p.

u: 2‘
.16. .J

Emu

“3..
£3?
. Eduﬁz.

‘7~TY

.3

.. r v.
.: .o.

pg i. ud‘ .
4 v1 .‘anun
mix. 3

33¢;
L 5.
x1 . . .

 

 

 

 

:. I.
#5:. .n»

.1.

a.“

 

THESlS

This is to certify that the

dissertation entitled

INDUCED RESPONSES OF POPLARS TO DEFOLIATION
AND THEIR EFFECTS ON LEAF—FEEDING LEPIDOPTERA

presented by
Dylan Parry

has been accepted towards fulﬁllment
of the requirements for

Ph.D. Entomology

(Lg/ﬂ. Aim

Major professor

degree in

 

 

 

Date 30 November, 2000

 

MSU is an Afﬁrmative Action/Equal Opportunity Institution 0—12771

LIBRARY
Michigan State
University

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 C‘JCIRC/DatoDmpGS-p.“

 

 

 

INDUCED RESPONSES OF POPLARS TO DEFOLIATION AND
THEIR EFFECTS ON LEAF-FEEDING LEPIDOPTERA

By

Dylan Party

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

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

2000

 

 

ABSTRACT

INDUCED RESPONSES OF POPLARS TO DEFOLIATION AND THEIR EFFECTS
ON LEAF -FEEDING LEPIDOPTERA

By
Dylan Parry
An enduring hypothesis for explaining cyclical populations of forest Lepidoptera has been
phytochemical resistance to herbiv01y elicited in trees subjected to defoliation. Severe
defoliation of trees elicits both rapid and delayed changes in host quality for herbivores,
known collectively as 'induced-resistance’. While rapid-induced resistance (RIR), which
affects the generation of insects causing the damage, is thought to contribute to population
stability, delayed-induced resistance (DIR), lasting a year or more, may lead to unstable
dynamics and generate population cycles. DIR has typically been investigated in trees
following a single defoliation event. However, many outbreak folivores defoliate trees for
consecutive years and single year studies may not capture the full effects of DIR. Thus,
cumulative effects of successive years of defoliation on tree physiology and insect

herbivores are poorly understood.

I conducted two long—term studies designed to emulate the defoliation intensity and
temporal scale of natural outbreaks. Chapter 2 describes an outbreak population of gypsy
moth (Lymantria dispar L.) established experimentally in large—scale stands of hybrid
poplar (Populus sp.), 1996—1999. In Chapter 3, I experimentally established forest tent
caterpillar (Malacosoma disstria Hubner) populations and defoliated two stands of

trembling aspen (Populus tremuloides) over a four-year period (1997-2000).

Gypsy moth defoliation increased total phenolics and condensed tannins in leaves

while having only marginal effects on phenolic glycosides and other minor secondary

 

compounds (Chapter 2). There were strong within-year, rapid-induced effects on
secondary metabolites, and condensed tannins and total phenolics remained high one year
after the cessation of defoliation. Nitrogen was reduced within the year that defoliation
occurred, especially in late season leaf samples. Effects on nitrogen did not extend to the
following season and levels had recovered to predefoliation levels within a year of
cessation of defoliation. Gypsy moth pupal mass, development time and fecundity were
not affected by the treatments in 1996, whereas female pupal mass and fecundity were
marginally reduced in 1997 and signiﬁcantly reduced in 1998. Gypsy moth defoliation
reduced the performance of competing species of Lepidoptera in 1998. Forest tent
caterpillar, poplar tent maker (Clostera inclusa), and big poplar Sphinx (Pachysphz’nx
modesta) all suffering reduced pupal mass, fecundity, and/or increased development time.
Fall webworm (Hyphantria cunea), was unaffected by defoliation or fertilization.
Defoliation in previous years did not affect gypsy moth and three other species in 1999,

suggesting that herbivore performance was independent of phenolic and tannin levels.

