrena spp.
(An. carolina, An. carlini, and An. vicina), the species complex of Ceratina
calcarata/dupla and four halictid species (A ugochlorella aurata, Lasioglossum (L. )
coriaceum, L. (Dialictus) imitatum, and L. (D. ) pilosum. All were present prior to bloom
and five were also present after bloom (Figure 3.5). Their long activity period indicates
that floral resources available to bees beyond the bloom period of the crop can help
support populations of these bees that are likely to be contributing to blueberry
pollination.

New ecological records. Direct observation and collection of bees foraging on
blueberry, as well as pollen load analysis, revealed three new species and two new genera
never before recorded on Vaccinium spp. (Table 3.4 and 3.5) Also, I report seven new
species range extensions. Intensive studies such as this can reveal changes in species

distributions and are essential for conservation planning.

70

Response of bees to trap position. Native bee abundance, richness, and diversity
were all greater in traps placed at the edge of the field. This finding corresponds to
previous studies of native bees in agricultural systems in which it has been repeatedly
noted that bee abundance and diversity is highest at field edges, where presumably most
bee nesting will be found (Cane 2001). However, although abundance of the blueberry
specialist Andrena carolina followed this pattern, it was not significantly confined to the
field edge and individuals were collected inside the field. The timing of its emergence
and activity as an adult coincided closely with blueberry bloom, and this may provide an
important advantage related to survival in commercial blueberry farms. Whereas bees
that are present longer before or after bloom could be negatively affected by insecticide
applications and other management practices in the field, An. carolina may be able to
build nests in blueberry fields before the time when more intensive pest management
practices begin. I explore this idea in later chapters.

Response of bees to trap color. Honey bees were more likely to be captured in
white traps, regardless of their position in the field. Likewise, native bees associated with
Vaccinium species were also more likely to be captured in white pan traps. This could be
due to the apparent similarity in color between the white blueberry flowers and the white
traps, since individual bees tend to remain constant in their foraging effort, collecting
nectar and pollen from a single species of flower (Wilson and Stine 1996). As stated in
the previous chapter, Leong and Thorp (1999) found that male and female Andrena
limnanthis, an oligolectic bee of white-flowering Limnanthes douglasii rosea (Benth.)
Mason, were most attracted to white pan traps over blue or yellow. The glaring exception

to the floral constancy hypothesis for explaining pan trap color preference is An.

71

carolina, a known specialist of Vaccinium spp. This species was often seen foraging on
blueberry, with almost pure Vaccinium pollen on specimens collected in pan traps, but it
was captured more often in yellow pan traps. More research is needed to explore the
degree to which flower color affects bee response to pan traps.

Conclusion. When the National Academy of Sciences published a recent
pollinator status report (2007), the message was clear: we know alarmingly little about
the status of pollinators in North America Agricultural land far exceeds the acreage
currently in wildlife reserves, and the potential for conservation in agroecosystems of
beneficial organisms is great. For native crops such as blueberry, pollen collecting native
bees are likely be important for yields, particularly when weather conditions are ill-
favored for foraging by the European honey bee, Apis mellifera. In this study, I found that
30—50% of all the bees captured in pan traps during bloom were non-Apis bees, and that
the proportion of bees in each family and nesting guild remained stable across the three
years.

Conservation efforts need to begin with a faunal survey. This three year study of
the bee community associated with highbush blueberry agriculture demonstrated the
utility of pan traps to monitor bee populations toward that end. Pan trapping revealed the
level of rarity of many of the bee species and gave a good estimate of the relative
abundance of the common species. Compared to other faunal surveys such as those by
MacKenzie and Eickwort (1996) and Drummond and Stubbs (1997) on lowbush
blueberry, studies that rely on netting or observations alone may be biased towards rare

species. Future studies aimed at conservation of native pollinators in blueberry should

72

target the several Andrena spp. that emerge prior to bloom, as well as Osmia bucephala,

which may turn out to be a good candidate for solitary bee management.

73

CHAPTER 4:
RESPONSE OF NATIVE BEES TO HABITAT QUALITY AND PRODUCTION

PRACTICES IN HIGHBUSH BLUEBERRY

74

INTRODUCTION

Insect communities endemic to agricultural landscapes are subjected to regular
disturbances associated with crop production. Growers face the challenge of producing a
product fiee of pest damage, while doing as little harm as possible to beneficial insects
such as pollinators. Unlike managed honey bees, wild bees are entirely dependent upon
the agricultural landscape and adjacent habitat, nesting and foraging in and around crop
fields (Free 1993, Williams and Kremen, 2007). Habitat features within and adjacent to
crop fields, and the management practices applied to crop fields are all expected to affect
the native bee community providing crop pollination.

Highbush blueberry (Vaccinium corymbosum L.) is native to North America and
requires bee-mediated pollination for economically viable yields (Free 1993, Delaplane
and Mayer 2000). Managed honey bees are typically rented by growers each year, but
they are less efficient pollinators than many native bees (Sampson and Cane 2000,
Javorek et al. 2002) and are declining due to diseases and mites (Watanabe 1994). Native
bee behavior and ecology are better adapted to blueberry flower morphology and cooler
weather conditions common during bloom (MacKenzie and Eickwort 1996, Batra 1997,
Heinrich 2004). Therefore, native bees are likely to contribute to the pollination of this
crop, particularly when honey bees are inhibited fiom foraging during inclement weather.

Blueberry bloom lasts 4-6 weeks depending on the number of cultivars on a

particular farm. In contrast, native bee life cycles extend beyond blueberry bloom and are

75

likely to be affected by management practices and surrounding habitat features (Kremen
et al. 2002, Kim et al. 2006, Holzschuh et aL 2007). Bee diversity is positively correlated
with flowering plant species richness in natural habitats such as temperate grasslands
(Hegland and Boeke 2006) and Mediterranean landscapes (Potts et aL 2003). Various
studies in agricultural systems have suggested that uncropped, flower-rich habitats
directly adjacent to crop fields will increase diversity and abundance of beneficial insects
in the field (Long et al. 1998, Kells et al. 2001, Croxton et a1. 2002, waell et aL 2005,
Marshall et a1. 2006), and that hedges adjacent to agricultural fields in particular can
support high arthropod diversity (Pollard and Holland 2006). Conversely, habitats in
which pest management practices are more intensive are likely to have a negative impact
on the structure of the endemic bee community (Shuler et al. 2005, Gabriel and
Tschamtke 2007).

Application of insecticides with high toxicity to bees has been shown to have
direct negative impacts on pollinators and other non-target insects that are found in crop
fields (Kevan and Plowright 1989, Johansen and Mayer 1990, Riedl et al. 2006). While
the direct toxicity may be known for honey bees, the lethal and sublethal effects of
various pesticides on native bees are not as well understood, and measurements are
typically taken under controlled laboratory conditions (Stark and Banks 2003, Desneux et
al. 2007). Field studies to determine how a typical insecticide program may be impacting
the endemic bee community are uncommon. Kremen et a1. (2004) lumped insecticides
into 4 categories based on the LD50 for honey bees and their known residual activity,
then used field areas and number of applications to create an index of insecticide use on

watermelon farms. They found no significant relationship between their index and

76

pollination services to the crop. In another study, a binomial variable of pesticides/no-
pesticides was used to examine how the density of bees visiting squash flowers was
related to insecticide use, and again, there was no significant relationship (Shuler et a1.
2005). In general, management intensity has been regarded as a combination of pest
management practices, including cultivation to reduce weeds in and around fields, and
proximity to semi-natural habitat, with a pattern of greater abundance and diversity
associated with nearby semi-natural habitats and a reduction in diversity associated with
conventional versus organic production systems (Kremen et al. 2002, Kremen et al. 2004,
Shuler et aL 2005, Kim et al 2006).

I have previously identified a community of native bees that pollinate highbush
blueberry in southwest Michigan (Chapter 3). Here my objective was to investigate
which components of the conventional highbush blueberry cropping system are driving

native bee abundance, richness, and diversity in blueberry agroecosystems.

MATERIALS AND METHODS

Thirteen commercial and two semi-abandoned highbush blueberry farms located
at least 3 km away fiom one another in southwest Michigan were sampled for bees using
pan traps in 2004-06. Due to varying weather conditions fiom year to year, trapping was
conducted two (2004, 2006) or three (2005) times during bloom in each field. Five pairs
of white and yellow pan traps mounted on 1.2 m PVC poles were placed 5 m apart along
each of two transects running perpendicular to the orientation of the rows. One transect
was established within 1 m of the field edge and the other was established 25 m into the

field. Traps were set out between 8:00-12:00 h and were collected between 16:00-20:00 h

77

for a minimum trapping period of 6 hours on days when weather conditions met the
following criteria: minimum temperature of 13°C with clear or partly cloudy skies or
17°C with any sky condition other than rain (waell et a1. 2005).

Pan traps filled halfway with a 2% unscented soap solution (Dawn® dish soap,
Procter & Gamble, Cincinnati, OH), were constructed fi'om 355 ml white and yellow
plastic bowls (Amscan, Inc., Elmsford, NY) mounted onto 2.7 diameter PVC poles
stabilized with rebar (see Chapter 2, page 37). After the sampling period concluded, pan
trap contents were strained into plastic bags and stored in a -12°C freezer for later
processing. Specimens were thawed at room temperature prior to washing in a 70%
ethanol solution. Honey bees were separated out and counted, then stored in 70% ethanol
solution. All other bees were placed in a mesh bag through which they were fluffed and
dried with a hairdryer before pinning and identification.

Species identifications. Preliminary identifications of bees to the lowest possible
taxonomic group were made using two published dichotomous keys (Mitchell 1960,
Michener et a1. 1994) and the online key available through www.Discoverlife.org.
Further identifications and verifications were made by J .S. Ascher of the American
Museum of Natural History, Division of Invertebrate Zoology. Voucher specimens are
held in the Albert J. Cook Arthropod Research Collection at Michigan State University.

Habitat features. Habitat features around each field were characterized at 45
degree intervals starting at the northern edge of the field, for a total of 8 areas sampled
(Figure 4.1). This method was used because the habitat bordering the blueberry fields
often was different fi'om one end to the other of a particular field edge. Features were

assigned a 1 if present and a 0 if absent and summed over the eight directions (i.e. if tree

78

lines were present in three 45 degree directions, the score for tree lines would be 3).
There were 11 categories of attributes used (Table 4.1).

Flowering plant species in the adjacent habitats were identified and recorded four
times throughout the season each year between April and August in 2004-6. In addition,
five randomly placed 5 m transects were used to assess the abundance of potential
foraging resources in the field perimeter adjacent to bee sampling sites in 2005-6, by
multiplying the number of flowers in a contiguous patch touching each transect by the
area of a single bloom in the patch, and then summing all the patch areas per transect.
The area of a single bloom, including bracts in the case of members of the family
Asteraceae, was estimated by measuring the widest diameter of an open flower

perpendicular to the stigma and using the diameter to calculate the area of a circle.

 

 

 

 

 

 

 

 

 

N
NW NE
450
W E
SW SE
S

 

Figure 4.1. Diagram of habitat feature sampling method around the perimeter of blueberry
fields, depicted here as the shaded square.

79

Table 4.1. Categories of habitat features used to characterize the habitat immediately
surrounding highbush blueberry fields where bees were sampled in southwest Michigan

 

 

from 2004-06.

Habitat feature Description

Blueberry Commercial or semi-abandoned blueberry fields

Other perennial crop Fruit orchards, usually apple

Annual crop Field crops

Meadow/scrubland/ fallow Any area that contained non-crop vegetation that was rarely
disturbed during the growing season (i.e. mown once or twice)

Tree line Trees planted as wind breaks along and between crop fields

Ditch Drainage ditches that may be dry for part of the season

Pond Manmade and used for irrigation

Deciduous woods Woodlots varying fiom open to semi-closed canopy

Other woods Mixed deciduous and coniferous, or coniferous only woodlots
varying from open to semi-closed canopy

Settlement Houses surrounded by yards that may or may not contain flower
gardens; some commercial property.

Road Blacktop or gravel

 

80

Management practices. The management intensity of non—crop vegetation was
characterized under the crop canopy, between crop rows, in the perimeter immediately
adjacent to the crop, and in the surrounding adjacent habitats. Intensity was based on
categorical divisions of management with greater vegetation disturbance assigned higher
numerical values (Table 4.2). The overall score of vegetation management intensity
within each category was summed to obtain a total vegetation management score for each
field.

Insecticide application records for 2003-05 were obtained for each field sampled
for bees. The insecticide products used, their chemical name, their chemical class, their
targeted pests, and their published LDso for honey bees are listed in Table 4.3. Kilograms
of active ingredient (AI) applied per hectare was calculated by dividing the application
rate by the percent Al for each product used. To obtain an insecticide program toxicity
(IPT) score for each field for each season, the AI per hectare for each application was
divided by its LDso for honey bees, and this was summed for all insecticide applications

applied to each field during each year.

Z amount of active ingredient (kg) l Ha

insecticide Program WWW = LD for honeybees
50

The LDso for honey bees was used because it is the most complete data set and should be
generally representative of the response of other bees in the community to insecticides.
With this equation, as the LDso decreases, the IPT score increases. Likewise, if more AI
is used, the IPT score increases. This score was used to determine the relationship
between insecticide use and wild bee abundance and diversity during crop bloom of the

following year. A second toxicity score was calculated by dividing kg/Ha of AI by l, 2,

81

or 3 based on the toxicity to bees rating listed in the 2007 Michigan Fruit Management

Guide (Table 4.3), where Highly Toxic = 1, Moderately Toxic = 2, and Relatively Safe =

3. This order was chosen to match the relationship between the bee variables described

above for the first IPT score.

Table 4.2. Areas in which non-crop vegetation was managed in and around focal
blueberry fields in southwest Michigan and how they were scored for intensity.
Areas of non-crop vegetation Description of categories for scoring vegetation

management
under crop canopy

between crop rows

crop field perimeter

adjacent habitat

management intensity

1 = 0-50% bare ground

2 = 50-75% bare ground

3 = 75-90% bare ground

4 = 90-99% bare ground

5 = 100% bare ground

1 = mown vegetation

2 = herbicide

3 = tilled

1 = untended

2 = mown vegetation

3 = herbicide

4 = tilled

l = untended

2 = management of strip adjacent to field
3 = management to edge of adjacent habitat
4 = management into adjacent habitat

5 = clearing of adjacent habitat

82

£8 30332 n m 6:8“ 388088

N E .83 9 2x8 mama u H 62:0 858%ng mam emmEouz boom 2: .3 Bnmwmmm 36¢on mo 35.: .38 o: u we ”moon 328: com 83.:
6:50.50 SEQ H Um E88 x888 Bx§-833 M $3.23 63% b.5353 u <mm 6:39 08525. n E. uowwma

3883 n 2mm E=ee§ 888838 u memo women: 358% n 2mm afiee 365%. u m: ”$83.5 page H 3.6 ”€03.58
3858 u 3.50 88.09.-Seameeeeaeeeeemaeaa. Seineea 830 “888932 gem gene: 88 a: a 3863 we...

 

 

m 8 «.50 3.6 .330 Ramses aeeeaetee 233% 85
m :8 Mime .35 .380 53%.: £32» 385 888822 a N £80
2 Rod 2mm ‘ eéeaee 9.85% 8 N seem
a. 3 Se <mm .3 .23 2038288 engages: a 3 838m
H 88 <mm demo .3 .2mm .35 .330 2252.3 38588 em 88 an .83.
e EN meme .8 .2mm 89:38 Lasso $3“ 35¢.
a 2 we Jame .m: .Emm aeaeeee Esae .53 8 .3 on esem
e 8.8 2923 4mm cease”. 38858 mm cm .5 am 2893
H Rd <mm .E .28 aeeemoeeeeewe 83338 >4 . a w .8. w 83%: see.
8. Ge memo .E .Emm 2888888 e888 .3 S 55
e 88 5mo .38 .330 223889.30 353-88%... 83 on 88.6
9 Ed <mm demo 2% .38 3me 03888888 eoenee .3 on 8:55
Veg. 33%: $269 3893—. man—U 3332—0 «5:3qu «>33. 3.832
35

 

.8458 E moon com BEES 083 gm 80:3 newEomz 603538 E mb3om§nofin .3 com: 823883 .mé 035.

83

Statistical analyses. The Mantel test (‘fi/egan” package for R 2.3.1) was used to
compare pairwise bee community similarity indices (Jaccard, Bray-Curtis, and Morisita-
Horn) with pairwise geographic distances between each of the fifteen blueberry farms.
Spatial autocorrelation in non-pairwise variables (bee abundance, species richness,
Shannon-Weiner and Simpson’s diversity indices) was assessed with Moran's I (“ape”
package for R 2.3.1).

Simple linear regression analysis was conducted between wild bee abundance,
species richness, diversity (Shannon-Wiener H’), abundance of Andrena carolina,
Lasioglossum coriaceum, Ceratina calcarata/dupla females, Andrena carlini, and
Augochlorella aurata were regressed and 4 different habitat features: flowering plant
richness, adjacent blueberry, adjacent deciduous woods, and adjacent ditches (PROC
REG, SAS 9.1). Bees in the categories described above were also regressed separately
against 3 different management intensity indexes: the IPT score based on the LDso values,
the IPT score based on the toxicity rankings, and vegetation management intensity
(PROC REG, SAS 9.1).

For each year, a redundancy analysis (RDA in CANOCO 4.5) was conducted
using all the bee species collected for which there is a Vaccinium floral record and the
following environmental variables: insecticide program toxicity, vegetation management
intensity, the number of flowering plant species, and adjacent ditches, treelines,
deciduous woods, other woods (mixed or conifer), blueberry fields, annual crop fields,
other flowering perennial crops, ponds, meadows, roads, and settlements. RDA is a form
of canonical analysis that is an extension of multiple linear regression, which assumes Y

(species data) and X (in this case habitat attributes and management intensity variables)

34

are linearly related. It may also be seen as an extension of principal components analysis
with the ordination of Y constrained in such a way that the resulting ordination vectors
are linear combinations of the variables in X (Legendre and Legendre 1998). Thus, it
enables analysis of the effect of multiple potential explanatory variables on a community

of many bee species.

RESULTS

Independence of sites. Bee communities at the sampled fields were considered to
be independent based on the results of the Mantel test for community similarity (Z =
1729.3, df = 14, p = 0.143). Likewise, bee abundance, species richness, Shannon-Weiner
and Simpson’s diversity indices assessed with Moran’s I were not significant indicating
no spatial autocorrelation among sites (I < 0.514, df= 13, p > 0.05).

Response of wild bee communities to habitat features. At least 84 flowering
plant species were found in the habitats adjacent to the blueberry fields in which bees
were trapped, with 32% of the species being found at single sites (Table 4.4). Increasing
bee species richness was associated with increasing plant species richness in 2005 only
(Figure 4.2e). Bee abundance, species richness, and diversity did not vary significantly
with flowering plant abundance in 2005 or 2006.

Bee abundance, species richness, and diversity also did not vary significantly with
vegetation management intensity, proximity to ditches, or deciduous woodland in all
years. However, with more blueberry fields in the surrounding habitat, bee species
richness and diversity declined significantly in 2004 and 2005 (Figure 4.3). Also, bee

abundance and species richness declined with increasing values of both measures of IPT

85

in 2004 and 2005 (Figures 4.4 a-b, d-e and 4.5 a-b, d-e). Likewise, bee diversity declined

significantly with increasing IPT scores in 2005 (Figure 4.5h).

86

Table 4.4. Plant species found in the perimeter of blueberry fields sampled for bees in
southwest Michigan. Nomenclature and US. nativity based on the USDA-NRCS Plants
Database at httJH/plantsusda. gov (last accessed 15 July 2007).

