V . 2:"? ,. i; ' ‘ h“ :27:
m '- :34 r - iii! ma— 21%;:
...' £5

3 {I t #3 .
. ‘ , 2 ~ ‘ balm
5?. “if- a. . ' _ . . e“! ' ’ ”aim
H23 5 :t’ .. .9 . . g“ 5223.; '
:2

i
3 3%?!{2539 0::

 

[mus

 

1001}

 

 

This is to certify that the
thesis entitled

HOST RANGE AND PREFERENCE OF THE EMERALD ASH
BORER, AGRILUS PLAN/PENNIS FAIRMAIRE
(COLEOPTERAIBUPRESTIDAE), IN NORTH AMERICA

presented by

ANDREA CHRISTINE ANULEWICZ

has been accepted towards fulﬁllment
of the requirements for the

MS. degree in FORESTRY

 

 

l ALVA 1.440 44/.

Major“? r-ssor’ 5 Si neatur

13 JULY 2006

 

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

.--.-.-.-.—.- -.-.-._ -.- -.-.: -_._--.... .d

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
JANO 7 2008“” 3 8 200?.
I21 .1 .0 .o 8

 

. EW&§°°8

 

 

19f§0€360@3

 

04222 my

 

NOV 2 3 2009

 

1124(Iy

 

§§1m9ﬂm0

 

MAR 2 8 2014

 

 

 

 

 

2/05 pleIRCIDaleDuojndd-pJ

HOST RANGE AND PREFERENCE OF THE EMERALD ASH BORER, AGRIL US
PLANIPENNIS FAIRMAIRE (COLEOPTERA: BUPRESTIDAE),
IN NORTH AMERICA
By

Andrea Christine Anulewicz

A THESIS

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

MASTER OF SCIENCE
Department of Forestry

2006

ABSTRACT
HOST RANGE AND PREFERENCE OF THE EMERALD ASH BORER, AGRIL US
PLANIPENNIS FAIRMAIRE (COLEOPTERA: BUPRESTIDAE),
IN NORTH AMERICA
By .
Andrea Christine Anulewicz

Oviposition and early instar development of A grilus plam'pennis Fairmaire
(Coleoptera: Buprestidae) was evaluated for four North American ash and six potential
alternate host species from 2003 to 2005. In no-choice tests, female A. planipennis
oviposited and larvae developed to at least the second instar on green ash, white ash,
black ash, blue ash, and privet. Unsuccessful larval feeding attempts were found on
black walnut, Japanese tree lilac, America elm, and hackberry, but not on hickory. With
the possible exception of privet, the non-ash species appear to be unsuitable for larval
development.

In multiple-choice tests with cut logs and live trees in the ﬁeld, female A.
planipennis consistently laid more eggs on ash than non-ash species. Larvae completed
development on all ash species, although green and white ash logs may be more preferred
than black or blue ash logs. Unsuccessful larval feeding attempts were found on black
walnut and Japanese tree lilac.

Field surveys of A. planipennis density and canopy dieback were conducted to
assess host preference among three North American ash species. Canopy dieback and
density of A. planipennis were signiﬁcantly greater on green ash street trees than white
ash street trees and increased in both species over time. A. planipennis density was also

greater in white ash woodlot trees than blue ash woodlot trees.

DEDICATION

I would like to dedicate this thesis to my mother, Frances Elizabeth Agius, who
passed away before I began graduate school. She was a remarkable woman who showed
me how to appreciate the beauty of nature and taught me to ﬁght for things that couldn’t

ﬁght for themselves. She has always been and always will be a guiding force in my life.

iii

ACKNOWLEDGEMENTS

I thank, ﬁrst and foremost, my advisor, Deborah G. McCullough. I couldn’t have
chosen a better educator, professional, or mentor to help guide me through this adventure.
Deb mixed the perfect blend of guidance and direction, without pressure and
overcrowding. At the same time, I always felt a sense of independence, without feeling
abandoned. I would also like to thank Therese Poland and Bert Cregg for serving on my
committee. They provided thoughtful insight on projects and always encouraged me to
“keep at it.”

I could not have successfully completed this program without the help of many,
many people. David Cappaert was indispensable when it came to executing ﬁeld studies
and was always ready to pick my brain about everything from experimental design to life
in general. I can’t imagine how impossible projects would have seemed without the loyal
help of Sarah Smith, Chris Pell, Bob McDonald, and Elizabeth Grisham.

I thank Mike Hommell and all the staff at Matthaei Botanical Gardens where I
conducted lab work in 2003, 2004, and 2005. They were always willing to ﬁnd just a
little more space for us to work, help us plant trees, and move mulch. Chris Pargoff, City
of Livonia, Paul Bairley, City of Ann Arbor, Paul Muelle, Kensington Metropark, John
Ringholz, Western Golf Course, and Scott Miret, Asplundh Tree Care Company, were
immensely helpful in providing ﬁeld sites in southeast Michigan. Debbie Miller was
always obliging when it came to answering questions about beetle rearing and
oviposition.

I would like to thank Greg Kowalewski and the other staff at Kellogg Forest for

the repeated harvesting of ash material for my studies. I thank Mike Klein, Nathan

iv

Schiff, and Jana Vanderhaar for donating privet material and Fredric Miller for providing
Asian elm branches. Poplar Farms Nursery was kind enough to discount 80 nursery trees
for use in my studies.

I must thank my fellow graduate students, especially Anna F iedler and Jeff Evans,
who were there for me from the beginning to keep my sanity in check. As lab mates,
Nate Siegert, Amy Kearney, and Andy Klein, were always a valued source of advice and
inspiration. I could not have done this without the love, support, and patience of my
husband, Trevor. He was always there to help celebrate the highs and commiserate in the
lows. I must also thank all of my other family and friends for their constant
encouragement over the last three years.

This research is funded by the USDA Forest Service, Northeastern Area, Forest
Health Protection, USDA Agricultural Research Service, and Michigan State

University’s Project GREEEN.

PREFACE

Emerald ash borer was discovered attacking ash trees in southeast Michigan in
June of 2002. Reared beetles were immediately identiﬁed as a member of the genus
A grilus, but experienced taxonomists could not determine the species. Specimens were
sent to specialists around the United States and Europe and, within two weeks of its
initial discovery, it was identiﬁed as A. planipennis Fairmaire by an Agrilus specialist in
Bratislava, Slovakia.

Initially, literature available on this species was limited to taxonomic descriptions
and several paragraphs published in Chinese literature. Research efforts by scientists
across the state of Michigan immediately focused on increasing our understanding of the
beetle’s life history traits and ecology, developing tools for detection and monitoring, and
minimizing the impact of this pest.

This thesis focused on evaluating the host range and preference of A. planipennis.
A grilus plam'pennis in Michigan had only been observed attacking ash trees (Fraxinus
sp.), but previous literature reported other hosts in several genera including Fraxinus,
Ulmus, Juglans, and Pterocarya. In Chapter One, no-choice bioassays were conducted to
evaluate A. planipennis oviposition and early instar development on cut branches of
green, white, black, and blue ash, American elm, black walnut, privet, Japanese tree lilac,
hickory, and hackberry and on live specimens of green and white ash, black walnut, and
Japanese tree lilac.

Studies reported in Chapter Two were designed to evaluate A. planipennis
oviposition preference and larval development on the same ten tree species, using logs

attached to t-posts and infested ash trees in multiple-choice ﬁeld experiments. Host

vi

preference was also evaluated using nursery trees planted in an area heavily infested with
A. planipennis. Chapter Three reports on ﬁeld surveys of A. planipennis density and
canopy dieback at multiple sites with a mix of ash species. My goal was to obtain a
better understanding of A. planipennis and some of the factors that may be inﬂuencing its
host range and preference in North America. Each chapter in this thesis is intended to be
prepared as a manuscript for publication.

Appendix 2 contains the methods and results of another study designed to
evaluate A. planipennis adult leaf-feeding preference among four North American species
of ash using two-choice bioassays. I felt that this study didn’t quite ﬁt in with the
contents of the other three chapters and lacked substantial, deﬁnitive, or conclusive
results, but still warranted reporting in some format. In 2005, we had difﬁculty with
beetle rearing and experienced substantial mortality and reduce fecundity. Laboratory
bioassays, therefore, were unsuccessful and Appendix 3 includes the methods and limited

results from no-choice and two-choice oviposition bioassays conducted that year.

vii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... x
LIST OF FIGURES ................................................................................. xii
CHAPTER 1

Oviposition and Early Instar Development of the Emerald Ash Borer
(Agrilus planipenm's Fairmaire) (Coleoptera: Buprestidae) in No-Choice

Laboratory Bioassays ................................................................................. 1
Introduction ................................................................................... 1
Materials and Methods ....................................................................... 4

Cut Branch Bioassays .............................................................. 4
Caged Stem Bioassays ............................................................. 8
Statistical Analysis ................................................................ 10
Results ....................................................................................... 12
Cut Branch Bioassays ............................................................ 12
Caged Stem Bioassay ............................................................. 15
Discussion ................................................................................... 16
Tables ........................................................................................ 21
Figures ....................................................................................... 24
CHAPTER 2

Oviposition Preference and Larval Development of the Emerald Ash Borer
(Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae) in Multiple-

Choice Field Experiments .......................................................................... 28
Introduction ................................................................................. 28
Materials and Methods ..................................................................... 31

Logs Attached to T-Posts ......................................................... 31
Logs Attached to Ash Trees ...................................................... 33
Nursery Trees ...................................................................... 37
American Elm Survey ............................................................ 38
Statistical Analysis ................................................................ 39
Results ....................................................................................... 41
Logs Attached to T-posts ......................................................... 41
Logs Attached to Ash Trees ...................................................... 43
Nursery Trees ...................................................................... 45
American Elm Survey ............................................................ 47
Discussion ................................................................................... 48
Tables ......................................................................................... 55
Figures ....................................................................................... 60
CHAPTER 3

Field Surveys of Emerald Ash Borer (Agrilus planipennis Fairmaire)
(Coleoptera: Buprestidae) Density and Canopy Dieback in Three

viii

North American Ash Species ...................................................................... 61

Introduction ................................................................................. 61
Materials and Methods ..................................................................... 64
Study Sites ......................................................................... 64
Sampling and Data Collection ..................................................... 65
Statistical Analysis .................................................................. 68
Results ................................................... ‘ .................................... 70
Green versus White Ash Street Trees .......................................... 70
White versus Blue Ash Woodlot Trees ......................................... 73
Discussion ................................................................................... 75
Tables ........................................................................................ 82
Figures ....................................................................................... 83
APPENDIX 1
Record of Deposition of Voucher Specimens ................................................... 89
APPENDIX 2
Leaf-Feeding Preference of A grilus planipennis (F airmaire) (Coleoptera:
Buprestidae) on Four North American Ash Species ............................................ 91
2004 Two-Choice Leaf-Feeding Bioassays ............................................. 91
2005 Two-Choice Leaf-Feeding Bioassays ............................................. 94
Figures ....................................................................................... 97
APPENDIX 3
Unsuccessful No-Choice and Two-Choice Laboratory Bioassays .......................... 101
Two-Choice Oviposition Bioassays .................................................... 101
No-Choice Cut Branch Bioassays ...................................................... 105
No-Choice Caged Stem Bioassays ...................................................... 107
Tables ....................................................................................... 109
LITERATURE CITED ........................................................................... 111

ix

LIST OF TABLES

CHAPTER 1

Table 1.1. Mean (:tSE) diameter, length, and surface area for branch
sections used in no-choice bioassays in 2003 and 2004 .............................. 21

Table 1.2. Total number of Agrilus planipennis eggs and galleries,
number of eggs and galleries on branch sections, and mean
(:tSE) number of eggs/beetle/day in no-choice bioassays in
2003 and 2004. Within columns, means followed by the same
letter are not signiﬁcantly different (Kruskal-Wallis test and
nonparametric multiple comparison procedure; p<0.05) .............................. 22

Table 1.3. Mean (iSE) tree diameter at breast height (DBH), height,
and number of A grilus planipennis eggs/m2 and galleries/m2
for four species of trees used in no-choice caged stem
bioassays in 2004. Within rows, means followed by the same
letter are not signiﬁcantly different (Kruskal-Wallis test and
nonparametric multiple comparison procedure; p<0.05) .............................. 23

CHAPTER 2

Table 2.1. Mean (iSE) diameter, length, and surface area for logs of six
species attached to t-posts in A grilus planipennis-infested areas in
2003 ................................................................................................ 55

Table 2.2. Mean (iSE) diameter, length, and surface area for logs of ﬁve
species attached to A grilus planipennis-infested ash trees in 2003,
2004, and 2005. 'Incomplete block design .............................................. 56

Table 2.3. Mean (iSE) tree diameter at breast height (DBH), surface area
sampled, and total and mean (iSE) number of Agrilus planipennis
galleries per m2 for four species of trees used in multiple-choice
nursery tree studies in 2004 and 2005. Within columns, means
followed by the same letter are not signiﬁcantly different (Kruskal-
Wallis test and nonparametric multiple comparisons procedure;
p<0.05) ....................................................................................... 57

Table 2.4. Total and mean (iSE) number of A grilus planipennis adults,
eggs per m2, and galleries per m2 for logs and drain pipes attached
to t-posts at four ﬁeld sites in Agrilus planipennis-infested areas in
2003. Within rows, means followed by the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric
multiple comparisons procedure; p<0.05) ............................................... 58

Table 2.5. Mean (iSE) number of Agrilus planipermis eggs per m2 and
galleries per m2 for logs attached to Agrilus planipennis-infested ash
trees in 2003, 2004, and 2005. Within rows, means followed by the
same letter are not signiﬁcantly different (Kruskal-Wallis test and
nonparametric multiple comparisons procedure; p<0.05). ........................... 59

CHAPTER 3

Table 3.1. Number of trees (n) and mean (:t SE) diameter at breast
height (DBH) for green, white, and blue ash trees at survey sites
in Agrilus planipennis-infested areas. 1The same trees were
re-surveyed each year. 2Trees were felled for sampling and new
trees were selected each year ............................................................. 82

APPENDIX 3

Table A3.1. Mean (iSE) diameter, length, and surface area for each species
used in the 2004 and 2005 two-choice oviposition bioassay. n=15 for
each species ................................................................................ 109

Table A3.2. Mean (:tSE) diameter, length, and surface area for each species

used in the 2005 no-choice cut branch bioassays. n=8 for each
species ...................................................................................... 110

xi

LIST OF FIGURES

CHAPTER 1

Figure 1.1. Mean (:tSE) number of Agrilus planipennis eggs per
100 cm2 on three known host species (ash) and ﬁve potential
alternate host species in the 2003 no-choice cut branch bioassay.
n=4 for black ash and walnut, n=5 for privet, elm, hickory and
hackberry, n=7 for green and white ash ................................................. 24

Figure 1.2. Mean (iSE) number of Agrilus planipenm’s galleries per
100 cm2 on three known host species (ash) and ﬁve potential
alternate host species in the 2003 no-choice cut branch bioassay.
=4 for black ash and walnut, n=5 for privet, elm, hickory and
hackberry, n=7 for green ash and white ash. Means with the
same letter are not signiﬁcantly different (Kruskal-Wallis test
and nonparametric multiple comparison procedure; p<0.05) ......................... 25

Figure 1.3. Mean (:tSE) number of Agrilus planipennis eggs per 100 cm2
on four known host species (ash) and six potential alternate host
species in the 2004 no-choice cut branch bioassay. n=8 branches
per species. Means with the same letter are not signiﬁcantly
different (Kruskal-Wallis test and nonparametric multiple
comparison procedure; p<0.05) ........................................................... 26

Figure 1.4. Mean (iSE) number of Agrilus planipennis galleries per
100 cm2 on four known host species (ash) and six potential
alternate host species in the 2004 no-choice cut branch bioassay.
n=8 branches per species. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric
multiple comparison procedure; p<0.05) ................................................ 27

CHAPTER 2

Figure 2.1. Mean (i SE) number of A grilus planipennis galleries per 1112
on logs of ﬁve species attached to Agrilus planipennis-infested
a) green ash street trees, b) green ash woodlot trees, c) white ash
street trees and d) white ash woodlot trees in 2005. Within site
and tree types, means with the same letter are not signiﬁcantly
different (Fisher’s protected LSD test; p<0.05) ......................................... 60

CHAPTER 3

Figure 3.1. a) Mean (:h SE) percent canOpy dieback and b) mean (t SE)
number of Agrilus planipennis exits holes (exits) and woodpecker

xii

attacks (WPs) per m2 for the small green (GR) and white (WH) ash

trees at Butzel in 2003, 2004, and 2005. Three dead green ash and

three dead white ash trees with 100% dieback were removed in 2004

and 2005. Means within years were not signiﬁcantly different

between species (Wilcoxon Rank Sum test; p<0.05) .................................. 83

Figure 3.2. a) Mean (:1: SE) percent canopy dieback and b) mean (E SE)
number of Agrilus planipennis exits holes (exits) and woodpecker
attacks per m2 (WPs) for green (GR) and white (WH) ash trees at
Lamed in 2003, 2004, and 2005. Four dead green ash trees with
100% dieback were removed in 2004. * Indicates signiﬁcant
differences between green and white ash trees (Wilcoxon Rank
Sum test; p<0.05) ........................................................................... 84

Figure 3.3. a) Mean (:t SE) percent canopy dieback and b) mean (:I: SE)
number of Agrilus planipennis exits holes (exits) and woodpecker
attacks (WPs) per m2 for green (GR) and white (WH) ash trees at
Windemere in 2004 and 2005. Two dead green ash trees with 100%
dieback were removed in 2005. One outlier for exits and woodpecks
per m2 was removed from the 2005 white ash tree data set. * Indicates
signiﬁcant differences between green and white ash trees
(Wilcoxon Rank Sum test; p<0.05) ...................................................... 85

Figure 3.4. Percent canopy dieback and number of A grilus planipennis
exit holes (exits) and woodpecker attacks (WPs) per m2 for 25
green ash and 24 white ash trees (49 trees total) at three sites
in the A grilus planipennis-infested areas ................................................ 86

Figure 3.5. Mean (:2 SE) number of A grilus planipennis exit holes (exits)
and woodpecker attacks (WPs) per m2 for white and blue ash trees
at Superior Twp in 2004 and 2005. * Indicates signiﬁcant differences
between white and blue ash trees (Wilcoxon Rank Sum test; p<0.05) .............. 87

APPENDIX 2

Figure A2. 1. Mean (:t SE) leaf area consumed for four ash species by
Agrilus planipennis beetles in ﬁve days during a two-choice
bioassay in 2004. n=36 replicates of six beetles feeding on one
leaf for each species. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric
multiple comparison procedure; p<0.05) ................................................. 97

Figure A2.2. Mean (i SE) proportion of leaﬂets of four species of ash fed
on by Agrilus planipennis beetles in each of three feeding
categories in ﬁve days during a two-choice bioassay in 2004.
n=36 replicates of six beetles feeding on one leaf for each species.

xiii

Within feeding types, means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric
multiple comparison procedure; p<0.05) ............................................... 98

Figure A2.3. Mean (i SE) leaf area consumed for four ash species by
A grilus planipennis beetles in ﬁve days during a two-choice
bioassay in 2005. n=29 replicates of six beetles feeding on one
leaf for green ash, n=4] for white ash, n=40 for black ash, n=40
for blue ash. Means with the same letter are not signiﬁcantly
different (Kruskal-Wallis test and nonparametric multiple
comparison procedure; p<0.05) ........................................................... 99

Figure A2.4. Mean (i SE) proportion of leaﬂets of four species of ash fed
on by A grilus planipennis beetles in each of three feeding
categories in ﬁve days during a two-choice bioassay in 2005.
n=45 replicates of six beetles feeding on one leaf for each species.
Within feeding types, means with the same letter are not
signiﬁcantly different (Kruskal—Wallis test and nonparametric
multiple comparison procedure; p<0.05) ............................................... 100

xiv

CHAPTER 1

Oviposition and Early Instar Development of the Emerald Ash Borer
(Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae) in No-Choice

Laboratory Bioassays

INTRODUCTION

Emerald ash borer (Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae), an
Asian phloem borer discovered in June 2002, is established in much of southeastern
Michigan and Windsor, Ontario (Cappaert et al. 2005c). More than 35 distinct, localized
infestations have been found since 2003 in other areas of Michigan, Ohio, and Indiana.
Agrilus planipennis is native to northeastern China, Korea, Japan, Mongolia, the Russian
Far East, and Taiwan (Yu 1992). A grilus planipennis was recently synonymized with A.
marcopoli Obenberger (in China), A. feretrius Obenberger (in Taiwan), and A. marcopoli
ulmi Kurosawa (in Korea and Japan) (Jendek 1994). This phloem-feeding species was
likely introduced into North America in solid wood packing material (Cappaert et al.
2005c), such as wooden containers, pallets, or dunnage made from ash wood containing
bark strips.

The life cycle of A. planipenm's in North America is generally completed in one
year, although some individuals may require two years to complete development (Siegert
et al. 2005). In southeast Michigan, A. planipennis adult emergence begins as early as

mid- to late May or early June (Brown-Rytlewski and Wilson 2005, Cappaert et al.

2005c). Adult beetles live for 3-6 weeks (Cappaert et al. 2005c). Beetles feed on ash

(F raxinus sp. [Oleaceae]), foliage for 5-7 days before mating and females feed for
another 5-7 days before oviposition begins (Bauer et al. 2004, Lyons et al. 2004).
Females generally lay 30-60 eggs (Chinese Academy of Science 1986, Yu 1992),
although egg viability declines as females age (D.L. Miller, pers. comm). Eggs are laid
individually on or just under the bark surface and inside bark cracks and crevices of live
ash trees, from late June to late July (Bauer et al. 2004). Eggs are initially creamy white
and gradually change to reddish brown after a few days (Bauer et al. 2004). Eggs hatch
in two to three weeks. Larvae feed under the bark in the carnbium and phloem from July
to October, in distinctive galleries that eventually score the outer sapwood. Larval
galleries disrupt vascular tissue, effectively girdling and killing branches or the entire
tree. Larvae pass through four instars (Cappaert et al. 2005c) and most overwinter as pre-
pupal larvae in the outer cm of the sapwood or in the thick bark of large trees (Cappaert
et al. 2005a).

Although the limited reports available suggest that A. planipenm's is not a major
pest in Asia (Akiyama and Ohmomo 2000), in North America, A. planipennis is
aggressively attacking healthy, as well as stressed ash. In Michigan, an estimated 12-15
million green ash (Fraxinus pennsylvanica Marsh), white ash (F. americana L.), black
ash (F. nigra Marsh), and blue ash (F. quadrangulata Michx.) have been killed as of
March 2006. In China, host range of A. planipennis reportedly includes Fraxinus sp.,
including native species of F. chinensis Roxb., F. mandshurica Rupr., and F.
rhynchophylla Hance (Chinese Academy of Science 1986, Yu 1992). In Japan, A.

plam'pennis has been recorded from F. mandshurica var. japonica Maxim, Juglans

mandshurica var. sieboldiana (Maxim) CK Schneid (Juglandaceae), J. mandshurica var.
sachalinensis (Miyabe et Kudo) Kitamura, Pterocarya rhoifolia Sied. et Zucc.
(Juglandaceae), and Ulmus davidiana var. japonica (Rehd) Nakai (Ulmaceae) (Akiyama
and Ohmomo 1997, Sugiura 1999). North American members of these genera or
families, including American elm (U. americana L), black walnut (J. nigra L), and
hickory (Carya sp.) (J uglandaceae), are common in landscapes and wooded areas in
North America, including areas with high density A. planipennis populations. Other
members of the ash family (Oleaceae), including privet (Ligustrum sp.) and Japanese tree
lilac (Syringa reticulata B1), are also commonly used in landscape settings.

Attacks on non-Fraxinus genera have not been observed in North America, but
there is concern about potential host switching as A. planipenm's densities increase and
ash trees die in the infested area. If A. planipennis can colonize and develop in additional
species, environmental and economic impacts on forest resources could increase
dramatically in North America. Given the extensive damage that A. planipennis has
already caused to the Fraxinus resource in southeastern Michigan, there is an urgent need
to evaluate the susceptibility of other commonly planted landscape and forest trees to
serve as alternate hosts to A. planipennis. I conducted two no-choice studies using cut
branches and live trees to determine if A. planipennis would oviposit on potential

alternate host species and to evaluate early instar development on these species.