Chronic defoliation of aspen (Chapter 3) had larger effects on tent caterpillars than
defoliation of poplars had on gypsy moth (Chapter 2). Relative growth rates of ﬁnal instar
females were signiﬁcantly reduced although growth of male ﬁfth instars was unaffected.
Mean fecundity was reduced by 8-21% in the first year, 13-16% following three years,
and 20-21% on trees with four consecutive years of defoliation. Development time and
survival from hatch to adult was unaffected by the treatments. Defoliation affected
parasitism rates by two tachinid ﬂies. Parasitism appeared to be mediated by caterpillar
density rather than induced host tree responses to defoliation, and also varied signiﬁcantly
among aspen clones. Intraspecific differences in phytochemistry among clones may alter

host detection abilities in tachinids using olfactory cues from herbivore damaged foliage.

 

Copyﬁghtby
DYLANPARRY

2000

 

 

I dedicate this, with love and utmost and heartfelt thanks, to my mother, Ruth Powell,
who has never wavered in her support for me, despite the many twists and turns of my
chosen career path.

ACKNOWLEDGEMENTS
First and foremost, I thank Dan Henns and Bill Mattson for reawakening my interest in
plant-herbivore interactions. I also thank my committee members Dan Herms, Mark
Scriber, Bill Mattson, Deb McCullough, and Suzanne Thiem for being rigorous and
critical in a positive mamier. My thanks to the Department for support and for being a
good place to call home for nearly 5 years. I would especially like to thank Jill Kolp for
she is not just a superb graduate secretary, but a great person and friend who has always
looked out for my best interests.

To those who keep my feet on the ground, provide support, and are friends in the
truest sense of the word; thanks Nate (N atron) Siegert, Cheryl Frankfater, Piera Giroux,
Gabe Ording, Tim Tibbetts, Linsey White, Cam & Gail Currie, Jim & Lisa Weber, Laura
Lazarus, Heather Govenor.

No doctoral program. would be complete without activities to divert ones attention
and to help in the maintenance of sanity. In no particular order I thank The Peanut Barrel,
Specialized Mountain Bikes, John Labatt Inc., the Delton Tuques Hockey Team, The
Chimes, and all those various and sundry people with whom I spent much money and
time shooting pool at various ﬁne Michigan establishments.

This research was funded by NSF grant DEB 95-10044, and the following grants to
D.P.: NASA-TECO (William J. Mattson, co-PI), Graduate Research Training Group at
the Kellogg Biological Station (NSF grant DIR-9602252), a Hutson Grant, Department of
Entomology, Michigan State University, Sigma Xi Grant—in-Aid of Research, and a

Dissertation Completion Fellowship, The Graduate School, Michigan State University.

vi

 

Summer salary support in 2000 was provided by the Michigan Agricultural Experiment
Station (Project # MICL 01644 to J. Mark Scriber).

I can not possibly thank enough all the individuals who have helped me out, lent me
equipment, and assisted with various aspects of my projects. It is, however, these little

contributions that add up and ensure the success of any project.

 

TABLE OF CONTENTS
LIST OF TABLES ..................................................................... x
LIST OF FIGURES ................................................................... xiii

CHAPTER 1. Biological and Ecological Signiﬁcance of Induced Resistance in Trees

Introduction .................................................................... 1
Effects of RIR on folivorous Lepidoptera .............................. 3
Effects of DIR on folivorous Lepidoptera .............................. l3
Mechanisms underlying induced resistance in trees .................. l9
Indirect effects of induced resistance — competition .................. 23
Indirect effects of induced resistance — the third trophic level ....... 25
Induced-resistance and population dynamics .......................... 28
Overview of dissertation .................................................. 30

Literature Cited ................................................................. 32

CHAPTER 2. Emulating insect outbreak: Experimentally testing the effects of

cumulative defoliations on induced-resistance and caterpillar herbivory

Introduction ...................................................................... 46
Materials and Methods ........................................................ 52
Experimental design .................................................... 52

Tree responses to defoliation ........................................... 58
Effects of defoliation on caterpillars .................................. 63
Statistical analyses ....................................................... 69
Results ........................................................................... 70
Effects of defoliation and fertilization on trees ....................... 68

viii

Effects of defoliation and fertilization on caterpillars ...............
Discussion .........................................................................
Effects of treatments on caterpillars .....................................
Effects of treatments on trees .............................................
Indirect-interactions between induced-resistance and herbivores
Literature Cited ..................................................................
CHAPTER 3. Induced-responses of aspen to successive years of defoliation: The
effects on an outbreak folivore and its natural enemies.
Introduction .....................................................................
Materials and Methods .........................................................
Experimental system ......................................................
Implementation of treatments ............................................
Insect bioassays ............................................................
Statistical analyses .........................................................
Results
Effects of treatments on caterpillars ....................................
Effects of treatments on survival and parasitism ......................
Discussion ........................................................................
Literature Cited ................................................................
CHAPTER 4. Conclusions

Conclusions ...........................................................