 

 

No. of farms

Family Scientific Name Native vs. exotic where present
Apiaceae Cicuta maculata native l
Apiaceae Daucus carota exotic 3
Apocynaceae Apocynum androsaemifolium native 1
Asclepiadaceae Asclepias syriaca native 2
Asteraceae Achillea millefolium exotic 10
Asteraceae Centaurea maculosa exotic 1
Asteraceae Chrysanthemum leucanthemum exotic 8
Asteraceae Cichorium intybus exotic 1
Asteraceae Crepis capillaris exotic 5
Asteraceae Erigeron annuus native 6
Asteraceae Erigeron philadelphicus native 2
Asteraceae Eupatorium perfoliatum native 1
Asteraceae Hieracium aurantiacum exotic 2
Asteraceae Hieracium sp. native and exotic 12
Asteraceae Hypochoeris radicata exotic 1
Asteraceae Rudbeckia hirta native 6
Asteraceae Senecio vulgaris exotic 1
Asteraceae Solidago sp. native 15
Asteraceae Sonchus sp. exotic l
Asteraceae Taraxacum oflicinale exotic 15
Balsaminaceae Impatiens capensis native 1
Bi gnoniaceae Campsis radicans native l
Brassicaceae Alliaria ofi‘icinalis exotic 1
Brassicaceae Barbarea vulgaris exotic 9
Brassicaceae Berteroa incana exotic 4
Campanulaceae, Lobelia sp. native 3
Caprifolaceae Sambucus sp. native 9
Caprifoliaceae Lonicera sp. native and exotic l
Caryophyllaceae Cerastium vulgatum exotic 14
Caryophyllaceae Dianthus armeria exotic 5
Caryophyllaceae Silene pratensis exotic 6
Clusiaceae Hypericum perforatum exotic 7
Cornaceae Camus sp. native 1
Elaeagnaceae Elaeagnus umbellata exotic l
Fabaceae Medicago lupulina exotic l3
Fabaceae Melilotus alba exotic 2
Fabaceae Melilotus oflicinalis exotic 3
Fabaceae T rifolium arvense exotic 2
Fabaceae T rifol ium dubium exotic 5
Fabaceae T rzfolium hybridum exotic 3
Fabaceae Trifolium incamatum exotic 1
Fabaceae Trifolium pratense exotic 11
Fabaceae T rifol ium procumbens exotic 2
Fabaceae T rifolium repens exotic 13
Fabaceae Vicia cracca exotic 2

87

Gentianaceae Centaurium pulchellum exotic 1
Geraniaceae Geranium sp. native and exotic 1
Iridaceae Iris sp. native and exotic 1
Iridaceae Sisyrinchium sp. native 3
Laminaceae Lamium purpureum exotic 5
Laminaceae Monardafistulosa native 1
Laminaceae Prunella vulgaris native 3
Lauraceae Sassafras albidum native 2
Liliaceae Lilium sp. native and exotic 1
Liliaceae Maianthemum canadense native 3
Lythraceae Lythrum salicaria exotic 1
Onagraceae Oenothera biennis native 2
Oxalidaceae Oxalis stricta or europa native 8
Plantaginaceae Plantago major native 3
Polygonaceae Polygonum persicaria unknown 2
Polygonaceae Rumex acetosella exotic l3
Polygonaceae Rumex crispus exotic 4
Ranunculaceae Ranunculus abortivus native 2
Ranunculaceae Ranunculus acris native and exotic l
Rosaceae Fragaria virginiana native 4
Rosaceae Potentilla recta exotic 6
Rosaceae Potentilla simplex native 2
Rosaceae Prunus sp. native and exotic 4
Rosaceae Rosa palustris native 3
Rosaceae Rubus sp. native 15
Rosaceae Spiraea alba native 1
Rubiaceae Galium sp. exotic 1
Salicaceae Salix sp. native and exotic 8
Scrophulariaceae Ntztiallanthus canadensrs (formerly: native 1
mana canadenszs)
Scrophulariaceae Mimulus sp. native 1
Scrophulariaceae Penstemon sp. native 2
Scrophulariaceae Verbascum blattaria exotic 3
Scrophulariaceae Verbascum thapsus exotic 2
Solanaceae Solanum carolinense native 3
Solanaceae Solanum dulcamara exotic 3
Solanaceae Solanum nigrum exotic 4
Violaceae Viola kitaibelliana exotic 6
Violaceae Viola sp. native and exotic 7

 

88

2004 2005 2006

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

300 40 3.5
o A 35, ° B C
g 250‘ ns 3‘ 0 ’8
‘5 ns 3°“ . 0 25« o .
E 200. 25- o ‘ I o 3‘
B a o 2‘ ’
co 150‘ 2°" ' 15
O 15J .0 . ' ‘
§ 100 ° ’."° 10« 3 1* "S
o
50 0... 9 5i 05‘
o l I I 0 r I o I
O 10 20 30 40 0 10 20 30 4O 0 10 20 30 40
250 50 3.5
o 45_ E 3 , F
g 200 o 40* O O .
35- 2.5-
ns ° 9 8
E 150‘ 3 30‘ 2~ 9. ¢
.2 25“ Q 0
h 100- o 20~ 1,5.
15— 1.
§ 501 ’00... 10'1 HS
. 5__4 0.5'
9 o
o If r o T I 0
0 10 20 30 40 O 10 20 30 40 0 1O 20 30 40
160 35 3.5
A 140‘ 0 G o H I
.1 30" 3'1 . ‘
V 120‘ ns 25‘ . ..’ 25.. . ..O.:
2* 100* O o
.E . 20-4 O . 2" .
0 80‘ e 8 .
% 60< . O 15" ns 0 .. ‘ 1.5:"J .
4o- ..‘ 10" O 1 ns 0
g 20* . 5‘ e 05‘
0.. .
0 I I I o 17 I I 0 I I I
0 1O 20 30 40 0 10 20 3O 4O 0 10 20 30 40

flowering plant richness

Figure 4.2. Regression analyses of wild (non-Apis) bee abundance (A, B, C), bee species
richness (D, E, F), and bee diversity (G, H, I) with the number of flowering plant species
found in the field margin at 15 farms in southwest Michigan. Except for bee richness in
2005, flowering plant species richness was not a significant factor in explaining bee
abundance, richness or diversity. ns = not significant at P < 0.05.

89

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

number of field margins adjacent to other blueberry fields

2004 2005 2006
300 250 160
O O
8 250 « ns 200 . . ns 14° 1 ns
3 . 120 «
'2 200 d 1501 ’ 10° ‘ .
3 150 " 80 ‘ Q
'3 e 100 - 0 ii . 0
60
100 ‘ . O . ‘
a . 50 q . e 3 4o 4
50 - 9 °
0 i 0 20 . O
o 0 . O
0 I T o I o T I
0 2 4 0 2 4 0 2 4
4O 50 35
’ .. D . o
m 354 R::3.3? l 31 . R2=031 30*
g 301} . . - . 35 ‘ P = 0.03 25 -1 O O
25 ‘ . O 0 . O
.c , 3° 4 g 20 ~ g
.g 20 ~ 0 25 ~
8 15 - . . $1531 0 . . 15 , e ,
n 10 q . 10 _‘ . 10 " .
5 . 51 o 5 . : ns
0 I o I I I
O 2 4 0 2 4 0 2 4
3.5 3.5 3.5
5‘ o
a 3 i.\;.\’ 3 4 : 3 ii I z
- 1 -
g, 2.5 . . . . 2.5 . 2.5 . .
g 2 « o 2 « ‘ . 2 ~ . 0
g 1.5 ~ 1.5 . ° 1.5 « o
§ 1 - 2 1 1 2 1 < .
0.5 R = 0.28 0.5 _. R = 0.26 0.5 ‘
P = 0.04 P = 0.048 "s
0 T : 0 1 O T I
0 2 4 0 2 4 0 2 4

 

 

 

Figure 4.3. Regression analyses of wild bee abundance (A, B, C), bee species richness
(D, E, F), and bee diversity (G, H, I) in relation to the number field margins bordered by
blueberry fields in eight 45 degree directions (see Materials and Methods for details). ns
= not significant at P < 0.05.

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

250
R2= 0.28
g 200 t P = 0.04
'8 150 0 .
3
a 100 4 0
§ 50- ’0 1
o
0 ° .
0 20 40
40 l D 50
351 R2: 033 45 0 R2: 0.30
g 301 o [3:003 432) P=0.03
g 25 . . O 304
c O O O . Q
.9 2° ° 15‘ o .
I— 15 4 O 0 ~ .
§ 10 « . o ’ 15 4 0 0 o
10 ~ 0
5 — 5 _ o
0 . 0 .
0 20 40 0 20 40
A 3.5 3.5
5 3 1; : o 0 G 3 9 . o
. 9 S
Q 2'5 , : ° 0. o 2.5 ’ . o . o
2 2 c 2 , o o o 3
d) O o
g 1.5 4 ns 1.5 .
1 .. 1 .. ns
§ 0.5 - 0.5 ‘
0 f 0 I
0 20 40 0 20 40

2006

 

 

 

 

 

 

9 ns

 

 

1..
0.54

0

20 40 60

 

 

ns

 

 

0

20 40 60

insecticide program toxicity score (kgAl/Ha/LD50)
from the previous year

Figure 4.4. Regression analyses of wild bee abundance (A, B, C), bee species richness
(D, E, F), and bee diversity (G, H, I) in relation to the insecticide program toxicity score
based on LDso for honey bees fiom the year previous to bee sampling (see Materials and
Methods for details). as = not significant at P < 0.05.

91

2004 2005 2006

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

300
D
g 250
g 200 ~
3 150 ~
‘° 1001
D
§ 50-
0
0
40 D
m 35
§ 3°‘
'5 25 -
-: 20 «
3 15 .
.o 10 4
5 ..
0
0
A 3.5
5 3 0
g; 25
2 .
o 2
2 151
.0 .
§ ‘ ‘
0.5-
o I l
0 5 10 0 5 10 0 5 10

insecticide program toxicity score (kg Al/Ha/T ox)
for the previous year
Figure 4.5. Regression analyses of wild bee abundance (A, B, C), bee species richness
(D, E, F), and bee diversity (G, H, I) with the insecticide program toxicity score based on
ratings in the 2007 Michigan Fruit Management Guide from the year prior to bee
sampling (see Materials and Methods for details). ns = not significant at P < 0.05.

92

Response of the most common blueberry foraging bees to management
practices and habitat features. The 5 most common Vaccinium foraging bees responded
differently to management practices and habitat features within and among years.
Augochlorella aurata was less abundant at sites with more intensive insect pest
management programs and more intensive management of vegetation in and around
fields (Table 4.5, 2004). However, it was more abundant at sites surrounded by a greater
number of ditches. Likewise, Andrena carlini was less abundant at sites with more
intensive insect pest management, but was either more or less abundant with increased
vegetation management intensity depending on the year (Table 4.5, 2004 and 2006).
Greater captures of An. carlini were found with increasing numbers of field margins
containing ditches (2004), but not with greater flowering plant richness (2006).
Abundance of Lasioglossum coriaceum increased positively with the proportion of
adjacent woodland (2005), and with flowering plant richness (2006). Ceratina
calcarata/dupla was less abundant in fields surrounded by more blueberry fields (2005).
One of the most common bees found in blueberry fields, Andrena carolina, did not

respond significantly to any of the variables tested.

93

Table 4.5. Regression coefficients for simple linear regressions of the 5 most abundant
native bee blueberry foragers and management or habitat variables at 15 blueberry farms
in southwest Michigan. R values that are significant are in bold followed by their
associated P-value in parentheses and +/- indicating the slope of the linear regression.

 

 

Management or habitat variables 2004 2005 2006
a) kgAI/Ha/LDSO
Andrena carolina 0.15 0.04 0.20
Lasioglossum coriaceum 0.23 0.001 0.05
Ceratina calcarata or dupla (9 only) 0.10 0.16 0.02
Andrena carlini 0.24 0.002 0.002
Augochlorella aurata 0.35 (0.02) - 0.16 0.04
b) kgAI/Ha/ToxRate
Andrena carolina 0.10 0.04 0.01
Lasioglossum coriaceum 0.23 0.01 0.005
Ceratina calcarata or dupla (S? only) 0.06 0.18 0.08
Andrena carlini 0.29 (0.04) - 0.03 0.001
Augochlorella aurata 0.27 (0.046) - 0.21 0.03
c) Vegetation management intensity
Andrena carolina 0.0004 0.19 0.16
Lasioglossum coriaceum 0.30 (0.04) - 0.04 0.0005
Ceratina calcarata or dupla (9 only) 0.09 0.15 0.09
Andrena carlini 0.32 (0.03) - 0.02 0.39 (0.01) +
Augochlorella aurata 0.34 (0.02) - 0.14 0.0004
d) Adjacent deciduous woods
Andrena carolina 0.02 0.01 0.23
Lasioglossum coriaceum 0.14 0.29 (0.04) + 0.04
Ceratina calcarata or dupla (9 only) 0.04 0.03 0.02
Andrena carlini 0.17 0.23 0.20
Augochlorella aurata 0.18 0.20 0.22
e) Adjacent ditches
Andrena carolina 0.04 0.16 0.12
Lasioglossum coriaceum 0.32 (0.03) + 0.04 0.07
Ceratina calcarata or dupla (9 only) 0.001 0.12 0.11
Andrena carlini 0.29 (0.04) + 0 0.03
Augochlorella aurata 0.20 0.27 0.13
t) Adjacent blueberry fields
Andrena carolina 0.01 0.04 0.004
Lasioglossum coriaceum 0.02 0.008 0.09
Ceratina calcarata or dupla (9 only) 0.24 0.30 (0.03) - 0.04
Andrena carlini 0.02 0.08 0.007
Augochlorella aurata 0.06 0.03 0.02
g) Flowering plant species richness
Andrena carolina 0.11 0.0013 0.04
Lasioglossum coriaceum 0.03 0.12 0.27 (0.047) +
Ceratina calcarata or dupla (9 only) 0.0001 0.24 0.9
Andrena carlini 0.06 0.21 0,41 (0,01) -
Augochlorella aurata 0.05 0.09 0.02

 

94

Comparison of bee species and habitat attributes using RDA. In 2004, the IPT
score (kgAI/Ha/T ox) was the most significant explanatory variable for the species
abundances in that year (Table 4.6). The abundance of most bee species was negatively
correlated with this variable (Figure 4.6). Although not statistically significant in the
RDA, the negative trend between bee species and greater pest management intensity
appeared again in both 2005 and 2006 (Figures 4.7-8). In 2005, bee species abundance
varied significantly with 5 variables: most bee species were negatively correlated with
increasing vegetation management intensity and proximity to other flowering perennial
crops, but positively correlated with the number of flowering plant species found in field
perimeters and proximity to deciduous woods and ditches (Table 4.6, Figure 4.7). The
number of flowering plant species in field perimeters and proximity to deciduous woods
were also significantly correlated with species abundance in 2006 (Table 4.6, Figure 4.8).
Increasing vegetation management intensity was negatively correlated with most bee
species in all years (Figures 4.6-8). Most species were positively correlated with habitat
variables associated with foraging or nesting resources, such as woodland habitat and
ditches, and were negatively correlated with habitats in which pest management was

more intensive (Figure 4.6-8).

95

Table 4.6. Summary of the RDA analyses of bee community abundance in blueberry
fields by year. The i. values from greatest to least refer to the order in which the variables
were added based on how well they explain the species data. Results of the Monte Carlo
permutation test for each addition of an explanatory effect is listed in the P column along
with its corresponding F-statistic. Significant probability values (< 0.05) are highlighted
in bold.

 

 

 

. 2004 2005 2006

Variable A P F 1. P F 2. P F
Insecticide program score

(k 1 1,11 'tox-rating) 0.19 0.006 3.01 0.03 0.526 0.88 0.06 0.506 0.94
Treelines 0.11 0.070 1.92 0.04 0.378 1.08 0.07 0.31 1.20
Ditches 0.1 0.074 1.81 0.09 0.04 2.24 0.05 0.58 0.76
Meadows 0.08 0.122 1.65 0.05 0.358 1.13 0.07 1.00 0.00
Ponds 0.07 0.248 1.32 0.04 0.41 1.41 0.06 0.438 1.09
N°° 0f ‘l°“’e"ing plant 0.06 0.276 1.28 0.17 0.02 2.6 0.13 0.01 2.15

specres
Deciduous woods 0.06 0.330 1.29 0.13 0.006 2.68 0.14 0.02 2.08

Vegetatim management 0.05 0.342 1.22 0.11 0.02 2.39 0.07 0.242 1.27

intensity
Blueberry fields 0.06 0.344 1.11 0.03 0.48 0.93 0.05 0.426 1.05
Annual crops 0.05 0.406 1.04 0.03 1.00 0.0 0.04 0.552 0.69

Floweringperennial crop 0.04 0.492 0.95 0.13 0.014 2.21 0.07 0.23 1.30

Other woodland 0.04 0.566 0.8 0.04 0.468 0.99 0.09 0.094 1.48
Settlement 0.03 0.708 0.45 0.04 0.506 0.91 0.05 0.386 1.09
Road 0.06 1.00 0.00 0.07 0.1 2.2 0.05 0.518 0.89

 

96

 

  
     
  

 

 

 

 

 

 

W
q 2004
‘_
g8
veg. mngt 9‘3 treeline
intenSIty annual
IPT
other woods J”
itcha
x—é-O’tfi
meadow ti j 7d\r
settlemeM K \\>e§deciduous woods
\ .
blueberry oad 1‘
00
fl. lant
fl. per. slim. 1
crop
b
O.
‘—
l
-10 1'0

Figure 4.6. Redundancy analysis of the abundance of 38 bee species known to forage on
Vaccinium and 13 environmental characters, including two measurements of crop
management intensity at 15 blueberry farms in southwest Michigan in 2004. Blue lines
represent difi‘erent bee species. Red lines indicate environmental characters. See text for
details on environmental characters. Key to species: a = Agapostemon sericeus, b =
Andrena carlini, c = An carolina, d = An. crataegi, f = An. hippotes, g = An. imitatrix or
morrisonella, i = An. nivalis, j = An. pruni, k = An. rugosa, 1=An. vicina, m =
Augochlora pura, n = Augochlorella aurata, o = Au. gratiosa, p = Augochloropsis
sumptuosa, q = Bombus bimaculatus, r = B. citrinus, t = B. griseocollis, u = B. impatiens,
v = B. perplexus, w = Ceratina calcarata/dupla, x = C. strenua, a = C. thoracicus, cc =
Halictus confusus, dd = H. rubicundus, ee = Lasioglossum acuminatum, if = L.
coriaceum, g = L. cressonii, hh = L. imitatum, ii = L. pilosum, jj = L. quebecense, kk =
Osmia atriventris/pumila, 11 = 0. bucephala, mm = Xylocopa v. virginica. This figure is
presented in color.

97

 

  
   
       

 

 

 

 

 

 

O c
,_' 2005
I b deciduous
ditches 2 g w
other woods
veg. mngt h
. . fl. plant
IntenSIt
k 2
J xx
We“
‘mp 7%“
blueberry
settlement fl" per. crop
IPT meadow
road
C!
V-
I l v
-1.0 1.0

Figure 4.7. Redundancy analysis of 30 bee species known to forage on Vaccinium and 13
environmental characters, including two measurements of crop management intensity at
15 blueberry farms in southwest Michigan in 2005. Blue lines represent different bee
species. Red lines indicate environmental characters. See text for details on
enviromnental characters. Key to species: a = Agapostemon sericeus, b = Andrena
carlini, c = An. carolina, d = An. crataegi, f = An. hippotes, g = An. imitatrix or
morrisonella, h = An. miserabilis, j = An. pruni, k = An. rugosa, 1= An. vicina, m =
Augochlora para, 11 = Augochlorella aurata, s = Bombusfervidus, u = B. impatiens, w =
Ceratina calcarata/dupla, x = C. strenua, z = Colletes inaequalis, aa = C. thoracicus, bb
= C. validus, cc = Halictus confusus, dd= H. rubicundus, ee = Lasioglossum acuminatum,
if = L. coriaceum, gg = L. cressonii, hh = L. imitatum, ii = L. pilosum, jj = L. quebecense,
kkl = Osmia atriventris, kk2 = 0. pumila, 11 = 0. bucephala, mm =Xylocopa v.
virginica. This figure is presented in color.

98

 

 

 

0- 2006
‘_.
n fl. plant spp.
fl. per. crop
hh
road
settlement meado _ other woods
IPT
blueberry

  

annual crop

veg. mngt intensity
W

 

 

 

Q
‘T

 

-1 .0 1 .0
Figure 4.8. Redundancy analysis of 28 bee species known to forage on Vaccinium and 13
environmental characters, including two measurements of crop management intensity at
15 blueberry farms in southwest Michigan in 2006. Blue lines represent different bee
species. Red lines indicate environmental characters. See text for details on
environmental characters. Key to species: a = Agapostemon sericeus, b = Andrena
carlini, c = An. carolina, d = An. crataegi, e = An. forbesii, f = An. hippotes, g = An.
imitatrix or morrisonella, h = An. miserabilis, k =An. rugosa, 1= An. vicina, m =
Augochlora para, 11 = Augochlorella aurata, q = Bombus bimaculatus, r = B. citrinus, s =
B. fervz'dus, t = B. griseocollis, w = Ceratina calcarata/dupla, z = C. inaequalis, aa =
Colletes thoracicus, cc = Halictus confusus, dd = H. rubicundus, ee = Lasioglossum
acuminatum, ff = L. coriaceum, gg = L. cressonii, hh = L. imitatum, ii = L. pilosum, jj =
L. quebecense, kk = Osmia atriventris or pumila, 11 = 0. bucephala. This figure is
presented in color.

99

DISCUSSION

Within a conventional cropping system, it is expected that there will be variation
among farms in terms of soils and hydrology, the distribution of habitat features
surrounding crop fields, and in the way individual growers choose to manage their
farmland. Previous studies have shown that broad categorizations of crop management
systems, such as the division of management practices into conventional versus organic,
reveal effects of this farm variation on the accompanying bee communities, with lower
bee diversity and abundance in conventional farms (Kremen et al. 2002, Shuler et al.
2005, Kim et a1. 2006, Holzschuh et al. 2007, Gabriel and Tschamtke 2007). This study
aimed to determine which components of conventional crop systems are driving bee
community structure in highbush blueberry.

It is clear from toxicological studies that insecticides can have both lethal and
sub-lethal effects on bees (Johansen and Mayer 1990, Desneux et al. 2007) and that
insecticides applied for pest management can have negative impacts on non-target
arthropods at the community level (Kevan and Plowright 1989, Stark and Banks 2003).
While growers of crops that require bee pollination typically avoid spraying insecticides
during bloom (Riedl et al. 2006), field studies to determine how a typical insecticide
program may be impacting native bee communities are uncommon and have been unable
to show direct relationships between insecticide use and the structure of the bee
community (Kremen et al. 2004, Shuler et a1. 2005).

Previous attempts to use indices or simple binomial categories of insecticide use
intensity have not found any relationship between insecticide application intensity

throughout the growing season and the endemic bee community (Kremen et a1. 2004,

100

Shuler et al. 2005). The two indices of insecticide program toxicity developed here
revealed significant negative relationships between the intensity of insecticide programs
applied to blueberry fields and the bee community present during bloom the following
year, in two out of three years. Either of these indices could be applied to pest
management programs to allow growers to assess the risk of their management practices
to bees, and to identify pesticide applications that could be removed or replaced with a
less toxic alternative.

Crop management intensity is a combination of pest management practices,
including cultivation to reduce weeds in and around fields and proximity to semi-natural
habitat, with a pattern of greater abundance and diversity of bees in fields near semi-
natural habitats and lower bee diversity in conventional versus organic production
systems (Kremen et al. 2002, Kremen et a1. 2004, Shuler et al. 2005, Kim et a1. 2006,
Holzschuh et al. 2007). I found that blueberry fields surrounded by similar crop habitat,
i.e other blueberry fields, were likely to have bee communities that were less speciose
and less diverse. This effect could be due to landscape homogeneity, and is likely to be an
important part of the explanation for variation in bee community structure.