MATERIALS AND METHODS
Cut branch bioassays

2003 bioassay. Branches of green ash, white ash, American elm, black walnut,
shagbark hickory [Carya ovata (Mill) K. Koch.] (Juglandaceae), and hackberry (C eltis
occidentalis L.), a relative of elm in the Ulmaceae family, were harvested from Michigan
State University’s WK. Kellogg Forest, Kalamazoo Co., M1 on 16 June 2003. Black ash
was collected from Kensington Metropark, Oakland Co., M1 on 28 May 2003. Swamp,
Chinese, and glossy privet [F orestiera acuminate (Michx.) Poir.], Ligustrum sinense
Lour., and L. lucidum Ait., respectively] branches were collected in Stoneville,
Mississippi from the Delta Experimental Forest on 4 June 2003 and sent by overnight
mail to Michigan. Branches were harvested from multiple trees of each species. The
ends of the freshly cut bolts were waxed after cutting to reduce desiccation. Bolts were
kept in cold storage at 1-2° C with a minimum of 80% RH. On 2 July, seven branches of
green and white ash and ﬁve branches of the remaining species were cut to approx. 17 cm
in length, then cut in half vertically down the center (Table 1.1). All out surfaces were
again waxed with melted parafﬁn to slow desiccation.

Each branch section was placed in a clear, plastic box (14 X 20 X 10 cm), along
with one green ash leaf collected from an untreated, infested green ash tree in Washtenaw
County. Ash leaf stems were submerged in vials of water to slow desiccation. Prior to
the start of bioassays, boxes were washed with soapy water, rinsed in a dilute bleach
solution and rinsed again in distilled water. Boxes were ventilated with four small holes

in two sides and six holes in the lid.

Adult beetles were collected on 1 July 2003 from wild populations at Bicentennial
Park, in Livonia, Wayne Co., MI. Adults were kept in a screened box overnight and
provided with green ash leaves for feeding. The following day, beetles were sexed and
one male/female pair was placed in each plastic box (44 pairs total). There were seven
replicates (consisting of one branch section) of the ash species and ﬁve replicates of the
non-ash species. Boxes were kept in a growth chamber at 24°C, 70% RH, and 16:8
lightzdark photoperiod.

Boxes were checked and foliage was replaced twice a week until the female
beetle in the box died. Condensation and frass were wiped from boxes and grth
chamber humidity was eventually reduced to 40% to help prevent grth of Penicillium
sp., Aspergillus sp., and Beauveria sp. mold. One black ash and one black walnut branch
section were removed and disposed of due to excessive mold. Twenty-two of the 44
pairs of beetles were observed mating and mating dates were recorded. Number of eggs
laid on each box was recorded and eggs were removed each time foliage was replaced.
Adults lived from eight to 43 days in the boxes. Upon death of the female beetle, the
branch section was removed, gently rinsed with water to remove any mold, air-dried, then
stored on a lab bench at 24°C. The total number of eggs laid on each box was recorded at
this time. The ﬁrst branch section was removed from its box on 9 July and the last
section was removed on 11 August 2003.

Branch sections were left undisturbed for 34 to 52 days after removal, allowing
time for eggs to hatch and ﬁrst instar larvae to develop. After that time, the entire surface
of each branch section was inspected for eggs and bark was peeled to quantify larval

feeding. Total bark surface area was calculated using the vertical length and horizontal

width of the outer bark surface on each branch section (Table 1.1). (Because branches
were originally cut in half to ﬁt in boxes, the “back” or cut portions of the branch without
bark were not included in the bark surface area). The bark surface of each branch section
was ﬁrst inspected with a magnifying lens to locate eggs on the surface. Hatched and
unhatched eggs were removed with forceps to prevent them from being recounted. White
non-viable eggs were distinguished from brown viable eggs and were not included in the
overall total. Less than ten non-viable eggs were found. Forceps or a small knife were
used to chip off bark ﬂakes to reveal eggs hidden in bark layers. After 15 min of
searching, total number of eggs was recorded and standardized in two ways: by 100 cm2
of bark surface area and by the number of eggs laid per beetle per day. For example, if a
female lived for 15 days in the box and laid 12 eggs on the branch section during that
time, then the number of eggs per beetle per day would be equal to 0.8. Bark was
carefully peeled down to the wood using a small knife or chisel and diameter of each
branch was measured (Table 1.1). Number and stage of larvae on each branch section
and a visual estimate of the percentage of the carnbium/phloem area covered by larval
galleries was recorded. Galleries were also standardized by 100 cm2 of surface area.
2004 bioassay. I repeated this bioassay three times in 2004 with the same eight
species used in 2003, plus the addition of blue ash and Japanese tree lilac. Wood from
several trees of each species for the bioassays was collected from the same sites used in
2003 on 24 May and again on 12 July 2004. Blue ash was harvested from a private
woodlot in Superior Township, Washtenaw Co., M1 on the same dates. Branches of
Japanese tree lilac were collected from the campus of Michigan State University, Ingham

Co., M1 on 20 May 2004. A different species of privet, Ligustrum amurense Carr., was

collected from Wooster, Wayne Co., OH on 25 May 2004 and sent by overnight mail to
Michigan. All branches were maintained in cold storage at 1-2° C with a minimum of
80% humidity. Eight branches of each species were cut to approx. 17 cm in length
(Table 1.1) at the beginning of each bioassay, but the branches were not cut vertically as
they were in 2003. Cut ends were waxed with parafﬁn. Boxes were set up using the
same methods from 2003, except that 20 ml vials of water with cotton wicks were
included in each box to provide drinking water for the beetles. Additional holes were
added to the boxes to increase ventilation.

Instead of collecting wild beetles from the ﬁeld, adult beetles were reared from
bolts of infested ash trees collected in the core of the southeast MI infestation. Logs were
kept in cold storage at 1-2° C until needed. Once emerged from logs, similarly aged
groups of beetles (approx. 200) were placed in screen cages (60 X 60 X 60 cm) to feed on
green ash foliage collected from an untreated, infested green ash tree. Cages were kept in
growth chambers at 24°C, 60% humidity, and 16:8 lightzdark photoperiod. Beetles were
allowed to feed and mate for two weeks before bioassays began.

On each day that bioassays were initiated, beetles were sexed and one
male/female pair was placed in each box (80 pairs total). The ﬁrst bioassay contained
two replicates of each species and was conducted from 18 June to 5 August. The second
bioassay was comprised of three replicates and was conducted from 14 July to 14 August,
while the third had three replicates and ran from 20 August to 14 September 2004. Boxes
were kept in growth chambers at 24°C, 60% humidity, and 16:8 lightzdark photoperiod.
Thirty-seven of the 80 pairs were observed mating. Adults lived from eight to 63 days in

the boxes. If a female beetle died during the ﬁrst three days of the bioassay, she was

replaced with a similarly aged beetle. Thereafter, upon death of the female beetle, the
branch was removed and stored on a lab bench at 24°C. The total number of eggs laid on
each box was recorded at this time.

The entire surface of each branch was inspected for eggs and bark was peeled to
assess larval feeding 42-45 days after the branch was removed from its box. Diameter of
each branch was measured and total bark surface area was calculated (Table 1.1).
Branches were inspected, peeled, and larval density assessed using the same methods as
in 2003.

Caged stem bioassay

On 12 May 2004, green ash, white ash, Japanese tree lilac, and black walnut
balled-in-burlap trees, with an average caliper of 5.8 ($01 SE) cm and height of 4.0 (10.2
SE) m (Table 1.3), were delivered from Poplar Farms Nursery, Waterman, IL and planted
at Matthaei Botanical Gardens, Washtenaw Co., MI. Ten blocks of the four species of
trees were set up in a randomized complete block design (40 trees total). Trees were
planted in an open ﬁeld slightly above ground, 3 m apart, in ﬁve rows of eight trees. The
partially exposed root balls were covered with composted wood mulch and drip line
irrigation was installed for frequent watering.

Screen cages, approx. 90 cm tall and 30 cm in diameter, were constructed to
conﬁne beetles around the stem of each tree. Two 120 cm pieces of contractor’s lathe
were driven into the root ball to provide structure for the cage. A hole the same diameter
as the tree was cut into a white, plastic disc (25 cm diam.) spray-painted green. The disc
was placed around the stem of the tree, serving as the top of the cage. A small piece of

carpet was ﬁrst wrapped around the stem to ensure that any gaps between the plastic disc

and the stem were closed to prevent beetles from escaping. The plastic disc was stapled
to the lathe and aluminum screen was wrapped around the circumference of the cage,
stapled to the plastic disc, and sealed closed with caulk. Excess screen above the disc
was wrapped around the stem of the tree and secured with a plastic cable tie. Excess
screen around the bottom of the cage overlapped the root ball and was covered with
mulch to seal the bottom. The lower 25 cm of the cage, above the root ball, was wrapped
in plastic wrap and ﬁlled with approx. 20 cm of sand to prevent beetles from reaching the
mulch. Approximately 50 cm above the base of the cage, a small ﬂap (approx. 15 cm X
20 cm) was cut into the screen and secured with Velcro to provide access into the cage.
A 5 cm long piece of garden hose was attached vertically to one of the pieces of lathe to
hold shoots of green ash foliage collected from untreated, infested green ash trees. Stems
of ash foliage were placed in plastic ﬂoral water pics to slow desiccation and foliage was
changed every 2-3 days as needed.

Beetles used for this study were reared from infested logs using methods
described in the 2004 cut branch bioassays. After two weeks of feeding, beetles were
sexed and three male/female pairs were place in each tree cage (six beetles per cage).
Cages were checked and foliage misted with water daily. Observed mating dates were
recorded. Thirty beetle pairs on 25 trees were observed mating. As beetles died, they
were typically replaced 2-3 times between 10 June and 13 September 2004 (586 beetles
total). The number of female beetles in each cage per day was recorded. For example, if
three female beetles were alive in a cage for ﬁve days, the number of female beetles per
day for that cage was 15 days. Overall, the mean number of female beetles per day per

cage was 40 (322 SE) days. Individual female beetles lived from one to 34 days in cages.

High densities of beetles were present in the area where the nursery trees were
planted. The wild beetles had access to the un-caged upper portions of the trees for
feeding and oviposition. This provided us with an opportunity to evaluate oviposition
preference and larval development in a multiple-choice setting, as part of a related study
described in Chapter Two.

Four of the ten blocks of trees were randomly selected for harvesting on 1 and 2
November and the stems were returned to the laboratory and dissected between 1 and 7
November 2004. The remaining trees were cut and dissected in March 2005. Cages
were removed and stems were cut with a chainsaw just above the cage and just above the
root ball. In the lab, each stem was carefully inspected for eggs for 10 min and the total
number of eggs found was recorded. Bark was then carefully peeled down to the wood
using a drawknife. Number and stage of larvae were recorded. Eggs and galleries were
standardized by m2 of surface area.

For all trees, relative roughness of the bark on the caged portion of the stem was
qualitatively estimated as low (very few to no cracks or crevices in the bark), medium
(moderate abundance of cracks and crevices for oviposition), and high (abundant cracks
and crevices). All of the black walnut and four of the ten green ash stems had high bark
roughness. All of the Japanese tree lilac stems had medium bark roughness and all of the
white ash and the remaining six green ash stems had low bark roughness.

Statistical analysis

Data were tested for normality using the Shapiro-Wilk test (Shapiro and Wilk

1965) and residual plots. Variables were not normal and could not be normalized with

transformations. Therefore, the nonparametric Kruskal-Wallis test was used to determine

10

signiﬁcant differences among species (Kruskal and Wallis 1952) using SAS statistical
software (SAS Institute, Inc. 1989). When the Kruskal-Wallis test was signiﬁcant,
nonparametric multiple comparisons tests (p< 0.05) were used to identify differences

among species (Conover 1971, Zar 1984).

11

RESULTS
Cut branch bioassays

2003 bioassay. Agrilus planipennis laid eggs on all eight tree species included in
this study (Table 1.2). A total of 525 eggs were. found on 29 of the 42 branch sections.
Number of eggs found in 15 min of searching per branch section ranged from 0 to 65
with an average of 12.5 (12.5 SE). A total of 202 eggs were laid on the lid, sides, and
bottom of 20 of the 44 plastic boxes. Number of eggs found on the boxes (not on the
branch sections) ranged from O to 93 with an average of 4.8 (£22 SE). The average
number of eggs on all ash sections combined was 20.5 (i4.6 SE). On average, there were
at least three times as many eggs/100 cm2 on the ash sections as on the other species, with
the exception of privet, which had nearly as many eggs/ 100 cm2 as green ash (Figure 1.1).
Densities of eggs (number of eggs/ 100 cm2) did not differ signiﬁcantly among species
(x’=9.62, df=7,34, p=0.21) (Figure 1.1), presumably because of relatively small sample
size and high variation.

Beetle fecundity ranged from O to 2.8 eggs per day, with an average of 0.46
($0.10 SE) (Table 1.2). Beetle fecundity varied widely and did not differ signiﬁcantly
among species 06:10.18, df=7,34, p=0.17) (Table 1.2).

Larval galleries were most abundant on ash sections; galleries were present on ten
of the 18 sections (Table 1.2). All larvae found were either dead or dying due to
desiccation of the small sections. Overall, mean gallery density on all three ash species
was 8.5 (i2.2 SE) galleries per branch section. A total of 22 galleries were excavated on
four of the ﬁve privet sections (Table 1.2). Four of the 22 galleries on privet contained

second instar larvae, while the rest contained ﬁrst instar or neonate larvae. Percentage of

12

the cambium/phloem area covered by galleries ranged from 0 to 85% with an overall
average of 10.5 (i3.1 SE) percent (Table 1.2). Black ash had three to ﬁve times more
surface area covered by galleries than green ash, white ash, and privet. The amount of
cambium/phloem area covered by galleries did not differ signiﬁcantly among species
(26:13.38, df=7,34, p=0.063) (Table 1.2). Gallery density did not differ signiﬁcantly
among species when branch sections with normal galleries were analyzed (i.e., ash and
privet) (xz=1.19, df=3,19, p=0.75). Overall, white ash, black ash, and privet branches had
signiﬁcantly higher densities of normal galleries than elm, walnut, hickory, and
hackberry ()8: 17.80, df=7,34, p=0.0129) (Figure 1.2).

First instar larvae attempted to feed on one elm, two walnut, and one hackberry
sections (Table 1.2). These feeding attempts were less than 1 mm wide, ranged from 5
mm to 10 cm long, and consistently resulted in larval death. No galleries or feeding
attempts were found on hickory sections. Feeding attempts were distinguished from
normal galleries for purpose of analysis. Species did not differ signiﬁcantly in the
density of feeding attempts (f: 3.53, df=3,15, p=0.09).

2004 bioassay. Similar to 2003, A. planipennis laid eggs on all ten species
included in this study (Table 1.2). A total of 577 eggs were found on 48 of the 80 branch
sections. Number of eggs found per branch section in 15 min of searching ranged from 0
to 73, with an average of 7.2 ($1.3 SE). A total of 79 eggs were laid on the lid, sides, and
bottom of eight of the 80 boxes. Number of eggs found on the boxes (not on branches)
ranged from 0 to 47 with an average of 1.0 (i0.6 SE).

The average number of eggs per branch on all ash combined was 12.3 (i2.7 SE).

On average, egg density was two times as high on the ash branches as on the other

13

species, with the exception of privet, which averaged more eggs/100cm2 than green,
white, and black ash (Figure 1.3). Egg density was signiﬁcantly greater on the green ash,
white ash, black ash, and privet branches than on the elm, walnut, and hickory branches
(25:21.78, df=9,70, p=0.0096) (Figure 1.3).

Beetle fecundity ranged from O to 1.6 eggs per day, with an average of 0.23
(3004 SE) (Table 1.2). There were signiﬁcantly more eggs laid per beetle per day on
black ash and privet than on tree lilac, elm, walnut, and hickory; other species had an
intermediate number of eggs (x’=23.35, df=9,70, p=0.0054) (Table 1.2).

Twenty-four of the 32 ash branches, along with six of the eight privet branches
had galleries (Table 1.2). All larvae that were recovered were either dead or dying due to
desiccation of the branch sections. Overall, mean gallery density on all four ash species
was 12.3 (i2.3 SE) galleries per branch. A total of 77 galleries were excavated on six
privet branches (Table 1.2). Seven of the 77 galleries contained second instar larvae and
14 contained third instar larvae. Larval galleries covered 0 to 90% of the carnbium
surface area, with an overall average of 6.9 ($2.1 SE) percent (Table 1.2). Larval
galleries covered signiﬁcantly more surface area on green and black ash branches than on
hickory (13:31.30, df=9,70, p=0.0003) (Table 1.2).

Branch sections from four tree lilac, one elm, three walnut, and one hackberry had
A. planipennis feeding attempts made by ﬁrst stage larvae (Table 1.2). No galleries or
feeding attempts were found on hickory branches. As in the 2003 bioassay, when larvae
attempted to feed on other species, galleries were abnormally small and resulted in larval
death. Similarly, the density of feeding attempts did not differ signiﬁcantly among

species for the branches with only abnormal galleries (yr/2:805, df=5,35, p=0.08). The

14

four ash species and privet had signiﬁcantly higher normal gallery densities than tree
lilac, elm, walnut, hickory, and hackberry (x’= 38.11, df=9,70 p<0.0001) (Figure 1.4).
Caged stern bioassay

I found a total of nine eggs on one of the ten stem sections of green ash that had
been caged and two eggs on two of the ten caged portions of white ash stems. No eggs
were found on walnut or tree lilac stems (Table 1.3). Egg density did not differ
signiﬁcantly among species (x2=3.72, df=3,36, p=0.29) (Table 1.3).

A total of 49 galleries were excavated on four green ash stems and a single gallery
was excavated on one white ash stem. No galleries or feeding attempts were found on
the other six green ash, nine white ash, or any of the walnut or tree lilac stems (Table
1.3). Green ash stems had signiﬁcantly greater gallery densities than black walnut and
tree lilac stems, but did not differ signiﬁcantly from white ash stems ({=10.08, df=3,36,
p=0.0179) (Table 1.3). Three of the four green ash stems that had eggs and galleries also
had rough, ﬂaky bark. The remaining ash stem with galleries had smooth bark. The six

green ash without eggs or galleries and all of the white ash stems had smooth bark.

15

DISCUSSION

The ﬁrst goal of these bioassays was to determine if A. planipermis would
oviposit on non—ash species in a no-choice situation. Results from the cut branch
bioassays showed that A. planipennis will oviposit on species other than ash, even on the
plastic boxes. It is not surprising, therefore, that egg density or beetle fecundity did not
vary among species in these bioassays.

My ability to ﬁnd eggs was often dependent upon certain bark characteristics,
especially when egg density was low. For example, privet, tree lilac, and hackberry had
relatively smooth bark with few crevices for females to hide eggs; eggs were easily
detected with little manipulation of the bark. In contrast, the ash, elm, walnut, and
hickory sections had rough bark and ﬁnding eggs at low density was challenging.

Beetle fecundity was almost two times greater in 2003 than in 2004. This may
have been due to several factors. Wild beetles were used in 2003 while lab reared beetles
were used in 2004. Although it was impossible to determine age and mating status of the
wild beetles, they still produced more eggs than the lab reared beetles. Interestingly,
while I found more eggs in 2003 than 2004, the 2003 eggs yielded fewer galleries than in
2004. Agrilus planipennis fecundity decreases as females age (D.L. Miller, pers. com.)
and females used in 2003 may have been older than females used in 2004. The eggs I
found on branch sections and on the plastic boxes all appeared solid and intact. I was
able to distinguish between reddish-brown viable eggs and non-viable eggs which remain
white. Less than 3% of the eggs I found were non-viable.

Fewer eggs than expected were laid on the caged stems in 2004 and beetles in this

study were often observed trying to escape their cages. Previous studies have also

16

reported oviposition problems with caged buprestids. Barter (1957) reported A. anxius
expended so much energy trying to escape cages, that they spent little time doing
anything else, resulting in decreased longevity of beetles in ﬁeld cages. Nash et al.
(1951) also encountered difﬁculties in obtaining borer attacks in cages and “eggs laid
were negligible and did not develop.” Akers and Nielson (1990) similarly reported low
fecundity and reduced longevity with caged A. anxius.

Most eggs were laid on the green ash stems with rough, ﬂaky bark where they
could be hidden in cracks and under bark ﬂakes. Only one of the two eggs laid on
smooth bark white ash produced a larval gallery and that egg was positioned at the site of
an old, healed-over branch junction where there were a few punctate creases in the bark.
Barter (1957) reported that A. anxius also favored rough bark for oviposition and Loerch
and Cameron (1984) found that almost 70% of A. anxius eggs were laid in rough crevices
in the dark triangular patch at branch origins. Additional research needs to be done to
further examine the inﬂuence of bark texture or tactile stimuli on A. planipennis
oviposition.

The second goal of these bioassays was to determine whether early instar A.
planipennis larvae would develop in non-ash species. In nearly all cases, pseudo-
replication was avoided by collecting branch samples from multiple trees of each species
used in the bioassays. I was unable to collect from multiple trees of Japanese tree lilac,
but in related studies of this species, I found no evidence of intra-speciﬁc variation. The
majority of larvae I recorded on ash were second instars when branches were peeled and
all larvae were dead or dying due to desiccation of the small branch sections. Four larvae

reached the fourth instar on four black ash branches and two fourth instar larvae were

17

found on one blue ash branch. Black and blue ash branches may have retained moisture
better than green or white ash, prolonging survival of the larvae.

Twenty-ﬁve percent of the galleries on privet branches had larvae that developed
to the second or third instar, suggesting that privet may be a suitable host. Previous
studies of adult A. planipennis feeding found that male longevity was 20 days on privet
(Ligustrum) and 13 days on swamp privet (F orestiera) (Haack et al. 2004). Female
longevity in theses studies was not reported. In two-choice and multiple-choice tests, A.
planipennis fed readily on privet foliage in the presence of ash foliage (Haack and Petrice
2005). Other adult A. planipennis feeding trials found no signiﬁcant difference in the
mortality rates of beetles feeding on green ash or privet foliage over a nine day bioassay
(p=0.59), but beetles consumed more green ash than privet foliage (p<0.01) (Pell,
Anulewicz, and McCullough 2005, unpubl. data). Additional research is needed to
further assess whether A. planipennis larvae can complete development on live privet.

Neonate larvae attempted to feed on black walnut, American elm, and hackberry,
but invariably failed to develop. These species, along with hickory that had no feeding
attempts, appear to be unsuitable hosts. While it is possible that females may oviposit on
hosts that are actually unsuitable for larval development (Thompson 1988), the artiﬁcial
design of a no-choice experiment must be taken into consideration. These results show
that, given no other choice, females will oviposit on non-host species. As reported in
Chapter Two, however, I also observed A. planipennis ovipositing on non-ash species in
the ﬁeld.

Foliage may also inﬂuence host selection by adult beetles who must feed on

leaves throughout their life span (Thompson 1988). In my studies, only green ash foliage

18

was available to females. If A. planipenm's host selection is inﬂuenced by the species of
foliage available for feeding due to conditioning, there would have been a stronger
preference for the green ash branches than was actually observed.

Asian literature includes reports of A. planipenm's attacks in Japan and Korea on
Juglans, Ulmus, and Pterocarya (Akiyama and Ohmomo 1997, Sugiura 1999), but
attacks in North America have only been observed on Fraxinus species. The North
American population of A. planipennis may have experienced a genetic bottleneck
because it is likely that only a small number of beetles were introduced into North
America. The small founding population probably represented only a fraction of the
parent population’s genetic variation. As a result of this founder effect, North American
populations may lack the genetic breadth of their parent populations (e.g., the genotypes
associated with development on other tree genera) and this may limit their host range to
F raxinus. Other examples of successful invasions after genetic bottlenecks include the
Argentine ant (Linepithema humile) (Suarez et al. 1999) and the imported ﬁre ant
(Solenopsis invicta) (Ross et al. 1993). Genetic changes in the introduced ant populations
resulted in changes in behavior (Suarez et al. 1999) and social organization of colonies
(Ross et al. 1993).

Since A. planipennis host range is limited to Fraxinus in China (Chinese
Academy of Science 1986, Yu 1992), it could be that the original source of the North
American population was from China. Future studies of A. planipennis host range should
include Asian tree species, originally listed as hosts in the literature, to evaluate

development of North American A. planipennis populations on these species.