Literature Cited ................................................................

 

95

120

120

127

134

141

158

164

164

165

170

177

178

190

195

210

223

229

 

 

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

LIST OF TABLES
A literature review of the effects of rapidly-induced resistance (RIR) in
deciduous trees on tree-feeding Lepidoptera ........................ 4
A literature review of the effects of delayed-induced resistance (DIR) in
deciduous trees on tree-feeding Lepidoptera ........................ 14
ANOVA table for the effects of defoliation and fertilization treatments on tree
growth 1996-1999. ...................................................... 72
ANOVA table for the effects of defoliation and fertilization on concentrations
of foliar total phenolics and condensed tannins in 1996. ............ 74
ANOVA table for the effects of defoliation and fertilization on concentrations
of foliar nitrogen, total phenolics and condensed tannins in June and July
1997. ........................................................................ 77
Foliar concentrations (mean i SE) of minor phenolic compounds in June and
July 1997. .................................................................. 79
ANOVA table for the effects of defoliation and fertilization on concentrations
of minor phenolics in June and July 1997. ........................... 81
ANOVA table for the effects of defoliation and fertilization on concentrations
of foliar nitrogen, total phenolics and condensed tannins in May, June and
July 1998. ................................................................. 83
Foliar concentrations (mean i SE) of minor phenolic compounds in May, June
and July 1998. ............................................................ 86
ANOVA table for the effects of defoliation and fertilization on concentrations

of minor phenolics in May, June and July 1998. ..................... 88

Table 11.

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

ANOVA table for the effects of defoliation and fertilization on water content
in June and July 1998, and on leaf area and leaf toughness in June 1998.
............................................................................... 92
ANOVA table for the effects of defoliation and fertilization on concentrations
of foliar nitrogen, total phenolics and condensed tannins in June 1999.
............................................................................... 94
ANOVA table for the effects of defoliation and fertilization on water content
and leaf area in June 1999. ............................................. 97
ANOVA table for the effects of defoliation and fertilization treatments on
gypsy moth pupal mass (g) and development time (days) in 1996-1999.
............................................................................... 99
ANOVA table for the effects of defoliation and fertilization treatments on
gypsy moth relative growth rate (RGR), relative consumption rate (RCR),
approximate digestibility (AD), efﬁciency of conversion of ingested food
(ECI), and efﬁciency of conversion of digested food (ECD) in 1997 and 1998.
................................................................................ 104
Pearson correlation coefﬁcients between foliar parameters and gypsy moth
pupal mass and development time 1996-1999. ...................... 106
ANOVA table for the effects of defoliation and fertilization treatments on
forest tent caterpillar, white-marked tussock moth, poplar tent maker, big
poplar Sphinx, and fall webworm pupal mass (g) and development time (days)

in 1998. ................................................................... 110

xi

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Table 23.

Table 23.

Table 24.

ANOVA table for the effects of defoliation and fertilization treatments 0:
forest tent caterpillar, white-marked tussock moth, poplar tent maker, bi,
poplar sphinx, and fall webworm pupal mass (g) and development time (days
in 1998. ................................................................... ll;
Pearson correlation coefﬁcients between foliar parameters and pupal mass anc
development time for forest tent caterpillar and poplar tent maker in 1998
1999, big poplar Sphinx and fall webworm in 1998, and white-marked tussocl
moth in 1999. ............................................................ 11‘
Treatment implementation in the aspen plots 1997-2000 .......... l6!
ANOVA table for the effects of aspen defoliation treatments on forest ten
caterpillar pupal mass and development time. ....................... 184
ANOVA table for the effects of aspen defoliation treatments on second insta
forest tent caterpillar. ................................................... 18‘
ANOVA table for the effects of aSpen defoliation treatments on fifth insta
forest tent caterpillar .................................................... 181
ANOVA table for the effects of aspen defoliation treatments on forest ten
caterpillar survival and parasitism. .................................... 191
The effects of induced-resistance and parasitism on approximate reproductiv

rates of forest tent caterpillar. .......................................... 19'

xii

 