This study also revealed a reduction in the abundance during bloom of bees that
are known to forage on blueberry with increasing crop management intensity
measurements, including the intensity of weed management and the abundance of other
adjacent crop fields. These findings suggest that the insecticide program intensity, while a
key variable affecting the structure of native bee communities in crop fields is one of

many factors that combine to affect the suitability of crop fields for these insects.

101

F lower-rich habitats and other semi-natural landscapes adjacent to crop fields are
associated with a greater abundance and diversity of beneficial insects in agricultural
landscapes (Long et al. 1998, Kells et al. 2001, Croxton et aL 2002, Ricketts 2004,
waell et al. 2005, Marshall et a1. 2006). A similar pattern in the bee community was
observed in the fields sampled here, with the abundance of blueberry foraging bees being
higher in fields having greater floral species richness. In addition, fields adjacent to
habitats likely to contain floral and nesting resources, such as woodlands and ditches,
were positively correlated with the abundance of blueberry foragers during bloom

Conclusions. A number of indices have been developed to determine or predict
the effects of pesticides on human health and the environment (e. g. Kovach et al. 1992),
but not with specific respect to their potential impact on bees (although see Kremen et al.
2004). There is evidence that accidental drift of insecticides into non-target areas can be
detrimental to native bee populations (Kevan and Plowright 1989), and there are a
number of studies that have compared conventional vs. organic crop management
systems showing that bees are typically more abundant and more diverse on organic
farms (Kremen et a1 2002, Kremen et al. 2004, Shuler et al. 2005, Kim et al. 2006). But
most agricultural land is managed conventionally and there is a wide variation in
management styles within the broad category of “conventional.”

Further work still needs to be done to examine what it is about conventional farms
that make them more or less hospitable to bees. We need metrics to measure the response
of native bees to insecticide use in conventional agricultural landscapes. For conservation
of diverse bee communities, habitat heterogeneity appears to be an important piece of the

puzzle. For conservation of specific crop pollinators, attention to their phenology with

102

regard to floral resources beyond crop bloom and nesting resources in protected areas
will be important. Agricultural landscapes have great potential for the protection of
pollinators, particularly in cropland dependent upon them. Further studies are needed to
better understand where native bees are nesting in agricultural lands and which sets of
flowering plants should be used to supplement floral resources throughout the growing

season.

103

CHAPTER 5:

COMPARISON OF NATIVE PLANTS FOR USE IN AGRICULTURAL BEE

CONSERVATION PROGRAMS IN THE MID-WESTERN US.

104

INTRODUCTION

Pollination is critical to the productivity of many agricultural crops (McGregor
1976, Free 1993, Kearns 1998, Kevan and Phillips 2001). A recent review found that of
115 cultivated plants grown for fruit, vegetable, or seed production, 87 depend upon
animal-mediated pollination, comprising 35% of global crop production yields (Klein et
al. 2006). Non-managed wild bees are estimated to be responsible for pollination
contributing $3.07 billion of fruits and vegetables in the United States annually (Losey
and Vaughan 2006). Until recently, little attention was paid to pollinators in biodiversity
conservation programs for managed and non-managed ecosystems (Buchmann and
Nabhan 1996, Kearns and Inouye 1997). However, concerns over declines in insect
biodiversity, thought to be caused by habitat loss and fragmentation as a result of
agricultural intensification and other anthropogenic land use changes, have emphasized
the importance of natural ecosystem services, including pollination (W estrich 1996,
Kremen et al. 2002, Tschamtke et al. 2005, Biesmeijer et al. 2006).

Plant productivity in natural and agricultural systems has been linked to pollinator
abundance and diversity, leading to an increased awareness of the services pollinators
provide in these ecosystems and greater attention on strategies that can support their
populations (Allen-Wardell et al. 1998, Kevan and Phillips 2001, Javorek et al. 2002,
Kremen et al. 2002, Klein et al. 2003, Potts et al. 2003, Fontaine et al. 2006). At the same

time, the availability of managed Apis mellifera colonies that are used to provide

105

pollination services is declining because of diseases and parasites (Torchio 1990,
Watanabe 1994). This increases the importance of conserving wild bees, as part of
growers’ strategies for achieving sustainable crop pollination (Southwick and Southwick
1992, Kevan and Phillips 2001, Klein et al. 2006).

The suitability of an ecosystem for bees depends on the ecology of each bee
species in the community, including bee phenology, foraging range, and the availability
of suitable foraging and nesting resources within that range (Kearns and Inouye 1997,
Cane 2001). For some bees, foraging resources are needed during a brief window in time,
with these bees typically emerging in synchrony with specific plant species. Those that
are multivoltine (e. g. many halictine bees) or social (6. g. bumble bees) require resources
throughout the season, and benefit fiom floral resources that are distributed in time as
well as in space (Michener 2000). Hence, conservationofplant-pollinator interactions
requires a community rather than an individual species approach (Kearns 1998), in which
appropriate plant species are selected to provide resources for bees with diverse
ecological attrlbutes (Potts et aL 2003).

Conservation programs that increase farmland biodiversity are expected to
achieve greatest adoption if they are designed to address multiple needs. Most bee species
require flowering plant resources through a longer time period than when the crop is‘ in
bloom and so flowering plants have been evaluated for supporting crop pollination by
insects (Patten et al. 1993, Kearns and Inouye 1997, Carreck and Williams 1997).
Flowering plants have also been evaluated for use in agricultural settings to help enhance
biological control (Bugg et a1. 1989, Maingay et al. 1991, Bugg and Waddington 1994,

Pontin et al. 2006, see also reviews by Landis et a1. 2000 and Gurr et a1. 2003). Plants

106

that can provide resources for both groups of beneficial insects should provide greater
economic return to growers and should also increase the likelihood that they will be
included in conservation programs designed to enhance arthropod-mediated ecosystem
services.

Typically, non-native flowering annuals are recommended for attracting
beneficial insects in agricultural settings to reduce pest populations (Baggen and Gurr
1998, Baggen et al. 1999, Begum et a1 2006). However these often require yearly
sowing, and would not be suitable for projects that also aim to conserve or restore native
plants and the beneficial insects associated with them. Perennial plants offer the potential
of creating a more stable habitat within and around farm land to enhance beneficial
insects, and a selection of plant species that bloom throughout the growing season is
expected to support beneficial insect communities better than a single sowing of an
annual plant species.

A few studies in North America have evaluated native plants for their attraction to
bees (Patten et a1. 1993, Frankie et al. 2005), and some studies in the United Kingdom
have evaluated pollinator attraction to cultivated (Comba et al. 1999a) and to native or
naturalized (Comba et al. 1999b) flowering plants. However, selection of native plants
from these studies as part of a conservation or restoration project aiming to enhance
pollinator populations is challenging because different plant species were tested in
different years and at different sites. Given the variability in weather, soils, and climate
found between study sites, direct comparison of plant species at the same site is expected
to provide a more robust comparison of relative plant suitability for pollinator

conservation (Patten et al. 1993, Gustafson et al. 2005). As part of a project designed to

107

evaluate native Midwestern USA prairie and savanna plants for their support of natural
enemies (Fiedler and Landis 2006a), I compared 43 native flowering plants for their
attraction to bees. The goal of the combined projects was to identify plants that could be
used in a multi-purpose ecosystem services enhancement program. Here I report on
which plants were visited most frequently by bees and whether simple floral

characteristics can be used as indicators of a plant’s degree of attraction to bees.

MATERIALS AND METHODS

Study site and plants. The study site was established on a former agricultural field
with Marlette fine sandy loam, previously managed in a corn and soybean rotation, at the
Michigan State University Entomology Research Farm in Ingham County, Michigan,
USA. Forty three native plant species were established in 1 m2 blocks spaced 6 m apart
with a background planting of orchard grass (Dactylis glomerata L.). The plots were
established using a randomized complete block design with five replicates of each plant
species. These plants were evaluated for their relative attractiveness to bees. Plant
nomenclature follows Voss (1996) and plant taxonomy follows Judd (2002). Native plant
species were selected for study using the following criteria: 1) native Michigan perennial
plant, 2) adapted to agricultural field conditions (e.g. full sun, moderate drought
tolerance), 3) species representing a diversity of bloom periods, 4) species from a variety
of plant families, with varied flower color and morphology easily accessible by natural
enemies, 5) forb or shrub species formerly found in Michigan oak savanna and prairie,
and 6) local genotypic plants commercially available in Michigan.

Three, five or eight plugs of the perennials were planted per plot, depending on

108

the grth habit of each species, to maximize plant density within the plot. Planting
occurred during the fall of 2003. Plots were maintained as described in Fiedler and
Landis (2006a).

Plant measurements. Floral area per meter square, corolla width, and corolla
depth during peak bloom were recorded from each plant species evaluated. To estimate
floral area per meter square of each plot, the number of open flowers per plot was
counted weekly and multiplied by the average area often representative flowers or
clusters based digital images taken at the site (Coolpix 4800, Nikon, Melville, NY), with
a ruler for reference in each image. Digital images were prepared for analysis by
converting flower images into white space (Knoll 2000) using Adobe Photoshop 6.0
software. Scionlmage fieeware (Alpha 4.0.3.2, www.scioncorp.com) was used to
calculate individual floral area based on the converted images.

Floral morphology was measured on young, open flowers with intact stamens
using a Spot Imaging System (v.3.5.9 Diagnostic Instruments, Inc. Sterling Heights,
Michigan) in combination with an Olympus SZX12 stereoscope. Corolla width and depth
were measured on five flowers per species to the nearest 0.01 cm. For plants in the
Asteraceae, one young, open disc flower was measured per flower head, and for species
with florets, one floret was measured. Width was measured at the point where the corolla
fused and depth was measured from the point of corolla fusion to nectaries (exceptions
are described in F iedler and Landis (2006b)). Corolla depth was recorded as zero in
species with nectaries located at the point where petals attach to the gynoecium.

Vacuum sampling for bees. All floral visitors were collected weekly fi'om 4 May

— 27 September 2005 fi'om flowering plants one week before, the week of; and one week

109

after peak bloom between 0930-1330 EST on calm, sunny days. Samples collected prior
to, during, and after the week of peak bloom (hereafter called “full bloom period”), based
on the weekly counts of the number of open flowers on each plant species, were used in
the analyses. A fine white mesh bag (Kaplan Simon Co., Braintree, MA) was placed over
the intake on a leaf blower (Stihl BG55, Norfolk, VA) modified into a vacuum and plots
were vacuumed until all flowers were sampled. Each sample was frozen, and bees were
subsequently sorted and identified to the lowest taxonomic level using the key of
Michener et al. (1994) and the online key to eastern North American bee species at
www.Discoverlife.org. The number of bees per sample was recorded and averaged over
the number of collections made during peak bloom per plot for analyses. For eight plant
species, one or more of the plots were not in bloom during the three sampling visits and
so the average was taken across the total number of plots sampled.

Bee observations. Timed observations of bees visiting each plot in bloom were
conducted from 1 June — 17 August 2005 between 1000 - 1700 EST on sunny, calm days
when vacuum sampling was not taking place. Each plot was observed once during peak
bloom for 5 minutes, for a total of 5 replicate observations per plant species. Bees visiting
the plants during this time were either recorded and identified to the lowest taxonomic
level (usually genus) in situ or collected with a modified DustbusterTM insect vacuum
(BioQuip Products, Rancho Dominguez, CA) for subsequent identification using the keys
described above. No samples were taken from the earliest blooming species (Sambucus
racemosa L.) and the last four blooming species (Solidago riddellii Frank ex Riddell, S.
speciose Nutt., Aster novae-angliae L., and A. laevis L.).

Statistical analysis. Analysis of variance with Tukey-Kramer adjusted means

110

separation (PROC MIXED, SAS v 8.02) was used to examine differences among plant
species in the number of non-Apis bees that visited plants within early, middle, and late
blooming periods for both the vacuum samples and timed observations. Simple linear
regression analyses (PROC REG, SAS 9.1) were conducted with each pair of floral
characters to check for autocorrelations, then a multiple linear regression analysis (PROC
REG, SAS v 8.02) was conducted on the bees obtained during vacuum sampling to
determine whether bee abundance (honey bees, bumble bees, wild bees other than
bumble bees, and all wild bees) and richness (number of different bee taxa represented in
the samples collected from each plant species) varied with any of the three floral

characteristics (average floral area during full bloom, corolla width, and corolla depth).

RESULTS

Attractiveness of plants. The number of bees collected during vacuum sampling
increased over the course of the 2005 growing season. Across all the plants in each of the
three seasonal groupings, there was an average of 3.5 d: 0.1, 22.6 i 5.7, and 69.8 i 20.2
bees per plant species in the early, mid, and late season samples, respectively (Figure
5.1). There was an associated increase in the richness of bees collected, with average
number of bee taxa collected of 2.3 :t 0.5, 4.8 d: 0.8, and 7.1 d: 1.2 per plant species in the
early, mid, and late season groups, respectively (Figure 5.1).

A total of 875 honey bees and 1393 wild bees was collected via vacuum
sampling. The most abundant wild bee was Bombus impatiens, comprising 62% of the
wild bees collected. Lasioglossum admirandum (93, 6%), Hylaeus aflinis (71, 5%),

Agapostemon virescens (66, 5%), Halictus ligatus (50, 4%), Ceratina calcarata/dupla

lll

females (38, 3%), and Xylocopa virginica virginica (34, 3%) were the next most
abundant wild bee species (Table 5.1). Honey bees were assumed to be from seven

managed hives that were within 200 m of the study site.

 

 

 

 

 

90.0 9.0

A 80.0 - ° abundance " - 8.0 8

1.11“ o ' '

"'5 70.0 - richness ? - 7.0 3°;
0 "E q - w 5-
§% 60.0 I 6.0 g 3
g; 50.0- .. -5.0 35;.
1:3400- 4.08:
8330.0- ~3.0§"‘
-° "5 I a

6 20.0 - i - 2.0 «’71

5 10.0 - - 1.0 jg

0
0.0 r 1 0.0
early middle late

bloom period

Figure 5.1. Average abundance (number of bees per plant species) and richness (number
of bee taxa per plant species) of all wild (non-Apis) bees collected at native plants in 2005
in Ingham Co., Michigan via vacuum sampling during peak bloom Samples are grouped
by peak bloom periods: early (mid-May — June), middle (July — mid-August), and late
(mid-August — September).

112

 

   
   

Andreni
A ndrena
heliant
Andrena

Apidae
Bombus

 

.m
S
m
h
p
m
a
m
e
h
n
w
0
h
S
n.
3
.WO
h
.m
Vu
t
n
u
0
C
m
m
g
1
h
n
S
.m
C
e
p
S
.m
pm
e
m
a
r
p
C
V
.U
a
H
3
4
Fl.
0
t
b
p
h
C
r
a
C
S
C
r
a
1
a
1
n
0
10
b
k
a
e
p
g
n
.n
u
d
S
d
0
.H
C
p
g
.m
m.
m
a
s
m
u
u
C
ma
\
3
g
n
.r11
U
10
d
e
1.
C
k
10
C
S
C
e
B
L
5
k
b
a
T

I - . . r‘ . . ) V .
’ ‘ ’ f g 2 ( ) 7 ( '

. " , H) Geraniaceae, (I) Hydrophyllaceae, (J)
' F) Cap11folaccae, (G) Fabaceae, (
l ‘ ‘ D) Asclepladaceae, (E) Campanulaceae, (
fam111es (A) Asteraceae, (B) Ap1aceae, (C) Apocynaceae, (
after each plant name indicates, when it, was in peak bloom in 2005.

3:2; =a L8 :33

17

858

10

 

Retard .Eo> 33me Safe:

5

 

R

.8 Stan 64V ECGNEME... 33:33:03;

£3. 34 .4 uSESESE EKEEEQEW

€3.43 42:3 A.4V 2:553 23:3ka

33¢,

66

21

 

9348 .4 unanimie exmxosmt

 

P

£3.34 .4 $823430 msfigNSEoU

 

Awe/$8 ESQ 33 632.3%
:34“: 4222 Samar: 30%
9:794 .4 :o: .63 eweotaaxxethemaem

CAN—2&8 28:85 35,453.: SEWER

 

23.4.. 3 .4 benzeneasu VSEQEBD

 

M

E378 .4 mESwSEc 332er

23.34 .4 EEEEEEc 25:32?

 

$442704 50% 52:53 :3va

 

3:330: .4 552:5 SCEESQ

A: V 05:34 4.44 33033: EESQMV

3,34

 

2:78 .4 45:455me 4342.339341

 

$22. _ Q .4 43:32qu 22:45.30

 

3:<-c@ .EoEoI 4.4L 5:: cuehekfifi
354-3 .oo .3;

QWNNQEBVLPXM ~:._....4~3.¢:.\MDMN

Am:<.~4 :23; 25.48439 51.554.

 

332.0 3 .4 36:32: itzfizzm

 

 

Am:<.mmv .4 35241;.3:234
:3.m C .4 36.3%: 9.25%:

3364-5 .4 52:55 3.538.“.

 

”ms/$9.4 and S.>:;e..:= 3452.3;

 

Plant familv"

237.2 4 .4 5:232:55 Ezzaflcomw
C573 5.0.0.44 .4v amaze SEN
E348 .Emm _E::.:,E= SEES»:
:53“: .4 sizmrzmehe eczema?

Eumém. .::Z Edie“? 862437.

-m: :UUUE we xcmi 4.53%.: o~w~§New
Am:<.mmv .4 :EEcebem52:32.4
Aw:<.mv .LEnm Cco>v ::::Ee§3£3~\
215(732 2:22 93%: 74:34
3.2.33 .4 22222.5. 94:24:24.3:
3:64.34 .4 :::EE\..ER EEEEEQ
CST _ my .4 353.8445 2,438.59
333.3 .3 £§Ee§ Essa
Eomémv .4 25m§¢95= 32..

fiaoméimv .4 742.2: .34.;

 

 

Bow

 

S ecies

Andrenidae
Andrena

7.

Ire/iaer/ziformis
Andrena spp.

Apidae

Bombus

birrlacu/atus

Barr
Borr

busfervidus

bus citrinus

seocollis

Borr bus
gr
BO)

57 7 233 19 69 20

6

13 12 37

bus impatiens

I

calcarata/dupla

Ceratina

(9:)

Ceratina dupla

(6‘)

Triepeolus lunatus
Hylaeus amzulatus
Hylaeus mesillae
Hylaeus modestus
Hylaeus

Triepeolus
Hylaeus affinis

Melissodes
bimaculata
Melissodes spp.
Nomada spp.
Triepeolus
concavus
pectoralis
Triepeolus
remigatus
Xylocopa v.
virginica
Colletidae

rudbeckiae
Hylaeus sp.
Halictidae
Agaposzemon

splendens

 

113

 

 

.l

 

 

 

%
:33 3
1
8

QNUNKNM ~v=mnwkm§

 

5:85ng SEESEEN:

2
7
aofinumxufi 333.33%. E

”Sufihhfi SOEmumfimk 7

 

BQGUNKNSQ ka£93mm 6

 

38:33.80 mxfixfignwb 6

 

~33 Gustav. 0
Emmcmm 30m. 6
umoot=§£2=§8k 1

33:43.2. SSMEK 3

 

h3=BUNkm2~Q M3~N~Q=Uw9 2

 

$253423 Emcmmzuw 2

wwwtmndtfib wtbswtvx 3

 

§z3tkmw {KENNNV

 

12 24

38.23% uthoE

@3083: 258:3va

97

 

555.453.:Eztbfioxwkm 4

 

ESSBNSUBE 234.4 :3ka

1

 

Eh: 338a? 7

NhENNSUEQD QEEEQNQMNQ 2

thbmwtab B£Qx©§~vx B

 

BMQNQWQQK hzunmwwtfim 11

 

Emmi; 333 %

 

333»: 36.4%th 4

33:35 34%th

 

3:233:28 23:39?

 

3&3 SEN B

:CZEkfiN: ~t3W$kamm 3

1

3.33430ch nocwmix

 

)
a n4m3 9 3 m

_ _ __ _ _ . _

. . .. . l . 4

. . .. . l l 8

. . .. . . . l

_ 3 __ . _ . 3
l . .. . . . l

. . .. . . . l
l . .. . . . .

. .33. l _ 3

. 7. .. . . . .

. . .. . . . l

. . .. . . . 2

. . .. . . . 2

_ _ .. _ . l 7

. . .. . . . .
l _ .. . . . 2

. l _. . . . _

_ . .. . . . .

. . .. . . _ .

. . .. _ . . l

. . ..l . l 00

. _ .. . . . .
2 S. .. _ 1;. . M

_ l _. _ _ . 7k

_ _ __ _ . _ 3

. . ._ . . _ l

. .I. .. _ _ . 7.

. . ._ . . . l

. . .. _ . . l
to _ .. . . . .

_ . ..l . _ 8
0.1) l 17.. l . m
B . .l . . . .

7‘. 1|

. .3. _ .

l . .. . _ . .
7. . .l _ . _ _
.5. 2 .5l 7. . 3

_ . .. . . . l

. . .5 . _ _ l
4 . _. . . . l
13. . .. . . . _

,.\.

VS

1 .mu 1 l

m H. Wm . mmmmmm
l C 02$. 8. timid!
)5.) .I u (7.31.513
WHO (I /. [3(3I5
RCIGXSSQSUOWOHO
SC/Iluuufluvfluflfl.
USCH17111H<ANKHUA
”WUFKRHHRIDDHDND
aflgUd/L/d/HMCWJHM
WVMahkMPMruaflaa
I [III (I l. I L

555 0 .Su. 53.33%:
g .