19

The taxonomy of A. planipennis is complex and this complicates the delineation
of its speciﬁc host range. A grilus planipennis is known in many parts of the world and in
the literature as A. marcopoli (Yu 1992), A. planipennis ulmi (Akiyama and Ohmomo
2000, Sugiura 1999) and A. feretrius (J endek 1994) each with individual and overlapping
host species. Several examples of species subpopulation variations in host range can be
found in the literature. Parry and Goyer (2004) found that the forest tent caterpillar,
Malacasoma disstria, originally thought to be an extreme generalist, is actually a
composite of regionally adapted populations. Similarly, Sunnucks et al. (1997) found
that the aphid pest species Therioaphis trifolii is a generalist in Western Palaearctic
region on alfalfa, clovers, and related legumes, but displays a particular form in Australia
that feeds almost exclusively on alfalfa. Fumanal et al. (2004) found that host
conditioning can inﬂuence the host range of the collar gall weevil Ceutorhynchus
assimilis. Host speciﬁcity tests in lab and ﬁeld experiments showed heterogeneity in the
host range of the weevils reared from different host plants as determined by larval
development. Genetic analysis of A. planipennis is underway to compare populations of
A. planipennis in the US. with populations in Asia, including China where the host range
is limited to Fraxinus, and Japan and Korea where hosts include Juglans, Ulmus, and

Pterocarya (Bray et al. 2005).

20

Table 1.1. Mean (iSE) diameter, length, and surface area for branch sections used
in no-choice bioassays in 2003 and 2004.

 

 

Species 11 Diameter (cm) Lengﬂcm) Surface area (cm‘)
2003 Bioassays

Green ash 7 8.3 i: 0.2 16.7 i 0.3 256.6 :E 12.0
White ash 7 8.4 i 0.1 17.3 :1: 0.3 247.0 i 16.8
Black ash 4 6.6 i 0.3 17.3 :t 0.5 232.0 :E 13.0
Privet 5 4.8 i 0.4 17.0 i 0.3 176.0 d: 13.5
American elm 5 7.8 i 0.1 17.2 i 0.4 243.2 i 21.3
Black walnut 4 8.0 i 0.1 17.0 d: 0.1 251.0 i 22.4
Hickory 5 9.2 :l: 0.1 15.2 :EO.6 251.4 :I: 11.5
Hackberry 5 8.9 i 0.2 16.4 :t 0.6 258.8 1 6.0
2004 Bioassays

Green ash 8 9.0 :t 0.6 16.8 :t 0.3 409.4 i 25.4
White ash 8 8.8 :1: 0.5 16.8 :1: 0.3 408.5 E 17.3
Black ash 8 7.7 :1: 0.2 16.6 :t 0.3 351.5 i 9.2
Blue ash 8 7.2 :t 0.3 16.5 i 0.3 316.5 i 20.8
Privet 8 6.4 i 0.3 16.9 d: 0.2 319.0 :1: 14.7
Japanese tree lilac 8 4.6 d: 0.3 16.9 i 0.4 224.8 i 16.1
American elm 8 9.3 i- 0.2 15.4 i 1.1 408.0 :1: 31.3
Black walnut 8 9.5 3: 0.1 15.73: 0.6 392.5 i 22.6
Hickory 8 9.7 i 0.1 17.0 d: 0.3 475.5 d: 8.4
Hackberry 8 9.3 3: 0.2 16.6 d: 0.3 405.6 :E 22.7

 

21

.Godv Q 65:82: 5:89:80 29:2: 0:38.830: :8 H8: m=_m>»-_8_m=§v 2.20.ch >_.:moc_:w_m 8: 08 3:2 2:3 0:: x: 3323.: £82: .8528 55:5

 

 

 

 

Ea :3 H a _ _ c o B? 85 H 3 m cm m 938.8:
a c o o o o B 85 H S m : a .032:
a: d m cm 0 d a 8d H d m m w 5583 :85
:a d _ w d d Edd H _d m d. w E? 58:0an
an d v dm d o to: mod H .d e mm m 8.: 8: omoggﬁ
:: dN.~ H Ev d d o E. a nod H vd w 3 w 83:—
:a and: H mdm d d 0 ma 9:: 2d H ed m m: w :8 can
a d3. H dd: d d N. mm a 2d H vd \. mm m :8 :85
a: 5.: H ed d d m 2: :a 3d H vd m 2: w :8 823
a 2:2 H odm d d o d: :a 3d H vd d mo w :8 520
33
o _ m o o Cd H 3 m R m 938:
o o c o o :6 A no m a m :32:
d m 2 d d Ed H Nd _ dm 9. 5583 :85
d _ m d d 8d H Nd m _m m E3 585:?
mad H ms o d v mm Ed H Nd m mm m 82::
Sdm H om d d M 5m Rd H 5d :4 3 v :8 :85
86 H _d o d v Q. and H 2 d m: N. :8 02:3
cod H 2: d d m 3 omd H dd 0 wm. N. :8 580
i i i Mdda
mote—Eu 5.3 39:93: 39:93: mote—Eu mote-Eu 83033£mwwe ammo 5.3 in: 8283.
:2: ..\e :82 “:63.— _a>..a_ widen.— :._3 .8388 .38. .o: :32 8233. :o wwwo 552:
:33 8283 33:. .38. 5:33 .e Z 5:2: 5: .Soh 62
5:8: 62 .oz

 

.Vddm v8 mddm :_ 8882: 86:96: E
83280:)wa :0 5:83: AmmHv :35 :8 .8283 :oSt: :o mote—Em v8 wwwo .«o ..o:E=: ﬁning—mm v8 wwwo ﬁSSRESQ Smamwuo 39:3: 130,—. .N._ 03:.

22

Table 1.3. Mean (iSE) tree diameter at breast height (DBH), height, and number of Agrilus

planipennis eggs/m2 and galleries/m2 for four species of trees used in no-choice caged stem
bioassays in 2004.

 

 

Green ash White ash Tree lilac Black walnut
DBH (cm) 6.0 i 0.2 6.3 :t 0.2 4.6 i 0.1 6.4 :1: 0.2
Height (m) 4.3 i 0.1 4.3 d: 0.1 2.5 :1: 0.1 5.1 i 0.2
Total no. of eggs on bark 9 2 0 0
Mean no. of eggs/m2 6.2 a: 6.2 1.5 i 1.0 0 0
Total no. of galleries 49 1 0 0
Mean no. of galleries/m2 34.8 i 24.9 a 0.8 i 0.8 ab 0 b 0 b

 

Within rows, means followed by the same letter are not signiﬁcantly different (Kruskal-Wallis test
and nonparametric multiple comparison procedure; p <0.05).

23

 

No. eggs/100cm‘
O

 

 

Green ash White ash Black ash Privet Elm Walnut Hickory Hackberry

Figure 1.1. Mean (:tSE) number of Agrilus planipennis eggs per 100 cm2 on three known
host species (ash) and ﬁve potential alternate host species in the 2003 no-choice cut
branch bioassay. n=4 for black ash and walnut, n=5 for privet, elm, hickory and
hackberry, n=7 for green and white ash.

24

 

§

No. galleries/100cmz
N u

b b b b
o ‘ r I r ﬁr

 

 

 

Green ash White ash Black ash Privet Elm Walnut Hickory Hackberry

Figure 1.2. Mean (iSE) number of Agrilus planipennis galleries per 100 cm2 on three
known host species (ash) and ﬁve potential alternate host species in the 2003 no-choice
cut branch bioassay. n=4 for black ash and walnut, n=5 for privet, elm, hickory and
hackberry, n=7 for green ash and white ash. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric multiple comparison
procedure; p<0.05).

25

 

No. eggs/100cm'

 

 

 

Green ash White ash Black ash Blue ash Privet Tree lllac Elm Walnut Hickory Hackberry

Figure 1.3. Mean (:ESE) number of Agrilus planipennis eggs per 100 cm2 on four known
host species (ash) and six potential alternate host species in the 2004 no-choice cut
branch bioassay. n=8 branches per species. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric multiple comparison
procedure; p<0.05).

26

 

N U ¥
1 1 1
II

No. galleries/100cm2

A
1

 

b b b b b
o_

 

 

r ﬁr

Green ash Whlte aah Black ash Blue ash Privet Tree lilac Elm Walnut HIckory Hackberry

Figure 1.4. Mean (iSE) number of A grilus planipennis galleries per 100 cm2 on four
known host species (ash) and six potential alternate host species in the 2004 no-choice
cut branch bioassay. n=8 branches per species. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric multiple comparison
procedure; p<0.05).

27

CHAPTER TWO

Oviposition Preference and Larval Development of the Emerald Ash Borer
(Agrilus planipennis Fairmaire) (Coleoptera: Buprestidae) in Multiple-Choice

Field Experiments

INTRODUCTION

The relationship between oviposition preference and host suitability is critical in
understanding the evolution of host-plant interactions in herbivorous insects (Futuyma
and Peterson 1985, Thompson 1988). The adaptation of insect populations to their host
plant involves two major classes of characters: a) behaviors that inﬂuence the choice of a
plant for feeding or oviposition and b) physiological traits that affect growth and
reproduction on a particular host (Via 1986). Presumably, adult females presented with a
variety of potential hosts will select the host species that bestows the greatest ﬁtness to
their offspring. In some insect populations, however, host selection by females results in
varying degrees of performance by offspring. Thompson (1988) concluded that adult
host preference does not always correlate with larval performance because some females
oviposit on hosts unsuitable for larval development.

Because the vast majority of wood-boring larvae are legless and cannot disperse
to alternate or more suitable hosts, oviposition choices by adult females are critical
(Hanks et al. 1993). Wood-borers typically rely on olfaction to locate potential hosts that

are often patchily distributed within forested areas (Haack and Slansky 1986). After

28

oviposition, survival of offspring is driven by host suitability rather than host preference,
since larvae must feed and develop on the host chosen by the female (Hanks 1999).

Emerald ash borer (A grilus planipennis Fairmaire), an Asian buprestid discovered
in June 2002, is established in much of southeastern Michigan and Windsor, Ontario
(Cappaert et al. 2005c). This phloem-feeding species is native to northeastern China,
Korea, Japan, Mongolia, the Russian Far East, and Taiwan (Yu 1992) and was likely
introduced into North America through solid wood packing material (Cappaert et al.
2005c)

The life cycle of A. planipennis in North America is generally completed in one
year, although some individuals may require two years to complete development (Siegert
et al. 2005). Females can lay 30-60 eggs from late June to late July. Eggs are laid on or
just under the bark surface and inside bark cracks and crevices (Chinese Academy of
Science 1986, Yu 1992). Larvae feed under the bark in the carnbium and phloem from
July to October, eventually scoring the outer sapwood. Larval galleries disrupt vascular
tissue, eventually girdling and killing the host plant. Larvae pass through four instars
(Cappaert et al. 2005c) and most overwinter as pre-pupal larvae in the ﬁrst cm of outer
sapwood or in the bark of thick-barked trees (Cappaert et al. 2005a).

Although limited reports suggest that A. planipennis is not a major pest in Asia
(Akiyama and Ohmomo 2000), in North America, A. planipennis is aggressively
attacking healthy, as well as stressed, green ash (F raxinus pennsylvanica Marsh), white
ash (F. americana L), black ash (F. nigra Marsh), and blue ash (F quadrangulata
Michx.). As of March 2006, an estimated 12-15 million ash trees in southeastern

Michigan were dead or dying.

29

In its native range, A. planipennis reportedly attacks other species including
Ulmus davidiana var. japonica Planch., Juglans mandshurica var. sieboldiana and var.
sachalinensis, and Pterocarya rhoifolia Sieb. & Zucc. (Akiyama and Ohmomo 1997,
Sugiura 1999). North American members of these genera, including American elm (U.
americana L), black walnut (J. nigra L), and hickory (Carya sp.), are common in
landscapes and wooded areas in North America. Other members of the ash family
(Oleaceae), including Japanese tree lilac (Syringa reticulata B1), are also commonly used
in landscape settings.

Attacks on these genera have not been observed in North America, but there is
concern about potential host switching as A. planipennis population densities increase
and ash trees die in the infested area. If A. planipennis can attack additional species,
environmental and economic impacts on forest resources could increase dramatically in
North America. Given the extensive damage that A. planipennis has already caused to
the Fraxinus resource in southeastern Michigan, there is an urgent need to evaluate the
susceptibility of other commonly planted landscape and forest trees that may serve as
alternate hosts to A. planipenm's.

I monitored adult A. planipennis landing rates, oviposition preference, and early
larval instar development on North American congeners of potential alternate hosts,
including American elm, black walnut, Japanese tree lilac, hickory, and hackberry (Celtis
occidentalis). In three multiple choice studies, I placed freshly cut logs and nursery trees
of ash and potential alternate host species at multiple sites in the core area of A.

planipennis infestation.

30

MATERIALS AND METHODS

Logs attached to t-posts

Logs of six different species, including green ash (Fraxinus pennsylvanica), white
ash (F. americana), black walnut (Juglans nigra), American elm (Ulmus americana),
shagbark hickory (Carya ovata), and hackberry (Celtis occidentalis), a member of the
elm family, were harvested from eight to ten trees of each species at Michigan State
University’s W.K. Kellogg Forest, Kalamazoo Co., M1 on 16 June 2003. As of 2006, ash
trees at the Kellogg Forest were not infested by A. planipennis. Cut ends of the logs were
waxed on 17 June 2003 to prevent desiccation. Logs were approx. 60 cm long and
diameter ranged from 7.3 to 18.2 cm with an average of 13.1 (i 0.2 SE) cm (Table 2.1).
We pounded t-posts, 2 m tall, into the ground, approximately 10 m apart at four sites in
the core area of A. planipenm’s infestation (locations given below). One log was attached
vertically to the top of each t-post with 14 gauge baling wire, positioning half of the log
above the top of the post. Logs were set out in a randomized complete block design with
4-8 blocks at each site (see below). There were a total of 24 logs per species, with the
exception of white ash that only had 23 logs (24 blocks, 143 logs total). Black plastic
drain pipe, approx. 12 cm in diameter and 60 cm long, was included in half of the blocks
as a “control” to see if A. planipennis adults would land on non-wood material. The drain
pipes were attached vertically to the t-posts, giving a dark silhouette similar to the logs.
Seventy-ﬁve percent of the logs of each species were randomly selected to be wrapped
with clear, plastic, stretch wrap and covered with Tanglefoot® (109 logs total). All 16 of

the drain pipes were wrapped. The top half of the logs was covered with wrap, leaving

31

the bark on the bottom half of the logs exposed for oviposition. Logs and drain pipes
remained on posts from 17 June to 14 August 2003.

Study sites included Bicentennial Park in Livonia, Wayne Co., MI which had a
high A. planipenm’s population density in 2003. There were many heavily infested ash
trees in woodlots within and immediately surrounding the park. Six blocks were set up at
this site along the border of an infested woodlot, surrounding open ﬁelds. Kensington
Metropark, Oakland Co., MI was selected for its moderate A. planipennis population
density. At this site, six blocks were set up at the north end of the park near a wooded
area where infestation levels in individual trees ranged from low to very high. Two
blocks were set up in two parallel lines, one block along the edge and one 4 m inside the
woodlot. The remaining four blocks were placed in an open ﬁeld ﬂanked on two sides by
wooded areas. The third site, Western Golf Course and Country Club in Redford, Wayne
Co., M1, was expected to have a low A. planipennis population density because of the
extensive removal of infested ash trees in the fall and winter of 2002. Approximately 20
lightly to moderately infested ash trees remained on the course, along with seemingly
healthy, untreated ash trees. Additional ash trees with A. planipennis occurred in
residential areas surrounding the course. Eight blocks were set up in a linear fashion in
unmowed areas of the course, near live, presumably infested ash trees. A fourth site on
Eckles Road, a disposal site for infested ash material in Plymouth Township, Wayne Co.,
MI, received large quantities of infested ash logs and wood chips from the surrounding
area. Infested logs were stockpiled until they could be chipped, providing a source of

newly emerged adult beetles during much of the summer. There were no other ash trees

32

on the site or in the immediate surroundings which were comprised of warehouses and
vacant buildings. Four blocks were arranged linearly around the perimeter of the site.

Logs were checked weekly from 20 June to 8 August and adult beetles were
counted and removed from the Tanglefoot® to assess landing rates. Tanglefoot® bands
were removed on 7 and 8 August. Logs from all sites were retrieved on 14 August 2003
and stored in an unheated shed until dissected.

Logs were returned to the laboratory for inspection and dissection between 5
November 2003 and 16 March 2004. Total bark surface area was calculated from the log
length and circumference at the middle of each log (Table 2.1). The number of eggs
found in 5 min of inspection was recorded. The total number of eggs laid on each species
of log was considered an expression of host preference. Bark was carefully peeled down
to the wood using a drawknife. Number and stage of larvae was recorded and
standardized by m2 of log surface area.

Logs attached to ash trees

2003 study: Ten black walnut and ten American elm logs were harvested from
three to four trees of each species at Hudson Mills Golf Course, Washtenaw Co., M1 on 9
June 2003 and stored outdoors. Because these logs came from an infested area, several
logs of walnut and elm were peeled preliminarily to conﬁrm that they were uninfested.
Ten green ash logs were harvested three trees at Kellogg Forest on 23 June 2003. Logs
were cut to approx. 60 cm long and diameter ranged from 9.2 to 14.6 cm with an average
of 11.9 (i 0.3 SE) cm (Table 2.2). Both ends of all logs were waxed on 24 June 2003.

Three logs, one of each species, were attached to the stem of a large, infested

green ash tree, 5-7 m aboveground (one group per tree). Logs were attached vertically

33

with plastic cable ties, in random order, one on top of the other, rotating around the
circumference of the stem. Five green ash trees were selected at Bicentennial Park. On
three of the trees (approx. 25 cm DBH and 17 m height) growing 5-10 m inside the edge
of a woodlot, I placed logs below the canopy. Logs were attached to the stem at mid-
canopy of two other trees (approx. 20 cm DBH and 12 m height) growing in the open on
the edge of a recreation ﬁeld, next to a paved parking area. Logs were also placed on the
stem, just below the canopy, of ﬁve green ash trees with average DBH of 25 cm and
height of 20 m, inside a wooded area at Kensington Metropark. Both sites were selected
because of the high number of infested ash trees in the area, determined by canopy
dieback, woodpecker attacks along the stems of the trees, and presence of adult beetles.
Logs remained in place from 24 June to 14 August 2003, then were retrieved and stored
outdoors in an unheated shed until dissected.

Logs were inspected and dissected between 26 August and 4 September 2003.
Total bark surface area was calculated from the length and circumference of each log
(Table 2.2). Inspection and peeling methods were as described earlier.

2004 study: Eighteen green ash, nine white ash, nine black walnut, and nine
American elm logs were harvested from three to six trees of each species at Kellogg
Forest and nine blue ash logs were harvested from three trees at an infested woodlot in
Superior Township on 15 June 2004. A subset of blue ash logs were peeled prior to the
start of the study and found to be uninfested. Logs (54 total) were approx. 60 cm long
with a diameter range from 9.3 to 16.7 cm, averaging 13.0 (i 0.3 SE) cm and waxed on

both ends on 16 June 2004 (Table 2.2).

34

Two sites with moderately high levels of A. planipennis densities were selected in
the city of Ann Arbor, Washtenaw Co., MI: Parker Mills and Marshall Parks. Both parks
were heavily wooded with moderate densities of white ash trees. At each park, nine
white ash trees (18 trees total, approx. 15 m in height, DBH ranging from 10 to 20 cm,
averaging 17.0 (i0.6 SE) cm) were selected. Trees had low to moderate infestation, as
determined by canopy dieback and presence of woodpecker attacks and A. planipennis
exit holes. All trees were inside a wooded area.

Groups of three logs were attached to live trees in an incomplete block design.
Each group consisted of a green ash log (control), a second ash log (white or blue ash),
and a non-ash log (elm or walnut). The three logs were attached, using plastic cable ties,
to the stem of a large, infested white ash tree, 5-7 m aboveground (one group per tree).
Green ash logs were always placed in the middle position of the group along the stem and
the two other logs alternated between the top and bottom positions. Logs remained in
place from 17 June to 11 August 2004, then were retrieved and stored outdoors until
dissected. Logs were inspected and dissected between 17 September and 10 December
2004. Inspection and peeling methods were as described earlier.

2005 study: In 2005, we attached both green and white ash logs to both green and
white ash trees. Forty green ash, 40 white ash, and 13 black walnut logs were harvested
from four to 15 trees of each species at Kellogg Forest and 14 black ash logs from four
trees at the Nan Weston Nature Preserve at Sharon Hollow, an apparently uninfested site
in Washtenaw Co., MI on 13 June 2005. Thirteen blue ash logs were harvested from four

trees at the Strait Creek Nature Preserve, another uninfested site in Adams Co., OH on 10

35

June 2005. Logs (120 total) were approx. 60 cm long, ranging from 48 to 66 cm in
diameter, averaging 12.5 (i0.2 SE) cm and waxed on both ends (Table 2.2).

Ten infested ash trees at each of four sites with moderately high densities of A.
planipenm's were selected in the core A. planipennis infestation area in southeast MI (40
trees total). Two sites were composed of street trees growing in the open and two sites
had trees growing in woodlots. Two other sites had green ash trees, one with green ash
street trees and the other with green ash woodlot trees. The design was similar for the
two white ash sites; one had white ash street trees and the other had white ash woodlot
trees.

Ten green ash trees growing along the edge of a wooded area were selected at
Kensington Metropark. Ten open grown green ash street trees located in the Wellington
neighborhood of Ann Arbor were also chosen. Ten white ash street trees were chosen in
a second Ann Arbor neighborhood, Windemere, where white ash was the predominant
street tree species. Lastly, ten white ash trees growing in the Dicken Woods woodlot in
Ann Arbor were also chosen for the study. DBH ranged from 15 to 70 cm, with an
average of 3 1 .0 (3:19 SE) cm and heights ranged from 12 to 18 m.

Groups of three logs were attached to trees using an incomplete block design.
Each group consisted of a green ash log (control), a white ash log (control), and a third
log (black ash, blue ash, or black walnut). On 15 and 16 June 2005, logs were placed in
random order along the stem of the trees and attached using plastic cable ties. Logs
remained in place until 17 August 2005 when they were retrieved and stored outdoors

until they were debarked between 30 September and 1 December 2005. Logs were not

36

inspected for eggs, but were measured and peeled down to the wood using methods
described earlier.
Nursery trees

2004 study: On 12 May 2004, green ash, white ash, Japanese tree lilac and black
walnut balled-in-burlap trees were delivered from Poplar Farms Nursery, Waterman, IL
and planted at Matthaei Botanical Gardens, Washtenaw Co., MI. Ten blocks of trees
were set up in a randomized complete block design (40 trees total). Trees were planted in
an open ﬁeld, slightly above ground, 3 m apart, in ﬁve rows of eight trees. The exposed
root balls were covered with wood mulch and drip line irrigation was installed for
frequent watering. Trees had an average DBH of 5.8 ($0.1 SE) cm (Table 2.3).

Screen cages were constructed around each tree to conﬁne lab reared beetles
around the lower 1 m of the stem for no-choice bioassays described in Chapter One. The
un-caged portions of the trees, however, remained exposed to high density A. planipenm's
populations in the area. Trees were in place throughout the 2004 A. planipennis ﬂight
season and were harvested late in the fall after larvae had ﬁnished feeding. F our of the
ten blocks of trees were randomly selected for harvesting on 1 and 2 November 2004.
Trees were out just above the cage, returned to the lab, and dissected between 1
November and 10 December 2004. The remaining six blocks were cut and dissected in
March 2005. In the laboratory, the main stem and branches greater than 0.5 cm in
diameter were carefully peeled down to the wood. Number and stage of the larvae and
the diameter of the stem or branch the larvae was found on was recorded. The average

area sampled for each tree was 0.55 ($0.06 SE) m2 per tree (Table 2.3).

37

For all trees, relative smoothness or roughness of the bark was estimated. Three
categories of bark roughness were designated. Low bark roughness indicated very few to
no cracks or crevices in the bark (smooth bark). Medium bark roughness described bark
with a moderate abundance of cracks and crevices for oviposition. High bark roughness
indicated a high abundance of cracks and crevices. Branch junctions in trees can also
have cracks and crevices, even on smooth bark trees. Therefore, the number of galleries
originating within 5 cm of a branch junction was also recorded.

2005 study: The study was repeated in 2005 using the same four species
purchased from the same nursery in 2004. Root balls from the previous study were
removed and new trees were planted in the existing holes on 13 May 2005. Exposed root
balls were covered with wood mulch and drip line irrigation was re-installed. Trees had
an average DBH of 7.1 ($0.2 SE) cm (Table 2.3).