 

 

LIST OF FIGURES

Figure 1. Defoliation levels in treatment and control plots 1996—1999. 59
Figure 2. Approximate larval feeding periods of the six species of Lepidoptera used in
the study in relation to important phenological events. ............ 65
Figure 3. Annual diameter growth (mean i SE) for poplars subjected to defoliation and
fertilization treatments over the course of the study. ............... 71
Figure 4. The effects of defoliation and fertilization treatments on concentrations (mean
i SE) of total phenolics and condensed tannins in June 1996. 73
Figure 5. The effects of 1997 defoliation and fertilization treatments on concentrations
(mean i SE) of nitrogen, total phenolics and condensed tannins. .. 76
Figure 6. The effects of 1998 defoliation and fertilization treatments on concentrations
(mean i SE) of nitrogen, total phenolics and condensed tannins. .. 82
Figure 7. The effects of 1998 defoliation and fertilization treatments on measures (mean
i SE) of water content, leaf area, and leaf toughness. .............. 91
Figure 8. The effects of previous defoliation and fertilization treatments on
concentrations (mean i SE) of total phenolics and condensed tannins in 1999.
................................................................................. 93
Figure 9. The effects of previous defoliation and fertilization treatments on water
content and leaf area (mean i SE) in 1999. ........................... 96
Figure 10. The effects of defoliation and fertilization treatments on mean (:1: SE) male
and female gypsy moth pupal mass and development time for 1996-1999
.............................................................................. 98

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

The effects of defoliation and fertilization treatments on mean (i SE) relative
growth and consumption rates, and feeding efﬁciency of 5th instar gypsy moth
in 1997 and 1998. Relative growth rate (RGR), relative consumption rate
(RCR), approximate digestibility (AD), efﬁciency of conversion of ingested
food (ECI), and efﬁciency of conversion of digested food (ECD). 102
The effects of defoliation and fertilization treatments on mean (i SE) pupal
mass and development time of forest tent caterpillar (M. disstria), poplar tent
maker (C. inclusa), big pOplar Sphinx (P. modesta), and fall webworm (H.
cunea) in 1998. ............................................................ 109
The effects of previous defoliation and fertilization treatments on mean (i SE)
pupal mass and development time of forest tent caterpillar (M. disstria),
white-marked tussock moth (O. leucostigma) and poplar tent maker (C.
inclusa) in 1999. ............................................................ 112
Mean (i SE) pupal mass for female and male forest tent caterpillars reared
from hatch to pupation on trees from each of the treatments. ......... 179
Mean (i SE) development time for female and male forest tent caterpillars
reared from hatch to pupation on trees from each of the treatments. 182
Mean (i SE) relative growth rates (RGR) and relative consumption rates
(RCR) for second instar forest tent caterpillars. ........................... 183
Mean (i SE) relative growth rate (RGR), relative consumption rate (RCR),

approximate digestibility (AD), efﬁciency of conversion of ingested foliage
(ECI), and efﬁciency of conversion of digested foliage (ECD) for ﬁnal (5th)

instar female and male forest tent caterpillars. ........................ 186

xiv

Figure 18. Mean (i SE) survival of forest tent caterpillar reared from hatch to pupation
on trees from each of the treatments. ................................... 191
Figure 19. Mean (i SE) parasitism by tachinids of forest tent caterpillar reared from

hatch to pupation on trees from each of the treatments. .............. 194

XV

   

CHAPTER 1:
THE ECOLOGICAL SIGNIFICANCE OF INDUCED RESISTANCE TO INSECT

HERBIVORES IN DECIDUOUS TREES

INTRODUCTION
Outbreaks of forest Lepidoptera are a conspicuous feature of temperate and boreal
forests. Of particular fascination for ecologists are the cyclical or nearly cyclical changes
in density that characterize the population dynamics of many outbreaking species. The
relative contributions of density-dependent and density-independent regulating
mechanisms in the population dynamics of these insects, and indeed that of many other
organisms, engaged researchers in fractious debate for several decades (e.g., Nicholson
1933, Andrewartha and Birch 1954, Dempster 1983, Hassell et a1. 1989, Murdoch, 1994,
Turchin 1995). Identiﬁcation of delayed negative feedback in time series data from many
forest insects (Turchin 1990, 1995) shifted the focus away from whether density
dependence regulates populations, to questions regarding which of several potentially
delayed—density dependent factors can best account for the observed cycles. While the
autocorrelation analysis used by Turchin (1990) is sufﬁcient to detect delayed-density
dependent regulation, it can not determine causal agents. Thus, several factors including
maternal effects, natural enemies, and defoliation-induced changes in host plant quality
are all theoretically capable of producing second-order dynamics in forest insect
population’s (Berryman et a1. 1987, 1996, Ginzburg and Taneyhill 1994, Underwood