 

leucozonium
nymphaearzmz
Lasioglossum

coriaceum
Lasioglossum
cressonii
Lasioglossmn
Lasioglossum
pilosum
Lasioglossmn
rolzweri
Lasioglossum
tegulare
Megachilidae
Ant/zidl'um
manicatum
Megachile
pugnata
Me ac/zile s .

sp. 2
Lasioglossmn

sp. 1
Lasioglossum

84.433422 32923.
323% Quantum.
223%.: QMEEQW
5583me 23.33%
SEER NSESK

Swag MISS

 

19 58 58 272 90 84 32

WNWQEVQE Mzfiutwtmm

33

8582:: $3830 2

EBEENAIE 8339 %
mcawmeogo: .35 B
$.83 52% M
s :
.l. s
p w
4,. b
n
o
n
f
0
0.
N

 

5

23 106 -

8

 

 

7
9

l7 7
I78 89 .
* Not shown are Senecio obovatus Muhl. ex Willd (Asteraceae) and Oenothera biennis L. (Onagraceae), from Wthh no bees were collected.

13
10

114

 

12
2

2

5

2

1

2 - 2

2

1

2
2

 

 

 

No. of non-Apis
bee species:
No. of Apis
melli era:

All but two of the plant species evaluated in this study had at least one wild bee
visitor collected from or observed on them No bees were collected from Senecz'o
obovatus and Oenothera biennis. Most of the plants were visited at low frequency by
bees (Table 5.1), while a smaller sub-set were visited by relatively greater numbers of
bees. Bees were more often collected or observed on plants in the following families:
Asteraceae, Asclepiadaceae, Campanulaceae, Laminaceae, Liliaceae, Rosaceae, and
Scrophulariaceae (Table 5.1). The average number of bees sampled was determined for
each plant species from the five observational samples and the fifteen vacuum samples.
From these values, plant species that were visited by 5 or more bees on average during
the samples were considered highly attractive species (Frankie et al. 2005). Based on this
criterion, wild bees were most attracted to 9 of the native plants: Potentillafruticosa auct.
non L., Scrophularia marilandica L., Veronicastrum virginicum (L.) Farw., Ratibida
pinnata (Vent) Bamh., Agastache nepetoides (L.) Kuntze, Silphium perfoliatum L.,
Lobelia siphilitica L., Solidago riddellii Frank ex Riddell, and Solidago speciosa Nutt.
(Figure 5.2A). Honey bees visited 26 out of 43 native plants and based on the same
criteria developed by Frankie et a1. (2005), they were most attracted to 5 plant species:
Asclepias incarnata L., V. virginicum, Allium cemuum Roth, S. riddellii, and S. speciosa
(Figure 5.2B).

When the three periods of the growing season were considered separately, the
most attractive plants that bloomed early, middle and late in the season were identified.
Vacuum sampling of plants that bloomed during the early season revealed that relatively
few bees were collected, but that wild bees were most abundant at Zizia aurea (L.) Koch

(F1143 = 3.46, P = 0.001) (Table 5.2). The most attractive mid-season blooming plants

115

using this method were P. fruticosa, A. incarnata, V. virginicum, R. pinnata, and Spiraea
alba Duroi (F1331 = 9.93, P < 0.0001) (Table 5.2). The most attractive late season plants
were A. nepetoides, S. perfoliatum, L. siphilitica, and S. riddellii, and S. speciosa (F1557 =
16.83, P < 0.0001) (Table 5.2).

From timed observations during the early season, wild bees were most attracted to
Fragaria virginiana Duchesne and Coreopsis lanceolata L. (F1135 = 7.38, p < 0.0001)
(Figure 5.3A). Of the mid-season plants, wild bees were most fiequently seen at P.
fruticosa, S. man'landica V. virginicum, and R. pinnata (F18,62 = 10.65, p < 0.0001)
(Figure 5.3A). During the late season bloom period, wild bees were most attracted to A.
nepetoides, S. perfoliatum and L. siphilitica (F1137 = 16.07, P < 0.0001) (Figure 5.3A).
Honey bees were observed visiting Scrophularia marilandica at much higher rates than
fiom samples taken with the vacuum (Figure 5.3B).

When the proportion of wild bees captured during vacuum sampling was
regressed against the proportion of wild bees observed during timed observations, in
order to compare methods, the methods appeared to be moderately similar with a
regression coefficient of 0.51 (Figure 5.4A). However, examination of the data in Figure
5.4 shows that the observational method was biased toward recording more bees on
Potentillafruticosa, and that the vacuum sampling method was biased toward collecting
more bees on Lobelia siphilz'tica. Excluding these two outliers increased the regression
coefficient to 0.86 with a slope of 1.24. Scrophularia marilandica was the outlier for
honey bees, with more honey bees recorded during observations than collected while
vacuum sampling. Excluding this outlier provided a regression coefficient of 0.86 with a

slope of 1.27.

116

Relationship between floral characteristics and bee abundance. The native
plants evaluated in this study ranged in their peak bloom period fiom the first week in
May to the first week of October. The range of peak bloom covered by these plants
indicates the temporal range of mostly herbaceous flowering resources achievable with a
combination of native plants (Table 5.2). Early blooming plants typically had the smallest
average floral area, with the overall size of floral area increasing toward the end of the
season among species (Table 5.2). Average corolla width and depth of the flowers tested
did not vary among species throughout the season.

Floral characteristics only explained 14% of the variation in all wild bee
abundance, 14% of the variation in bumble bee abundance, and 13% of the abundance of
wild bees other than bumble bees, with floral area being the significant parameter (Table
5.3). Almost none of the variation in honey bee abundance could be predicted by floral
characteristics (Table 5.3). However, the number of different wild bee species visiting the
tested plants could be explained by floral characteristics, which explained 33% of the
variation in bee diversity, suggesting that floral area may be used to indicate potential bee

diversity at particular plants (Table 5.3).

117

late

mid

early

 

 

 

 

  

 

 

 

I Apis mellifera

 

«Sow: $u<
8.35-33: :2m<

saggy: E=E§w
quQQE «EOE?»
880:3: £53:
83283 gauge,
dew 38%
8:338 Eebozuon
EaaEco E:.s<
3:53 E05050
“:38ch 230$
83$: €3th
E2859: 53:38:83,
22:35 «8:38.‘
805m .3255
8%:EtmE €33.36
3:30th 3553: o0
Boga «not
8835 mwfioBo.‘
3:358: ”250530
E::3m::wu E35003
.8093: «$389:
£30683 £30900
m:wo.:2:m Eocene:
E3§me EzoBEoI
3:383:93 wo:ou:<
«.8322 $52301
$32553 e:oEo:<
SEESPS Eaicqofixl
nausea £85m.
was: mEN
wu:obm:wo Emotau<
EEwScmE 53:800
3859.: mtwufil
89:89 ozoanmm

 

 

oocwucanw eon 2.3

W 0..

m m m

.m m We)
mm m wumw
.3.“ .W BIS
MmManMw
mmmmmwme
“13mg
m.wmm.wmmm
mHAchm
Inn-lulu
000000
050505
32211

200 n
150 ‘
100 ‘

plant species

Michigan in

and (B) honey bees collected during 15, 30

9

Total number of (A) wild bees
sec vacuum samples for floral visitors at 43 native plants in Ingham Co.,

2

5

lgure

2005. Plants are organized from left to right by bloom phenology in 2005. Blue lines

F

represent different bee species. Red lines indicate environmental characters. This figure is

presented in color.

118

 

 

 

to had I 1| I I I I . I I od m6 2: O 336:3 33.65.,
. 5m M I _. . I W . I M m 2 :o 93%. < ucgsmfifié
e. 23. m l u I m u I m m an .2 3: m 5335.: 5.583228.
2. an m I . I m . . I I m m we do 35: a 332385.35
.3 mg I I I . I u . I I 3 ad 2% m were. 33$.
3.. «2 ml I l I . WI . . I ml m 2. 3. $2 a 835:9: Ensign
3 86 m I I m . I . m m o: 5 En a és§s§ EEuEeG
.5 23 m m . I ._ m m ed an E? 0 Saga 53
to .36 I I I I .I I I I W o; md 13% o 86:35 weifiumw
.3 “3 m | MI I I . ml m od 2 32 2 38.5% 3588.6
3 85 m I ml . I . m . I m 3 3 3? o Engages 55:83..
3.. m; I I I IIIII u . I u .. I | ed 2. 3.8 o §§E§§§£
m m m m seem as
._ m; I I I I I MI I I I m _. I ._ m 3 3 3.2. < ss§§33§8
e. 93 l I . I .. I 3 9m 22 o Baurmswufiaam
an 36 m m m I I I I m 0.9 m._ an: m 535.38.: 5:38me
a :3 m m ml .. I .m Qo 2 «in m entities 8.3%
a. :3 m m m . I . m S: m.“ was m 3:22 Egg:
an end I I . I I I I WI 0.0 mm mdmm 2 mmgwhescb gosmfix
a. 38 m m m _. I m u on 3. 33 _ §§e§e 233:83:
a o I .. I u I :o 3 :8 < 3223 £85.
a :3 m m m | . I m . I | od 3 28 m 8% saw
._ m; m m ml I . I m . I 2m 3 an: 2 engage i233
._ 5o m I I . WI . oo 5 N San : sesame 55:28
e. and m m I .. I ._ I o.o :m 35 o 333%.: crews:
tram? I . I . od 3 2% a 96282 332%
.e: oun..u>< II :35 been
.3 M £3. m .3. m E... m 32 M 0 mm m m u. «8.8.5
IIIBEOA— Soc-fl m. .m" T. Wu (u MI M"
w, .. m .. .. ..
m m m .

 

Eco—sue 520 5 Ba: 03 $53 .38 E “ca—m name we 885 M89 “:96
£933 38¢ Ho.“ mam—9:8 Ego? maize c9828 moon 2:5 :0 bass: omega; 05 98 date: 883 .859tta ESE .N.m view.

119

.3039398 838 >835 “gamma. hum—353mm 2a 8032 ~88me 3 BRO—Em 8w80>< ._.

acoz 33% n I .885 .5 u _. .23 883 fix. I I ”8.5a a8: 5 hi 2....

uncommon—ht, u m .uaooatsanobm n m .oaooawgmxdm

.I. 0 .08853— n m .oaoonmom n O .owooagnm N Z Janos—sodas u 2 680838 u A 688:5 u M 63838.3 u _. ages—Samoa»: H H 6383800
u m dawning n O .onooflflwfiuo n m 688352580 u m .oanofino—ofix M Q .oaooaaoom< n U .28an< n m 638834 N < ”858mm aqua you 83D .2
42.32: a m_ €33 6.3.3.5 .0 8a .326 £33.23 03 838: =< I

 

 

 

.5 2: . I . I m m m m S. no 38 < ”Eats?
.3 «.2 I . I . Im m m m S 3 032 < wémzigoffiw
.5 86 I ._ I .. I ma 2. 3:; < u§8q€m§3m
2. 8.9 I ._ I . m m m m 3 9o 33 < unfit $35
.8 :2 I I . mI . I m m m 3 2 3; < 8&8 €33
a :8 . I . I 9o 3 m3 0 Si canfig
.5 E; I I . mI . I I I I m m .2 1m 32». < Ecsgaéasmm
a. 8.2 I l I m . I . I I ml m m 3 3 2% m 853% £83
3 OS I I m .. I . I I m m m 3 2 323 < sagabfisarsfifi
.5 can I I . I a u | m to 3 was < SBEEEQ £88
. m3: I m . I n I I ml m m 2 3 25; < saszbfisfififiw
.3 m; I I . I _. Im " m 3 3 3% < SEES uses»;
to 86 I I I. .I m I I I I I mil in fig cane _. SUBSRNSEEQE
.3 $6 I m . n I u .. ml m m S E v.88 . $333 38.8%
m m m m :88» 33
2. mom I I . . I . m M ed 3 new 0 33885
a 25 m I . I . I m M od 3 :2 o mamngcu ESEEQQ
mason? I I . . I . I cflo mu no; x 53.58 23.5w
6:@uc.5>< WI I I I I I I o mm gm 0 m2 \— fltzmfi 522»sz
.5 m §< m .2. m E... u :2 m m w m a u u. Inseam
$.85.— ESE m. m m m m. m m.
) fl ) m I .M.
m m u u
m m u

 

.Bsaaoo .3 as;

120

 

late

mid

eafly

u%,

7%

i fly/4m,“
,7,:,E_r,:_:s,_:. 3,—m
:'

: ,M, ,, _W

3, Tiff/é

In :H w///A

t, n
, ,

,/

,, Acre-\I‘otte'evun rrrrrr t.

-////// aaaaaaaaaaaaaa H «H45:

7 nunrnou'nln'nm. n .....

\
t

”/4 -... r _

’l/l 7/% % III!

’///1

 

 

‘ Sodom .0st

£2.22 $6833
£892.23 32.233:
8.25.2223 c.2323
“253323 Eatemqam
wagoESm gmomo
5532228 53222.5
octane? £20229)
£80232 mEm2o§
322252 m2oEmmu<
mfim 88.22%
om2mbm2mo 25.592on
22.2280 E22?
£2222 892028
3882.8 m2BoE<
3852 museum
92.380.220.23,
$32.82.. mmEEom<
$928 m2efio>
8.829.292 m2m32qoam
2.22232qu

@323 meet

omega mmEoBm<
323.28% «250280
E=2Em22mo E22282?
38.232 £22201
830823 £38.60
38.525 @6203:
2522.2me anBEoI
eBaangm 82mu2<
2.39.22 2229651
$38.38 e2oEm2<
$232qu:
35620 282%
8.26 SEN

$55828 £522.32
55338.2 2.22.2288

121

Plant species

C] All other Apidae
I Apis mellifera

I Bombus spp.
12 ~ I Megachilidae

Halictidae
I Andrenidae
a Colletndae

392.29.: «238m

 

 

 

fi . A _

q
0 8 6 4 2 0
1

1

0
4
12‘

2 a _
6 4 2

16

14-

10‘
8‘

2252030 5.: m .on moon *0 .oz

observations. Observations were not made at the earliest (Sambucus racemosa) and latest

(two species each of Solidago and Aster) blooming species in the study. Plants are

Figure 5.3. Average number of (A) wild bees, and (B) honey bees noted during 5 timed
organized fi'om left to right by bloom phenology in 2005.

.0
N
or

 

y = 0.8322x + 0.0044 9
R2 = 0.51

9 s:
—l N
01 O
O

P
_L
O

  
  

.0
o
01

 

during vaccum sampling

9

 

Proportion of wild bees collected

 

0.00 I
0.00 0.05 0.10 0.15 0.20 0.25

Proportion of wild bees recorded
during timed observations

V l i f

0.35

 

y=1.123x-0.0032 ’ B
0-30 ‘ R2: 0.74

0.25 ~
0.20 -
0.15 -
0.10 -
0.05 ‘

0.00 , I r .
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Proportion of honeybees recorded
during timed observations

 

during vaccum sampling

 

 

 

I

Proportion of honeybees collected

Figure 5.4. Comparison of bee sampling methods using simple linear regression of (A)
the proportion of wild bees and (B) the proportion of honey bees caught or recorded
using each method at 38 of the plant species tested in Ingham County, Michigan in 2005.

122

Table 5.3. Results of multiple linear regressions of the abundance and diversity of bees
collected at native flowering plants during peak bloom against three floral characters.
Significant regression coefficients (P < 0.05) and probability values less than 0.05 are
highlighted in bold.

 

 

 

Overall model Parameter estimate probabilities
floral corolla corolla

Variable R2 F3,39 P area width depth
Bee abundance

Honey bees (A. mellifera) 0.05 0.69 0.56 0.54 0.29 0.65

Bumble bees (Bombus spp.) 0.14 2.18 0.11 0.02 0.57 0.99

“ng “her than bumble 0.13 2.01 0.13 0.03 0.86 0.44

All wild bees 0.14 2.08 0.12 0.03 0.69 0.61
Bee diversity

No. of wild bee species 0.28 5.09 0.005 0.001 0.45 0.42

DISCUSSION

With increasing concern about the suitability of agricultural landscapes for wild
pollinators (Buchmann and Nabhan 1996, National Academy of Sciences 2006) and other
beneficial insects (Baggen and Gurr 1998, Landis et a1. 2000, Begum et al. 2006),
conservation activities are expected to increase in agricultural lands. Agricultural habitats
can be inhospitable for beneficial insects during much of the growing season due to
intensive management practices. Intensification of agricultural systems over the past fifty
years has led to declines in native bee populations through various mechanisms (Osborne
et al. 1991, Matheson 1994, Allen-Wardell et a1. 1998, Stubbs and Drummond 2001b).
The most important of these are the use of agrochemicals for pest control and the loss of
field margins and hedgerows, resulting in habitat fiagmentation and a reduction in flower
abundance and diversity in farm landscapes (Buchmann and Nabhan 1996, Kearns 1998,

Steffan-Dewenter and Tschamtke 1999). This loss of plant diversity translates into both

123

spatial and temporal gaps in the availability of floral resources (Matheson 1994). By
integrating flowering plants that support native bees into farms, growers of pollination-
dependent crops may receive greater pollination services when the crop is in bloom
(Matheson 1994).

Most species of native bees that are endemic in agricultural landscapes require
nectar and pollen resources beyond those that a crop plant may provide. Because of this,
conservation of bees on farmland will require that they have access to plants that provide
suitable flowers throughout the growing season. The investment required to create a
managed season-long area of flowering plants would suggest that optimizing the
suitability of the plant species for local pollinators will give the greatest return on that
investment in terms of pollinator conservation and benefit to the crop. Based on two
sampling methods, this study has identified 14 native perennial plants to which wild bees
in southern Michigan show affinity. These plants were originally selected to be suitable
for use by natural enemies, so their use in an agricultural setting could promote
pollination and biological control, the two main ecosystem services provided to
agriculture by arthropods.

Plants in this study were divided into early, middle, and late-blooming groups,
and we found increasing bee abundance and diversity as the season progressed. This
temporal pattern in bee abundance at flowers mirrors the availability of floral resources,
variation in weather (waell et a1. 2005) and population growth of multivoltine and social
bees later in the season (Buchmann and Nabhan 1996). By taking this approach, plants
that attracted relatively few bees in the spring were not being compared to those in bloom

during the warmer summer months when social bee colony size was greatest. In their

124

peak bloom order, the plants fi‘equented the most by bees were: F ragaria virginiana,
Zizia aurea, Coreopsis lanceolata, Potentillafruticosa, Scrophularia marilandica,
Asclepias incarnata, Veronicastrum virginicum, Ratibida pinnata, Spiraea alba,

A gastache nepetoides, Silphium perfoliatum, Lobelia siphilitica, Solidago riddellii, and
Solidago speciosa (Table 5.2, Figure 5.3). These plants are representatives from seven
different plant families: five species of Asteraceae, three Rosaceae, two
Scrophulariaceae, and one each of Apiaceae, Asclepiadaceae, Campanulaceae, and
Laminaceae. All of these families contain species of plants that have been shown to be
attractive to bumble bees and other wild bees (Corbet et al. 1994, Frankie et al. 2005,
Carvell et al. 2006).

Because this study was designed primarily to address the issue of providing
attractive floral resources for natural enemies, a vacuum sampling method was used to
collect insects. While this is an unconventional method for monitoring bees, results fi'om
timed observations were similar for most plant species to those obtained during vacuum
sampling. Vacuum sampling could be considered for use across a large set of field sites
as a rapid and reproducible method to obtain a fairly accurate representation of the bee
community.

It was unexpected that so few bees in the families Andrenidae and Megachilidae
were collected in the samples. However, the time of year when this study was conducted
may account for their absence. For instance, many andrenid and megachilid bees emerge
early in spring, and are solitary univoltine species, thus their period of activity may not
have overlapped with this study. Future studies should include blooming woody plants

that provide floral resources early in the spring to help support early-season pollinators

125

(Matheson 1994). Also, floral resources alone are not enough to sustain populations of
bees; nesting resources are also needed. It may be that this site is depauperate of the kind
of nesting resources required by cavity nesting bees in the family Megachilidae.

Of the three floral attributes measured in this study, floral area was the mo st
explanatory factor related to the abundance of bees other than honey bees. This finding
suggests that unlike honey bees, that receive information from hive mates about
rewarding patches, wild bees maximize reward for their foraging efforts by seeking
patches with greater floral area. This finding agrees with previous studies showing that
pollinating insects concentrate their foraging in dense patches of flowers (Thomson 1981,
Westphal et a1. 2003, Hegland and Totland 2005, Hegland and Boeke 2006). Plants with
greater average floral area were also more likely to have greater wild bee diversity (Table
5.2). Together, these results suggest that floral area might be a simple indicator of a bee’s
potential attraction.

Recent studies have linked plant community diversity to pollinator community
diversity in natural systems (Potts et al. 2003), and long term declines in bee pollinated
plants have been linked to declines in pollinators (Biesmeijer et aL 2006). Further
evidence comes from experimental studies that have shown that pollinator diversity is
linked to the persistence of plant communities (Fontaine et al. 2006), In agricultural
systems, diverse pollinator communities can increase productivity in crops such as
sunflowers (Greenleaf and Kremen 2006), watermelon (Kremen et a1. 2002), and coffee
(Klein et al. 2003), so enhancing pollinator diversity is a worthwhile goal for managers of
land on which pollinator-dependent crops are grown.

The link between pollinator and plant diversity support the continued

126

development of native perennial plants for use within beneficial insect conservation
programs in agricultural settings. Perennial plants may have higher initial planting costs
than annuals and take some time to mature and reach their potential floral area, but there
are long-term benefits. In addition to providing resources for pollinators (waell et al.
2005, Carvell et al. 2006) and insect natural enemies (Landis et a1. 2000, Colley and Luna
2000, Gurr et al. 2003), these plant species are adapted to the local environment
(Gustafson et al. 2005), and can also provide aesthetic value to the landscape (Goulder
and Kennedy 1997).