Screen cages were again constructed for no-choice tests as described earlier. Un-
caged portions of the trees were harvested and debarked two blocks at a time in
December 2005 and January 2006. Peeling methods and estimates of bark roughness
were as described earlier. Total surface area sampled per tree (Table 2.3) and number of
galleries originating within 5 cm of a branch junction was similarly recorded.

American elm survey

The objective of this survey was to determine if A. planipennis adult females
would oviposit on elm trees adjacent to or in close proximity to infested ash trees.
Bicentennial Park was chosen for sampling because of the high density A. planipennis
population present in 2003, rapidly decreasing availability of suitable host material (live

ash trees), and the high number of elm and green ash trees growing together.

38

On 19 December 2003, nine American elm trees, each growing within 5 m of 28
highly infested green ash trees were selected for sampling. Three of the nine elms had
trunks that were physically in contact with the trunks of infested ash trees. Diameter at
breast height was recorded for each ash and elm tree. Distance from each ash to the
nearest elm was recorded. Distance of elms from the nearest ash ranged from 0 to 4.6 m,
with an average of 2.8 ($0.3 SE) m. Average DBH for green ash and elm was 25.8 ($1.4
SE) cm and 15.6 ($1.3 SE) cm, respectively.

Percentage of phloem girdled by A. planipennis larvae was visually estimated by
removing an approx. 30 X 30 cm window of bark at 1.5 m above ground on each ash tree.
Twenty-ﬁve of the 28 ash trees had apparently died in 2003. Canopy dieback and heavy
A. planipennis density on the remaining three trees suggested that these trees would be
dead within a year.

The nine elm trees were felled (6 to 8 cm aboveground), and carefully debarked
with drawknives to look for A. planipennis galleries. F our of the elms were completely
peeled and 50-80% of the bark on the trunk and branches of the other ﬁve trees was
removed in long, vertical strips from top to bottom. A total of 28.4 m2 of elm surface
area was sampled, averaging 3.2 ($0.5 SE) m2 per tree.

Statistical analysis

All variables were tested for normality using the Shapiro-Wilk test (Shapiro and
Wilk 1965) and residual plots. Landing rates and gallery density for logs attached to ash
trees in 2005 were normalized by log (x+1) and square-root (x+1) transformations,
respectively (Ott and Longnecker 2001). Gallery density for logs attached to ash trees

was analyzed using the MIXED procedure for mixed models in SAS statistical software

39

(SAS Institute, Inc. 1989) with site type (street tree or woodlot tree), tree species (green
or white ash), and log species as ﬁxed effects and tree number as a random effect.
Differences among treatment means were tested as unplanned comparisons and tests were
applied only when overall analysis of variance (ANOVA) was signiﬁcant (p<0.05). Two
and three-way ANOVAs were performed to determine signiﬁcant effects and interactions
among site location, tree type, and species factors using SAS statistical software (SAS
Institute, Inc. 1989). When signiﬁcant, Fisher’s protected least signiﬁcant difference
(LSD) test was used to determine which species differed within the appropriate marginal
comparisons (Ott and Longnecker 2001).

Other variables, including eggs/m2 and galleries/m2 were not normalized by
transformations. For non-normal variables, Friedman’s two-way nonparametric ANOVA
was used to determine signiﬁcant interactions between site location and species factors
(SAS Institute, Inc. 1989). When signiﬁcant, nonparametric Wilcoxon Rank Sum and
Kruskal-Wallis tests were used within each site location to determine if egg and gallery
densities differed among species (Conover 1971, Kruskal and Wallis 1952). When the -
Kruskall-Wallis test was signiﬁcant, nonparametric multiple comparisons (p<0.05) were
used to identify differences among species (Zar 1984, Conover 1971). The Wilcoxon
Ranked Sum test was also used to evaluate bark roughness as an effect on gallery density.

All analyses were conducted at the p<0.05 level of signiﬁcance.

40

RESULTS
Logs attached to t-posts

A total of 602 A. planipennis adults were collected from the 109 logs and 16 pipes
with Tanglefoot®, ranging from 0 to 48 with an average of 4.8 ($0.7 SE) beetles per log.
Thirty-four percent of the beetles were caught on the six blocks of logs at Bicentennial,
33% on the four blocks at Eckles, and 21% on the six blocks at Kensington. Only 12% of
the beetles were captured on the eight blocks at Western, which was signiﬁcantly fewer
than the number captured at the other three sites (F =21.28, df=3,121, p<0.0001). Black
walnut logs captured the most beetles; 133, followed by white ash and green ash, which
accounted for 39% of all beetles (Table 2.4). Signiﬁcantly more A. planipennis adults
landed on ash, elm, walnut, and hickory logs than on drain pipes, which accounted for
only 3% of adults (F =3.04, df=6,118,p=0.0091) (Table 2.4). Number of adults collected
from hackberry logs did not differ signiﬁcantly from the other logs or the drain pipes
(Table 2.4). The interaction between site location and species was not signiﬁcant
(F=0.92, df=18,p=0.56).

Green and white ash logs combined had four times as many eggs as the non-ash
species combined (Table 2.4). Seventy-eight percent of the 256 eggs I found were on ash
logs. Fifty-ﬁve percent the ash logs had eggs compared to 23% of the non-ash logs
(Table 2.4). Twelve of the 23 eggs found on walnut were on a single log at Western. No
eggs were found on the drain pipes. The number of eggs found in 5 min of searching
ranged from 0 to 32 with an average of 1.8 ($ 0.3 SE) eggs per log.

Site location and log species signiﬁcantly affected egg density 06:12.69,

df=3,139, p=0.0054 and x’=31.18, df=5,l37, p<0.0001, respectively). Logs at

41

Kensington had signiﬁcantly higher egg densities than logs at Western, but egg densities
on logs at Bicentennial and Eckles were not signiﬁcantly different from egg densities on
logs at Kensington or Western. White ash logs had signiﬁcantly more eggs/m2 than elm,
walnut, and hickory logs (Table 2.4). There was a signiﬁcant interaction between site
and species factors (F =2.0, df=15, p=0.0208), but this interaction was related to very low
egg densities on logs at Eckles and Western and higher densities at Bicentennial and
Kensington. Therefore, marginal comparisons among species within sites are not
presented.

A total of 432 A. planipennis larval galleries were found on 19 of the 47 green
and white ash logs (Table 2.4). The number of galleries per ash log ranged from 0 to 75
with an average of 9.2 ($2.7 SE). Green and white ash logs at Bicentennial had the
highest gallery density with approx. 110 galleries/m2, followed by Kensington with
approx. 50 galleries/m2. Galleries were only found on white ash logs at Eckles and
Western.

Four walnut logs had discemable A. planipennis feeding attempts made by ﬁrst
instar larvae (Table 2.4). Forty-seven of the 51 larval feeding attempts on walnut were
on a single log. All feeding attempts on walnut, however, were small, less than 2 cm
long, and consistently resulted in larval death (Table 2.4). Elm, hickory, and hackberry
logs had no feeding attempts (Table 2.4). Feeding attempts were distinguished from
normal galleries for purpose of analysis and logs with feeding attempts were assigned a
gallery density of zero.

Site and log species signiﬁcantly affected gallery density (x’=11.64, df=3,139,

p=0.0087 and 12:45.77, df=5,l37, p<0.0001, respectively). There were signiﬁcantly

42

more galleries/m2 on logs at Bicentennial and Kensington than at Western, but gallery
densities on logs at Eckles did not differ signiﬁcantly from the other sites. The ash logs
had signiﬁcantly more galleries/m2 than the non-ash logs (Table 2.4). The interaction
was signiﬁcant (F =4.79, df=15, p<0.0001) probably because of the extremely low gallery
densities on logs at Eckles and Western and higher densities at Bicentennial and
Kensington. Of the 72 logs inspected from Eckles and Western, only 39 galleries were
found on 3 white ash logs. Therefore, marginal comparisons among species within sites
are not presented.
Logs attached to ash trees

2003 study: The number of A. planipenm's eggs found in 5 min of searching
ranged from 0 to 35 per log, with an average of 3.8 ($1.4 SE) eggs per log. Site location
did not signiﬁcantly affect egg density (S=227.0, df=1,28, p=0.83), but effects of log
species were signiﬁcant (f=15.45, df=2,27 p=0.0004). Agrilus planipennis females laid
signiﬁcantly more eggs/m2 on green ash logs than on walnut and elm logs (Table 2.5).
The interaction between site and species factors was not signiﬁcant (F =1 .57, df=2,
p=0.23).

Overall, the mean number of galleries per green ash log was 40.6 ($10.1 SE).
Only three feeding attempts, each less than 1 cm long, were identiﬁed on one of the ten
black walnut logs. These feeding attempts were not included in the analysis and logs
with feeding attempts were assigned a gallery density of zero. No galleries were found
on any of the elm logs (Table 2.5). Gallery density varied signiﬁcantly among species

(x’=27.49 df=2,27, p<0.0001), but not among sites (S=240.0, df=1,28, p=0.72). Green ash

43

logs had signiﬁcantly more A. planipennis galleries/m2 than walnut and elm (Table 2.5).
The interaction between site and species factors was insigniﬁcant (F =0.0, df=2, p=1.0).

2004 study: The number of eggs found in 5 min of searching ranged from 0 to 13
with an average of 1.5 ($ 0.4 SE) eggs per log. Egg densities of white ash logs were
almost ﬁve times that of logs of other species, including green ash and blue ash (Table
2.5). Only one egg was found on walnut and no eggs were found on elm. Egg density
varied signiﬁcantly by species (28:23.79, df=4,49, p<0.0001), but not by site (S=658.5,
df=1,52, p=0.11). Egg density did not differ signiﬁcantly among the three ash species,
however, white ash logs had signiﬁcantly more eggs/m2 than elm and walnut logs (Table
2.5). The interaction between site and species factors was not signiﬁcant (F =1 .22, df=4,
p=0.32).

A total of 389 galleries were found on the three ash species. Gallery density
varied signiﬁcantly among sites (S=618.0, df=1,52,p=0.0231). On average, there were
almost ﬁve times more galleries/m2 on the logs at Parker Mills than the logs at Marshall
Park. Gallery density also varied signiﬁcantly by species (f=35.61, df=4,49, p<0.001).
White ash logs had signiﬁcantly more galleries/m2 than the other species with more than
three times more galleries than the ash other species (Table 2.5). Only one small feeding
attempt was found on a walnut log and it was not included in the analysis. No feeding
attempts were found on elm (Table 2.5). The interaction between site and species was
not signiﬁcant (F=0.32, df=4, p=0.87).

2005 study: Agrilus planipennis larvae excavated 800 and 621 galleries on 36
green ash logs and 35 white ash logs, respectively (Table 2.5). I found 148 small larval

feeding attempts were made on seven of the 12 black walnut logs and those logs were

44

assigned a gallery density value of zero. Two main effects, site type (street tree vs.
woodlot trees) and tree species (green ash vs. white ash), were marginally insigniﬁcant
(F=3.97, df=1,61, p=0.0509 and F=3.66, df=1,61, p=0.0604, respectively). Gallery
density varied signiﬁcantly among species (F =16.23, df=4,61, p<0.0001). Green ash logs
had signiﬁcantly more galleries/m2 than black ash, blue ash, and black walnut logs, but
gallery density in white ash logs did not differ signiﬁcantly from the other ash logs
(Table 2.5). None of the two-way interactions were signiﬁcant: site type and log species
(F=2.28, df=4,61, p=0.07), tree species and site type (F=0.78, df=1,61,p=0.38), and tree
species and log species (F=0.43, df=4,61, p=0.79).

A signiﬁcant three-way interaction occurred for tree species, site type, and log
species (F =3.33, df=4,61, p=0.0156). Therefore, means were separated and analyzed
within each site type and tree type. At the Wellington site with green ash street trees,
there were signiﬁcantly more galleries/m2 in the green ash, white ash, and black ash logs
than on the walnut logs (Figure 2.1a). At Kensington, where logs were attached to green
ash woodlot trees, green ash and white ash logs had signiﬁcantly more galleries/m2 than
blue ash, black ash, and walnut logs (Figure 2.1b). Green and blue ash logs also had
signiﬁcantly more galleries/m2 than white ash, black ash, and walnut logs at Windemere,
were logs were attached to white ash street trees (Figure 2.1c). At Dicken Woods, logs
were in white ash woodlot trees and there were signiﬁcantly more galleries/m2 on green
ash, white ash, and black ash logs than on the walnut logs (Figure 2.1d).

Nursery trees
2004 study: Similar numbers of galleries were found on the upper portion of the

green and white ash trees in the nursery at the Matthaei Botanical Gardens (Table 2.3).

45

Sixty-eight percent of the galleries found on green ash were on the four trees with high
bark roughness, while 32% of the galleries were on green ash with smooth bark. A single
larval feeding attempt was found on one Japanese tree lilac and that feeding attempt was
not included in the statistical analyses. No feeding attempts were found on black walnut.
An average of 146.9 ($46.3 SE) galleries/m2 was calculated for green ash trees (Table
2.3). Galleries density did not differ signiﬁcantly between green and white ash trees, but
the ash trees had signiﬁcantly more galleries/m2 than tree lilac and walnut (overall
x"=33.82, df=3,36, p<0.0001) (Table 2.3).

On smooth-barked trees, rough bark texture can still be found around branch
junctions. Ten white ash and six green ash trees had smooth bark on the stem and 63%
and 71% of galleries, respectively, originated within 5 cm of a branch junction on those
smooth-barked trees. Four green ash trees had high bark roughness along the main stem
and had 53% galleries within 5 cm of a branch junction. Japanese tree lilac had medium
bark roughness and the one small larval feeding attempt was found within 5 cm of a
branch junction. Black walnut also had high bark roughness, but I did not observe larval
feeding. Ash stems with rough bark and ash stems with smooth bark had an average of
458 ($51 SE) and 137 ($13 SE) galleries/m2, respectively. Rough bark ash stems had
signiﬁcantly higher gallery densities than the ash stems with smooth bark (S=74.0,
df=1,18, p<0.0001). The difference in the percentage of galleries originating within 5 cm
of a branch junction between smooth bark and rough bark green ash trees was not
signiﬁcant (f=2.93, df=1,8, p=0.09) (Table 2.3).

2005 study: Green ash trees had 365 more galleries than white ash trees. Over

70% of the galleries on green ash trees were on the ﬁve trees with rough bark. Two small

46

larval feeding attempts were found on one Japanese tree lilac and were not included in
the analysis. No feeding attempts were found on black walnut. There were signiﬁcantly
more galleries on green ash and white ash trees than on tree lilac and walnut (£533.46,
df=3,36, p<0.0001) (Table 2.3).

Ten white ash and ﬁve green ash trees had smooth bark. On these trees, 75% and
93% of galleries, respectively, originated within 5 cm of a branch junction. Five green
ash trees had high bark roughness along the main stem and had 74% of the galleries
originating within 5 cm of a branch junction. Japanese tree lilac had medium bark
roughness and the two small larval feeding attempts were both within 5 cm of a branch
junction. Black walnut also had high bark roughness, but no feeding attempts were
found. Ash stems with rough bark and ash stems with smooth bark had an average of 255
($26 SE) and 163 ($21 SE) galleries/m2, respectively. Rough bark ash stems had
signiﬁcantly higher gallery densities than the ash stems with smooth bark (S=80.0,
dﬁ1,18, p<0.0146). The smooth bark green ash trees had signiﬁcantly more galleries
originate within 5 cm of a branch junction than the rough bark green ash trees (f=6.32,
df=1,8,p=0.0119).
American elm survey

Each of the nine elm trees that were peeled appeared healthy and contained no
apparent insect infestation or pathogens of any kind. On two of the elm trees I peeled, at
least 200 cm2 of bark on the trunk was directly contacting bark on the trunk of ash trees
that had been completely girdled by A. planipennis larvae. In this situation, adults could
have readily laid eggs on elm bark and larvae would have been able to feed from ash into

elm with little exposure.

47

DISCUSSION

The primary goal of these studies was to examine how A. planipennis adult
females would respond in the ﬁeld when presented with a choice between known hosts
and potential alternate hosts for oviposition. Logs and trees were dissected to evaluate
suitability of the host species for larval development. Monitoring A. planipennis landing
rates on the logs on t-posts showed that, in the ﬁeld, adults will oﬁen land on species
other than ash. A grilus planipennis adults distinguished between the cut logs, in general,
and black plastic drain pipe. Although not measured in this study, volatiles associated
with cut logs presumably affect the ability of females to distinguish between the logs and
the plastic pipe (Dunn et al. 1986a, 1986b, Kimmerer and Kozlowski 1982).

My ability to ﬁnd A. planipennis eggs varied among species and depended largely
on bark texture. Bark of some species, like hackberry, lacked ﬂaky layers of periderrn
and eggs were apparent because they were on the surface. Other species, including ash,
required meticulous examination and careful removal of the ﬂaky, outer layers of bark to
reveal hidden eggs. On ash, egg densities were likely underestimated and were
considerably lower than gallery densities for this reason. Therefore, estimates of egg
density should not be considered an accurate measure of host preference.

As predicted, population densities of A. planipennis varied considerably among
sites, ranging from a low-level infestation at Western, a moderate-level at Kensington,
and high-level infestations at Bicentennial and Eckles. Thirty-three percent of adults
captured in this study came from the logs at the Eckles disposal yard, but these logs had

the lowest egg and gallery densities of all four sites. Low egg and gallery densities at

48

Eckles were probably caused by the lack of ash foliage in the surrounding area. Females
need to feed for at least two weeks before mating and ovipositing (Bauer et al. 2004).
Beetles that emerged from material at this site may have landed on the logs, but no
oviposition would have occurred because they needed to ﬂy off site to ﬁnd live ash trees
for feeding. Similarly, several other species of Agrilus are known to require foliage for
maturation feeding (Carlson and Knight 1969).

At Kensington, the two logs wrapped with Tanglefoot® in one block placed 4 m
inside the woodlot only caught three adults throughout the duration of the ﬂight season
compared to 86 adults caught on the other 11 logs in sunlight at that site. This suggests
that beetles preferred for sunny, open areas over shaded areas. Other studies have found
that traps or trap trees located in sunny, open conditions were more attractive than those
within closed canopy woodlots (Francese et al. 2006, McCullough et al. 2006, Poland et
al. 2005). Many A grilus sp. and other buprestids show a similar tendency to select hosts
which are exposed to sunlight (Barter 1957, Carlson and Knight 1969).

By attaching logs to the trunks of infested ash trees in the second study, artiﬁcial
conditions were created to facilitate females ovipositing within the ash trees to oviposit
on logs of other species. In this setting, logs of potential alternate host species were
surrounded by ash volatiles emitted from foliage, wood, and bark. In 2003, I attached
green ash logs to green ash trees and found very high gallery densities in green ash logs.
In 2004, green and white ash logs were both attached to white ash trees and I found high
gallery densities in the white ash logs, but considerably lower densities in the green ash
logs. Though highly debated, Hopkins’ host selection principle is based on the

observations that show many adult insects demonstrate a preference for the host species

49

on which they themselves developed as larvae (Barron 2001). Since the majority of the
A. planipennis adults probably emerged from the surrounding white ash trees and there
were signiﬁcantly more galleries on the white ash logs, I considered the possibility of
host conditioning inﬂuencing results in 2004. Therefore, I repeated the study for a third
year in 2005, and attached both green and white ash logs to green and white ash trees. If
a conditioning effect had occurred, I would have expected higher gallery densities in the
green ash logs that were attached to green ash trees and lower gallery densities in green
ash logs attached to white ash trees. Similarly, I expected the higher gallery density on
white ash logs to occur in sites with white ash trees.

My results, however, were not consistent with a conditioning effect. At three of
the four sites in 2005, green ash logs had more galleries/m2 than white ash logs and
differences were only signiﬁcant at Windemere, where the logs were in white ash street
trees. In both 2004 and 2005, there were more galleries in white ash logs than green ash
logs only in forested sites dominated by white ash. In the 2005 site with open-grown
white ash street trees, green ash logs had more galleries than white ash logs.

As in the t-post study, exposure to sun or shade probably played an important role
in host selection by A. planipennis females. In 2003, two of the ten trees used in my
study were open-grown, while the other eight trees were in closed canopy woodlots. In
2004, all 18 trees were in closed canopy woodlots and, in 2005, I initially selected open-
grown and shaded trees. The Kensington trees were along the edge of woodlots, while
trees at Dicken Woods were within a closed canopy. Logs on trees in the shade, overall,

had roughly 20% fewer galleries/m2 than those in the sun.

50

While eggs were often laid on non-ash species, larvae attempted to feed only on
black walnut logs, though always unsuccessfully. In both the t-post and logs attached to
ash tree studies, the only non-ash species to sustain any larval feeding was black walnut.
These feeding attempts on black walnut consisted of very thin galleries that were usually
less than a few centimeters long. A few were 10-20 cm long, but there was never any
evidence of larval development through successive instars before feeding stopped. The
highest number of feeding attempts was on black walnut logs attached to ash trees. Here,
these non-host logs would have been engulfed in ash volatiles and females may have
been more likely to mistakenly oviposit on these logs than on logs on t-posts. I also
noted that larval feeding attempts were only found on black walnut logs that were
attached to open-grown ash street trees and not on logs attached to trees in woodlots.

This trend is most likely explained by greater beetle activity at sunny sites than shaded
sites.

Gallery densities were higher, overall, on the logs attached to ash trees than on the
logs on t-posts, probably due to the fact that females spend time feeding on leaves and
hence lay more eggs when foliage is available. In addition, higher concentrations of leaf
and bark volatiles would have been associated with the logs in the trees than compared to
logs on the t-posts. These volatiles were, presumably, an important stimulus for A.
planipennis oviposition, and were either greatly reduced or completely lacking around the
cut logs on t-posts or chemically different when produced by the cut logs. Future studies
should focus on determining variations among the chemical proﬁles of volatiles produced

by ash species under varying conditions, including cut logs.

51

Bicentennial Park had a woodlot with heavily infested green ash trees with a large
number of American elm trees growing in close proximity. Several of these elms were in
physical contact with the infested ash. I hypothesized that if A. planipennis could utilize
American elm as a host, either by intentionally or accidentally ovipositing on elms in
close contact with ash, I would ﬁnd evidence of this at this site. More than 28 m2 of elm
phloem surface area was inspected and no sign of A. planipennis attack was found. This,
along with results from the no-choice bioassays and ﬁeld studies, suggests that American
elm is not a suitable host for A. planipennis.

The study with nursery trees provided beetles with live trees with intact vascular
systems and foliage rather than cut logs. Moisture loss in cut logs can inﬂuence both host
selection and larval development (Haack and Slansky 1986, Hanks et al. 1999), but
would not have been a factor in the nursery tree study. I hypothesized that larval
feeding attempts would not be found on live black walnut trees because living trees have
natural defense mechanisms in place that are no longer fully intact or functional in cut
logs (Hanksl999, 1993). There were no A. planipennis galleries or feeding attempts
found on any of the black walnut trees in 2004 or 2005. A total of three small,
unsuccessﬁil feeding attempts were made on two of the Japanese tree lilac nursery trees.
These consisted of narrow galleries approx. 1 mm wide and less than 3 cm long and there
was no evidence to suggest that larvae could successfully complete development on this
species.

In 2004 and 2005, green ash nursery trees had higher gallery densities than the
white ash trees, but differences were not statistically signiﬁcant. There appeared to be

only a weak preference for green ash over white ash in these studies. Field surveys of

52

green and white ash trees showed that green ash trees consistently had more canopy
dieback and higher densities of A. planipennis exit holes and woodpecker attacks than
white ash trees growing in the same area (see Chapter 3). Interestingly, there were twice
as many galleries found on the ash nursery trees in 2005 than in 2004. This was probably
due to the decrease in the number of suitable ash trees, as most ash trees in the
surrounding area succumbed to A. planipennis by 2004. In 2005, the 20 ash nursery trees
were probably the only live host trees in a 1.5 km radius. The incident of higher attack
densities may be indicative of what could happen in the core area of A. planipennis
infestation. As A. planipennis population densities increase and suitable ash host
population densities decrease, small pockets of ash that had initially escaped infestation
in previous years, may be attacked at high rates.

One factor that could be driving A. planipennis preference for green ash trees over
white ash trees in this size class could be bark texture. Other studies on Agrilus species
have also reported a preference for rough-barked portions of trees for oviposition (Barter
1957, Loerch and Cameron 1984). Rough bark provides more oviposition sites where
females can hide eggs and where eggs are more likely to escape predation by other
insects. On smooth-barked trees, branch junctions provide bumps and rough patches in
the bark where eggs can be hidden. If there was a positive relationship between bark
roughness and oviposition preference, I would have expected more galleries to originate
near branch junctions on smooth-barked trees than on rough-barked trees. This occurred
in 2005; signiﬁcantly more galleries originated near branch junctions on smooth-barked

green ash than on rough-barked green ash trees. In 2004, there were also more galleries

53

near branch junctions on smooth-barked than rough-barked green ash, but the difference
was marginally insigniﬁcant (p=0.09), probably due to the small sample size.