1999)

 

 

 

 

More than two decades ago, investigators ﬁrst recognized that reductions in survival,
growth, and fecundity occurred in herbivorous insects feeding on the foliage of defoliated
deciduous trees (e.g., Haukioja and Niemela 1977, Haukioja 1980). Because food quality
is integral to the success of herbivores, defoliation—induced changes in foliage have been
an appealing explanation for ﬂuctuating population dynamics in insects that outbreak on
hardwood trees. Any response by plants to herbivory that reduces the ﬁtness of
herbivores has been deﬁned as ‘induced-resistance’ (Karban and Baldwin 1997). In
deciduous trees, two types of induced-resistance have been recognized, based on both
differences in temporal manifestation, and on their postulated effects on herbivore
population dynamics (Haukioja and Niemela 1977, Haukioja 1980, 1982). Rapid-
induced resistance (RIR) occurs within minutes or days of herbivory. Relatively little
damage may be required to elicit RIR responses and the duration of the effects can be
transient. Because of the short induction time, RIR affects the generation of herbivore
responsible for the herbivory and thus is thought to exert a stabilizing effect on
population dynamics (Haukioja 1982). In contrast to RIR, delayed-induced resistance
(DIR) occurs in the year or years following herbivory and generally is elicited only by
severe defoliation. Because of the time-lags in induction and relaxation, DIR acts on
future generations of folivorous insects, and thus is hypothesized to have a destabilizing
rather than a regulating effect on population dynamics (Haukioja 1982). The postulated
effects of DIR on herbivore population dynamics has led a number of authors to propose
similar host-quality driven models for explaining cycles in forest Lepidoptera (e. g., Benz

1974, Haukioja 1980, Rhoades 1983a).

 

 

 

 

The quality of foliage for herbivorous insects is determined by a surfeit of physical
and biochemical traits including nitrogen, water content, carbohydrates, ﬁber, toughness,
and a host of secondary compounds such as alkaloids, phenolics, and terpenoids (e.g.,
Feeny 1970, Mattson 1980, Scriber and Slansky 1981, Schultz 1988, Koricheva et al.
1998). These parameters are often altered quantitatively by defoliation thus changing the
value of leaves as food for caterpillars (e.g., Bryant et al. 1988, Karban and Myers, 1989,
Haukioja 1990). Identiﬁcation of the individual leaf traits responsible for reductions in
herbivore performance following defoliation has proven difﬁcult, leading some to
speculate that it is suites of characters rather than individual components that are

responsible for the observed effects (Kause et al. 1999).

Rapid-induced resistance — direct effects on herbivorous insects

Herbivores ranging from mollusks to mammals have been used to investigate the effects
of RIR. However, herbivorous insects, particularly the Lepidoptera, have been the
organism of choice in most investigations. Studies have shown that RIR can reduce
growth, fecundity, and survival of insects, as well as inﬂuence behaviors such as
preference and foraging (e.g., Haukioja and Niemela 1977, Haukioja 1980, 1982,
Haukioja and Hanhimaki 1985, Bergelson et al. 1986, Rossiter et al. 1988, Edwards et al.
1991, Hanhimaki and Senn 1992). Effects of RIR on herbivores have been variable (see
Table 1) generating lengthy debate about the generality of RIR responses, as well as their
effectiveness (e.g., Fowler and Lawton 1985, Hartley and Lawton 1991, Edwards et al.

1991, Karban and Baldwin 1997).