A first step toward conservation of native bees on farmland is to determine which
plants are most suitable for providing foraging resources at different times of the growing
season. The results from this direct comparison of co-blooming plants can be combined
with the findings of Fiedler and Landis (2007a, 2007b) related to natural enemy
attraction. Using these two studies, future research should evaluate a combination of
highly suitable plants to provide overlapping bloom periods from late spring through the
rest of the season. Such a combined floral planting can then be tested for its utility in
conserving beneficial insects within agricultural settings, with the ultimate aim of
improving sustainable pollination of crops that depend on bees for this important

component of yield.

127

CHAPTER 6:

RESPONSE OF NATIVE BEES TO LAND USE PATTERNS IN BLUEBERRY
AGROECOSYSTEMS

128

INTRODUCTION

Concern over perceived worldwide pollinator declines and their relationship to
anthropogenic land use change have been the driving force behind a number of landscape
level studies of bee abundance and diversity. The response of bee communities to
anthropogenic land use and proximity to natural areas has been documented around the
world. In California, Kremen et al. (2002) found that both conventional and organic
watermelon fields far fiom adjacent natural habitat had low wild bee community diversity
and had to be supplemented with honey bees, whereas fields near natural habitat had
sufficient pollination by native bees. In Germany, potted flowering plants were placed
next to cereal fields within landscape matrices containing varying proportions of semi-
natural habitat. Bees were attracted to these plants according to their body size and
nesting guild, with long-range foraging bees responding to the degree of semi-natural
habitat at larger spatial scales than bees with shorter foraging ranges (Steffan-Dewenter et
al. 2002). However, Westphal et al. (2003) found that bumble bees were more likely to
respond to the percentage of mass-flowering crops than to nearby semi-natural habitat
and found that bumble bee foraging duration decreased and colony grth increased
where forage resources were abundant near their nests (Westphal et al. 2006). In Costa
Rica, pastures near forest patches had a greater abundance and diversity of native
meliponine bees than those distant fiom the forest (Brosi et al. 2007), and fi'uit set of

coffee plants was greater near larger adjacent tropical forest fragments (Ricketts 2004,

129

Ricketts et al. 2004). In Argentinian grapefi'uit plantations, flower-visiting insects also
were most abundant in fields bordering premontane subtropical forests (Chacoff and
Aizen 2006).

The effect of land use on pollinator communities in urban settings was explored
recently in Arizona, California, and New Jersey. Desert pollinator communities were in
residential areas with xeric landscaping were found to be more diverse than turf-based
yards, but most diverse in areas on the fringe of the metropolitan area of Phoenix
(McIntyre and Hostetler 2001). In northern California, both native and exotic bees (e. g.
Apis mellifera) preferred native over exotic flowering plants on a percent basis, even
though exotic plants far outnumbered the natives in the two urban areas studied (Frankie
et al. 2005). In southern New Jersey, more bee species were found in agricultural and
residential developments than in extensive pine-barren forests (W infree et al. 2007).

Several studies have examined the response of cavity-nesting bees and wasps to
spatial scale. Tylianakis et al. (2006) found that most cavity nesting bees and wasps
responded to landscape quality at small scales in Ecuador, while Klein et al. (2006) found
that diversity and parasitism of trap-nesting Hymenoptera were greater in agroforestry
adjacent to rainforests in Indonesia. In Germany, the abundance of trap-nesting bees,
wasps, and their natural enemies increased with greater percent semi-natural habitats
within a radius of 750 m, but decreased with increasing landscape scale up to 3 km
(Steffan-Dewenter, 2002).

The underlying pattern in most of these studies is that land use affects bee
abundance and/or diversity at various spatial scales based on the bees’ potential use of

resources, its mobility or foraging range, and proximity to nesting habitat. In southwest

130

Michigan, where most highbush blueberry (Vaccinium corymbosum) production occurs,
land use has changed dramatically over the past 3040 years, towards a greater percentage
of developed land. Because native bees are a significant component of the pollinating
community in blueberry fields, and blueberry is a long-lived perennial crop, the system is
well suited to native bee community studies in a landscape context. My objectives were:
(1) to examine the pattern of bee abundance and diversity across 15 highbush blueberry
farms in the context of landscape structure and scale, and (2) to determine whether

blueberry pollination was affected by landscape characteristics.

METHODS

Thirteen commercial and two semi—abandoned highbush blueberry farms located
at least 3 km away from one another in southwest Michigan, were sampled during bloom
in 2004-06 using pan traps. Due to varying weather conditions from year to year, trapping
was conducted two (2004, 2006) or three (2005) times during bloom in each field.
Trapping occurred between 16 May - 3 June in 2004, 16 — 25 May in 2005, and 17 - 31
May in 2006.

Five pairs of white and yellow pan traps mounted on 1.2 m PVC poles were
placed 5 m apart along each of two transects running perpendicular to the orientation of
the rows. One transect was established within 1 m of the field edge and the other was
established 25 m into the field. Traps were set out between 8:00-12:00 h and were
collected between 16:00-20:00 h for a minimum trapping period of 6 hours on days when
weather conditions met the following criteria: minimum temperature of 13°C with clear

or partly cloudy skies or 17°C with any sky condition other than rain (waell et al. 2005).

131

Pan traps filled halfway with a 2% unscented soap solution (Dawn® dish soap,
Procter & Gable, Cincinnati, OH), were constructed from 355 ml white and yellow
plastic bowls (Amscan, Inc., Elmsford, NY) mounted onto 2.7 diameter PVC poles
stabilized with rebar (see Chapter 2, page 37). After the sampling period concluded, pan
trap contents were strained into plastic bags and stored in a -12°C freezer for later
processing. Specimens were thawed at room temperature prior to washing in a 7 0%
ethanol solution. Honey bees were separated out and counted, then stored in 70% ethanol
solution. All other bees were placed in a mesh bag through which they were fluffed and
dried with a hairdryer before pinning and identification.

Species identifications. Preliminary identifications of bees to the lowest possible
taxonomic group were made using two published dichotomous keys (Mitchell 1960,
Michener et al. 1994) and the online key available through www.Discoverlife.org.
Further identifications and verifications were made by J .S. Ascher of the American
Museum of Natural History, Division of Invertebrate Zoology. Voucher specimens are
held in the Albert J. Cook Arthropod Research Collection at Michigan State University
(see Appendix B).

Yield assessments. To compare fi'uit set and yield on blueberry clusters exposed
or not exposed to pollinators, five unopened flower clusters on two separate branches on
ten blueberry bushes within 1 m of the field edge were tagged. On each cluster, the
number of flowers were counted, then one set of clusters was covered per bush with fine
mesh netting (bridal veil) to exclude floral visitors. When bloom was finished, the mesh
was removed and the number of fruit set per cluster was counted on all the marked

branches. After the terminal berry was ripe and the other berries were starting to turn

132

blue, all the berries from the clusters on the marked branches were harvested. Berries
were counted and weighed to obtain an average berry weight. The diameter of the largest
berry from each sample was measured with a caliper, then the berry was squashed to
extract its seeds. This was done in a plastic zip bag so that seeds could easily be felt and
seen through the bag and counted.

Quantification of landscape features. Aerial photos taken during the summer of
2005 at a precision level of 1:5000 were downloaded fiom the USDA Geospatial Data
Gateway (httpz/ldatagateway.nrcs.usda.gov/GatewayHome.html) and imported into
ArcMap (Arc G18 9, ESRI). A handheld GPS unit (SportTrak Pro, Magellan Navigation,
Inc., Santa Clara, CA) was used to record coordinates for each of the field sites, and then
coordinates were added to the ArcGIS file. Because bumble bees are very rare in my bee
collection and honey bees are presumed to be from managed hives, I choose to make the
maximum radius of the area of concern 1500 meters because most native bees are
assumed not to forage fiirther than this. Thus, site inspections were conducted within a
1500 m radius of the focal field and aerial photos were digitized and labeled accordingly
(see Table 6.1 for description of landscape categories used). Circles with radii of 250,
500, 750, and 1000 were overlaid on the images, and the area calculated for each digital
piece of land in the resulting nested areas (Figure 6.1). Landscape categories used in
regression analyses were percent: (1) forest margins, (2) settlement, (3) annual cropland,
(4) blueberry crop habitat, (5) open uncultivated (pasture/meadow/fallow/scrubland,
ditches/tree lines, field margins, and vegetation near water), and (6) semi-natural habitats

(which are groups 1 and 5 added together) (Table 6.1).

133

Statistical analyses. The Mantel test (“vegan” package for R 2.3.1) was used to
compare pairwise bee community similarity indices (Jaccard, Bray-Curtis, and Morisita-
Hom) with pairwise geographic distances between each of the fifteen blueberry farms.
Spatial autocorrelation in non—pairwise variables (bee abundance, species richness,
Shannon-Weiner and Simpson’s diversity indices) was assessed with Moran's I (“ape”
package for R 2.3.1).

Simple linear regressions (PROC REG, SAS 9.1) at each spatial scale were
conducted between the bees collected during bloom (overall abundance, richness, and
diversity) and to each of the 6 land use groupings (see above). Similarly, the abundances
of each the 5 of the most abundant blueberry foragers were also regressed against the 6
different groupings. The difference between the fruit set, weight, diameter, and seed
count from clusters that were open pollinated and those from which pollinators were
excluded was used in regression analyses (PROC REG, SAS 9.1) to determine the
relationship between the change in those aspects of fruit development and landscape

context using the 6 land use groups above.

134

 

Figure 6.1. Examples of the aerial photographs used to digitize landscape features; (A) is
a site with a high proportion of annual and nursery crops and depicts the 5 different
radius (meters) sectors used in the analyses, (B) is a site near Lake Michigan (to the west)
with a high proportion of settlement area, and (C) is a site with a high proportion of
blueberry plantations. Images from USDA-NRCS Geospatial Data Gateway
(http://datagateway.nrcs.usda.gov/GatewayHomehtml last accessed 8/14/2007). This
figure is presented in color.

135

Table 6.1. Categories of landscape types used in the digitization of aerial photos.

 

 

Habitat type Description

Blueberry plantations commercial and semi-abandoned highbush blueberry
Perennial crops other perennial crops, including vineyards and nurseries
Annual crops annual crops

Pastures grazing pastures

Open uncultivated including meadows, scrubland, fallow and other ruderal areas
Ditches and treelines running along or bisecting agricultural land

Other field margins margins along agricultural land other than ditches and tree lines
Forest/woodland <10 m from forest edge

margrn

Forest/woodland >10 m fi'om forest edge

interior

Settlement suburban development including golf courses and a landfill
Road paved or dirt

Train tracks abandoned or still in use

Utility areas cleared for powerlines

shoreline vegetation along Lake Michigan

wetlands usually vegetation along a river

Other vegetation near
water

open water

vegetation along inland bodies of water such as ponds and lakes

open bodies of water including lakes, ponds, and rivers

136

RESULTS

Independence and landscape composition of sites. Bee communities at the
sampled fields were considered to be independent based on the results of the Mantel test
for community similarity (Z = 1729.3, df = 14, p = 0.14). Likewise, bee abundance,
species richness, Shannon-Weiner and Simpson’s diversity indices assessed with Moran’s
I were not significant indicating no spatial autocorrelation among sites (I < 0.514, df =
13, p > 0.05).

Farms selected for this study varied in the proportion of each measured land use
type in the surrounding landscape (Table 6.2). The most abundant landscape types were
forest margins and settlements, which accounted for an average of 22 and 21%,
respectively, of the total area of all habitats at the 1500 m spatial scale (Table 6.2).
Annual cr0pland and blueberry plantations each comprised 13% of the land within a 1500
m radius of the fields (Table 6.2). All other landscape types accounted for the remaining
area at the 1500 m spatial scale (Table 6.2).

Response of bees to landscape scale. Total bee abundance (log n+1), species
richness, and diversity did not vary significantly with any of the 6 landscape categories at
any of the spatial scales (Table 6.3). However, 34% of the variation in the abundance of
Andrena carolina, a Vaccinium specialist, at the 1500 m scale was explained by the
proportion of settlements, with there being more A. carolina with a greater proportion of
settlement area (Figure 6.2). Variation in the abundance of bees in the species group
Ceratina calcarata/dupla was explained by the proportion of blueberry fields at the 250
and 500 m spatial scales (35 and 28% respectively), with fewer bees associated with

higher proportions of blueberry farmland (Table 6.3, Figure 6.3b). Also, C. calcarata/

137

dupla increased in abundance with more semi-natural habitat at the 250 m spatial scale

with 28% of the variation explained by this landscape category (Figure 6.3d).

Table 6.2. Composition and quantification of the 1500 m radius landscape sectors in
southwestern Michigan. Data were gathered by field inspection. Means 3: S.E., minimum,
and maximum are given for 15 study sites.

 

 

 

Area (%) of landscape

Landscape type Average Minimum Maximum
Forest margin (10 m deep) 22.48 i 3.11 5.39 44.56
Settlement 20.64 i 2.53 6.00 36.99
Annual crops 13.08 i 3.05 0 37.79
Blueberry plantations 12.71 :t 3.39 0.08 45.06
Forest interior (>10 m) 7.98 i 0.85 3.35 13.69
Open uncultivated‘l' 7.33 :1: 0.81 0.78 12.50
Perennial crops and nurseries 5.25 i 3.17 0 38.82
Open water 3.04 i 1.53 0.25 23.51
Road 2.07 i 0.18 1.25 3.50
Ditches and tree linesT 1.65 :t 0.38 0.02 4.18
Other field margins‘l’ 1.54 i 0.31 0.58 5.70
Pasture'l' l .06 i 0.60 0 8.62
Vegetation near water‘l' 0.63 d: 0.49 0 7.39
Wetland 0.49 i 0.49 0 7.30
Conifer plantation 0.30 d: 0.24 O 3.65
Utility and train tracks 0.25 :1: 0.21 0 3.22

 

1‘ Landscape categories that were grouped together as semi-natural habitat.

138

Table 6.3. Regression coefficients for wild bee abundance, species richness, diversity,
and the 5 most abundant Vaccinium-foragers averaged over three years and the
proportion of (a) forest margin, (b) settlement, (c) annual cropland, (d) blueberry

plantations, (e) semi-natural habitat, and (f) semi-natural and forest margins together at 5

spatial scales. Significant (P < 0.05) r'2 values are in bold and the slope indicated in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

parentheses.
250 m 500 In 750 m 1000 In 1500 m
a) Forest margin (10 m deep)
log (wild bee abundance +1) 0.08 0.02 0.05 0.06 0.06
wild bee species richness 0.03 0.02 0.05 0.02 0.02
wild bee diversity (H’) 0.0002 0.03 0.09 0.07 0.04
Andrena carolina 0.12 0.008 0.005 0.02 0.0003
Lasioglossum coriaceum 0.0002 0.19 0.16 0.08 0.05
Ceratina calcarata or dupla (S? only) 0.16 0.02 0.003 0.006 0.003
Andrena carlini 0.01 0.02 0.004 0.004 0.02
Augochlorella aurata 0.001 0.05 0.04 0.04 0.06
b) Settlement
log (wild bee abundance +1) 0.02 0.002 0.006 0.02 0.08
wild bee species richness 0.04 0 0.002 0.007 0.03
wild bee diversity (H’) 0.14 0.10 0.007 0.005 0.005
Andrena carolina 0.0002 0.008 0.02 0.08 0.34 (+)
Lasioglossum coriaceum 0.05 0.04 0.02 0.01 0.0004
Ceratina calcarata or dupla (9 only) 0.02 0.007 0.0009 0.002 0.005
Andrena carlini 0.10 0.001 0.001 0.003 0
Augochlorella aurata 0.004 0.06 0.08 0.09 0.13
9 Annual crop
log(wild bee abundance +1) 0.004 0.001 0.01 0.04 0.10
wild bee species richness 0.0004 0.003 0.02 0.04 0.09
wild bee diversity (H’) 0.05 0.04 0.0002 0.008 0.05
Andrena carolina 0.004 0.0001 0.0002 0.0007 0.009
Lasioglossum coriaceum 0.06 0.04 0.09 0.11 0.17
Ceratina calcarata or dupla (92 only) 0.0006 0.07 0.02 0.006 0.004
Andrena carlini 0.0006 0.0001 0.02 0.04 0.05
Augochlorella aurata 0.02 0.0004 0.02 0.02 0.005
d) Blueberry
log(wild bee abundance +1) 0.06 0.007 0.0002 0.0002 0.0001
wild bee species richness 0.17 0.04 0.01 0.006 0.003
wild bee diversity (H’) 0.14 0.01 0 0.007 0.02
Andrena carolina 0.0004 0.01 0.006 0.002 0.006
Lasioglossum coriaceum 0.0005 0.003 0.009 0.02 0.02
Ceratina calcarata or dupla (9 only) 0.35 (-) 0.28 (-) 0.17 0.12 0.10
Andrena carlini 0.005 0.006 0.02 0.03 0.04
Augochlorella aurata 0.04 0.09 0.08 0.08 0.08
e) Semi-natural
log (wild bee abundance +1) 0.03 0.004 0.002 0.007 0.03
wild bee species richness 0.11 0.01 0.03 0.008 0
wild bee diversity (H’) 0.06 0 0.04 0.03 0.02
Andrena carolina 0.005 0.002 0.02 0.11 0.21
Lasioglossum coriaceum 0.008 0.0001 0.05 0.02 0.005
Ceratina calcarata or dupla (9 only) 0.13 0.02 0.10 0.15 0.24
Andrena carlini 0.01 0.002 0.006 0.0001 0.02
Augochlorella aurata 0.01 0.003 0.007 0.03 0.05

 

139

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l) Semi-natural + forest margins
log (wild bee abundance +1) 0.08 0.01 0.006 0.04 0.06
wild bee species richness 0.17 0.002 0.002 0.0007 0.005
wild bee diversity (H’) 0.07 0.005 0.0001 0.0005 0.0002
Andrena carolina 0.04 O 0.03 0.10 0.09
Lasioglossum coriaceum 0.01 0.04 0.0003 0.006 0.03
Ceratina calcarata or dupla (9 only) 0.28611 0.03 0.10 0.11 0.08
Andrena carlini 0.005 0.0001 0.001 0.0007 0.03
Augochlorella aurata 0.008 0.001 0.03 0.06 0.08
0.4 45
A - B
0.35 r e i 40 R2 = 0.34
0.3 - r 35 P = 0.02
a 30 «
025 - g 25 ,
at 0.2 7 is 20 _
0.15 “ g 15 4
0.1 ‘ . 10 1
0.05 r g 5 r
0 3 I . r 1 < 0 I 3: l
0 500 1000 1500 0 0.1 0.3 0.4
Radius (m) Proportion of settlement

 

 

 

 

 

In a 1500 m radius sector

Figure 6.2. Scale-dependent effect of landscape structure on the number of Andrena
carolina bees collected in pan traps at 15 highbush blueberry fields in southwestern

Michigan. The graph on the left (A) shows the regression coefficients for the average
number of bees and the proportion of human settlement at each of the 5 spatial scales.

The graph on the right (B) shows a simple linear regression of the spatial scale with the

highest r2 value for A. carolina and the proportion of human settlement.

140

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.4 A ‘2 25 B
O
0.35 - . i 20 R2 = 0.35
0.3 . , .. P = 0.02
0.25 - g 15 1
0.2 r o
0.15 — ’ § 10 -
_ ° a
0.1 0 a 5 -
O
0.05 « 3
0 i if r 0
0 500 1000 1500 2000 0 0.2 0.4 0.6
Radius (m) Proportion of blueberry habitat
in a 250 m radius sector
. ,_ 5
O 3 e C a 2 . D
0.25 — E20 R2 = 0.28
a P = 0.04
7 I
0.2 § 15 _
r2 0.15 7 3' 10 -
0.1 1 . ° . g
0
0.05 ~ . g 5 ‘
0 i r 1 < 0
0 500 1000 1500 2000 0 0.2 0.4 0.6
Radius (m) Proportion of semi-natural habitat

In a 250 m radius sector

Figure 6.3. Scale dependent effects of landscape structure on the number of Ceratina
calcarata or dupla (they are morphologically indistinguishable) female bees collected in
pan traps at 15 highbush blueberry fields in southwestern Michigan. The graphs on the
left show the regression coefficients for the average number of bees and the proportion of
(A) blueberry habitat or (C) semi-natural habitat at each of the 5 spatial scales. The
graphs on the right show simple linear regressions of the spatial scales with the highest r2
value for C. calcarata/dupla and the proportion of (B) blueberry habitat or (D) semi-
natural habitat.

141

Response of fruit set, size, and weight to landscape scale. Three of the four fruit
attributes, used as a measure of pollination, varied significantly with the proportion of the
landscape containing blueberry fields. Regression coefficient (R2) values for the
proportion of the difference between open-pollinated and pollinator-excluded fruit set at
harvest were highest at the 500 m scale (Figure 6.4b), while fi'uit weight per berry (Figure
6.4d) and diameters of the largest berry (Figure 6.41) were highest at the 250 m scale, all
increasing with the proportion of blueberry habitat (Figure 6.4). Berry weight and fruit
diameter decreased with a greater proportion of open uncultivated habitats at the 250 and
500 m scales respectively (Figure 6.5b,d). Finally, berry weight and hit diameter
decreased with a greater proportion of semi-natural habitat at the 500 m spatial scale
(Figure 6.6b,d). This result was found even though there was no significant correlation

between land used for blueberry production or in semi-natural habitat..