In both years, galleries on green ash trees with rough bark were so dense that most
larvae ran out of phloem and died before reaching the fourth instar. I have observed
these high attack rates on other ash trees in the ﬁeld. In some cases, trees have few or no
A. planipennis exit holes on the trunk, but are completely infested by larvae that died
from intra-speciﬁc competition for the limited phloem resource. Additional studies are
needed to further examine differences in ash phloem quality and bark characteristics and
how these factors affect host suitability and preference.

In each of my studies, A. planipennis appeared to prefer ash species over the non-
ash species and there appeared to be a preference for green ash over the other ash species.
In addition to green and white ash trees, I have observed full development and emergence
of A. planipennis from black and blue ash trees in the ﬁeld, so there is no question that
these species are suitable hosts. What is unclear is the order of preference among all four
species, whether this preference is consistent, and how host preference is affected by A.
planipennis density and host tree exposure to sun. While females occasionally oviposited
on black walnut and Japanese tree lilac, these species do not appear to be suitable hosts

for larval development.

54

Table 2.1. Mean ($SE) diameter, length, and surface area for logs of six species
attached to t-posts in A grilus planipennis- infested areas in 2003.

 

 

Species 11 Diameter (cm) Length (cm) Surface area (m’)
Green ash 24 12.0 $ 0.4 62.0 $ 0.2 0.23 $ 0.01
White ash 23 13.0 $ 0.5 61.1 $ 0.2 0.24 $ 0.01
American elm 24 13.2 $ 0.4 61.4 $ 0.2 0.25 $ 0.01
Black walnut 24 14.1 $ 0.5 59.8 $ 0.9 0.26 $ 0.01
Hickory 24 13.5 $ 0.4 60.7 $ 0.2 0.26 $ 0.01
Hackberry 24 13.0 $ 0.4 61.5 $ 0.3 0.25 $ 0.01

 

55

Table 2.2. Mean ($SE) diameter, length, and surface area for logs of ﬁve species
attached to Agrilus planipennis- infested ash trees in 2003, 2004, and 2005.

 

 

Species 11 Diameter (cm) Length (cm) Surface area (m’)
2003 Logs attached to green ash trees

Green ash 10 10.8 $ 0.2 62.0 $ 0.5 0.21 $ 0.01
American elm 10 12.2 $ 0.6 59.1 $ 1.7 0.22 $ 0.01
Black walnut 10 12.6 $ 0.6 57.2 $ 1.4 0.23 $ 0.01
2004 Logs attached to white ash trees‘

Green ash 18 13.6 $ 0.5 60.3 $ 0.5 0.23 $ 0.01
White ash 9 12.4 $ 0.8 60.2 $ 0.1 0.21 $ 0.01
Blue ash 9 11.4 $ 0.4 60.3 $ 0.8 0.20 $ 0.01
American elm 9 12.3 $ 0.5 60.7 $ 0.5 0.20 $ 0.01
Black walnut 9 14.4 $ 0.6 60.0 $ 0.2 0.23 $ 0.01
2005 Logs attached to green and white ash treesI

Green ash 39 11.8 $ 0.3 61.4 $ 0.2 0.21 $ 0.01
White ash 39 13.0 $ 0.3 61.3 $ 0.3 0.24 $ 0.01
Black ash 14 11.8 $ 0.3 58.5 $ 1.3 0.21 $ 0.01
Blue ash 12 12.5 $ 0.6 58.4 $ 1.2 0.22 $ 0.01
Black walnut 12 14.3 $ 0.5 60.9 $ 0.6 0.24 $ 0.01

 

lIncomplete block design.

56

.Amddv Q 08.00008 88.89800
9:28 0.808880: .08 80H 0.83-80.03.05 80:08.0 bp8om8wu do: 08 8030. 08m 0:: x: 832.8 808 £83.00 8:55

 

 

d d : d d mod H 3d Nd H 5N d. 8:.83 0.85

md H md N : d d mod H .md .d H Om d. 8.: 00: 0888a

d d a :.N H c..m mow ddd H E... Nd H We d. :8 08.3

d d : dd. H ddm me. hNd H no: Nd H dd d. :8 80:0

80.: 808:7. mddN

d d : d d mod H cmd Nd H vd d. 8:.83 0.03m.

vd H vd . : d d Nod H cNd .d H :6 d. 8... 00: 088:3

d d 8 ed H mdh NNm bdd H mud Nd H md d. :8 0::3

d d : ode H Name. mmm ddd H dvd Nd H dd d. :8 80:0

80.: 808:7. VddN

"88880:: 02.80:: N8325.80.18 8.80:8 F8. A80. Ema 80.: 57.

u:..e00.« widow. .e: 802 .ﬁeh @0388 «0.8
.e: 803. .88.: 008...; 805.

 

.mddN :8 VddN 8 8:88 00.: 8088 00.80.07.288 8 :00: 80: .«o 8.00% .50.: :8 N8 :0: 8.80:8 82208803 828V
mo 80:83: AmmHv 808 .08 .88 .08 60380 808 08:80 Ammﬂv “:80: “80:: 8 80:08.: 00.: $53 80.). .m.N 038.“.

57

.Amddv a: 68:000.:

m:0m.8:800 0.:538 0.50.8880: :8 :8: m...:>>-.:v.m_:v.v 80:0::.: .580E:w.m :0: 08 .030. 08:0. 05 A: :030..0: :808 £30.. 85.3

 

: md H ...
mN
N.
o.

0:.:
£05.80:

o
o
:o
o
d

0:: 0d H 5M

MN
N.
:N

:: hd H 5N
o:
N.
w.

.0:—05.03:

d
d
:d
d
o

0 :d H md
M

N

:N

: M.. Ho:V
Mm
..
w.

:88:

c... H:...
.m

: d

d

v

0: :d H .
MN

M

:N

: 0N H v.5
MM.
m.
w.

85:3
0.28

d d
d d
: d : w.:. H mdM
d m.N
d ..
0: deN.. :m.:H:...N
w dN.
m w.
VN MN
:M.. HM:V :m.NH...:
2: .N.
o. v.
w. w.
8.0
::0...08< :8 03:3

d N88:80:: 8800: :0 0: 80.).
d 8:808 8800: :0 .0: .80...

: 2.. H 98 ”88:08: :0 .0: 802
n . N m0.:0..:m :0 .0: .80...
w 3:803: w:.:00:
.0 850:8 5.3 $0. :0 .02
8.8.8» .88.:
:: 0M H .8. N:.\mww0:0 .0: 80$.
on :30 :0 .0: .Soh
N. :80 5.3 $0. :0 .02
:N 80. :0 .07.
mu“:
: m: H ow 3.3:: :0 .0: 802
N. . 33:: :0 .0: .80...
c. 3.2:: 5.3 mwo. :0 .07.
o. 80. :0 .02
8:: 8:5: :32
:8 80:0

 

.88 5 $2: .0288 $885: 8.2:: 5 we: 22.: so: a 383
0: :0:0::: m0:.: £8: :8 8.0. .8: N8 :0: 8.85% :8 m8 :0: :30 .3.2:: 880828: 8.2MV :0 8:88 AMmHv 808 :8 .80... .v.N 03:8

58

.Amddv A 68:08::

m:0m.:::800 0.:.:.:8 0.30880: :8 :8: 3.834835: ::0:0::.: 582.38% :0: 0:: :030. 08: 05 a: :030..0: 8:08 $.30: 85.3

 

wd. Hvdv
xv.

0 o

o

h

N.

: :d H :d

: W. H W.
M
.
d.

85:3
:0:...

o
o
: :6. H N.Nv
m.—
o.
N.
o o
o o
0 o 0: 06 H Q:—
o wN
o o
: o :: 5N H N.m
o o.
o v
o o
d
o
: o
o
o
: TN H a:
N.
m
d.
8.0
::0.:08< :8 0:...

o o o

o o d

: N: H :.:M :: a... H 9.2. : M.M. H v.3
NN. .N: cow

v. WM :M

:. 0M 0M

N85:803.: w:.:00: :0 .0: 803.

3:803: w:.:00: :0 .0: .80...

N830.85% :0 .0: 80.).

m0.:0..:w :0 .0: .80.:

3:803: w:.:00: :0 8.8.8: 5.3 80. .02
«:0. :0 .02

80:: :8 03:3 :8: 80:» 0: 8:83: 80‘. 33

o d

o d

: m.MM H 0.2». : M.M H d...
::N m:

a v.

: w: H ...N :: M.. H «0.:
o: o.

w o.

m w.

o

o

: m6: H 93.
:9:

o.

: Wm. H .8:

:0

o

d.

:8 «.0:...

Ema Gama—B Ema Evy—U

N85:82:: w:.:00: :0 .0: 80:).

8:803: 8500: :0 .0: .50...

N830.85% :0 .0: 802

8.8.8: :0 .0: .88.

3:803: 9:80: :0 m0.:0..:w 5.3 80. .02

8330 :0 .0: 80.).

:80 :0 .0: .80...

:80 5.3 mwo. :0 .02

80. :0 .02

80:: 5.: 8...; 8 855:: 83 2:3

N83:80:: 8500: :0 .0: 80.).

3:803: 8500: :0 .0: .80...

N::>.0.:0..:w :0 .0: 80.).

88:8 :0 .0: .80...

3:803: w:.:00: :0 8.85% 5.3 mwo. .02

N83%: :0 .0: 80$.

:80 :0 .0: .80...

:30 5.3 8.0. :0 .02

mwo. :0 .02

80:: :8 :09.» 0: :0:0:3: 80: MddN

 

.88 :5... .38 .88 5 82: 8 38:5

23:888.: .32me 0: 8:83: 80. :0: N8 :0: 850:8 :8 N8 :0: 8:0 38088.: 8.2%... :0 :0:88 .me. 80$. .m.N 0.5:...

59

 

Logs on green ash trees

250

 

a) b)

200 -

150 —

100 —

Galleries/m3

50~

 

 

ll Green ash
‘IWhite ash

 

 

 

Wellington (street trees) Kensington (woodlot trees) 0 Black ash

Site type .5 Blue ash
[1] Walnut

 

Logs on white ash trees

 

250

 

200 <

150 ~

100 -

Gallerieslm‘

50-

 

 

 

Windemere (street trees) Dicken Woods (woodlot trees)
Site type

 

 

 

Figure 2.1. Mean (i SE) number of Agrilus planipennis galleries per m2 on logs of ﬁve
species attached to Agrilus planipennis-infested a) green ash street trees, b) green ash
woodlot trees, c) white ash street trees and d) white ash woodlot trees in 2005. Within

site and tree types, means with the same letter are not signiﬁcantly different (Fisher’s
protected LSD test; p<0.05).

60

CHAPTER 3

Field Surveys of Emerald Ash Borer (Agrilus planipennis F airmaire) (Coleoptera:

Buprestidae) Density and Canopy Dieback in Three North American Ash Species

INTRODUCTION

The emerald ash borer (A grilus planipennis F airmaire), an Asian buprestid
discovered in June 2002, is established in much of southeast Michigan and Windsor,
Ontario, Canada (Cappaert et al. 20050). More than 35 distinct, localized infestations
have been found since 2003 in other areas of Michigan, Ohio, and Indiana. Agrilus
planipennis is native to northeastern China, Korea, Japan, Mongolia, the Russian Far
East, and Taiwan (Yu 1992). This phloem-feeding species was likely introduced into
North America through solid wood packing material (Cappaert et al. 2005c), such as
crating, pallets, or dunnage made from ash wood containing bark strips.

In southeast Michigan, A. planipennis adult emergence begins as early as mid- to
late May or early June (Brown-Rytlewski and Wilson 2005, Cappaert et al. 2005c).
Adult beetles live for 3-6 weeks (Cappaert et al. 2005c). Upon emergence, beetles feed
on ash foliage for 5-7 days before mating and females feed for another 5-7 days before
oviposition begins (Bauer et a1. 2004, Lyons et al. 2004). Adult activity peaks from late
June to early July (Cappaert et al. 2005c). Most eggs are laid in July and early August.
Eggs hatch in approx. two weeks (Lyons et al. 2004). Larvae feed on phloem and outer

sapwood for several weeks and complete their feeding in autumn (McCullough and

61

Katovich 2004). Most A. planipennis overwinter as pre-pupal larvae in shallow chambers
excavated in the outer sapwood or in the bark of thick-barked trees. Woodpecker
predation of pre-pupal larvae occurs mostly during winter and early spring and can range
from 9-95% (Cappaert et al. 2005b). Pupation begins in late April or May. Newly
eclosed adults may remain in pupal chambers for 1-2 weeks, before emerging head-ﬁrst,
leaving a D-shaped exit hole 3-4 mm in diameter (McCullough and Katovich 2004).

In North America, A. planipennis is aggressively attacking healthy, as well as
stressed, ash trees (Fraxinus sp.). As of early 2006, an estimated 12-15 million ash trees
southeast Michigan were dead or dying. However, limited reports suggest that A.
planipennis is not a major pest in Asia (Akiyama and Ohmomo 2000).

A grilus planipennis has the potential to spread throughout North America via
natural dispersal by adult beetles and artiﬁcial spread caused by humans moving infested
ash material. F raxinus species are widely distributed across the eastern U.S. and portions
of southeastern Canada, occurring in many different forest ecosystems (Harlow et al.
1991). There are at least 16 native species of Fraxinus in the U.S. (Harlow et al. 1991),
four of which are found in Michigan: green ash (F. pennsylvanica Marsh), white ash (F.
americana L.), black ash (F. nigra Marsh), and blue ash (F. quadrangulata Michx.)
(Barnes and Wagner 1981). More than 802 million ash trees occur on timberlands in
Michigan alone [USDA Forest Inventory and Analysis (F IA) database 2005]. Ash trees
are also one of the more widely planted trees in many urban areas of the U.S. (Giedraitis
and Kielbaso 1982, Ottman and Kielbaso 1976).

In natural forests, A. planipennis could substantially alter biodiversity and stand

dynamics (MacFarlane and Meyer 2005). Green ash is the most widely distributed of the

62

North American ashes, both naturally in forested areas and as a street tree (Harlow et al.
1991). White ash is the most abundant native ash species and is a component of more
than 24 forest cover types (Harlow et al. 1991). White ash is also highly valued as a
street tree for its form and fall color. Blue ash occurs on dry, limestone uplands.
Southeast Michigan is the northern edge of the range for blue ash and this species is
relatively rare in Michigan in comparison with other ash species. Blue and white ash will
often occur together on sites with better soils. Black ash, typically found in northern
regions, grows in swamps, peat bogs, stream and riverbanks (Harlow et al. 1991).

Given the extensive damage that A. planipennis has already caused to the
Fraxinus resource in southeast Michigan, an important question is whether A. planipennis
preferentially feeds or oviposits on certain North American species of Fraxinus.
Understanding the preference hierarchy and the underlying mechanisms of A. planipennis
attraction among the F raxinus species will increase our ability to accurately assess stand
susceptibility to this pest based on species composition. It is also important for
developing survey, detection, and management options and predicting spread and impacts
of A. planipennis throughout North America.

To assess host preference, I quantiﬁed the density of A. planipennis exit holes and
woodpecker attacks at four sites where green and white ash street trees were planted
together and at two woodlots where white and blue ash trees co-occurred. I also
monitored canopy dieback and changes in exit hole and woodpecker attack densities over

time.

63

MATERIALS AND METHODS
Study sites

Study sites were selected based on four main criteria. Sites had to have two
species of ash growing in close proximity, with similar range in DBH (diameter at breast
height). Trees within each site had to be untreated, growing under similar conditions, and
were assumed to be exposed to similar levels of A. planipennis pressure. Permission to
survey trees was acquired from the appropriate parties.

Green vs. white ash street trees: In 2003, I sampled green and white ash street
trees at two sites. The ﬁrst site, Butzel Elementary and Middle School, was in Detroit,
Wayne Co., MI. There were eight, small green ash and eight white ash trees growing
around the perimeter of the school yard. The second site on Lamed Boulevard, also in
Detroit, had eleven green ash trees located on the east side of the street and eleven white
ash trees in the grassy median. In 2004, a third site was added to the survey. The
Windemere neighborhood in Ann Arbor, Washtenaw Co., MI, had an abundance of green
and white ash street trees growing in the grass strip between the sidewalk and the street.
Seven green ash and seven white ash trees were selected for surveying at this site. A
fourth site, added in 2005, had clusters of green and white ash trees within a ﬁve mile
radius of the Lakeside Terraces condominium complex in Sterling Heights, Macomb Co.,
MI. Eight green ash and nine white ash trees growing within 30 m of one another were
selected for survey. Trees ranged in DBH from 4 to 38 cm (Table 3.1).

White vs. blue ash woodlot trees: In 2004, I identiﬁed two woodlots with A.
planipennis populations that had both white and blue ash trees growing in close

proximity. In Superior Township, Washtenaw Co., MI, DBH of white and blue ash trees

64

ranged from 16.5 to 32 cm and trees were growing 6—24 m apart (Table 3.1). In a
woodlot in Plymouth Township, Wayne Co., MI, DBH of trees ranged from 11 to 35 cm
and trees were growing 3-7 m apart (Table 3.1). Both woodlots were surveyed in 2004,
and Superior Twp. was surveyed again in 2005 (Table 3.1).

Sampling and data collection

Green vs. white ash street trees: For each tree, I made visual estimates of canopy
dieback in late summer and quantiﬁed the number of A. planipennis exit holes and
woodpecker attacks on larvae in the fall, after adult emergence was complete.
Woodpecker attacks occur primarily in winter and early spring when pre-pupae are
present (Cappaert et al. 2005b, c). Fall surveys were cumulative and included A.
planipennis that had developed and emerged or pre-pupae that had been preyed upon by
woodpeckers through that year’s survey.

Trees with DBH greater than 5 cm were surveyed around the entire trunk or
branch circumference in four areas: at 1.5 and 3 m aboveground, at the base of the lowest
ﬁrst order branch, and 1 m out on that branch. On each tree, an area of approx. 0.26 m2
was inspected at 1.5 m above ground and 0.13 m2 at the other three sample areas [average
total area of 0.65 ($0.01 SE) m2 per tree]. Areas above 1.5 m were accessed using a
ladder. On 11 trees at Butzel with a DBH less than 5 cm, I examined the entire stem and
all the branches. On average, I inspected 0.52 (i003 SE) m2 on each of the small trees.
Percentage of the canopy that was dead, number of A. planipennis exit holes, and number
of woodpecker attacks on A. planipennis pre-pupal larvae was recorded for each area

inspected on each tree. Mean A. planipennis density per tree was calculated by

65

combining the total number of A. planipennis exit holes and woodpecker attacks and
dividing it by the total area sampled in the tree.

Survey methods were modiﬁed in 2004 because related research showed that ﬁrst
attacks by A. planipennis typically occur in the upper canopy (Cappaert et al. 2005a).
Therefore, methods for subsequent surveys were adjusted to sample higher in the trees.
Trees with DBH greater than 5 cm were inspected by tree climbers who surveyed four to
six intervals between 1.5 and 8.5 m aboveground on each tree. An average area of 1.04
(21:0.07 SE) m2 was inspected on each tree. Small trees (<5 cm) were completely
surveyed as before. With the exception of the sample area at 1.5 m aboveground, sample
areas were randomly selected and may or may not have overlapped with areas sampled
the previous year.

In 2003, 19 green ash and 19 white ash (38 trees total) at Butzel and Lamed were
surveyed for canopy dieback on 12 August and A. planipennis exit holes and woodpecker
attacks on 12 August and 11 October. In 2004, canopy dieback estimates for 26 green
ash and 26 white ash (52 trees total) were made at Butzel, Larned, and Windemere on 13
September. During July 2004, seven dead trees were felled and removed from the survey
sites by the respective municipalities: four green ash trees at Larned and two green ash
and one white ash tree at Butzel. Because canopy dieback on these trees was greater than
90%, dieback was recorded for these trees as 100% in the subsequent year’s survey. Data
for exit holes and woodpecker attacks on these trees were collected on 6 August 2004 at
Butzel and Lamed and on 9 August at Windemere.

Seven additional dead trees were removed in 2005: one green ash and two white

ash trees at Butzel and two green ash and two white ash trees at Windemere. The two

66

white ash trees at Windemere were cut down after dieback estimates were made, but
before exit hole and woodpecker attack data could be collected. Dieback for these ﬁve
trees was recorded as 100% in 2005. Dieback estimates were made on 16 September
2005 at all sites, except Lakeside, where leaves began falling before I could estimate
dieback. Exit holes and woodpecker attacks on 25 green ash and 30 white ash (55 trees
total) were counted on 12 October at Butzel and Windemere, 19 October at Lakeside, and
19 and 21 October 2005 at Lamed. Data collected from 25 green ash and 26 white ash
trees (51 trees total) at Lamed and Butzel in 2003 and at Windemere in 2004 were used
to assess the relationship between canopy dieback and A. planipennis density.

White vs. blue ash woodlot trees: On 2 March 2004, two white ash and three
blue ash trees were surveyed by climbers in the Superior Twp. woodlot. Seven to 11
intervals on the main stem and branches between 1.5 and 15 m aboveground were
examined on each tree and A. planipennis exit holes and woodpecker attacks were
recorded for each area. An average of 2.9 (3:022 SE) m2 per tree was inspected. On 28
May 2004, before the onset of A. planipennis adult emergence, two additional white ash
and two blue ash trees were felled. Four to eight intervals between 1.5 and 15 m
aboveground were inspected on each tree. An average of 0.8 (i009 SE) m2 per tree was
examined. Trees that were felled were debarked and the percentage of phloem surface
area covered by A. planipennis galleries was estimated. Means were calculated by
combining the total number of exit holes and woodpecker attacks and dividing it by the
total area sampled in the tree. No signiﬁcant difference was found between A.‘

planipennis density data gathered by climbing trees or by felling them (S=26.0, df=1,7,

67

p=0.09). Because of the difﬁculty of accurately estimating canopy dieback on closed-
canopy woodlot trees, dieback estimates were not recorded.

On 20 January 2005, ﬁve white ash and ﬁve blue ash trees in the Plymouth Twp.
woodlot were felled. Five to 16 trunk and branch sections between 1.5 and 17 m
aboveground were removed from the trees and taken to the laboratory for inspection.
Average area sampled per tree was 1.4 (i021 SE) m2. Number of A. planipennis exit
holes, woodpecker attacks. and percentage of phloem surface area covered by galleries
were recorded.

On 21 November and 2 December 2005, another four white ash and four blue ash
trees in the Superior Twp. woodlot were felled and inspected using methods from the
2004 Plymouth Twp. survey. Five to eight sections were taken from each tree between
1.5 and 21.5 m aboveground, with an average sample area of 1.1 ($0.10 SE) m2 per tree.
Statistical analysis

All data were tested for normality using the Shapiro-Wilk test (Shapiro and Wilk
1965) and residual plots. Data were not normally distributed aﬁer transformations.
Therefore, the nonparametric Wilcoxon Rank Sum test (Ott and Longnecker 2001) was
used to determine if the mean number of A. planipennis exit holes and woodpecker
attacks density, percentage of phloem surface area covered by galleries, and percentage
of canopy dieback differed signiﬁcantly between species at each site.

Data from Lamed and Butzel in 2003 and Windemere in 2004 was pooled and
tested for normality. Data were normally distributed and simple linear regression was
used to determine if canopy dieback was a signiﬁcant predictor of A. planipennis exit

hole and woodpecker attack densities. Two outliers were identiﬁed in the data set and

68

were excluded. All analyses were conducted at the p<0.05 level of signiﬁcance, using

SAS statistical software (SAS Institute, Inc. 1989).

69

RESULTS
Green vs. white ash street trees

At Butzel, canopy dieback of green ash trees was, on average, 13-20% higher than
canopy dieback of white ash trees during the three year period, but differences were not
signiﬁcant in 2003 (S=83.5, df=1,14, p=0.0532), 2004 (S=79.0, df=1,14, p=0.13), or 2005
(S=73.0, df=1,14, p=0.31) (Figure 3.1a). Average canopy dieback in these small green
ash and white ash trees increased 32% and 25%, respectively, from 2003 to 2004, and
13% and 18%, respectively, from 2004 to 2005. (Figure 3.1a). In 2004, average canopy
dieback, excluding the three dead trees that were removed, was 43% ($10.2 SE) for green
ash and 30% (i182 SE) for white ash. In 2005, canopy dieback (excluding the dead
trees) was 53% (i11.4 SE) for green ash and 29% (i185 SE) for white ash.