142

Table 6.4. Regression coefficients for the proportion of fruit set at harvest, fiuit weight
per berry at harvest, diameter of the largest berry, and the number of seeds per largest
berry averaged over three years and the proportion of (a) forest margin, (b) settlement, (c)
annual cropland, (d) blueberry plantations, (e) semi-natural habitat, and (f) semi-natural
and forest margins together at 5 spatial scales. Significant (P < 0.05) R2 values are in

bold; slope indicated in parentheses.

 

 

250 m 500 m 750 m 1000 m 1500 m
a) Forest margin (10 m deep)
proportion of Mt set at harvest 0.08 0.0003 0.005 0.001 0.06
weight per berry 0.05 0.06 0.01 0.009 0.02
diameter of largest berry 0.006 0.05 0.03 0.006 0.0002
no. of seeds in largest berry 0.04 0.0003 0.0005 0 0.02
b) Settlement
proportion of fruit set at harvest 0.13 0.05 0.04 0.4 0.01
weight per berry 0.02 0.008 0 0.01 0.02
diameter of largest berry 0.005 0 0.01 0.07 0.06
no. of seeds in largest berry 0.006 0 0.003 0.03 0.03
c) Annual crop
proportion of fruit set at harvest 0.007 0.04 0.04 0.03 0.05
weight per berry 0.07 0.07 0.05 0.07 0.009
diameter of largest berry 0 0.004 0.008 0.03 0.001
no. of seeds in largest berry 0.02 0.01 0.02 0.005 0.01
d) Blueberry
proportion of fruit set at harvest 0.13 0.34 (+) 0.26 0.19 0.16
weight per berry 0.47 (+) 0.21 0.18 0.19 0.14
diameter of largest berry 0.27 (+) 0.17 0.20 0.24 0.21
no. of seeds in largest berry 0.06 0.009 0.02 0.05 0.04
e) Semi-natural
proportion of fruit set at harvest 0.06 0.05 0.13 0.13 0.12
weight per berry 0.58 (-) 0.58 (-) 0.32 (-) 0.26 0.10
diameter of largest berry 0.22 0.26 (-) 0.22 0.12 0.05
no. of seeds in largest berry 0.04 0.05 0.04 0.009 0.004
t) Semi-natural + forest margins
proportion of fruit set at harvest 0.13 0.05 0.09 0.06 0.008
weight per berry 0.50 (-) 0.58 (-) 0.21 0.18 0.08
diameter of largest berry 0.21 0.26 (-) 0.12 0.04 0.02
no. of seeds in largest berry 0.08 0.05 0.03 0.004 0.01

 

143

0.4
0.35 - .

0.3 ~
0.25 - ’

 

.0
or

 

.0 .o
co .2
l l

0.15 . ’ 0-2 J

0.1 -
0.05 ~
0 I l I 0 fi 1

0 500 1000 1500 2000 0 0.2 0.4 0.6

Proportion of blueberry
Radius 1'“) In a 500 m radius sector

0.1 -

 

,2
o
N
0
Prop. of fruit set at harvest

 

 

 

 

 

 

0.5

 

 

0.35 ‘

0.3 ‘

“L 0.25 ‘
0.2 ' O . e

0.15 r e

0.1 -

0.05 '

 

 

 

 

 

Weight (g) per berry at harvest

 

 

0 500 1000 1500 2000 0 0-2 0.4 0.6

Proportion of blueberry
Radius (ml In a 250 m radius sector

0.3
0.25 - .

0.2 - . ’

x 0.15 -

0.1 4

0.05 -

0 I I I
0 500 1000 1500 2000 0.2 0.4 0.6

Proportion of blueberry
Radius (ml In a 250 In radius sector

 

 

R2 = 0.27
P = 0.046

1

l

L

 

l

O—le-hUION
L

 

 

 

Diameter (mm) of largest berry

 

 

 

C

Figure 6.4. Scale-dependent effects of the proportion of the landscape in blueberry
production on three different fruit attributes at harvest at 15 highbush blueberry fields in
southwestem Michigan. Graphs on the left show the regression coefficients for (A)
proportion of hit set, (C) weight per berry, and (E) diameter of largest berry and the
proportion of blueberry habitat at each of 5 spatial scales. Graphs on the right show
simple linear regressions of the spatial scales with the highest r2 value for (B) proportion
of hit set, (D) weight per berry, and (F) diameter of largest berry and the proportion of
blueberry habitat.

144

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.7 0.5
06 - A 2’ B
' . . g 0.4
0.5 7 ‘
0.4 - g 0-3 ‘
x O
0-3 ‘ , g 0.2 4
02 - ‘3
‘J 0.1 4
0.1 ~ e g,
0 . e , i 0 . .
0 500 1000 1500 2000 0 02 0.4 0.6
Proportion of open uncultivated habitat
Radius (m) In a 250 m radius sector
0.3 7
C I? D
0.25 — ° 5 6 . R2=0.26
0 2 . . g 5 -1 P = 0.054
. - e
_ 4 . e
x 0.15 - E 3 _ . .
e E ‘ 0 e e
0.1 1 TE: 2 q . .
0.05 - 0 g 1 J . e
0 . . . 5 0 . .
0 500 1000 1500 2000 0 0.1 02 0.3
Radius (m) Proportion of open uncultivated habitat

in a 500 rn radius sector

Figure 6.5. Scale-dependent effects of the proportion of open uncultivated land on two
different fruit attributes at harvest at 15 highbush blueberry fields in southwestern
Michigan. Graphs on the left show the regression coefficients for (A) weight per berry
and (C) diameter of largest berry and the proportion of open uncultivated habitat at each
of 5 spatial scales. Graphs on the right show simple linear regressions of the spatial scales
with the highest r2 value for (B) weight per berry and (D) diameter of largest berry and
the proportion of open uncultivated habitat.

145

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. 0.5
0 7 A ~ B
0.6 ~ . g Q4 q
0.5 ‘ . u
0.4 - E 0-3 ‘
“k 0 3 _ 3
- g 0.2 -
0.2 ~ ’ , a
2' 0.1 4
0.1 ‘1 . g
0 T . . i 0 . .
0 500 1000 1500 2000 0 0.1 0.2 0.3
Proportion of semi-natural habitat
Radius (m) Ina500rnradlus sector
0.3 7
C E D
025 J ’ .8 6 - . R2=0.26
_ P = 0.054
0.2 - . g 5 e
_ 4 . .
0.15 - ‘5
“K e E 3 I . . 9 e e
0.1 r .5, 2 _ e e
5 e
0.05 - . g 1 - e
. -
0 . . . o 0 . .
0 500 1000 1500 2000 0 0.1 0.2 0.3

p

Radium) "twenties”:
Figure 6.6. Scale-dependent effects of the proportion of semi-natural land, including
woodland habitat, on two different fruit attributes at harvest at 15 highbush blueberry
fields in southwestern Michigan. Graphs on the left show the regression coefficients for
(A) weight per berry and (C) diameter of largest berry and the proportion of semi-natural
habitat at each of 5 spatial scales. Graphs on the right show simple linear regressions of
the spatial scales with the highest 1'2 value for (B) weight per berry and (D) diameter of
largest berry and the proportion of semi-natural habitat.

146

DISCUSSION

Bees and their relatives have been shown to respond to habitat features in the
landscape at various scales depending on their known or expected foraging range and the
quality of the surrounding habitat (Perfecto and Vandermeer 2002, Steffan-Dewenter
2002, Stefi‘an-Dewenter et al. 2002, Ricketts 2004, Klein et a1. 2006, Veddeler et al.
2006). The proportion of the landscape that is semi-natural has been the strongest
predictor of bee abundance, presumably because this provides nesting sites and non-
managed habitat for bees. Bees with long foraging ranges such as honey bees have been
shown to respond to landscape features up to 3000 m from sampled fields, whereas
smaller, solitary bees with shorter foraging ranges, respond to landscape features within
750 m (Steffan-Dewenter et al. 2002, Klein et aL 2004).

In this study, where any honey bees collected were presumed to be from managed
hives temporarily present in the landscape, and where bumble bees were infrequently
collected, most of the bee species collected during blueberry bloom were solitary or semi-
social. Measurements of abundance, richness, or diversity of wild bees showed no
significant relationship to any of the landscape features examined, which is unexpected
considering previous studies (Steffan—Dewenter et al. 2002, Klein et a1. 2006, Veddeler et
al. 2006.

Despite the lack of response of the whole bee community to landscape variation,
some individual bee species that are known to forage on blueberry varied significantly
with the surrounding landscape. Members of the species complex Ceratina
calcarata/dupla were more abundant in habitats with a higher proportion of semi—natural

landscape at the 250 m scale. This fits well with the idea that the smaller the bee, the

147

smaller its foraging range. This bee is one of the smallest species in Michigan (5-7 mm),
and is a semi-social carpenter bee that nests in pithy stems (Michener 2000). Thus, its
size and the fact that semi-natural habitat is likely to contain a high proportion of its
nesting resource, could explain this relationship. Ceratina calcarata/dupla were less
abundant in habitats with a high proportion of blueberry fields at the 250 m spatial scale.
This could be due to the intensive management of plants in crops margins, i.e. the
destruction of nesting resources, or that management practices such as insecticide use
after bloom are reducing the number of bees that survive to the following year.

In contrast to the expected response by solitary bees to landscape complexity, the
main blueberry specialist bee species found in this region, Andrena carolina, was more
abundant in habitats containing a greater proportion of human settlement at the 1500 m
spatial scale. At first, this appears to support the Winfree et al. (2007) finding that some
bees may benefit from the floral diversity and nesting opportunities provided by non-
natural, residential landscapes. However, because A. carolina is a solitary, soil-nesting
species (Michener 2000) and is a blueberry specialist (MacKenzie and Eickwort 1996),
one would expect this bee to be more highly correlated with blueberry field abundance
and to have a relatively short foraging range, considering the results of Steffan—Dewenter
et al. (2002). On the contrary, this result suggests that the foraging range of this species
may be farther than expected, allowing it to use habitat containing a wide variety of plant
species for forage. It is possible that A. carolina is capable of foraging up to a much
greater distance than expected by the average foraging range of solitary bees (Steffan-

Dewenter et a1. 2002).

148

Pollination of blueberry, measured as the difference in fi'uit set, berry weight, and
diameter of the largest berry, between open pollinated and pollinator-excluded flower
clusters, increased with the area of surrounding blueberry plantations. This seems to
indicate that blueberry habitat, typically containing a high proportion of managed honey
bees is beneficial to production of this crop. Honey bees were the most abundant species
of bee collected in the pan traps during bloom and were the most abundant kind of bee
observed visiting blooms (see Chapter 3).

Fruit weight and diameter decreased with the proportion of open-uncultivated and
semi-natural habitats within 500 m of the focal field. One explanation for this result is
that there may not have been enough bees either native, rented, or as spillover from
neighboring farms in fields that were surrounded by more semi-natural habitat (i.e. the
fields may have been too isolated). This result suggests that wild pollinators residing in
adjacent habitats may be contributing little to pollination in intensive highbush blueberry
production, contrary to other studies in which adjacent semi-natural habitat and its
associated diversity of wild bees resulted in higher fi'uit or seed set (Steffan—Dewenter
and Tschamtke 1999, Ricketts et al. 2004).

Conclusion. This is the first study of which I am aware to look at the response of
several solitary soil-nesting or semi-social dwarf carpenter bee species and their response
to landscape context at different spatial scales. In most of the sampled fields, honey bees
outnumbered any other kind of bee by two or three to one, and honey bees were unlikely
to be greatly affected by landscape context since they were only brought in for crop
bloom. In this situation, the presence of honeybees may make it difficult to measure the

effect of native bees on commercial blueberry pollination. Further studies to examine

149

resource partitioning between different bee species, similar to that by Greenleaf and
Kremen (2006) in sunflower, would be helpful in describing what role native bees play in
the pollination of this crop for which certain native bee species, such as Bombus spp., are

efficient pollinators.

150

LITERATURE CITED

Allen-Wardell, G., P. Bernhardt, R. Bitner, A. Burquez, S. L. Buchmann, J. H.
Cane, P. A. Cox, V. Dalton, P. Feinsinger, M. Ingram, D. W. Inouye, C. E. Jones, K.
Kennedy, P. Kevan, H. Koopowitz, R. Medellin, S. Medellin-Morales, G. P. Nabhan,
B. Pavlik, V. Tepedino, P. F. Torchio, and S. Walker. 1998. The potential
consequences of pollinator declines on the conservation of biodiversity and stability of
food crop yields. Conservation Biology 12: 8-17.

Aizen, M. A. and P. Feinsinger. 1994. Habitat fragmentation, native insect pollinators,
and feral honey-bees in Argentine Chaco Serrano. Ecological Applications 4: 378-392.

Baggen, L. R. and G. M. Gurr. 1998. The influence of food on Copidosoma koehleri
(Hymenoptera : Encyrtidae), and the use of flowering plants as a habitat management tool
to enhance biological control of potato moth, Phthorimaea operculella (Lepidoptera :
Gelechiidae). Biological Control 11: 9-17.

Baggen, L. R., G. M. Gurr, and A. Meats. 1999. Flowers in tri-trophic systems:
mechanisms allowing selective exploitation by insect natural enemies for conservation
biological control. Entomologia Experimentalis et Applicata 91: 155-161.

Batra, S. W. T. 1997. Solitary bees for Vaccinium pollination. Acta Horticulturae 446:
71-76.

Baum, K. A., W. L. Rubink, and R. N. Coulson. 2006. Trapping of feral honeybee
workers (Hymenoptera : Apidae) in a coastal prairie landscape: effects of season and
vegetation type. Canadian Entomologist 138: 228-234.

Begum, M., G. M. Gurr, S. D. Wratten, P. R. Hedberg, and H. I. Nicol. 2006. Using
selective food plants to maximize biological control of vineyard pests. Journal of Applied
Ecology 43: 547-554.

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

Bohart, G. E. 1972. Management of wild bees for the pollination of crops. Annual
Review of Entomology 17: 287-312.

Bosch, J. and W. P. Kemp. 2001. How to Manage the Blue Orchard Bee. Sustainable
Agriculture Network, Beltsville, MD.

Bosch, J., W. P. Kemp, and S. S. Peterson. 2000. Management of Osmia lignaria

(Hymenoptera: Megachilidae) populations for almond pollination: methods to advance
bee emergence. Environmental Entomology 29: 874-883.

151

Boyle, R. M. D. and B. J. R. Philogene. 1983. The native pollinators of an apple
orchard: variations and significance. Journal of Horticultural Science 58: 355-363.

Brewer, J. W., R. C. Dobson, and J. W. Nelson. 1969. Effects of increased pollinator
levels on production of the highbush blueberry, Vaccinium corymbosum. Journal of
Economic Entomology 62: 815-818.

Brosi, B.J., G.C. Daily, and P.R. Ehrlich. 2007. Bee community shifts with landscape
context in a tropical countryside. Ecological Applications 17: 418-430.

Buchmann, S. L. 1983. Buzz pollination in angiosperms, pp. 73-113 In C. E. Jones and
R. J. Little [eds], Handbook of Experimental Pollination Biology. Van Nostrand
Reinhold, New York

Buchmann, S. L. and G. P. Nabhan. 1996. The Forgotten Pollinators. Island Press,
Washington, DC.

Bugg, R. L., R. T. Ellis, and R. W. Carlson. 1989. Ichneumonidae (Hymenoptera)
using extrafloral nectar of faba bean (Viciafaba L., Fabaceae) in Massachusetts.
Biological Agriculture and Horticuture 6: 107-114.

Bugg, R. L. and C. Waddington. 1994. Using cover crops to manage arthropod pests of
orchards - a review. Agriculture, Ecosystems and Environment 50: 11-28.

Cane, J. H. 2001. Habitat fiagmentation and native bees: a premature verdict?
Conservation Ecology 5: 3.

Cane, J.H., R.L. Minckley, and L. Kervin. 2001. Sampling bees (Hymenoptera:
Apifonnes) for pollinator community studies: pitfalls of pan-trapping. )oumal of the
Kansas Entomological Society 73:208-214.

Cane, J. H. and J. A. Payne. 1993. Regional, annual, and seasonal variation in
pollinator guilds: intrinsic traits of bees (Hymenoptera: Apoidea) underlie their patterns
of abundance at Vaccinium ashei (Ericaceae). Annals of the Entomological Society of
America 86.

Cane, J. H. and V. J. Tepedino. 2001. Causes and extent of declines among native
North American invertebrate pollinators: detection, evidence, and consequences.
Conservation Ecology 5: 1.

Carreck, N. L. and I. H. Williams. 1997. Observations on two commercial flower
mixtures as food sources for beneficial insects in the UK Journal of Agricultural Science
128: 397-403.

Carvell, C., W. R. Meek, R. F. waell, D. Goulson, and M. Nowakowski. 2007.

Comparing the efficacy of agri-environment schemes to enhance bumble bee abundance
and diversity on arable field margins. Journal of Applied Ecology 44: 29-40.

152

Carvell, C., P. Westrich, W. R. Meek, R. F. waell, and M. Nowakowski. 2006.
Assessing the value of annual and perennial forage mixtures for bumblebees by direct
observatibn and pollen analysis. Apidologie 37: 326-340.

Chacoff, N. P. and M. A. Aizen. 2006. Edge effects on flower-visiting insects in
grapefi'uit plantations bordering premontane subtropical forest. Journal of Applied
Ecology 43: 18-27.

Chauzat, M.-P., J.-P. Faucon, A.-C. Martel, J. Lachaize, N. Cougoule, and M.
Aubert. 2006. A survey of pesticide residues in pollen loads collected by honey bees in
France. Journal of Economic Entomology 99: 253-262.

Chittka, L. and R. Menzel. 1992. The evolutionary adaptation of flower colors and the
insect pollinators color-vision. Journal of Comparative Physiology A - Sensory, Neural,
and Behavioral Physiology 171: 171-181.

Colley, M. R. and J. M. Luna. 2000. Relative attractiveness of potential beneficial

insectary plants to aphidophagous hoverflies (Diptera: Syrphidae). Environmental
Entomology 29: 1054-1059.

Comba, L., S. A. Corbet, A. Barron, A. Bird, S. Collinge, N. Miyazaki, and M.
Powell. 1999a. Garden flowers: insect visits and the floral reward of horticulturally-
modified variants. Annals of Botany 83: 73-86.

Comba, L., S. A. Corbet, L. Hunt, and B. Warren. 1999b. Flowers, nectar and insect

visits: evaluating British plant species for pollinator-friendly gardens. Annals of Botany
83: 369-3 83.

Corbet, S. A., N. M. Saville, and J. L. Osborne. 1994. Farmland as a habitat for bumble
bees, pp. 33-46 In A. Matheson [ed.], Forage for Bees in an Agricultural Landscape.
International Bee Research Association, Cardiff, UK.

Croxton, P. J., C. Carvell, J. O. Mountford, and T. H. Sparks. 2002. A comparison of
green lanes and field margins as bumblebee habitat in an arable landscape. Biological
Conservation 107: 365-374.

Daily, G. C. 1997. Introduction: What are ecosystem services?, pp. 1-10 In G. C. Daily
[ed.], Nature's Services. Island Press, Washington, DC.

Darrow, G. M. and D. H. Scott. 1966. Varieties and their characteristics, pp. 94-108 In
P. Eck and F. Childers [eds], Blueberry Culture. Rutgers -- The State University, New
Brunswick, NJ.

Darvill, B., M. E. Knight, and D. Goulson. 2004. Use of genetic markers to quantify
bumble bee foraging range and nest density. OIKOS 107: 471-478.

153

Dedej, S. and K. S. Delaplane. 2003. Honey bee (Hymenoptera: Apidae) pollination of
rabbiteye blueberry Vaccinium ashei var. 'Climax' is pollinator density-dependent.
Journal of Economic Entomology 96: 1215-1220.

DeGrandi-Hoffman, G. 1987. The honeybee component of horticultural crop production
systems. Horticultural Review 9: 237-272.

DeGrandi-Hoffman, G. 2003. Honey Bees in US. Agriculture: Past, Present, and
Future, pp. 11-20 In K. Strickler and J. H. Cane [eds], For Normative Crops, Whence
Pollinators of the Future? Entomological Society of America, Lanham, Maryland.

DeGrandi-Hoffman, G., M. Chambers, S.S. Schneider, D.R. Tarpy,and D. Smith.
2006. The African honey bee has arrived — so where do we go fi'om here? Bee Culture
134: 24-28.

Delaplane, K. S. and D. F. Mayer. 2000. Crop Pollination by Bees. CABI Publishing,
New York.

Desneux, N., A. Decourtye, and J.-M. Delpuech. 2007. The sublethal effects of
pesticides on beneficial arthropods. Annual Review of Entomology 52: 81-106.

Dogterom, M. H. 1999. Pollination by four species of bees on highbush blueberry. Ph.D.
Biological Sciences, Simon Fraser University.

Dogterom, M. H. and M. L. Winston. 1999. Pollen storage and foraging by honey bees
(Hymenoptera: Apidae) in highbush blueberries (Ericaceae), cultivar 'Bluecrop'. The
Canadian Entomologist 131: 737-768.

Dogterom, M. H., M. L. Winston, and A. Mukai. 2000. Effect of pollen load size and
source (self; outcross) on seed and fi'uit production in highbush blueberry cv. 'Bluecrop'
(Vaccinium corymbosum; Ericaceae). American Journal of Botany 87: 1584-1591.

Dorr, J. and E. C. Martin. 1966. Pollination studies on the highbush blueberry,
Vaccinium corymbosum L. Q. Bull. Mich. St. Univ. Agric. Exp. Stn. 48: 437-448.

Drummond, F. A., F. E. Groden, and R. J. Prokopy. 1984. Comparative efficacy and
optimal positioning of traps for monitoring apple maggot flies (Diptera: Tephritidae).
Environmental Entomology 13: 232-235.