Green ash trees at Butzel had an average of 17-39 more exit holes and
woodpecker attacks per rn2 than white ash trees throughout the study, but differences
were not signiﬁcant in 2003 (S=80.0, df=1,14,p=0.11), 2004 (S=47.0, df=1,l l, p=0.27),
or 2005 (S=33.0, df=1,8, p=0.31) (Figure 3.1b). Exit hole and woodpecker attack
densities in green ash more than tripled from 2003 to 2004 and increased ﬁve fold in
white ash (Figure 3. lb). Densities remained relatively consistent from 2004 to 2005
(Figure 3.1b). Average exit hole density ranged from 21-78 exit holes/m2 in green ash
and 9-43 exit holes/m2 in white ash from 2003 to 2005. Woodpecker attacks were found
on both branches and trunks of green and white ash trees. Average woodpecker attack
density ranged from 4-13 woodpecker attacks/m2 in green ash and 0-10 in white ash over
the three year period. On average, by 2005, woodpeckers had preyed upon 21% of the A.

planipennis in these trees.

70

At Lamed, green ash trees had signiﬁcantly more canopy dieback than white ash
trees in 2003 (S=182.0, df=1,20, p<0.0001), 2004 (S=l87.0, dﬁ1,20, p<0.0001), and
2005 (S=181.5, df=1,20, p<0.0001) (Figure 3.2a). Dieback increased approx. 25% each
year for both species and green ash trees averaged 49-69% more dieback than white ash
trees (Figure 3.2a). In 2004, six of the seven green ash trees remaining on site had 100%
canopy dieback and only 10% of the canopy was alive on the other. By 2005, all of the
green ash trees in the survey were dead.

Green ash trees at Lamed had 1.5 to four times as many A. planipennis exit holes
and woodpecker attacks per m2 as white ash trees throughout the study. Exit hole and
woodpecker attack densities were signiﬁcantly greater in green ash than in white ash in
2003 and 2004 (S=169.0, df=1,20, p=0.0019 and S=101.0, df=1,16, p<0.0004,
respectively), but differences were marginally insigniﬁcant in 2005 (S=85.0, df=1,16,
p=0.0521) (Figure 3.2b). Exit hole and woodpecker attack densities in green ash
increased by 45% from 2003 to 2004 and by 41% from 2004 to 2005. Densities in white
ash decreased slightly from 2003 to 2004, but almost tripled from 2004 to 2005 (Figure
3.2b). Average exit hole density ranged from 50-114 exit holes/m2 in green ash and 6-14
exits holes/m2 in white ash from 2003 to 2005. Average woodpecker attack density
ranged from 2-10 woodpecker attacks/m2 in green ash and 4-35 woodpecks/m2 in white
ash. By 2005, all trees surveyed had woodpecks and, on average, woodpeckers had
preyed upon 34% of the A. planipennis in these trees.

At Windemere, green ash trees had signiﬁcantly more canopy dieback than white
ash trees in both 2004 (S=77.0, df=1,12, p<0.0003) and 2005 (S=77.0, df=1,12,

p<0.0003) (Figure 3.3a). Green ash was trees were at or near 100% dieback both years,

71

while average dieback in white ash trees increased over 45% from 2004 to 2005. In
2005, two green ash trees were removed in mid-summer and all of the ﬁve remaining
green ash trees were dead by September.

In 2004, green ash trees at Windemere had signiﬁcantly more A. planipennis exit
holes and woodpecker attacks per m2 than white ash trees (S=66.0, df=1,12, p=0.0487)
(Figure 3.3b). In 2005, however, white ash trees had signiﬁcantly greater exit hole and
woodpeck densities than the dead or dying green ash (S=l 7.0, df=1,8, p=0.0159). One
white ash tree in 2005 with a recorded density of 229 exit holes and woodpecker attacks
per m2 was considered an outlier and removed from the data set. The data were
reanalyzed and white ash trees still had signiﬁcantly more exits holes and woodpecker
attacks per m2 than green ash trees (S=28.0, df=1,7, p=0.0317) (Figure 3.3b). Average
exit hole density ranged from 44-76 exit holes/m2 on green ash and 32-93 exit holes/m2
on white ash from 2003 to 2005. Average woodpecker attack density ranged from 6—14
woodpecker attacks/m2 on green ash and 13—46 woodpecker attacks/m2 on white ash over
the three-year period. By 2005, all trees had woodpecker attacks and, on average,
woodpeckers had preyed upon 27 % of the A. planipennis in these trees.

Green ash trees at Lakeside had signiﬁcantly more exit holes and woodpecks per
m2 than white ash trees, with an average of 22.8 (£6.86 SE) for green ash and 5.9 (i102
SE) for white ash (S=94.0, df=1,15, p=0.0180). On green ash trees, 56% of the average
total density was comprised of exit holes and 44% was comprised of woodpecker attacks.
On white ash trees, exit holes made up 29% of the average total density and woodpecker

attacks made up 71% of A. planipennis density.

72

Overall, using the Butzel, Lamed, and Windemere data, canopy dieback increased
linearly as the density of A. planipennis exit holes and woodpecker attacks increased. A
positive and signiﬁcant linear relationship existed between these two variables for green
ash, for white ash, and the two species combined (Figure 3.4). In general, if dieback was
less than 30%, less than 30 exit holes and woodpecker attacks per m2 were found. If
dieback was less than 50%, the density of exit holes and woodpecker attacks was
between 30 and 60 per m2 (Figure 3.4). Dead trees were found with densities as low as
37 to 50 exit holes and woodpecker attacks per m2. Overall, on trees with more than 90%
dieback, the density of exit holes and woodpecker attacks ranged from 60 to 155 per m2
(Figure 3.4).

White vs. blue ash woodlot trees

The white ash trees sampled in these studies were heavily infested and declining
or dying. Blue ash trees of similar size and growing in the same site conditions had much
lower exit hole and woodpecker attack densities than white ash. In 2004, white ash trees
in the Plymouth Twp. woodlot had signiﬁcantly more A. planipennis exit holes and
woodpecker attacks per m2 than blue ash trees, with an average of 80.1 ($20.65 SE) for
white ash and 11.5 (i783 SE) for blue ash (S=16.0, df=1,8, p=0.0079). Exit holes made
up 70% of the average total A. planipennis density in white ash trees and woodpecker
attacks comprised the remaining 30%. On blue ash trees, exit holes accounted for 61% of
the average total density and woodpecker attacks made up the other 39%. Exit holes and
woodpecker attacks were found on branches and trunks of both species. Percentage of
phloem surface area covered by galleries ranged from 0 to 100% in white ash, with an

average of 39% (21:10.9 SE) and from 0 to 40% in blue ash, with an average of 8% (3:57

73

SE). Signiﬁcantly more phloem surface area was covered by galleries in white ash than
in blue ash (S=18.0, df=1,8,p=0.0278).

Similarly, in the Superior Twp. woodlot, white ash trees had signiﬁcantly more
exit holes and woodpecker attacks per m2 than blue ash in 2004 (S=30.0, df=1 ,7,
p=0.0079) and 2005 (S=10, df=1 ,6, p=0.0143) (Figure 3.5). There was a signiﬁcant
increase in the density of exit holes and woodpecker attacks from 2004 to 2005 for the
white ash trees in the Superior Twp. woodlot (S=26.0, df=1,6, p=0.0143) (Figure 3.5), but
not for the blue ash trees (S=22.0, df=l,7, p=0.36) (Figure 3.5). For the two trees of each
species that were felled and debarked in 2004, the percentage of phloem surface area
covered by galleries ranged from 25 to 100% in white ash, with an average of 67% ($5.9
SE) and from 0 to 40% in blue ash, with an average of 15% ($7.6 SE). In 2005,
percentage of phloem covered by galleries ranged from 0 to 100% in white ash, with an
average of 73% ($8.1 SE) and from 0 to 30% in blue ash, with an average of 7% ($5.0
SE). The percentage of phloem covered by galleries did not differ signiﬁcantly between
species in 2004 (S=7.0, df=1,2, p=0. l 7), but were signiﬁcant in 2005 (S=10.0, df=1,6,
p=0.0143). Throughout all surveys, a large portion of the galleries found on blue ash
trees had callous wood growing around the edges of the galleries. Little to no callous
wood was observed on white ash trees.

At all sites surveyed, more than 86% of the green, white and blue ash trees had at
least one woodpecker attack. The percentage of A. planipennis larvae killed by
woodpeckers on urban street trees ranged from 0 to 100%, with an average of 51% ($6.5
SE). This predation rate was similar for woodlot trees, where woodpecker attacks killed

0 to 100% of A. planipennis larvae, with an average mortality of 48% ($6.6 SE).

74

DISCUSSION

At all sites, in all years, green ash trees showed a greater percentage of canopy
dieback related to A. planipennis infestation than white ash trees. As expected, dieback
in both species increased from year to year as anew generation of A. planipennis larvae
began to feed under the bark each year and populations continued to build. The larvae
progressively consumed more phloem each year, further limiting translocation of water
and nutrients throughout the tree. While clearly high densities of phloem-borers are
associated with canopy dieback (Ball and Simmons 1980, Haack and Slansky 1986,
Miller et al. 1991 , others), canopy dieback is merely a symptom of infestation and not a
direct measure of larval density. The data that I collected from these trees represents the
cumulative density of A. planipennis larvae that successfully developed and emerged or
were preyed upon by woodpeckers as pre-pupal larvae. It does not reﬂect total larval
density which would include live and dead larvae remaining in the tree. In the ﬁeld, I
debarked several heavily infested trees that had few or no exit holes or woodpecker
attacks at eye level, yet 100% of the phloem was consumed by larvae. Further
examination showed that virtually all of the larvae had died by the second or third instar
because of intra-speciﬁc competition for phloem.

Some tree species may be more “tolerant” of phloem-boring insects and sustain
less canopy dieback than other species with similar infestation levels. Only a small
portion of the sapwood actively conducts water: the outermost one or two growth rings in
ring-porous hardwoods (e. g., ash, chestnut, and oak), and several of the outermost rings

in diffuse-porous hardwoods (e. g., aspen, birch, and maple) and conifers (Kozlowski and

75

Pallardy 1997, Kozlowski and Winget 1963). These patterns of water conduction explain
in part why ring-porous hardwoods are generally more vulnerable to wilt diseases (e. g.,
chestnut blight, Dutch elm disease, and oak wilt) (Kozlowski and Pallardy 1997) and
girdling insects (e. g., Agrilus bilineatus on oaks) (Haack and Benjamin 1982), than
diffuse-porous hardwoods or conifers. The nearly identical linear models I found for
green and white ash imply a similar relationship between canopy dieback and exit hole
and woodpecker attack density. This is intuitive as the more vascular tissue destroyed by
larval galleries, the less foliage and root area that can be supported. The model for green
ash indicated a negative y—intercept because it’s plausible that some canopy dieback
occurs in the absence of A. planipennis exit holes or woodpecker attacks. Smitley et al.
(2006) also reported a signiﬁcant relationship (R’=0.59) between canopy dieback and A.
planipennis exit hole density in ash trees they surveyed at 33 sites in southeast Michigan.
However, they only surveyed the trunks of the trees and, therefore, had a lower density of
exit holes relative to canopy dieback than reﬂected in my data. For example, in Smitley
et al.’s model, trees with less than 30% dieback had almost no exit holes, while in my
model, trees with 30% or less dieback had less than 30 exit holes and woodpecker attacks
per m2. They did not indicate whether they surveyed green ash, white ash, or’both.

One of the key questions these studies addressed is whether or not A. planipennis
exhibits a host preference hierarchy among the predominant ash species native to
Michigan. Results suggest that, when growing together, A. planipennis preferred green
ash street trees to white ash street trees in the core area of infestation in southeast
Michigan. McCullough et al. (2005) reported similar differences in canopy dieback and

larval densities for green and white ash control (untreated) trees monitored in a three-year

76

insecticide study. At a residential site in Washtenaw Co., they recorded over 60%
dieback on green ash trees and less than 10% dieback on white ash trees.

Results over this three year period also illustrate the progression of the infestation,
which increased in density each year in both species studied. Agrilus planipennis
densities in white ash increased as available phloem in green ash declined. Throughout
the study, the only exception to this trend was at Windemere, in 2005, where white ash
trees actually had greater exit hole and woodpecker attack densities than green ash.
Unlike the other sites that had fairly equal numbers of each species in the surrounding
area, Windemere was unique in that its street trees were predominantly white ash with an
occasional green ash tree mixed in with the white ash trees. Singer (1983) points out that
insects may not use their most highly preferred host if they encounter another acceptable
host ﬁrst.

Differences between species were not signiﬁcant in any year at Butzel for canopy
dieback or exit hole and woodpecker attack density. This was probably due to the
smaller tree size at this site. The majority of the trees at this site were less than 5 cm
DBH and A. planipennis may not show a preference between species for trees in small
size classes because of the limited phloem available in small trees. In addition, there was
a small decrease (four exit holes and woodpecker attacks per m2) in exit hole and
woodpecker attack densities in the white ash from 2003 to 2004 at Lamed and a rather
large decrease (40 exit holes and woodpecker attacks per m2) in the green ash from 2004
to 2005 at Windemere. The removal of dead trees likely accounted for the apparent

decrease in the overall mean A. planipennis densities at Windemere.

77

Agrilus planipennis also clearly preferred white ash trees over blue ash trees in
the two woodlots I surveyed. A large portion of the white ash trees in both woodlots had
already been killed by A. planipennis in 2002 and 2003. Densities calculated from the
dead trees I sampled at all residential and woodlot sites, therefore, suggest a maximum
number of A. planipennis per m2 that a single tree can successfully host was approx. 107
($10.8 SE) for green ash and 109 ($10.6 SE) for white ash.

It is interesting that a large portion of the galleries on blue ash trees had callous
wood around the edges or covering a substantial proportion of the galleries. It was also
evident that most larvae would have taken two years to develop on these trees. Cappaert
et al. (2005a) reported that A. planipennis could have a two-year life cycle in lightly to
moderately infested trees. Future studies are needed to address potential differences
among ash species in their phloem chemistry and their initial resistance to A. planipennis
larval feeding.

Previous studies have shown that woodpeckers can play an important role
regulating populations of wood-boring beetle larvae (Akers and Nielsen 1990b, Anderson
1944, Barter 1957, Fayt et al. 2005, Nash et al. 1951). An overall average of 33% of the
A. planipennis were preyed upon by woodpeckers in trees I surveyed. Anderson (1944)
reported that 10% of A. anxius borer larvae were preyed upon by woodpeckers
overwinter, while Akers and Nielsen (1990b) found that almost 30% of the total holes on
European white birch resulted from woodpecker predation of A. anxius larvae. Two
other studies by Barter (1957) and Nash et al. (1951) found predation rates on A. anxius
larvae ranged from very little to up to 50%. Studies on other natural enemies of A.

planipennis in Michigan reported less than 2% of larvae were infected with pathogenic

78

fungi, and less than 1% of larvae and 0.5% of eggs were parasitized by parasitoids (Bauer
et a1. 2006). In its native range in China, A. planipennis larval parasitism rates by
hymenopteran parasitoids were less than 7% (Liu et al. 2003). Woodpeckers are
currently the best natural enemy of A. planipennis in North America (Cappaert et al.
2005b, Fuester and Schaefer 2006, Marshall et al. 2006, Strazana 2006). More research
needs to be conducted on woodpecker abundance, behavior, and habitat use and their
effects on A. planipennis.

These studies were observational and used exit hole and woodpecker attack
densities to measure A. planipennis host preference. They serve as a stepping stone to
initiate additional studies to identify mechanisms that underlie this apparent preference.
Attack of woody plants by phloem-feeding Coleoptera, especially Buprestidae and
Scolytidae, is closely associated with the condition of the host tree (Anderson 1944,
Larsson et al. 1983, Dunn et al. 1986a, b, 1987, 1990d, Waring and Pitrnan 1980). Trees
weakened by drought (Mattson and Haack 1987), disease (Dunn et al. 1990c, Potter and
Hartman 1993), lightening strikes (Hodges and Pickard 1971), defoliation (Houston
1981, Wright et al. 1984), phloem and xylem girdling (Dunn et al. 1986a, Haack and
Benjamin 1982) or stand competition (Hard 1985) are most likely to be attacked and
killed by phloem-feeding borers. While A. planipennis is capable of attacking healthy
trees, it appears to prefer stressed trees (McCullough et al. 2005, 2006). The ﬁrst year I
surveyed Lamed, there was one white ash tree that had 75% canopy dieback while the
other white ash trees had little to none. While I could not ﬁnd physical evidence of stress
in the form of mechanical damage to the trunk, obvious soil compaction, etc., it was

apparent that this tree was more attractive to A. planipennis than the other white ash.

79

While stress is important in host selection of many phloem and wood-boring
insects, it is unlikely that stress alone is driving this differential attack on ash species
because it appears to be so widespread. Other factors, such as differences in foliar
chemistry among the ash species, could be inﬂuencing A. planipennis host preference.
Other trees species, such as birch (Betula sp.) (Laitinen et al. 2005, Taipale et al. 1994)
pine (Pinus sp.) (Hanover 1966), and willow (Salix sp.) (Orians et al. 1996) exhibit inter-
and intraspeciﬁc variation in the types and concentrations of secondary compounds found
in leaves and bark. Female A. planipennis feed on foliage for at least 14 days before they
oviposit and the species they feed upon may inﬂuence the species they select for
oviposition.

Bark texture could be another factor driving host preference. Larval densities on
rough-barked ash trees were signiﬁcantly higher than on smooth-barked ash trees of the
same species (see Chapter Two). Other studies on A grilus species have also reported a
preference for rough-barked portions of trees for oviposition (Barter 195 7, Loerch and
Cameron 1984). Many varieties of green ash tend to have rough bark before and aﬂer
they reach maturity, while most varieties of white ash have smooth bark before they
reach maturity (pers. obs.). Blue ash trees tend to have rough bark (pers. obs.), however,
so bark roughness should not be considered a sole factor inﬂuencing host selection.

A combination of many factors may affect A. planipennis host preference and
should not be considered mutually exclusive. Controlled studies need to be designed to
increase our understanding of host selection behavior. More work is needed to
investigate differences in ash species in relation to factors such as foliar chemistry and

nutrition, phenolic concentrations and other chemicals in the bark and phloem, bark

80

texture, and stress-inducing variables such as drought, girdling, disease, and
transplanting. Future studies should also be designed to evaluate the relative preference
of other ash species, including black ash.

These surveys of A. planipennis host preference provide baseline data that can be
used to more thoroughly assess the susceptibility of both urban and forest trees based on
species composition and to increase efﬁciency in developing survey, detection, and
management options. For example, in areas where mixes of ash species occur, efforts
should focus on surveying the preferred species (i.e., green ash) when trying to detect
newly established or low density populations of A. planipennis. These data also support
regulations in other states that promote using green or white ash (versus the less preferred

blue ash) for trap trees to detect new infestations.

81

Table 3.1. Number of trees (11) and mean ($ SE) diameter at breast height (DBH) for
green, white, and blue ash trees at survey sites in A grilus planipennis- infested areas.

 

 

 

 

Green Ashl White Ash1
Site 11 DBH (cm) 11 DBH (cm) Survey year(s)
Butzel 8 8.1 $ 1.2 8 4.2 $ 0.2 2003, 2004, 2005
Larned 11 18.0 $ 0.5 11 23.3 $ 0.7 2003, 2004, 2005
Windemere 7 27.7 $ 2.5 7 23.3 $ 1.3 2004, 2005
Lakeside 8 32.0 $ 5.3 9 33.0 $ 4.7 2005
Blue Ash2 White Ash2
Superior Twp. 5 20.1 $ 3.9 4 20.6 $ 2.4 2004
Plymouth Twp. 5 22.2 $ 3.7 5 24.2 $ 4.6 2004
Superior Twp. 4 21.2 $ 2.4 4 23.4 $ 4.0 2005

 

lThe same trees were re-surveyed each year.
2Trees were felled for sampling and new trees were selected each year.

82

 

 

 

 

     

 

 

 

 

 

 

 

 

   

 

a)

100
g 80 - I Green ashﬂ
g D White ash
a so — ~-— . e
.2
'U
>. 40 -
a.
o
g 20 -

0 _
2003 2004 2005
n=8GR. 8WH n=BGR, BWH n=86R, 8WH

b) 120

10° « ".9913.
"E 80 4 I Exits}
3
“g 60 -
'I-
ﬁ 40 -
I“

20 -

0 ‘ i i l I I‘ l I
2003 2004 2005
n=BGR. BWH n=GGR TWH n=5GR. 5WH

 

 

 

Figure 3.1 a) Mean ($ SE) percent canopy dieback and b) mean ($ SE) number of
Agrilus planipennis exits holes (exits) and woodpecker attacks (WPs) per m2 for the small
green (GR) and white (WH) ash trees at Butzel in 2003, 2004, and 2005. Three dead
green ash and three dead white ash trees with 100% dieback were removed in 2004 and
2005. Means within years were not signiﬁcantly different between species (Wilcoxon
Rank Sum test; p<0.05).

83

 

 

 
 
     

 

 

 

 

 

 

 

a) 100
*
3 80 ll Green ash
°v E] White ash I
x _____._—J
E 60 ~
.2
O
>. 40 -
a
o
5
U 20 .
0 .
2003 2004 2005
n=11GR,11WH n=11GR,11WH n=11GR.11WH
bl 140
120 .
“E 100 *
g 80 <
3 *
E 60 . IWPS
“>5 40 , IEXIts
20 <
O .
2003 2004 2005
n=I1GR. 11WH n=TGR11WH n=TGR,11WH

 

 

 

Figure 3.2 a) Mean ($ SE) percent canopy dieback and b) mean ($ SE) number of
Agrilus planipennis exits holes (exits) and woodpecker attacks (WPs) per m2 for green
(GR) and white (WH) ash trees at Larned in 2003, 2004, and 2005. Four dead green ash
trees with 100% dieback were removed in 2004. * Indicates signiﬁcant differences
between green and white ash trees (Wilcoxon Rank Sum test; p<0.05).

84

 

100
90 < *
8O -
70
60 a
50 -
40 <
30 <
20 ~

 

  
 

 

TGreen ash
it] White ashj

Canopy Dieback (%)

 

 

 

 

 

2004 2005
n=7GR, 7WH n=7GR. 7WH

 

E

180
160 « *
140 <
120 —
100 4
80 -

60 <
40 .
20

0

2604 2005
n=7GR. 7WH n=SGR. 5W“

 

Exits + WPs I m’

 

 

 

 

 

 

Figure 3.3 a) Mean ($ SE) percent canopy dieback and b) mean ($ SE) number of
Agrilus planipennis exits holes (exits) and woodpecker attacks (WPs) per m2 for green
(GR) and white (WH) ash trees at Windemere in 2004 and 2005. Two dead green ash
trees with 100% dieback were removed in 2005. One outlier for exits and woodpecks per
m2 was removed from the 2005 white ash tree data set. * Indicates signiﬁcant differences
between green and white ash trees (Wilcoxon Rank Sum test; p<0.05).

85

 

 

160

 

 

 

 

   
 
    

 

 

 

o
0 Green ash '
140 ’ I White ash
120 , Green ash
y = .1.416 + 0.897x l
R2 = 0.6147 0
"E 100 . ' o
C White ash I
& y = 4.413 + 0.812x
3 80 R2 = 0.656
+
8 o
31 Combined
l“ 5° ‘ y = 2.863 + 0.846x . o
R‘ = 0.703 .
‘° " o
20 .
I
0 . . . . .
0 20 40 60 so 100
Canopy Dieback (%)

Figure 3.4. Percent canopy dieback and number of Agrilus planipennis exit holes (exits)
and woodpecker attacks (WPs) per m2 for 25 green ash and 24 white ash trees (49 trees
total) at three sites in the Agrilus planipennis-infested areas.

86

 

160

 

 

140 - IWPs
120 , * IExits
100 -
sol .
60 4
40 J
201

 

 

 

Exits + WPs I m'

         

W'lur'

_1. . . .‘yu
"AJA‘A'

White ash Blue ash | White ash Blue ash
; 2004 l 2005

 

 

 

 

 

Figure 3.5. Mean ($ SE) number of Agrilus planipennis exit holes (exits) and
woodpecker attacks (WPs) per m2 for white and blue ash trees at Superior Twp in 2004
and 2005. "‘ Indicates signiﬁcant differences between white and blue ash trees (Wilcoxon
Rank Sum test; p<0.05).

87

APPENDICES

88

APPENDIX 1
Record of Deposition of Voucher Specimens*
The specimens listed on the following sheet(s) have been deposited in the named
museum(s) as samples of those species or other taxa, which were used in this research.

Voucher recognition labels bearing the Voucher No. have been attached or included in
ﬂuid-preserved specimens. -

Voucher No.: 2006-05

Title of thesis or dissertation (or other research projects):

HOST RANGE AND PREFERENCE OF THE EMERALD ASH BORER, AGRILUS
PLANIPENNIS FAIRMAIRE (COLEOPTERA: BUPRESTIDAE), IN NORTH
AMERICA

Museum(s) where deposited and abbreviations for table on following sheets:
Entomology Museum, Michigan State University (MSU)

Other Museums:

Investigator’s Name(s) (typed)
Andrea Christine Anulewicz

 

 

Date 13 July 2006

*Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in North
America.
Bull. Entomol. Soc. Amer. 24: 141-42.

Deposit as follows:
Original: Include as Appendix 1 in ribbon copy of thesis or dissertation.

Copies: Include as Appendix 1 in copies of thesis or dissertation.
Museum(s) ﬁles.

Research project ﬁles.

This form is available from and the Voucher No. is assigned by the Curator, Michigan
State University Entomology Museum.

89

 

8-37m— 8me

 

 

>380ch 88m 5&222 2: E 58%
Low 2083on was: gone 05 3230M 83235.. 2.535 865x

 

 

moéoom .Z 8:025 $598 $3.52 {Sawzmgﬁ
c0888: .z 38% 1:82an 33

 

APPENDIX 1.]
Lof_1Pages

Voucher Specimen Data
Page

:3
m
2

Ego—=54. .U.< .36
8-30::
3-5.78

awo— was

5% :3 SE? was 58m Set cameo—cm

S€< :c< mo .05
wee—v.30 _3_:Som 62332
.00 328:3? “26550—2 oummcEam matnmkisﬁ §~2MV

 

deposhed
Other

Museum

where

 

uozmoaoo new vow:

co .5 o 8 883
8 386:8 wee—58% .8 Set :53 x8 5 . m

 

Adults 6‘ °°
Adults 92 "
Larvae

Pupae
Nymphs
Eggs

 

 

 

 

 

 

 

 

$-

ao 3:52

 

 

90

APPENDIX 2

Leaf-Feeding Preference of Agrilus planipennis (Fairmaire) (Coleoptera:

Buprestidae) on Four North American Ash Species

This appendix contains the methods and results of a study designed to evaluate
A grilus planipennis adult leaf-feeding preference among four North American species of
ash using two-choice bioassays. I felt that this study didn’t quite ﬁt in with the contents
of the other three chapters of this thesis and lacked substantial, deﬁnitive, or conclusive
results, but still warranted reporting in some format.

2004 bioassays: Foliage from four different species of ash was collected on each
of three set up days of the experiment: 22 July, 29 July, and 5 August 2004. White and
black ash leaves were collected from trees at Elm Park in Livonia, Wayne Co., MI. Blue
ash leaves were collected from a private woodlot in Superior Township, Washtenaw Co.,
MI and green ash leaves from Matthaei Botanical Gardens in Ann Arbor, Washtenaw
Co., MI. All trees were untreated, in the core area of A. planipennis infestation, and
showed varying degrees of foliar feeding.

Leaves randomly selected for the study were photocopied before the start of each
bioassay to capture, in an image, the amount of leaf area present. Two leaves of a single
species were paired and stems submerged in a vial of water to prevent desiccation. Each
vial contained three cotton wicks to stabilize the leaves and provide drinking water for
the beetles. Two vials, each with a different leaf species, were placed in opposite ends of

clear, plastic boxes (31 x 23 x 10 cm). Boxes were kept on a lab bench at 24-27°C.

91

Adult beetles were reared from bolts of infested ash trees collected in the core
area of A. planipennis infestation. Rearing logs were kept in cold storage at 1-2°C with a
minimum of 80% humidity until needed. Six newly emerged, unfed beetles were
released from the center of each box and allowed to feed undisturbed for ﬁve days. Six
pair-wise combinations of species were used: 1) green and white ash, 2) green and black
ash, 3) green and blue ash, 4) white and black ash, 5) white and blue ash, and 6) black
and blue ash. Bioassays were initiated on 22 July, 29 July, and 5 August 2004 as beetles
emerged from their logs. Each bioassay contained three to four replicates (12 replicates
total). Fresh foliage was collected on each day that the bioassays were initiated.

At the end of each ﬁve-day bioassay, leaves were removed from their boxes and
photocopied a second time. Before and after feeding images of leaves were scanned and
analyzed using WinFolia 2004a Leaf Analysis Software by Regent Instruments, Inc. The
amount of leaf area consumed was estimated by calculating the difference in leaf area
before and after the bioassay. Because there were two leaves of each species in the box,
total leaf area consumed for each species was calculated by combining the leaf area
consumed on each of the two leaves in the box, generating one value for each pair of
leaves in the box. The proportion of total leaﬂets fed upon was also recorded. Three
classes of feeding behavior were designated: ‘marginal’, ‘to the vein’, and ‘through the
vein’. ‘Marginal’ feeding described any feeding that took place along the margin of the
leaﬂet, less than halfway to the main vein of the leaﬂet. ‘To the vein’ feeding described
any feeding that occurred more than halfway to the main vein of the leaﬂet and ‘through
the vein’ feeding was any feeding that included the main vein of the leaﬂet. The

proportion of leaﬂets within each feeding class was calculated by combining the total

92

number of leaﬂets on the pair of leaves in each box that exhibited the feeding class of
interest and dividing it by the total number of leaﬂets in the pair of leaves.

Data were tested for normality using the Shapiro-Wilk test (Shapiro and Wilk 1965) and
residual plots. Data were not normal and could not be normalized with transformations.
Friedman’s nonparametric two-way analysis of variance (ANOVA) was used, therefore,
to determine signiﬁcant interactions between species and the pair-wise combination
factors (SAS Institute, Inc. 1989). The nonparametric Kruskal-Wallis test was then used
to determine if total leaf area consumed and proportion of leaﬂets fed upon within each
feeding category differed among species (Kruskal and Wallis 1952). When signiﬁcant,
nonparametric multiple comparisons (p<0.05) were used to identify differences among
species (Zar 1984, Conover 1971).

Differences in leaf area consumed among species were analyzed without the pair-
wise combination factor because there was not a signiﬁcant interaction between the two
effects (F =0.42, df=3, p=0.74). Average leaf area consumed was signiﬁcantly greater on
green ash, white ash, and black ash than on blue ash (X2: 25.35, df=3,140, p<0.0001)
(Figure A21).

The average proportion of leaﬂets that had one or more types of feeding was 0.71
($0.03 SE) for green ash, 0.69 ($0.03 SE) for black ash, 0.65 ($0.03 SE) for blue ash, and
0.63 ($0.03 SE) for white ash. Overall, the proportion of leaﬂets feeding in general did
not differ signiﬁcant among species (x’= 2.40, df=3,140, p=0.49). Within the ‘marginal’
feeding class, blue ash had a signiﬁcantly greater proportion of leaﬂets with ‘marginal’
feeding than green or white ash (x2= 19.69, df=3,140, p=0.0002) (Figure A2.2). Green

and black ash had signiﬁcantly greater proportions of leaﬂets with feeding more than half

93

way to the main vein than blue ash ()6: 50.93, df=3,140, p<0.0001) (Figure A2.2). Blue
ash also had a signiﬁcantly lower proportion of leaﬂets with feeding through the main
vein than green, white, and black ash ()f= 27.29, df=3,140, p<0.0001) (Figure A2.2).
‘Marginal’ feeding was the most common type of feeding and ‘through vein’ was the
least common. 0

2005 bioassays: This study was repeated again in 2005 using the same methods as
in 2004. Leaves from the lower canopy exposed to direct sunlight of four different
species of ash were collected on each of ﬁve set up days of the experiment: 28 June, 30
July, and 1 August 2005. Green and white ash leaves were collected from trees at Elm
Park. Blue ash leaves were collected from a tree in Plymouth Township, Wayne Co., MI,
and black ash leaves from Hudson Mills Metropark, Washtenaw Co., MI. All trees were
untreated, in the core area of A. planipennis infestation, and showed varying degrees of
foliar feeding.

Leaves were photocopied before the start of each bioassay. This time, only one
leaf of each species was selected and stems were submerged in a vial of water to prevent
desiccation. Adult beetles were reared from infested ash logs using the same methods as
in 2004. Six newly emerged, unfed beetles were released from the center of each box and
allowed to feed undisturbed for ﬁve days. The same six possible pair-wise combinations
of species were used. Bioassays were initiated on 28 June, 18 July, and 1 August 2005 as
beetles emerged from their logs. Each bioassay contained four to seven replicates (15
replicates total). Fresh foliage was collected on each day that bioassays were initiated.

At the end of each ﬁve-day bioassay, leaves were removed from their boxes and

photocopied again. Before and after feeding images of leaves were scanned and analyzed

94

using WinFolia Leaf Analysis Software. The total amount of leaf area consumed was
estimated by calculating the difference in leaf area before and after the bioassay. The
proportion of leaﬂets within each of the three feeding classes described in 2004 was
calculated. Data were analyzed using the same methods as in 2004.

Similar to the 2004 analysis, there was not a signiﬁcant interaction between
species and pair-wise combination, therefore, differences in leaf area consumed among
species were analyzed independently of pair-wise combination (F =0.42, df=3, p=0.74).
Signiﬁcantly more leaf area was consumed on white ash than on green or blue ash ()6:
38.75, df=3,146,p<0.0001) (Figure A23).

The average proportion of leaﬂets that had one or more types of feeding was 1.0
($0 SE) for black and blue ash, 0.96 ($0.02 SE) for white ash, and 0.93 ($0.02 SE) for
green ash. Overall, black and blue ash had signiﬁcantly greater proportions of leaﬂets
with feeding in general than green ash (x’= 30.44, df=3,176, p<0.0001). However, it
should be noted that for the 2005 bioassay, the majority of the black and blue ash leaves
were made up of only two to three leaﬂets, while green and white ash leaves had ﬁve to
seven. Similar to results in 2004, ‘marginal’ feeding was the most common type of
feeding pattern and ‘through vein’ feeding was the least common. Within the ‘marginal’
feeding class, black and blue ash had signiﬁcantly greater proportions of leaﬂets with
‘marginal’ feeding than green ash ()6: 30.44, df=3,176, p<0.0001) (Figure A2.4). White
and black ash had signiﬁcantly greater proportions of leaﬂets with feeding more than half
way to the main vein than blue ash (x): 14.41, df=3,176, p=0.0024) (Figure A2.4). Blue
ash also had a signiﬁcantly lower proportion of leaﬂets with feeding through the main

vein than green and white ash (x2: 26.62, df=3,176, p<0.0001) (Figure A24).

95

Data from both years suggest blue ash is least preferred for leaf feeding and

feeding is mostly marginal on blue ash.

96

 

Leaf area consumed (cm’)

I“
o
1

 

 

0.0 -
Green ash White ash Black ash Blue ash

Figure A2.1. Mean ($ SE) leaf area consumed for four ash species by Agrilus
planipennis beetles in ﬁve days during a two-choice bioassay in 2004. n=36 replicates of
six beetles feeding on one leaf for each species. Means with the same letter are not
signiﬁcantly different (Kruskal-Wallis test and nonparametric multiple comparison
procedure; p<0.05).

97

.0
‘1
1
l

 

a
0 6 . Fl— 1I Green ashl
‘ ab in White ash
b jEl Black ash
% 0.5 < b V j lDBlue ashmj
I:
8
z 0.4
0
S
'E 0.3
O
D.
O
5 0.2 -
0.1
0.0 -

 

 

 

 

 

 

 

Marginal To vein Through vein

Figure A2.2. Mean ($ SE) proportion of leaﬂets of four species of ash fed on by Agrilus
planipennis beetles in each of four feeding categories in ﬁve days during a two-choice
bioassay in 2004. n=36 replicates of six beetles feeding on one leaf for each species.
Within feeding types, means with the same letter are not signiﬁcantly different (Kruskal-
Wallis test and nonparametric multiple comparison procedure; p<0.05).

98

25.0

 

 

 

a

a" 20.0 ~
E
3
In ab
0:
E 15.0 ~
:3 b
in
r:
8
a 10.0 - °
2
a
Ii-
8
_I 5.0 -

0.0 -

Green ash White ash Black ash Blue ash

Figure A2.3. Mean ($ SE) leaf area consumed for four ash species by Agrilus
planipennis beetles in five days during a two-choice bioassay in 2005. n=29 replicates of
six beetles feeding on one leaf for green ash, n=4] for white ash, n=40 for black ash,
n=40 for blue ash. Means with the same letter are not signiﬁcantly different (Kruskal-
Wallis test and nonparametric multiple comparison procedure; p<0.05).

99

 

a a J Greene‘s—h]

 

1.01 b ab,._ jIWhiteash;

IE1 Black ash ‘
a a Cl Blue ash

.0
on

Proportion of leaﬂets
.0 .0
h 0)

0.2 <

 

 

 

 

 

 

0.0 <

 

Marginal To vein Thru vein

Figure A2.4. Mean ($ SE) proportion of leaﬂets of four species of ash fed on by Agrilus
planipennis beetles in each of four feeding categories in ﬁve days during a two-choice
bioassay in 2005. n=45 replicates of six beetles feeding on one leaf for each species.
Within feeding types, means with the same letter are not signiﬁcantly different (Kruskal-
Wallis test and nonparametric multiple comparison procedure; p<0.05).

100

APPENDIX 3

Unsuccessful No-Choice and Two-Choice Laboratory Bioassays

When working with new, exotic insects, many difﬁculties arise that can impede
productive research. The insect rearing process can be full of challenges from selecting
the appropriate diet to keeping the rearing system free of pathogens. In 2005, there was a
problem in our beetle rearing process that resulted in high mortality when beetles reached
approx. two weeks of age and reduced fecundity in beetles that lived long enough to be
used in bioassays. A nutrient deﬁciency in the feeding foliage was speculated early in the
process and was immediately addressed, but an unidentiﬁed and unconﬁrmed pathogen
may have contaminated the system, invalidating any of the 2005 oviposition bioassays.
While the results of these studies are unreliable and inconclusive, I feel that the methods
are still valid for use in future studies, therefore they are presented in this appendix.
Two-choice cut branch bioassays

2004 bioassays: Green and white ash bolts were harvested from Michigan State
Universities W.K. Kellogg Forest, Kalamazoo Co., MI. Black ash bolts were collected
from Kensington Metropark, Oakland Co., MI and blue ash bolts from the same private
woodlot in Superior Township were blue ash foliage was collected for leaf-feeding
bioassays. All bolts were harvested and waxed to slow desiccation on 24 May 2004 and
kept in cold storage at 1-2°C with a minimum of 80% humidity.

At the start of each bioassay, branches of each species were cut to approx. 20 cm
in length (Table A3.l). All cut surfaces were waxed with parafﬁn. Two branches, each a

different species of ash, and one green ash leaf collected from an untreated, infested

101

green ash were placed in opposing comers of a screened cage (30 x 30 x 30 cm). Ash
leaf stems were submerged in vials of water to slow desiccation. Each vial contained two
cotton wicks to provide drinking water for the beetles. Cages were kept outside in a
screened-in greenhouse, leaving them exposed to ambient temperatures and wind
conditions, but protected from rain.

Adult beetles were reared from bolts of infested ash trees collected in the core
area of A. planipennis infestation using the same methods described in the leaf-feeding
bioassays. Once emerged from rearing logs, similarly aged groups of beetles (approx.
200) were placed in screen cages (60 x 60 x 60 cm) and feed green ash foliage collected
from untreated, infested trees. Cages were kept in growth chambers at 24°C, 60%
humidity, and 16:8 lightzdark photoperiod. Beetles were allowed to feed and mate for
two weeks before bioassays began.

The same six pair-wise combinations of ash species used in the leaf-feeding
bioassay were used again here. Two bioassays were set up on 1 and 28 June 2004. On
each set up day, beetles were sexed and one male/female pair was placed in each box (60
pairs total). Each bioassay had ﬁve replicates (10 replicates total). Boxes were checked
twice a week and foliage was replaced weekly. Twenty-six of the 60 pairs were observed
mating. Adults lived from two to 49 days in the cages. If a female beetle died during the
ﬁrst three days of the bioassay, she was replaced with a similarly aged beetle. The ﬁrst
bioassay was conducted from 1 June to 27 June and the second from 28 June to 16
August. By the third day of the second bioassay, twenty-seven of the 60 beetles in the
bioassay had died. It was suspected that the unusually high mortality may have been

caused by bleach residue left on the cages during cleaning after the ﬁrst bioassay. On 1

102

July, beetles were removed from their cages and placed into Petri dishes. The cages were
washed with soapy water, rinsed thoroughly with water, dried and set up again. Dead
beetles were replaced with similarly aged beetles. Beetle mortality returned to expected
levels.

Upon death of the female beetle, the branches were removed and stored on a lab
bench at 24°C. The entire surface of each branch was inspected for eggs, larvae, and
galleries 30-45 days after it was removed from its cage. Diameter of each branch was
measure and total bark surface area was calculated (Table A3.l). The bark surface of
each branch was ﬁrst inspected with a magnifying lens to locate eggs on the surface, then
eggs were removed with forceps to prevent them from being recounted. Forceps or a
small knife were then used to chip off bark ﬂakes to revel eggs hidden in bark layers.
After 15 min of searching, total number of eggs was recorded. Bark was carefully peeled
down to the wood using a small drawknife or chisel. Number and stage of the larvae on
each branch was recorded.

A total of 258 eggs were found on 52 of the 120 branches with an average of 2.2
($ 0.4 SE) eggs per branch. This low number of eggs may indicate that branches in this
type of setting are not as attractive to females for oviposition. A total of 70 small
galleries excavated by ﬁrst instar larvae were found on 15 of the 120 branches, averaging
0.6 ($ 0.3 SE) galleries per branch. This low number of galleries may suggest that cut
branches will desiccate quickly and lose the necessary moisture content to support early
instar development.

2005 bioassays: Green and white ash branches were collected on 9 May and 13

June 2005 from the same locations as in the 2004 bioassay. Black ash branches were

103

harvested from the Nan Weston Nature Preserve at Sharon Hollow, Washtenaw Co., M1
on 12 May and 13 June 2005. Blue ash branches were also collected from the Nan
Weston Nature Preserve on 12 May and the Strait Creek Nature Preserve in Adams Co.,
OH on 10 June 2005. Cut ends were waxed and branches remained in cold storage at l-
2°C. At the start of each bioassay, branches of each species were cut to approx. 20 cm in
length (Table A3.l).

Two branches and one wooden dowel (1.9 cm diam. x 18 cm height) were placed
in random order in the left, middle or right positions in a clear, plastic box (31 x 23 x 10
cm). In order to slow branch desiccation, boxes were kept in growth chambers at 24°C,
60% humidity, and 16:8 lightzdark photoperiod. Each branch and dowel was wrapped
with thin, white, plastic Easy Gardener® Tree Wrap 513. The wrap was cut into 2.5 cm
wide strips and wrapped around the branch or dowel leaving 3-4 cm of exposed branch in
between the strips of tree wrap. This was done to provide additional crevices for egg
laying.

Adult beetles were reared from bolts of infested ash using the same methods as in
2004. Once emerged from rearing logs, similarly aged groups of beetles (approx. 20)
were placed in clear, plastic cylinders (10 x 20 cm) and fed foliage collected from
infested potted green ash saplings. Beetles were allowed to feed and mate for two weeks
before bioassays began. Approx. 35 holes were drilled into the lids and sides of the
cylinders to allow for ventilation.

The same six pair-wise combinations of ash species used in the 2004 bioassays
were initiated on 16 May and 20 June 2005. On each day that bioassays were inititated,

beetles were sexed and one male/female pair was placed in each box (32 pairs total).

104

Each bioassay had three replicates (6 replicates total). Boxes were checked twice a week
and foliage was replaced weekly. Eleven of the 32 pairs were observed mating. Adults
lived from three to 28 days in the boxes. If a female beetle died during the ﬁrst three
days of the bioassay, she was replaced with a similarly aged beetle. The ﬁrst bioassay
was conducted from 16 May to 13 June and the second from 20 June to 18 July. Upon
death of the female beetle, the branches were removed and stored on a lab bench at 24°C.
The entire surface of each branch was inspected for eggs, larvae, and galleries approx. 28
days after it was removed from its cage. Diameter of each branch was measure and total
bark surface area was calculated (Table A3. 1). InspeCtion and peeling methods were as
described in 2004.

A total of 112 eggs were found on 41 of the 72 branches with an average of 1.6 ($
0.2 SE) eggs per branch. A total of 61 small galleries excavated by ﬁrst instar larvae
were found on 13 of the 72 branches, averaging 1.3 ($ 0.4 SE) galleries per branch.

No-choice cut branch bioassays: In 2005, branches of green ash, white ash,
black ash, blue ash, privet, and tree lilac, along with two Asian elm species, Ulmus
davidiana Planch. and U. japonica (Sarg. ex Rehder.) Sarg.. These species were
previously reported as host for A. planipennis in Asia (Akiyama and Ohmomo 1997,
Sugiura 1999). Green ash, white ash, and Japanese tree lilac branches were collected
from the same locations as in the 2004 bioassays. Green and white ash branches were cut
on 9 May and 13 June 2005. Japanese tree lilac branches were cut on 17 May and17 June
2005. Black ash branches were harvested from the Nan Weston Nature Preserve at
Sharon Hollow, Washtenaw Co., M1 on 12 May and 13 June 2005. Blue ash branches

were also collected from the Nan Weston Nature Preserve on 12 May and the Strait

105

Creek Nature Preserve in Adams Co., OH on 10 June 2005. Both Asian elm species
were harvested from the Morton Arboretum Ware Research Field, Dupage Co., IL on 16
May and 27 June 2005 and sent by overnight mail to Michigan. Privet branches were
collected from Ann Arbor, Washtenaw Co, M1 on 18 May and 21 June 2005. Cut ends of
branches were waxed and remained in cold storage at 1-2° C until the start of each
bioassay. Eight branches of each species were cut to approx. 17 cm in length (Table
B.2). Cut ends were waxed with parafﬁn.

Boxes were set up and beetles were reared using the same methods as in 2004.
Once beetles were emerged from rearing logs, similarly aged groups of beetles (approx.
20) were placed in clear, plastic cylinders (10 x 20 cm) and fed foliage collected from
uninfested potted green ash saplings. Beetles were allowed to feed and mate for two
weeks before bioassays began. Approximately 35 holes were drilled into the lids and
sides of the cylinders to allow for ventilation.

One each day bioassays were initiated, beetles were sexed and one male/female
pair was placed in each box (64 pairs total). The two bioassays each contained four
replicates of each species and the ﬁrst ran from 18 May to 10 June and the second from
29 June to 20 July 2005. Boxes were kept in growth chambers at 24°C, 60% humidity,
and 16:8 lightzdark photoperiod. If a female beetle died during the ﬁrst three days of the
bioassay, she was replaced with a similarly aged beetle. Adults lived from three to 23
days in the boxes. Upon death of the female beetle, the branch was removed from the
box and stored in growth chambers under the above conditions.

The entire surface of each branch was inspected for eggs and bark was peeled to

assess larval feeding 28-32 days aﬁer it was removed from its box. Diameter of each

106

branch was measured and total bark surface area was calculated (Table A3.2). Bark was
carefully peeled down to the wood using a small knife or chisel and diameter of each
branch was measured. Number and stage of larvae on each branch section and a visual
estimate of the percentage of the cambium/phloem area covered by larval galleries was
recorded. Galleries were also standardized 100 cm”.

I only found 101 eggs on 15 of the 84 branches used in this bioassay, averaging
1.2 ($0.6 SE) eggs per branch. A total of 57 galleries were excavated on eight branches,
averaging 0.8 ($0.4 SE) galleries per branch.

No-choice caged stern bioassay: The study was a repeat of the caged stem
bioassay reported in Chapter 1 using the same four tree species, green ash, white ash,
black walnut, and Japanese tree lilac, purchased from Poplar Farms Nursery in
Waterman, IL. Ten blocks of trees were set up in a randomized complete block design
(40 trees total). Trees were planted in an open ﬁeld slightly above ground, 3 m apart, in
ﬁves rows of eight trees. The partially exposed root balls were covered with wood mulch
and drip line irrigation was installed for frequent watering. Root balls from the 2004
study were removed and new trees were put in the existing holes on 13 May 2005. Trees
had an average caliper of 7.1 ($0.2 SE) cm.