Drummond, F. A. and C. S. Stubbs. 1997a. Potential for management of the blueberry
bee, Osmia atriventris Cresson. Acta Horticulturae 446: 77-85.

Drummond, F. A. and C. S. Stubbs. 1997b. Sampling bee populations in lowbush
blueberry in Maine. Acta Horticulturae 446: 101-108.

Edwards, M. 1996. Optimizing habitats for bees in the United Kingdom - a review of

recent conservation action, pp. 35-45 In A. Matheson, S. L. Buchmann, C. O'Toole, P.
Westrich, and I. H. Williams [eds], The Conservation of Bees. Academic Press, London.

154

Elliott, N. C., R. W. Kieckhefer, G. J. Michels, Jr., and K. L. Giles. 2002. Predator
abundance in alfalfa fields in relation to aphids, within-field vegetation, and landscape
matrix. Environmental Entomology 31: 253-260.

Fiedler, A. K. and D. A. Landis. 2007a. Attractiveness of Michigan native plants to
arthropod natural enemies and herbivores. Environmental Entomology. In Press.

Fiedler, A. K. and D. A. Landis. 2007b. Plant characteristics associated with natural
enemy attractiveness to Michigan native plants. Environmental Entomology. In Press.

Fontaine, C., I. Dajoz, J. Meriguet, and M. Loreau. 2006. Functional diversity of
plant-pollinator interaction webs enhances the persistence of plant communities. PLoS
Biology 4: 129-135.

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

Frankie, G. W., R. W. Thorp, M. Schindler, J. Hernandez, B. Ertter, and M.
Rizzardi. 2005. Ecological patterns of bees and their host ornamental flowers in two
northern California cities. Journal of the Kansas Entomological Society 78: 227-246.

Frazier, M. T., J. Finley, W. Harkness, and E. G. Rajotte. 2000. A sequential
sampling scheme for detecting infestation levels of tracheal mites (Heterostigmata:

Tarsonemidae) in honey bee (Hymenoptera: Apidae) colonies. Journal of Economic
Entomology 93: 551-558.

Free, J. B. 1993. Insect Pollination of Crops, 2nd ed. Harcourt Brace Jovanovich,
London.

Gabriel, D. and T. Tschamtke. 2007. Insect pollinated plants benefit fi‘om organic
farming. Agriculture, Ecosystems and Environment 118: 43-48.

Gill, C. 2007. Honey bee die-off alarms beekeepers, crop growers and researchers
College of Agricultural Sciences, News Releases (online, last accessed 4/18/2007, ).

Goulder, L. H. and D. Kennedy. 1997. Valuing ecosystem services: philosophical bases
and empirical methods, pp. 23-47 In G. C. Daily [ed.], Nature's Services: Societal
Dependence on Natural Ecosystems. Island Press, Washington, DC.

Goulson, D. 2003. Bumblebees: Their Behavior and Ecology. Oxford University Press,
Oxford.

Greenleaf, S. S. and C. Kremen. 2006. Wild bees enhance honey bees' pollination of
hybrid sunflower. PNAS 103: 13890-13895.

Gumbert, A. 2000. Color choices by bumblebees (Bombus terrestris): innate preferences
and generalization after learning. Behavioral Ecology and Sociobiology 48: 36-43.

155

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.

Gustafson, D. J., D. J. Gibson, and D. L. N ickrent. 2005. Using local seeds in prairie
restoration: data support the paradigm. Native Plants Journal 6: 25-28.

Hegland, S. J. and L. Boeke. 2006. Relationships between the density and diversity of

floral visitor activity in a temperate grassland community. Ecological Entomology 31:
532-538.

Hegland, S. J. and 0. Totland. 2005. Relationships between species' floral traits and
pollinator visitation in a temperate grassland. Oecologia 145: 586-594.

Heinrich, B. 2004. Bumblebee Economics. Harvard University Press, Cambridge,
Massachusetts.

Holzschuh, A., I. Steffan-Dewenter, D. Kleijn, and T. Tschamtke. 2007. Diversity of
flower-visiting bees in cereal fields: effects of farming system, landscape composition
and regional context. Journal of Applied Ecology 44: 41 -49.

Hurd, P. D., Jr. 1979. Superfamily Apoidea, pp. 1741-2209 In K. V. Krombein, P. D.
Hard, Jr., D. R. Smith, and B. D. Burks [eds], Catalog of Hymenoptera in America North
of Mexico. Smithsonian Institution Press, Washington, DC.

J ahns, T. R. and G. D. J olliff. 1991. Survival rate and reproductive success of Osmia
lignaria propinqua Cresson (Hymenoptera: Megachilidae) in cages meadowfoam,
Limnanthes alba Benth. (Lirnnanthaceae). Journal of the Kansas Entomological Society
64: 95-106.

J avorek, S. K., K. E. MacKenzie, and S. P. Vander Kloet. 2002. Comparative
pollination effectiveness among bees (Hymenoptera: Apoidea) on lowbush blueberry

(Ericaceae: Vaccinium angustifolium). Annals of the Entomological Society of America
95: 345-35 1 .

Johansen, C. A. 1977. Pesticides and pollinators. Annual Review of Entomology 22:
1 77 -1 92.

Johansen, C. A. and D. F. Mayer. 1990. Pollinator Protection: a Bee and Pesticide
Handbook. Wicwas Press, Cheshire.

Judd, W. S. 2002. Plant systematics: a phylogenetic approach. Sinauer Associates,
Sunderland, Mass

Kearns, C. A. 1998. Endangered mutualisms: the conservation of plant-pollinator
interactions. Annual Review of Ecology and Systematics 29: 83-112.

156

Kearns, C. A. and D. W. Inouye. 1993. Techniques for Pollination Biologists.
University Press of Colorado.

Kearns, C. A. and D. W. Inouye. 1997. Pollinators, flowering plants, and conservation
biology. Bioscience 47: 297-307.

Kearns, C. A. and J. D. Thomson. 2001. The Natural History of Bumblebees: A
Sourcebook for Investigations. University Press of Colorado, Boulder, Colorado.

Kells, A. R., J. M. Holland, and D. Goulson. 2001. The value of uncropped field
margins for foraging bumblebees. Journal of Insect Conservation 5: 283-291.

Kevan, P. G. 1991. Pollination: keystone process in sustainable global productivity. Acta
Horticulturae 288: 103-110.

Kevan, P. G. 1999. Pollinators as bio indicators of the state of the environment: species,
activity and diversity. Agriculture, Ecosystems and Environment 74: 373-393.

Kevan, P., M. Giurfa, and L. Chittka. 1996. Why are there so many and so few white
flowers? Trends in Plant Science 1: 280-284.

Kevan, P. G., C. F. Greco, and S. Belaoussofl'. 1997. Log-normality of biodiversity and
abundance in diagnosis and measuring of ecosystemic health: pesticide stress on
pollinators on blueberry heaths. Journal of Applied Ecology 34: 1122-1136.

Kevan, P. G. and T. P. Phillips. 2001. The economic impacts of pollinator declines: an
approach to assessing the consequences. Conservation Ecology 5: 8.

Kevan, P. G. and R. C. Plowright. 1989. Fenitrothion and insect pollination, pp. 13-42
In W. R. Ernst, P. A. Pearce, and T. L. Pollock [eds], Environmental Effects of
Fenitrothion Use in Forestry: Impacts on Insect Pollinators, Songbirds, and Aquatic
Organisms. Environment Canada, Dartmouth, NS.

Kim, J ., N. Williams, and C. Kremen. 2006. Effects of cultivation and proximity to
natural habitat on ground-nesting native bees in California sunflower fields. Journal of
the Kansas Entomological Society 79: 309-320.

Klein, A., I. Steffan—Dewenter, and T. Tschamtke. 2003. Fruit set of highland coffee
increases with the diversity of pollinating bees. Proceedings of the Royal Society of
London B 270: 955-961.

Klein, A.-M., I. Steffan-Dewenter, and T. Tschamtke. 2004. Foraging trip duration
and density of megachilid bees, eumenid wasps and pmpilid wasps in tropical
agroforestry systems. Journal of Animal Ecology 73: 517-525.

Klein, A. M., I. Steffan-Dewenter, and T. Tschamtke. 2006. Rain forest promotes

trophic interactions and diversity of trap-nesting hymenoptera in adjacent agroforestry.
Journal of Animal Ecology 75: 315-323.

157

Klein, A.-M., B. E. Vassiere, J. Cane, I. Steffan-Dewenter, S. Cunningham, C.
Kremen, and T. Tschamtke. 2006. Importance of pollinators in changing landscapes
for world crops. Proceedings of the Royal Society of London B FirstCite.

Knoll, E. A. 2000. Adobe Photoshop V 6.0. Adobe Systems Incorporated.

Kovach, J., C. Petzoldt, J. Degnil, and J. T ette. 1992. A method to measure the
environmental impact of pesticides. New York’s Food and Life Sciences Bulletin, No.
‘ 139, New York State Agricultural Experiment Station, Geneva, NY.

Kremen, C., N. M. Williams, R. L. Bugg, J. P. Fay, and R. W. Thorp. 2004. The area
requirements of an ecosystem service: crop pollination by native bee communities in
California. Ecology Letters 7: 1109-1119.

Kremen, C., N. M. Williams, and R. W. Thorp. 2002. Crop pollination from native
bees at risk from agricultural intensification. PNAS 99: 16812-16816.

LaBerge, W. E. 1980. A revision of the bees of the genus Andrena of the western
hemisphere. Part X. Andrena. Transactions of the American Entomological Society 106:
395-525.

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:
1 75-201 .

LeBuhn, G. et al. 2002. The Bee Inventory Plot. http://on1ine.sfsu.edu/~beeplot/ (last
accessed 8/16/2007)

Legendre, P. and L. Legendre. 1998. Numerical Ecology, Second English Edition,
Elsevier Science B.V., Amsterdam

Leong, J. M. and R. W. Thorp. 1999. Colour-coded sampling: the pan trap colour
preferences of oligolectic and nonoligolectic bees associated with a vernal pool plant.
Ecological Entomology 24: 329-335.

Long, R. F., A. Corbett, C. Lamb, C. Reberg-Horton, J. Chandler, and M.
Stimmann. 1998. Beneficial insects move fiom flowering plants to nearby crops.
California Agriculture 52: 23-26.

Losey, J. E. and M. Vaughan. 2006. The economic value of ecological services
provided by insects. Bioscience 56: 311-323.

MacKenzie, K. E. 1997. Pollination requirements of three highbush blueberry

(Vaccinium corymbosum L.) cultivars. Journal of the American Society of Horticultural
Science 122: 891-896.

158

‘h—

 

MacKenzie, K. E. and A. L. Averill. 1995. Bee (Hymenoptera: Apoidea) diversity and
abundance on cranberry in southeastern Massachusetts. Annals of the Entomological
Society of America 88: 334-341.

MacKenzie, K. E. and G. C. Eickwort. 1996. Diversity and abundance of bees

(Hymenoptera: Apoidea) foraging on highbush blueberry (Vaccinium corymbosum L.) in
central New York. Journal of the Kansas Entomological Society 69: 185-194.

MacKenzie, K. E. and S. K. J avorek. 1997. The alfalfa leafcutting bee, Megachile
rotundata Fabr.: an alternative managed pollinator of lowbush blueberry. Acta
Horticulturae 446: 87-96.

MacKenzie, K. E. and M. L. Winston. 1984. Diversity and abundance of native bee
pollinators on berry crops and natural vegetation in the Lower Fraser Valley, British
Columbia. Canadian Entomologist 116: 965-974.

Maingay, H. M., R. L. Bugg, R. W. Carlson, and N. A. Davidson. 1991. Predatory and
parasitic wasps (Hymenoptera) feeding at flowers of sweet fennel (Foeniculum vulgare
Miller var. dulce Battandier & Trabut, Apiaceae) and Spearmint (Mentha spicata L.,
Lamiaceae) in Massachusetts. Biological Agriculture and Horticulture 7: 363-3 83.

Marshall, E. J. P., T. M. West, and D. Kleijn. 2006. Impacts of an agri-environment
field margin prescription on the flora and fauna of arable farmland in different
landscapes. Agriculture, Ecosyststems and Environment 113: 36-44.

Martin, E. C. and S. E. McGregor. 1973. Changing trends in insect pollination of
commercial crops Annual Review of Entomology 18: 207-226.

Marucci, P. E. and H. J. Moulter. 1977. Blueberry pollination in New Jersey. Acta
Horticulturae 95: 345-351.

Matheson, A. [ed.], 1994. Forage for Bees in an Agricultural Landscape. International
Bee Research Association, Cardiff, UK

Matheson, A., S.L. Buchmann, C. O’Toole, P. Westrich, and LB. Williams [eds]
1996. The Conservation of Bees. Academic Press, London.

McGregor, S. E. 1976. Insect Pollination of Cultivated Crop Plants. Agricultural
Research Service, USDA, Washington, DC.

McIntyre, N. E. and M. E. Hostetler. 2001. Effects of urban land use on pollinator
(Hymenoptera: Apoidea) communities in a desert metropolis. Basic and Applied Ecology
2: 209-218.

Meader, E. M. and G. M. Darrow. 1947. Highbush blueberry pollination experiments.
Proceedings of the American Society of Horticultural Science 49: 196-204.

159

Merrill, T. A. 1936. Pollination of highbush blueberry. Tech. Bull. Mich. (St. Coll.)
Agric. Exp. Stn. 151.

Michener, C. D. 2000. The Bees of the World. The Johns Hopkins University Press,
Baltimore.

Michener, C. D., R. J. McGinlcy, and B. N. Danforth. 1994. The Bee Genera of North
and Central America (Hymenoptera: Apoidea). Smithsonian Institution Press,
Washington, DC.

Mitchell, T. B. 1960. Bees of the Eastern United States. North Carolina Ag. Ex. St..

Morandin, L. A., T. M. Laverty, and P. G. Kevan. 2001a. Bumble bee (Hymenoptera:
Apidae) activity and pollination levels in commercial tomato greenhouses. Journal of
Economic Entomology 94: 462-467.

Morandin, L. A., T. M. Laverty, and P. G. Kevan. 2001b. Effect of bumble bee
(Hymenoptera: Apidae) pollination intensity on the quality of greenhouse tomatoes.
Journal of Economic Entomology 94: 172-179.

Morandin, L.A. and M.L. Winston. 2003. Effects of novel pesticides on bumble bee

(Hymenoptera: Apidae) colony health and foraging ability. Environmental Entomology,
33(3): 555-563.

Morrow, E. B. 1943. Some effects of cross-pollination versus self-pollination in the
cultivated blueberry. Proceedings of the American Society of Horticultural Science 42:
469-472.

Morse, R.A. and N.W. Calderone. 2000. The value of honey bees as pollinators of US.
crops in 2000. Bee Culture 128: 1-15.

N abhan, G. P. and S. L. Buchmann. 1997. Services provided by pollinators, pp. 133-
150 In G. C. Daily [ed.], Nature's Services: Societal Dependence on Natural Ecosystems.
Island Press, Washington, DC.

National Academy of Sciences. 2007. Status of Pollinators in North America. The
National Academies Press, Washington, DC.

Ockinger, E. and H. G. Smith. 2007. Semi-natural grasslands as population sources for
pollinating insects in agricultural landscapes. Journal of Applied Ecology 44: 50-59.

Osborne, J. L., S. J. Clark, R. J. Morris, 1. H. Williams, J. R. Riley, A. D. Smith, D.
R. Reynolds, and A. S. Edwards. 1999. A landscape-scale study of bumble bee foraging
range and constancy, using harmonic radar. Journal of Applied Ecology 36: 519-533.

Osborne, J. L., I. H. Williams, and S. A. Corbet. 1991. Bees, pollination and habitat
change in the European community. Bee World 72: 99-116.

160

Ostman, O., B. Ekborm, and J. Bengtsson. 2001. Landscape heterogeneity and farming
practice influence biological control. Basic and Applied Ecology 2: 365-371.

Patten, K. D., C. H. Shanks, and D. F. Mayer. 1993. Evaluation of herbaceous plants
for attractiveness to bumble bees for use near cranberry farms Journal of Apicultural
Research 32: 73-79.

Pelz-Stelinski, K. S., L. J. Gut, and R. Isaacs. 2006. Vertical position of traps

influences captures of eastern cherry fruit fly (Dipter: Tephritidae). Florida Entomologist
89: 80-82.

Perfecto, I. and J. Vandermeer. 2002. Quality of agroecological matrix in a tropical
montane landscape: ants in coffee plantations in southern Mexico. Conservation Biology
16: 174-1 82.

Peterson, RT. and M. McKenny. 1996. A Field Guide to Wildflowers: Northeastern
and Nortcentral North America. Peterson Field Guide Series, Houghton Mifflin
Company, New York, NY.

Pollard, K. A. and J. M. Holland. 2006. Arthropods within the woody element of
hedgerows and their distribution pattern. Agricultural and Forest Entomology 8: 203-211.

Pontin, D. R., M. R. Wade, P. Kehrli, and S. D. Wratten. 2006. Attractiveness of
single and multiple species flower patches to beneficial insects in agroecosystems.
Annals of Applied Biology 148: 39-47.

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

waell, R. F., E. A. Warman, C. Carvell, T. H. Sparks, L. V. Dicks, D. Bennett, A.
Wright, C. N. R. Critchley, and A. Sherwood. 2005. Providing foraging resources for
bumblebees in intensively farmed landscapes. Biological Conservation 121: 479-494.

Ricketts, T. H. 2004. Tropical forest fiagments enhance pollinator activity in nearby
coffee crops Conservation Biology 18: 1262-1271.

Ricketts, T. H., G. C. Daily, P. R. Ehrlich, and C. D. Michener. 2004. Economic value
of tropical forest to coffee production. PNAS 101: 12579-12582.

Riedl, H., E. Johansen, L. Brewer, and J. Barbour. 2006. How to Reduce Bee
Poisoning. Oregon State University, Pacific Northwest Extension Bulletin: PNW 591.

Roubik, D. W. 1996. Measuring the meaning of honey bees, pp. 163-172 In A.
Matheson, S. L. Buchmann, C. O'T 001e, P. Westrich, and]. H. Williams [eds], The
Conservation of Bees. Academic Press, London.

161

Roulston, T.H., S.A. Smith, and A.L. Brewster. 2007. A comparison of pan trap and
intensive net sampling techniques for documenting a bee (Hymenoptera: Apiformes)
fauna. Journal of the Kansas Entomological Society 80: 179-181.

Russell, K. N., H. Ikerd, and S. Droege. 2005. The potential conservation value of
unmowed powerline strips for native bees. Biological Conservation 124: 133-148.

Sampson, B. J. and J. H. Cane. 2000. Pollination efficiencies of three bee
(Hymenoptera: Apoidea) species visiting rabbiteye blueberry. Journal of Economic
Entomology 93: 1726-1731.

Sampson, B. J., S. J. Stringer, J. H. Cane, and J. M. Spiers. 2006. Screenhouse
evaluations of a mason bee Osmia ribifloris (Hymenoptera: Megachilidae) as a pollinator
for bluberries in the southeastern United States. Proceedings of the Ninth North
American Blueberry Research. pp. 381-392.

Schaub, I. O. and L. D. Baver. 1942. Blueberries earlier and larger with cross-
pollination. Rep. N. Carol. Agric. Exp. Stn. 65: 53.

Shepherd, M., S. L. Buchmann, M. Vaughan, and S. Hoffman Black. 2003. Pollinator
Conservation Handbook. The Xerces Society in association with The Bee Works,
Portland, OR. ~

Shuler, R. E., T. H. Roulston, and G. E. Farris. 2005. Farming practices influence wild

pollinator populations on squash and pumpkin. Journal of Economic Entomology 98:
790-795.

Southwick, E. E. and Jr. L. Southwick. 1992. Estimating the economic value of honey
bees (Hymenoptera: Apidae) as agricultural pollinators in the United States. Journal of
Economic Entomology 85: 621-633.

Stark, J. D. and J. E. Banks. 2003. Population-level effects of pesticides and other
toxicants on arthropods. Annual Review of Entomology 48: 505-519.

Steffan-Dewenter, I. 2002. Landscape context affects trap-nesting bees, wasps, and their
natural enemies. Ecological Entomology 27: 631-637.

Steffan-Dewenter, I., U. Munzenberg, C. Burger, C. Thies, and T. Tschamtke. 2002.
Scale-dependent effects of landscape context on three pollinator guilds. Ecology 83:
1421 -1432.

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

Stephen, W. P. and S. Rao. 2005. Unscented traps for non-Apis bees (Hymenoptera:
Apoidea). Journal of the Kansas Entomological Society 78: 373-3 80.

Stevenson, J. 2003. Hazards of pesticides to bees. Bee World 84: 51-52.

162

r...—

 

Stubbs, C. S. and F. A. Drummond. 1997a. Management of the alfalfa leafcutting bee,
Megachile rotundata (Hymenoptera: Megachilidae), for pollination of wild lowbush
blueberry. Journal of the Kansas Entomological Society 70: 81-93.

Stubbs, C. S. and F. A. Drummond. 1997b. Pollination of wild lowbush blueberry,

Vaccinium angustifolium by the alfalfa leafcutting bee, Megachile rotundata. Acta
Horticulturae 446: 1 89-196.

Stubbs, C. S. and F. A. Drummond. 2000. Pollination of lowbush blueberry by
Anthophora pilipes villosula and Bombus impatiens (Hymenoptera: Anthophoridae and
Apidae). Journal of the Kansas Entomological Society 72: 330-333.

Stubbs, C. S. and F. A. Drummond. 2001a. Bombus impatiens (Hymenoptera: Apidae):
an alternative to Apis mellifera (Hymenoptera: Apidae) for lowbush blueberry
pollination. Journal of Economic Entomology 94: 609-616.

Stubbs, CS. and EA. Drummond [eds], 2001b. Bees and crop pollination - crisis,
crossroads, conservation, pp. 1-156. Entomological Society of America, Lanham,
Maryland.

Suchail, S., D. Guez, and L.P. Belzunces. 2001. Discrepancy between acute and chronic
toxicity induced by imidacloprid and its metabolites in Apis mellifera. Environmental
Toxicology and Chemistry, 20(11): 2482—2486.

Svensson, B., J. Lagerliif, and B. G. Svensson. 2000. Habitat preferences of nest-
seeking bumble bees (Hymenoptera: Apidae) in an agricultural landscape. Agriculture,
Ecosystems and Environment 77: 247-255.

Thies, C., I. Steffan—Dewenter, and T. Tschamtke. 2002. Effects of landscape context
on herbivory and parasitism at different spatial scales. OIKOS 101: 18-25.

Thies, C. and T. Tschamtke. 1999. Landscape structure and biological control in
agroecosystems. Science 285: 893-895.

Thomson, J. D. 1981. Spatial and temporal components of resource assessment by
flower-feeding insects. Journal of Animal Ecology 50: 49-59.

Teixeira, L. A. and S. Polavarapu. 2001. Effect of sex, reproductive maturity stage and
trap placement, on attraction of the blueberry maggot fly (Dipter: Tephritidae) to sphere
and Pherocon AM traps. Florida Entomologist 84: 363-369.

Torchio, P. F. 1990. Diversification of pollination strategies for US. crops
Environmental Entomology 19: 1649-1656.

Torchio, P. F. 1994. The present status and future prospects of non-social bees as crop
pollinators. Bee World 75: 49-53.

163

‘h ‘

 

Tschamtke, T. and R. Brandl. 2004. Plant-insect interactions in fragmented landscapes.
Annual Review of Entomology 49: 405-430.

Tschamtke, 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.

Tylianakis, J. M., A. M. Klein, T. Lozada, and T. Tschamtke. 2006. Spatial scale of
observation affects alpha, beta and gamma diversity of cavity-nesting bees and wasps
across a tropical land-use gradient. Journal of Biogeography 33: 1295-1304.

USDA 2004. Michigan Fruit Inventory 2003-2004. USDA-NASS.

Vaknin, Y., S. Gan-Mor, A. Bechar, B. Ronen, and D. Eisikowitch. 2000. The role of
electrostatic forces in pollination. Plant Systematics and Evolution 222: 133-142.

Veddeler, D., A. M. Klein, and T. Tschamtke. 2006. Contrasting responses of bee
communities to coffee flowering at different spatial scales. OIKOS 112: 594-601.

Visscher, P. K., R. S. Vetter, and F. C. Baptista. 1997. Initial rapid invasion of
Afiicanized bees in the United States (1990—1995). California Agriculture 51: 25 [online]
http://bees.ucr.edu/ahb-spread.html

Voss, E. G. 1996. Michigan flora; a guide to the identification and occurrence of the
native and naturalized seed-plants of the State: Parts I:III. Cranbrook Institute of Science.
University of Michigan. University Herbarium, Bloomfield Hills, Michigan.

Walther-Hellwig, K. and R. Frank]. 2000. Foraging habitats and foraging distances of

bumblebees, Bombus spp. (Hym., Apidae), in an agricultural landscape. Journal of
Applied Entomology 124: 299-306.

Watanabe, M. E. 1994. Pollination worries rise as honey bees decline. Science 265:
1 170.

Westerkamp, C. 1991. Honeybees are poor pollinators - why? Plant Systematics and
Evolution 177: 71-75.

Westphal, C., I. Steffan-Dewenter, and T. Tschamtke. 2006. Foraging trip duration of
bumblebees in relation to landscape-wide resource availability. Ecological Entomology
31: 389-3 94.

Westphal, C., I. Stefl'an-Dewenter, and T. Tschamtke. 2003. Mass flowering crops
enhance pollinator densities at a landscape scale. Ecology Letters 6: 961-965.

Westrich, P. 1996. Habitat requirements of central European bees and the problems of
partial habitats, pp. 1-16 In A. Matheson, S. L. Buchmann, C. O'Toole, P. Westrich, and I.
H. Williams [eds], The Conservation of Bees. Academic Press, London.

164

Williams, N.M. and C. Kremen. 2007. Resource distributions among habitats determine
solitary bee offspring production in a mosaic landscape. Ecological Applications 17: 91 O-
921.

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

Willemstein, S. C. 1987. An Evolutionary Basis for Pollination Ecology. Leiden
University Press, Leiden, The Netherlands.

Wilson, P. and M. Stine. 1996. Floral constancy in bumble bees: handling efficiency or
perceptual conditioning? Oecologia 106(4): 493-499.

Winfree, R., T. L. Griswold, and C. Kremen. 2007. Effect of human disturbance on
bee communities in a forested ecosystem. Conservation Bioliogy 21: 213-223.

Winston, M. L. 1998. From Where I Sit: Essays on Bees, Beekeeping, and Science.
Cornell University Press, Ithaca.

Wood, G. W. 1965. Note on the activity of native pollinators in relation to the bloom
period of lowbush blueberry. Journal of Economic Entomology 58: 777.

165

APPENDIX A:

GPS COORDINATES FOR FIELD SITES

166

Table A. GPS coordinates for the 15 blueberry fields in southwest Michigan in which

bees were sampled between 2004-6.

 

 

Site Code N W
l GAL 42° 1 6.075 086° 1 3.997
2 DGJ 42°20.869 086°13.355
3 DGS 42°21.885 086°l 7.217
4 BOD 42°24.749 086°06.378
5 EAR 42°32.148 086° 12.896
6 FEL 42°34.9l 7 086°09.107
7 DOU 42°37.589 086°10.117
8 DJ S 42°41.677 086°08.920
9 DJP 42°43.631 086°06.638
10 BOW 42°49.122 086°10.236
1 l WAS 42°50.604 086°09.902
12 STA 42°52.051 086°07.578
13 CAR 42°52.905 086°09.369
14 TIL 42°57.208 086°06.568
15 RAN 42°59.413 086°09.451

 

167

APPENDIX B:

LIST OF VOUCHER SPECIMENS

168

Table B. List of voucher specimens. The ID column refers to the unique number assigned
to each specimen. Key to determinations: A = J .S. Ascher; L = J .C.W. Langdon; T = J .K

 

 

Tuell.
Date

ID Collected Family Genus species sex det
40506 5/ 16/2004 Andrenidae Andrena algida f A
40644 6/3/2004 Andrenidae Andrena alleghaniensis f A
52049 4/ 2 10005 Andrenidae Andrena arabis f A
50972 5/16/2005 Andrenidae Andrena barbilabris f A
52024 6/30/2005 Andrenidae Andrena bisalicis f A
40109 5/15/2004 Andrenidae Andrena carlini f A
40921 5/25/2004 Andrenidae Andrena carolina f A
41023 5/25/2004 Andrenidae Andrena ceanothi f A
53791 4/21/2005 Andrenidae Andrena clarkella f A
40181 5/19/2004 Andrenidae Andrena commoda f A
4031 1 5/19/2004 Andrenidae Andrena crataegi f A
41046 8/19/2004 Andrenidae Andrena cressonii f A
40707 5/26/2004 Andrenidae Andrena dunningi f A
50364 5/25/2005 Andrenidae Andrena erigeniae f A
50925 5/16/2005 Andrenidae Andrena erythrogaster f A
40389 5/16/2004 Andrenidae Andrena forbesii f A
50023 4/16/2005 Andrenidae Andrena frigida f A
61776 6/12/2006 Andrenidae Andrena geranii f A
41499 6/3/2004 Andrenidae Andrena hippotes f A
42461 8/19/2004 Andrenidae Andrena hirticincta m A
40988 5/25/2004 Andrenidae Andrena """a’rfx 0’ f A

morrzsonella

41236 5/25/2004 Andrenidae Andrena integra f A
40169 5/15/2004 Andrenidae Andrena mandibularis f A
40385 5/16/2004 Andrenidae Andrena mariae f A
41 1 18 5/26/2004 Andrenidae Andrena milwaukeensis f A
40150 5/15/2004 Andrenidae Andrena miserabilis f A
40098 5/25/2004 Andrenidae Andrena morrisonella f A
4015 l 5/ 1 5/2004 Andrenidae Andrena nasonii f A
40992 5/25/2004 Andrenidae Andrena neonana f A
41289 5/25/2004 Andrenidae Andrena nigrae f A
40104 6/2/2004 Andrenidae Andrena nivalis f A
41300 5/26/2004 Andrenidae Andrena nuda f A
40572 5/16/2004 Andrenidae Andrena perplexa f A
60617 5/17/2006 Andrenidae Andrena persimulata f A
41768 8/19/2004 Andrenidae Andrena placata f A
61800 6/12/2006 Andrenidae Andrena platypaea f A
4 l 698 6/3/2004 Andrenidae Andrena pruni f A
40300 5/19/2004 Andrenidae Andrena rehni f A
40995 5/25/2004 Andrenidae Andrena robertsonii f A
40575 5/16/2004 Andrenidae Andrena rugosa f A
40306 5/19/2004 Andrenidae Andrena salictaria f A
62286 4/24/2006 Andrenidae Andrena Sigmundi f A
40979 5/25/2004 Andrenidae Andrena spiraeana f A
60030 5/26/2006 Andrenidae Andrena thespii f A

169

 

Date

 

II) Collected Family Genus species sex det
52666 4/16/2005 Andrenidae Andrena tridans f A
4040 l 5/ l 6/2004 Andrenidae Andrena vicina f A
52662 4/16/2005 Andrenidae Andrena violae f A
52232 4/21/2005 Andrenidae Andrena wellesleyana f A
61845 6/12/2006 Andrenidae Andrena wilkella f A
41395 9/4/2004 Andrenidae Calliopsis andrenzformis f A
53902 9/15/2005 Andrenidae Perdita octomaculata f A
41795 8/19/2004 Andrenidae Pseudopanurgus nebrascensis f A
41527 6/3/2004 Apidae (Eucera) atriventris f A
61726 6/12/2006 Apidae (Eucera) hamata f A
61272 7/6/2006 Apidae Anthophora tenninalis f A
40240 5/15/2004 Apidae Apis mellifera f T
41620 6/2/2004 Apidae Bombus bimaculatus f A
41515 5/25/2004 Apidae Bombus citrinus f A
42529 8/19/2004 Apidae Bombus fervidus f A
40320 5/19/2004 Apidae Bombus griseocollis f A
40321 5/29/2004 Apidae Bombus impatiens f A
4077 1 6/2/2004 Apidae Bombus perplexus f A
42396 8/ 19/2004 Apidae Bombus vagans f A
40528 5/ 19/2004 Apidae Ceratina calcarata m T
40364 5/19/2004 Apidae Ceratina dupla m T
40843 6/3/2004 Apidae Ceratina strenua f T
53333 9/15/2005 Apidae Melissodes agilis f A
42405 7/15/2004 Apidae Melissodes apicata m A
41156 7/29/2004 Apidae Melissodes bimaculata m A
41388 9/4/2004 Apidae Melissodes communis A
42420 8/ l 6/2004 Apidae Melissodes desponsa f T
53 897 9/15/2005 Apidae Melissodes druriella f A
42407 7/23/2004 Apidae Melissodes tridonis f A
53143 5/10/2005 Apidae Nomada cf armatella m A
53 104 5/ 2 1/2005 Apidae Nomada cressonii m A
53589 5/10/2005 Apidae Nomada denticulata f A
52838 4/16/2005 Apidae Nomada luteoloides f A
53586 5/10/2005 Apidae Nomada maculata f A
52833 4/16/2005 Apidae Nomada obliterata f A
53 103 5/21/2005 Apidae Nomada ovata m A
53573 4/21/2005 Apidae Nomada pygmaea m A
42398 8/22/2004 Apidae T riepeolus lunatus f T
42425 5/ 2 5/2004 Apidae X ylocopa virginica virgin ica m A
41403 5/25/2004 Colletidae Colletes thoracicus m L
41180 5/26/2004 Colletidae Hylaeus aflinis f L
50639 5/21/2005 Colletidae Hylaeus rudbeckiae f L
41790 8/19/2004 Halictidae Agapostemon sericeus m L
4088 1 6/3/2004 Halictidae Agapostemon splendens f L
40185 5/19/2004 Halictidae Agapostemon texanus f L
40993 5/25/2004 Halictidae Agapostemon virescens f L
42267 8/22/2004 Halictidae Augochlora pura f L
41220 5/ 2 5/2004 Halictidae Augochlorella aurata f L

170

 

Date

 

II) Collected Family Genus species sex det
42303 8/19/2004 Halictidae Augochlorella gratiosa f L
40239 5/19/2004 Halictidae Dufourea marginata f L
40238 5/19/2004 Halictidae Halictus confusus f A
42100 8/22/2004 Halictidae Halictus ligatus m L
40538 5/16/2004 Halictidae Halictus parallelus f A
42075 8/19/2004 Halictidae Halictus rubicundus m L
40923 5/25/2004 Halictidae Lasioglossum acuminatum f A
41309 9/4/2004 Halictidae Lasioglossum admirandum f L
41720 6/3/2004 Halictidae Lasioglossum anomalum f L
40726 5/16/2004 Halictidae Lasioglossum boreale f L
40355 5/19/2004 Halictidae Lasioglossum bruneri f L
41788 8/19/2004 Halictidae Lasioglossum coeruleum f L
4023 1 5/ 15/2004 Halictidae Lasioglossum coriaceum f L
40815 6/3/2004 Halictidae Lasioglossum cressonii f A
50245 5/25/2005 Halictidae Lasioglossum fattigi f L
40325 5/29/2004 Halictidae Lasioglossum fuscipenne f A
61399 7/6/2006 Halictidae Lasioglossum illinoense f A
41967 7/15/2004 Halictidae Lasioglossum imitatum f A
40326 5/16/2004 Halictidae Lasioglossum leucozonium f A
41657 5/25/2004 Halictidae Lasioglossum nelumonis f A
41286 6/3/2004 Halictidae Lasioglossum nigroviride f A
42527 8/19/2004 Halictidae Lasioglossum nymphaearum f L
52304 4/21/2005 Halictidae Lasioglossum nymphale f L
41637 5/25/2004 Halictidae Lasioglossum ‘ oblongum f A
40489 5/ 19/2004 Halictidae Lasioglossum pectorale f A
40825 6/3/2004 Halictidae Lasioglossum pilosum f A
40819 6/3/2004 Halictidae Lasioglossum quebecense f A
4071 1 6/3/2004 Halictidae Lasioglossum rohweri f L
51784 6/29/2005 Halictidae Lasioglossum suvianae m L
408 13 6/3/2004 Halictidae Lasioglossum tegulare f A
60621 5/17/2006 Halictidae Lasioglossum versans f L
42465 7/15/2004 Halictidae Lasioglossum vierecki f A
42322 7/15/2004 Halictidae Sphecodes confertus f A
40220 5/19/2004 Halictidae Sphecodes dichrous f A
40738 5/16/2004 Halictidae Sphecodes ranunculi f A
54122 6/9/2005 Megachilidae Anthidium manicatum f T
52104 5/21/2005 Megachilidae Ashmeadiella sp. m L
5 1465 6/29/2005 Megachilidae Dianthidium simile f T
61626 6/12/2006 Megachilidae Heriades leavitti m L
42495 8/ 16/2004 Megachilidae Heriades variolosus f A
5 148 1 5/2 1/2005 Megachilidae Hoplitis producta f A
51482 5/21/2005 Megachilidae Hoplitis spoliata m A
50414 5/25/2005 Megachilidae Megachile 0(anthosarus) sp. m A
53895 9/15/2005 Megachilidae Megachile albatarsis f T
41406 7/29/2004 Megachilidae Megachile campanulae f T
5 1466 6/29/2005 Megachilidae Megachile centuncularis f A
53 156 9/ 15/2005 Megachilidae Megachile mendica f T
415 l 1 7/29/2004 Megachilidae Megachile montivaga f A

171

 

Date

 

II) Collected FamilL Genus species sex det
41416 9/4/2004 Megachilidae Megachile mucida f T
42502 8/16/2004 Megachilidae Megachile pugnata f A
5 147 1 6/29/2005 Megachilidae Megachile rotundata f A
50062 4/16/2005 Megachilidae Osmia atriventris m L
40096 5/25/2004 Megachilidae Osmia bucephala f T
52 102 5/2 1/2005 Megachilidae Osmia conjuncta f L
61472 6/12/2006 Megachilidae Osmia distincta f A
61654 6/12/2006 Megachilidae Osmia felti f A
61350 7/6/2006 Megachilidae Osmia georgica f A
51490 4/15/2005 Megachilidae Osmia lignaria m A
51484 4/15/2005 Megachilidae Osmia mfcf’gm’fm 0" f L

rllrnoensrs
50678 5/21/2005 Megachilidae Osmia pumila f A
62608 4/26/2006 Megachilidae Osmia simillima m A
50011 4/16/2005 Megachilidae Osmia subfasciata? m L
60994 4/24/2006 Megachilidae Osmia virga m A

 

172

APPENDIX C:

A RECORD OF BEE SPECIES ASSOCIATED WITH VACCINI UM SPP. IN
MICHIGAN

173

\. 32.8.3 323222V
32M~2e232 3232232V
3.22.35 32935..
2323223... 323.2232V
.232 323222V
3232M23m32 3232232V
33MB 323.2232“,N
9.233%? 32m222V
223.222 323.2232V
M23922 323222V
32323222322 323232V
33323.22 323222V
. 323232.32 323222V
@3233333322 32933..
3322823322 32825..
3.353 323232V
52332.2 32322322.
EBRRE 323222‘
MESS 323232V
mmwoc23K323222V
$332323 323.2232V
3223233 323.322.
2.2233 22:23.3 323222V
33:33.5 323.2232“.‘

«33:96:
3.2K - a as: - we: $2 - m 83 - 2 sea...
2:53..—

.Geoom - b assuage e5 5 €883 amaze;

5 coma 9262—8 98 30333030 88m .2 Eco»: v.2 BE. .GeeTmQEv tonomm new 258Mer ,3 totem»: #8.» Bo Z 383: 5

258:3 5.5.33 a seam $5888 22% a see a 288 25 a: .8875 2:: e5 8836 =22§ 6% 8:38 use

3.808 02: “man 2:. 95 25.22.33.333 2o $3.808 203332, Econ 20 .258: 5:8 6:20 8w 805 £033 com 860% com .0 Bach.

\\\
\\ \\
\ \\\\\
\ \\\

\
\

\\\\\

\e
\

\\ \
\\\\\\\\\\

174

32333322 “3232230

 

\ \
2:53:30
\. \ \ \ g 5 33.22.2923 33.223323» 3322.23.22 32.3332wa
\. 333223232 3.8.2.2325 333E3223$~
\. \ 3323.23 32.23230
\ \ \. 33323 32.33230
\. 333233233 32.23230
\. 333~3223sxe32223m
\. 2233233 3223%
\. \ \. 323%?» 322.3%
\. \ \ 3233.22.23» 3228M
\ \. 332232232 3223M
\ \ 332.32.233.32 3223M
\ \' \v \; ngENQ gQEQm
\ \ 333223233223Q 3223M
\. \ \ 3322332233 3232323332 3223M
\. \ \ \ 23.33%22 3228M
\ \ \ 3.383332% 332223m
\ \. 33223332\332223m
2. 2. 2. 33223x§e§m
\ \. 3.233232 3223M
\. \ \ \ 33232333223 332223m
2. 2. 3.23% 3232322
2. 2. 2. 2. 3233232 232
\ 32.3.23 32.3.23 32323322232
823..
\. \ 3:33.33 323223222
.283 333295.32
TEEN . h. 03— . H332 are" - I 83 - 2 «333%

3:83...—

175

3323M2.3 33333.33.

3~32~3M 3~32~3M 333333:

3.33.53333Q 3.3323 33233M33<

3323333 33333M33<

3333233 3333233 333-3E
oacaguawoz

«33 3333333333.

22333332333€ E3333NM3333Q

2233323223333€ 223333~M3333N

«3.3323333333H» 3323333333 «33333393 223333333333

.333\ x3333~§m3 223333NM3333N

33.3333 «3333.33.3Q» E3333~M3333N

3333323> «3333.33.3Q» 223333~M3333N

3323333223 «3333.33.3Q» 223333~M3333N

«33333QH» 22333333 «3333.33.3Q3 E3333~M3N33N

«33333322?» 22333323 33333.33RQ 223333~M3333q

32333323 «3333.33.3Q» E3333~M3m33q

333233.532 3333.33E

. 3333\233 333333~.~

\ 333332233 3.33m323333mv33.

33.533322 3.33m323333M3V

333.33% 33323333M3V

«333.23?» 333233 33323333M3V

323m 32333 323333M3V

3333.233 23323333Q3MV
33353:“!

\ \. \ 333.33; 333330

\. 333.333.2333 3333330
.233 2:53:30

0.25" - h. wag . Haw—Z aha .. m :33 u 2 333on
E83..—

\\\\\
\\\\

\

\\\\
\\\\\\\\\\\\

\\\
\\
\\\
\\\

176

 

cm

:3

N3

"333% 33,—.

 

\\

xxxxxxx \;

\\\\

\\

\

3332 3.332333.
33.5532
33333323233 3.32236
3322333 3.3230
323332322 3.23%
3.322323 3.3230
3333323 3.23%
332233332 3.22236
33222332233 3.32230
.2333 3332232
3232\Nw32 332233333320» 33233.3:
332M233m. «322333333203 33233M33<
3.332323% 3.3.38 «32233333323» 332332333
33.232322 33.232322 33233M33<
.283 332—33332

 

euvocu . h.

333 - H322

aha . E

:33 - E

333%
3233

 

177