Screen cages were constructed using the same methods as in 2004. However,
prior to constructing the cages, the stem of each tree was wrapped with thin, white,
plastic Easy Gardener® Tree Wrap 513. The wrap was ﬁrst cut into 2.5 cm wide strips
and then wrapped around the stem leaving 4-5 cm of exposed bark in between the strips
of tree wrap. This was done to provide additional crevices for egg laying. Instead of

using foliage stems for feeding, small potted green ash seedlings were placed in each

107

cage and were watered twice a week. For detailed methods of cage construction, refer to
Chapter 1.

Because of the reduced fecundity and longevity of the beetles used in this
bioassay, I felt that no useful data would come from peeling these stems. Therefore, they

were left in place for use in future studies.

108

Table A3.]. Mean ($SE) diameter, length, and surface area for each species used in the
2004 and 2005 two-choice oviposition bioassay. n=15 for each species.

 

Species Diameter (cm) Length (cm) Surface area (cm’)

 

2004 Bioassay n=30 branches per species

Green ash 7.8 (0.2) 20.3 (0.1) 439.2 (11.6)
White ash 8.1 (0.2) 20.0 (0.1) 468.4 (8.3)

Black ash 7.4 (0.3) 19.9 (0.1) 424.5 (18.8)
Blue ash 8.9 (0.2) 20.1 (0.1) 506.1 (12.8)

2005 Bioassay n=27 branches per species

Green ash 6.5 (0.2) 17.1 (0.1) 305.4 (7.2)
White ash 7.2 (0.1) 17.0 (0.1) 338.7 (5.1)
Black ash 6.9 (0.1) 16.8 (0.1) 319.2 (9.4)
Blue ash 6.2 (0.1) 16.9 (0.1) 287.5 (4.9)

 

109

Table A3.2. Mean ($SE) diameter, length, and surface area for each species used in the
2005 no-choice cut branch bioassays. n=8 for each species.

 

 

Species Diameter (cm) Length (cm) Surface area (cm’)
Green ash 6.9 (0.4) 16.4 (0.4) 310.0 (18.6)

White ash 7.1 (0.4) 16.5 (0.2) 323.8 (18.7)

Black ash 6.5 (0.2) 16.6 (0.2) 301.4 (9.0)

Blue ash 7.8 (0.1) 16.8 (0.2) 364.3 (3.7)

Privet 4.8 (0.3) 16.9 (0.2) 230.5 (14.7)

Tree lilac 5.7 (0.3) 16.9 (0.2) 268.9 (10.9)

Ulmus davidiana 5.3 (0.3) 17.3 (0.2) 244.1 (11.5)

U. japonica 5.3 (0.4) 17.4 (0.3) 248.9 (19.9)

 

110

LITERATURE CITED

111

LITERATURE CITED

Akers, R. C., and D. G. Nielsen. 1990a. Reproductive biology of the bronze birch borer

(Coleoptera: Buprestidae) on selected trees. Journal of Entomological Science 25:
196-203.

Akers, R. C., and D. G. Nielsen. 1990b. Spatial emergence patterns of bronze birch
borer, (Coleoptera: Buprestidae) from European white birch. Journal of
Entomological Science 25: 150-157.

Akiyama, K., and S. Ohmomo. 1997. A checklist of the Japanese Buprestidae. Gekkan-
Mushi, Supplement 1. 67 pp.

Akiyama, K., and S. Ohmomo. 2000. The Buprestid Beetles of the World. Iconographic
Series of Insects 4. Gekkan-Mushi Co., Ltd. 341 pp.

Anderson, R. F. 1944. The relation between host condition and attacks by the bronzed
birch borer. Journal of Economic Entomology 37: 588-596.

Ball, J., and G. Simmons. 1980. The relationship between bronze birch borer and birch
dieback. Journal of Arboriculture 6: 309-314.

Banno, H., and A. Yamagami. 1991. Life cycle and larval survival rate of the redspotted
longicom beetle, Eupromus ruber (Dalman) (Coleoptera: Cerambycidae). Applied
Entomology and Zoology 26: 195-204.

Barnes, B. V., and W. H. W. Jr. 1981. Michigan Trees: A Guide to the Trees of
Michigan and the Great Lakes Region. The University of Michigan Press, Ann
Arbor.

Barron, A. B. 2001. The life and death of Hopkins' host-selection principle. Journal of
Insect Behavior 14: 725-737.

Barter, G. W. 1957. Studies of the bronze birch borer, Agrilus anxius Gory, in New
Brunswick. Canadian Entomologist 89: 12-36.

Bauer, L. S., H. Liu, R. Gao, and T. Zhao. 2006. Egg and larval parasitoids of emerald
ash borer from China: potential for biological control in North America, pp. 48-
49. In Mastro et al. (eds), Emerald Ash Borer Research and Technology
Development Meeting, Pittsburgh, PA, 26—27 Sept. 2005. U.S. Department of
Agriculture, Forest Service publication FHTET-2005-16, Morgantown, WV.

Bauer, L. S., R. A. Haack, D. L. Miller, T. R. Petrice, and H. Liu. 2004. Emerald ash
borer life cycle, p. 8. In Mastro and Reardon (eds), Emerald Ash Borer Research
and Technology Development Meeting, Port Huron, MI, 30 Sept.-1 Oct. 2003.

112

U.S. Department of Agriculture, Forest Service publication FHTET-2004-02,
Morgantown, WV.

Bray, A. M., L. S. Bauer, R. A. Haack, and J. J. Smith. 2005. Genetic analysis of
emerald ash borer to determine the point of origin of Michigan infestations, pp.
17-18. In Mastro and Reardon (eds), Emerald Ash Borer Research and
Technology Development Meeting, Romulus, MI, 5-6 Oct. 2004. U.S.
Department of Agriculture, Forest Service publication FHTET-2004—15,
Morgantown, WV.

Brown-Rytlewski, D. E., and M. A. Wilson. 2005. Tracking the emergence of emerald
ash borer adults, pp. 13-14. In Mastro and Reardon (eds), Emerald Ash Borer
Research and Technology Development Meeting, Romulus, MI, 5-6 Oct. 2004.
U.S. Department of Agriculture, Forest Service publication FHTET-2004-15,
Morgantown, WV.

Cappaert, D., D. G. McCullough, and T. M. Poland. 2005a. Emerald ash borer life
cycle: a reassessment, pp. 19-20. In Mastro and Reardon (eds), Emerald Ash
Borer Research and Technology Development Meeting, Romulus, MI, 5-6 Oct.
2004. U.S. Department of Agriculture, Forest Service publication FHTET-2004-
15, Morgantown, WV.

Cappaert, D., D. G. McCullough, and T. M. Poland. 2005b. The upside of the emerald
ash borer catastrophe: a feast for woodpeckers, pp. 69-70. In Mastro and Reardon
(eds), Emerald Ash Borer Research and Technology Development Meeting,
Romulus, MI, 5-6 Oct. 2004. U.S. Department of Agriculture, Forest Service
publication FHTET-2004-15, Morgantown, WV.

Cappaert, D., D. G. McCullough, T. M. Poland, and N. W. Siegert. 2005c. Emerald
ash borer in North America: A research and regulatory challenge. American
Entomologist 51 : 152- 1 65.

Carlson, R. W., and F. B. Knight. 1969. Biology, taxonomy, and evolution of four
sympatric A grilus beetles. Contributions of the American Entomological Institute
4: 1-105.

Chinese Academy of Science. Institute of Zoology. 1986. Agrilus marcopoli
Obenberger, Agriculture Insects of China. China Agriculture Press, Bejing, China.
p. 445.

Conover, W. J. 1971. Practical Nonparametric Statistics. Texas Technical University,
New York.

Dunn, J. P., T. W. Kimmerer, and G. L. Nordin. 1986a. The role of host tree condition

in attack of white oaks by the twolined chestnut borer, Agrilus bilineatus (Weber)
(Coleoptera: Buprestidae). Oecologia 70: 596-600.

113

Dunn, J. P., T. W. Kimmerer, and G. L. Nordin. 1986b. Attraction of the twolined
chestnut borer, Agrilus bilineatus (Weber) (Coleoptera: Buprestidae), and
associated borers to volatiles of stressed white oak. Canadian Entomologist 118:
503-509.

Dunn, J. P., T. W. Kimmerer, and D. A. Potter. 1987. Winter starch reserves of white
oak as a predictor of attack by the twolined chestnut borer, Agrilus bilineatus
(Weber) (Coleoptera: Buprestidae). Oecologia 74: 352-355.

Dunn, J. P., D. A. Potter, and T. W. Kimmerer. 1990c. Chestnut blight fungus
predisposes oaks to attack by the two—lined chestnut borer. Environmental
Entomology 19: 239-243.

Dunn, J. P., D. A. Potter, and T. W. Kimmerer. 199011. Carbohydrate reserves, radial
growth, and mechanisms of resistance of oak trees to phloem-boring insects.
Oecologia 83: 458-468.

F ayt, P., M. M. Machmer, and C. Steeger. 2005. Regulation of spruce bark beetles by
woodpeckers - a literature review. Forest Ecology and Management 206: 1-14.

Francese, J. A., J. B. Oliver, 1. Fraser, N. Youssef, D. R. Lance, D. J. Crook, and V.
C. Mastro. 2006. Effects of trap design and placement on capture of emerald ash
borer, Agrilus planipennis, pp. 57-58. In Mastro et al. (eds), Emerald Ash Borer
Research and Technology Development Meeting, Pittsburgh, PA, 26-27 Sept.
2005. U.S. Department of Agriculture, Forest Service publication FHTET-2005-
16, Morgantown, WV.

Fuester, R. W., and P. W. Schaefer. 2006. Research on parasitoids of buprestids in
progress at the ARS Beneﬁcial Insects Introduction Research Unit, pp. 53-55. In
Mastro et al. (eds), Emerald Ash Borer Research and Technology Development
Meeting, Pittsburgh, PA, 26-27 Sept. 2005. U.S. Department of Agriculture,
Forest Service publication FHTET-2005-16, Morgantown, WV.

Fumanal, B., J.-F. Martin, R. Sobhian, A. Blanchet, and M.-C. Bon. 2004. Host range
of Ceutorhynchus assimilis (Coleoptera: Curculionidae), a candidate for

biological control of Lepidium draba (Brassicaceae) in the USA. Biological
Control 30: 598-607.

Futuyma, D. J., and S. C. Peterson. 1985. Genetic variation in the use of resources by
insects. Annual Review of Entomology 30: 217-238.

Giedraitis, J. P., and J. J. Kielbaso. 1982. Municipal Tree Management. Urban Data
Service Reports 14 (1). Washington, DC: International City Management
Association.

Haack, R. A., and D. M. Benjamin. 1982. The biology and ecology of the twolined
chestnut borer, Agrilus bilineatus (Coleoptera: Buprestidae), on oaks, Quercus
spp., in Wisconsin. Canadian Entomologist 114: 385-396.

114

Haack, R. A., and F. S. Slansky. 1986. Nutritional ecology of wood-feeding Coleoptera,
Lepidoptera, and Hymenoptera, pp. 449-485. In F. Slansky and J. G. Rodriquez
[eds], Nutritional Ecology of Insects, Mites, Spiders, and Related Invertebrates.
John Wiley and Sons, Inc., New York. 1032 pp.

Haack, R. A., and T. R. Petrice. 2005. Host range of emerald ash borer, p. 27. In Mastro
and Reardon (eds), Emerald Ash Borer Research and Technology Development
Meeting, Romulus, MI, 5-6 Oct. 2004. U.S. Department of Agriculture, Forest
Service publication FHTET-2004-15, Morgantown, WV.

Haack, R. A., T. R. Petrice, D. L. Miller, L. S. Bauer, and N. M. Schiff. 2004. Host
range of emerald ash borer, p. 38. In Mastro and Reardon (eds), Emerald Ash
Borer Research and Technology Development Meeting, Romulus, MI, 5-6 Oct.
2004. U.S. Department of Agriculture, Forest Service publication FHTET-2004-
15, Morgantown, WV.

Hanks, L. M. 1999. Inﬂuence of the larval host plant on reproductive strategies of
cerambycid beetles. Annual Review of Entomology 44: 483-505.

Hanks, L. M., T. D. Paine, and J. G. Millar. 1993. Host species preference and larval
performance in the wood-boring beetle Phoracantha semipunctata F. Oecologia
95: 22-29.

Hanks, L. M., J. G. Millar, and T. D. Paine. 1995. Biological contraints on host range
expansion by the wood-boring beetle Phoracantha semipunctata (Coleoptera:
Cerambycidae). Annals of the Entomological Society of America 88: 183-188.

Hanks, L. M., T. D. Paine, and J. G. Millar. 2005. Inﬂuence of the larval environment
on performance and adult body size of the wood-boring beetle Phoracantha
semipunctata. Entomologia Experimentalis et Applicata 114: 25-34.

Hanks, L. M., T. D. Paine, J. G. Millar, C. D. Campbell, and U. K. Schuch. 1999.
Water relations of host trees and resistance to the phloem-boring beetle
Phoracantha semipunctata F. (Coleoptera: Cerambycidae). Oecologia 119.

Hanover, J. W. 1966. Genetics of terpenes. 1. Gene control of monoterpene levels in
Pinus monticola Dougl. Heredity 21: 73-84.

Hard, J. S. 1985. Spruce beetles attack slowly growing spruce. Forest Science 31: 839-
850.

Harlow, W. M., E. S. Harrar, J. W. Hardin, and F. M. White. 1991. Textbook of
Dendrology, 8th ed. McGraw-Hill, New York. 534 pp.

Hodges, J. D., and L. S. Pickard. 1971. Lightening in the ecology of the southern pine

beetle, Dendroctonusfrontalis (Coleoptera: Scolytidae). Canadian Entomologist
103: 44-51.

115

Houston, D. R. 1981. Oak decline and mortality, pp. 217-225. In: Doane C. C.,
McManus M. L. (eds), The gypsy moth: research toward integrated pest
management. USDA For Ser Tech Bull 1584. 757 pp.

Jendek, E. 1994. Studies in the East Palaearctic species of the genus Agrilus Dahl, 1823
(Coleoptera: Buprestidae) Part 1. Entomological Problems 25: 9-25.

Kimmerer, T. W., and T. T. Kozlowski. 1982. Ethylene, ethan, acetaldehyde, and
ethanol production by plants under stress. Plant Physiology 69: 840-847.

Kozlowski, T. T., and C. H. Winget. 1963. Patterns of water movement in forest trees.
Botanical Gazette 124: 301-311.

Kozlowski, T. T., and S. G. Pallardy. 1997. Physiology of Woody Plants. Academic
Press, New York.

Laitinen, M.-L., R. Julkunen-Tiitto, J. Tahvanainen, J. Heinonen, and M. Rousi.
2005. Variation in birch (Betula pendula) shoot secondary chemistry due to
genotype, environment, and ontogeny. Journal of Chemical Ecology 31: 697-717.

Larsson, S., R. Oren, R. H. Waring, and J. W. Barrett. 1983. Attacks of mountain
pine beetle as related to tree vigor of ponderosa pine. Forest Science 29: 395-402.

Lee, C. E. 2002. Evolutionary genetics of invasive species. TRENDS in Ecology and
Evolution 117: 386-391.

Liu, H., L. S. Bauer, R. Gao, T. Zhao, T. R. Petrice, and R. A. Haack. 2003.
Exploratory survey for the emerald ash borer, Agrilus planipennis (Coleoptera:
Buprestidae), and its natural enemies in China. The Great Lakes Entomologist 36:
191-204.

Loerch, C. R., and E. A. Cameron. 1984. Within-tree distribution and seasonality of
immature stages of the bronze birch borer, Agrilus anxius (Coleoptera:
Buprestidae). Canadian Entomologist 116: 147-152.

Lyons, D. B., G. C. Jones, and K. Wainin-Keizer. 2004. The biology and phenology of
the emerald ash borer, A grilus planipennis, p.5. In Mastro and Reardon (eds),
Emerald Ash Borer Research and Technology Development Meeting, Port Huron,
MI, 30 Sept.-1 Oct. 2003. U.S. Department of Agriculture, Forest Service
publication F HTET-2004-02, Morgantown, WV.

MacFarland, D. W., and S. P. Meyer. 2005. Characteristics and distribution of potential
ash tree hosts for emerald ash borer. Forest Ecology and Management 213: 15-24.

Marshall, S. A., S. M. Paiero, M. Buck, and B. D. Gill. 2006. Using Cerceris
ﬁtmipennis wasps to monitor the spread of emerald ash borer, p. 56. In Mastro et
al. (eds), Emerald Ash Borer Research and Technology Development Meeting,
Pittsburgh, PA, 26-27 Sept. 2005. U.S. Department of Agriculture, Forest Service
publication FHTET-2005-16, Morgantown, WV.

116

Mastro, V., and R. Reardon. (eds). 2004. Emerald Ash Borer Research and
Technology Development Meeting, Port Huron, MI, 30 Sept.-1 Oct. 2003. U.S.
Department of Agriculture, Forest Service publication F HTET-2004-02,
Morgantown, WV.

Mastro, V., and R. Reardon. (eds). 2005. Emerald Ash Borer Research and
Technology Development Meeting, Romulus, MI, 5-6 Oct. 2004. U.S.
Department of Agriculture, Forest Service publication F HTET-2004-1 5,
Morgantown, WV.

Mastro, V., R. Reardon, and G. Parra. (eds). 2006. Emerald Ash Borer Research and
Technology Development Meeting, Pittsburgh, PA, 26-27 Sept. 2005. U.S.
Department of Agriculture, Forest Service publication F HTET-2005-16,
Morgantown, WV.

Mattson, W. J., and R. A. Haack. 1987. The role of drought in outbreaks of plant-eating
insects. BioScience 37: 110-118.

McCullough, D. G., and S. A. Katovich. 2004. Emerald Ash Borer. Pest Alert. USDA
Forest Service, State and Private Forestry, Northeastern Area, NA-PR-02-O4.

McCullough, D. G., T. M. Poland, and D. Cappaert. 2006. Attraction of emerald ash
borer to trap trees: effects of stress agents and trap height, p. 61. In Maestro et al.
(eds), Emerald Ash Borer Research and Technology Development Meeting,
Pittsburgh, PA, 26-27 Sept. 2005. U.S. Department of Agriculture, Forest Service
publication FHTET-2005-16, Morgantown, WV.

McCullough, D. G., T. M. Poland, D. Cappaert, P. Lewis, and J. Molongowski. 2005.
Evaluation of trunk injections for control of emerald ash borer, pp. 38-99. In
Mastro and Reardon (eds), Emerald Ash Borer Research and Technology
Development Meeting, Romulus, MI, 5-6 Oct. 2004. U.S. Department of
Agriculture, Forest Service publication FHTET-2004-15, Morgantown, WV.

Miller, R. O., P. D. Bloese, J. W. Hanover, and R. A. Haack. 1991. Paper birch and
European white birch vary in growth and resistance to bronze birch borer. Journal
of the American Society of Horticultural Science 116: 580-584.

Nash, R. W., E. J. Duda, and N. H. Gray. 1951. Studies on the extensive drying,
regeneration, and management of birch. Maine Forest Service Bulletin 15. 82 pp.

Orians, C. M., B. M. Roche, and R. S. Fritz. 1996. The genetic basis for variation in
the concentration of phenolic glycosides in Salix sericea: An analysis of
heritability. Biochemical Systematic Ecology 24: 719-724.

Ottman, K. A., and J. J. Kielbaso. 1976. Managing Municipal Trees. Urban Data

Service Reports 8(11). International City Management Association. Washington,
DC. November 1976. 15 pp.

117

Parry, D., and R. A. Goyer. 2004. Variation in the suitability of host tree species for
geographically discrete populations of forest tent caterpillar. Environmental
Entomology 33: 1477-1487.

Poland, T. M., D. G. McCullough, P. d. Groot, G. Grant, and D. Cappaert. 2005.
Progress toward developing trapping techniques for the emerald ash borer, pp. 53-
54. In Mastro and Reardon (eds), Emerald Ash Borer Research and Technology
Development Meeting, Romulus, MI, 5-6 Oct. 2004. U.S. Department of
Agriculture, Forest Service publication FHTET-2004-15, Morgantown, WV.

Potter, D. A., and J. R. Hartman. 1993. Susceptibility of honeylocust cultivars to
Thyronectria austro-americana and response of A grilus borers and bagworms to
infected and non-infected trees. Journal of Environmental Horticulture 11: 176-

181.

Ross, K. G., E. L. Vargo, L. Keller, and J. C. Trager. 1993. Effect of a founder event
on variation in the genetic sex-determining system of the ﬁre ant Solenopsis
invicta. Genetics 135: 843-854.

Schaefer, P. W. 2005. Foreign exploration for emerald ash borer and its natural enemies,
pp. 67-68. In Mastro and Reardon (eds), Emerald Ash Borer Research and
Technology Development Meeting, Romulus, MI, 5-6 Oct. 2004. U.S.
Department of Agriculture, Forest Service publication F HTET-2004-1 5 ,
Morgantown, WV.

Siegert, N. W., and D. G. McCullough. 2005. Reconstructing the temporal and spatial
dynamics of emerald ash borer in black ash: a case study of an outlier site in
Rosecommon County, Michigan, pp. 21-22. In Mastro and Reardon (eds),
Emerald Ash Borer Research and Technology Development Meeting, Romulus,
MI, 5-6 Oct. 2004. U.S. Department of Agriculture, Forest Service publication
FHTET-2004-15, Morgantown, WV.

Smitley, D., E. Rebek, and T. Davis. 2006. Ash dieback in Michigan, 2003-2005, pp.
68-69. In Maestro et al. (eds), Emerald Ash Borer Research and Technology
Development Meeting, Pittsburgh, PA, 26-27 Sept. 2005. U.S. Department of
Agriculture, Forest Service publication FHTET-2005-16, Morgantown, WV.

Strazanac, J. S. 2006. Overview of Hymenoptera genera currently considered for
emerald ash borer biocontrol release, p. 51. In Mastro et al. (eds), Emerald Ash
Borer Research and Technology Development Meeting, Pittsburgh, PA, 26-27
Sept. 2005. U.S. Department of Agriculture, Forest Service publication F HTET-
2005-16, Morgantown, WV.

Suarez, A. V., N. D. Tsutsui, D. A. Holway, and T. J. Case. 1999. Behavioral and
genetic differentiation between native and introduced populations of the
Argentine ant. Biological Invasions 1: 43-53.

118

Sugiura, N. 1999. The family Buprestidae in F ukushima Prefecture: the genus Agrilus.
Online at http://wwwl .linkclub.or.jp/~sugirin/fukusima/nagatama2.htrnl
(accessed August 2002).

Sunnucks, P., F. Driver, W. V. Brown, M. Carver, D. F. Hales, and W. M. Milne.
1997. Biological control and genetic characterization of morphologically similar
Therioaphis trifolii (Hemiptera: Aphididae) with different host utilization.
Bulletin of Entomological Research 87: 425-436.

Thompson, J. N. 1988. Evolutionary ecology of the relationship between oviposition
preference and performance of offspring in phytophagous insects. Entomologia
Experimentalis et Applicata 47: 3-14.

U.S. Department of Agriculture Forest Inventory and Analysis Database (FIA).
2006. http://ﬁa.fs.fed.us.

Via, S. 1986. Genetic covariance between oviposition preference and larval performance
in an insect herbivore. Evolution 40: 778-785.

Waring, R. H., and G. B. Pitman. 1980. A simple model of host resistance to bark
beetles. Oregon State Univ, Forest Res Lab, Res Note 65. 8 pp.

Wright, L. C., A. A. Berryman, and B. E. Wickman. 1984. Abundance of the ﬁr
engraver, Scolytus ventralis, and the Douglas-ﬁr beetle, Dendroctonus
pseudotsugae, following tree defoliation by the Douglas-ﬁr tussock moth, Orgyia
pseudotsugata. Canadian Entomologist 116: 293-305.

Yu, C. 1992. Agrilus marcopoli Obenberger, pp. 400-401. In G. Xiao (ed.), Forest Insects
of China (2nd edition). China Forestry Publishing House, Beijing, China.

Zar, J. H. 1984. Biostatistical Analysis. Prentice Hall, Inc, Englewood Cliffs, New
Jersey.

119