 

 

 

 

 

 

€80

 

 

32:80:
Ammo: :0:0>:0 Z 0% «833$ 530% 8:32:00: 8.22%: :0 :020:00: $3-0 8358:: 8.02.32 80.30232 3802 A0\:V 3::«2
8.30:3 .m (:0 3803 3>:«2 3:2 3:30:02 203300.22
8 030:000 o\om ::«028w3-:0z 62:02:80. .2 (:0 082
032 «.080: Z 0% «83322 80800330 :0 mm«8 3:82 :0 30.20 0.32.233 02 3305:0Q 3023.5 w§0§23§ 3502 A85 3::«2
$3: 883%: a. 202%: 028:0 :03
Ammo: .3 :0 «.8333 08: 808002030 8 80: 8302 .20
83: «208022 0% «.333: 8 0308:: 30 m-N £38 30:: 8 :020:00: $3-0 80:58:: 8:22.32 3003023: 380% :xomNAv 3::«2
are «a
$3: «2080: Z 0% «.8333 :3 00305:: 082 808002030 .3333 8 :020:002 88:05:: 82.2.3.2 800.8230: 350% A35 3::«2
GwoS .3 :0 :0m20w:0m .330 m E 00308:: 08: 80800323: 0208233 020230209 £30200: 00:02 *A0\:V 3:32
3 $9.3
00:0:0202 0.28800 0: 03320: 30022 830% 3082 m::0w 00: .2. 0% 8320.232

0038030 :0: H 0\: * 000:000: .33 $082 0:022:02 20 A03: 205080”

.082 80800330 £38 3:23 0:3«08 0:«m0:::m :0 3:28:08 :0 520::00b 0.8308 0.0020 « a: 023: 02: 8 000208 3:0 0:03

m0208m 0030.3 830% 02: :«2: :080 830% 808: 0: 0:0 .33 :02320200 02: m0m«0 080m :2 33228030 223 0.00:: 20 82330::

38082080 :3 :0 .$305.80 :0082 383: w8::0 n0.0302 w830: :0 80828 m8? 33:38 :0220 002:3 33 8320200

.30:3 082: :2 008:0:00 0:03 $8305 02: m« 833 088 02: :2 00320200 30008800 :0 3323: 0:03 30:: .2 00:00:00

0:03 00208m 300:: 80:02:00 :0 32000.: 80800802 883202 20 00380.30: 02: :0 :02«20.:00 (:0 80080 8:03:02m A 0338

0.800

 

 

030: .0 0 25:29

:3: .0 0 8:0:

0:80 0:200 0 0:0:

Ammo: .3 :0 8:30.02

:32: .0 0 802.:

6mm: 8>:«O0«2 0% :0230:

0.8: .0 0 0800:

£80 «Sm 0 00800:

30> 82:08: 8 gem :3 :: 08:00:08 0:« :«0» 0:0 8

X0: :3 008 3::: 0380: 00m«0:000 8:320:00 n20:2

:38 :O .30 :0 .20: 0:« ﬁx: :3 Q38 3::: 0380:
000:00: :8000: 8:32:80 :0::«20:00 0300088:

000::

00:«20:00:: 0: 00:«:800 0023 08: 8088350

8 0880 0: 2:25 o\omm :3 038 3::: 0380: 000:00:
m0:«:m 30 00: 00:20:00 33:88 80:: 0w«20:

802880800
:88: 0:« 02802: 8282 08 20.83 000:: 00:«20:00
80:: m0>«0_ 00: :023 >:Z 80:: 33:08 00m«0:00D

0030: :0 80800 883: 0:« .02802: .:0::«20:00
2:25 00320800 30233: 00:08 3::: 0380 :

38083098 ::0:0:::0 28:88 8 >: Z 80::
323.58 :0 32:38:: 0:: :0 :0::«20:00 :0 :00::0 0 Z

.000:
-2 2 003808 33:88 .fowmé: 000:00: 008: :33:

00380: :0 :00::0 0:

08 82 83% 02: 8 :03: 808 3003:: 003380825
038 000:00: :0::«20:0Q 830% 2:08 3:03 3:08
830 :0 8:320:00 :«2:::0:«0 80: 30:0: :0 :00::0 0 Z

0:08:00 0:« 080830: 8:320:00 m w:08«
0«>:«: 3:08 am 0:« 50 8 0.38000 8828?: 02

8:08 0.0000003

.8820 0.280003

8:02: 0.0000203

003.2: 0.280003

.888 0.280003

0.20003: 00002003.:

050000000: 0.0920

8320 0.500203

3:6:53830 QEKKNQW

020030800: 0280::

00.088: .2:0:00®

0.50.: 000000®

80:00Q

0.80.: 000.:00Q

0.50.: 000.:00Q

030:0: 0:802

0:033:00: 880%

0000008: 0:802

008:: 020:):

002A: 08::

003-3: :82:

008-0 08::

008.00 :88:

003-3 0202

882: :88:

A08: 080830:
3:0>0m - 3::«2

 

 

«.08 : 20.:

 

 

n. :59

 

5a: $222

$3: Emacs q 22:2

A80: 5:33

$3: 8822

$3: 2&2 @ WEEEB

6mm: 532$ a. :25:

age .23: 52$

83: 8:55 a $535

.50» 550:“ E
@8st 530% 35888 E2 £8» 28 E Emzahzmo
EB so 38% o: 6832 BS 530% E03303

32%:
commEmw 8 Baum?“ can comemc :o ﬁn: cam XMN
E 3262 EEEE £22: .33 2mE£ :o 88% 02

5658'“ E sonosce $3 a can 3?: Gang
BEEN.“ E 55:62 o\ow EN @828 :oumzomow :mi

68883.
Ecczoﬂ ,wommoaooc BS 530% nEWSCE 3:352

BE 9803 E 3333 $32 $2 ,8; 0:0
3 Rita E BESEE oZ .26 958m 3 mmmE REE
E 388E Axum 63m 28 E 32: Gaza E oosohotﬁ OZ

£3 E beasoe Ea x3
3 mmmE 39a 22:8 @0268 cogsﬂuu acmmum 38m

omemU
mo EnoEm 08 no wEwaoaov £850 E 33386
39 £293 uEOm E commoSE bzatoE Gﬁmq

53.5832 E85
Em can Mao co €031 Ea fowl 3 $2: Egan ﬁght
@8st wEBﬂ wEHaEEo :ocmzoﬁc 0338on

5:350on UEQSUQB:

3:29 Eh§<$¢~ £23qu 53m.
5:82.588 qumeuSa: 3.35 325x
EmuEesEcu qumouEEé ES: 2E;
33953539 Etewcusa: 5?: ”Six

EEG EESEQXE =35 $.th

E22»; SANG wxmxefc Eimsq

5w ExmamEcD

in ESEQM NABSM mxu§x®
ENQKSEQQ 55mm
.25.me Eh§§3 3.23.»; mxuxmzm

€02 .Sv 835

$5 3232
Soot 835
”$2 .Sv 80E
€03»

o\o o: I wowaﬁmw
mu>m2 :8 883

$08-3 335

A82: 4258
885 €232

§ooc 3252

 

3:8 E 03¢

 

 

88$ «and a. :52

33: «tax a 52%

A30: «tam Ea samsom

$8: $59 a 522:

$3: .3 a 50m

:3: .3 3 Ease

6%: Exam

$3: E»: a 220m

.mmmE ESE Em E 835%
gm .355 Em 8 2:: “sumac—gov E 388E $2

.283 05 so wﬁwgaow 35538.“

:ocmzaoc 5 ﬁx; To ‘3 @85on $3 538w
£08 3&0 @253 EobﬁE NH macaw 3:385
053 3:25an 5303?: £59?» go 508% 2;

.o\owo 28 $3 3 @0268 53on
can 8:20 93 50¢ 852 wowﬁmc wow 332:. 988m

.292: no 3385

szm dogma co Ewﬁk S 388% 595 0:58

so Soﬁe o: dogma :o @8st 5380 .039: no
03883 dogma no 2:: EoEQBguw E 8885 Ewsm

.292: S 888% nxvomz
63% E 88 £3on so Soto oZ .292: E 0322:
o\om .596 S 2:: EoEQBoEw :32: av no Soto 02

50380908 .XEN was $2 3 53on :maaozmam
3268 cothohow 85388 E8 588 fame was oi:

.292: Ho :85 no :88
xmgw 28 85988 E3 $88 338 “L WC 3233
S 53on so gumbo Hamocamﬁ on was 5523on

.o\o~h xmo
so was ﬁcw mm 888% 3:586 03% :0 .qonazowow

(«0 £33 wﬁmaobﬁ at? 333% 36:38,»

Sag» Eh=a§3

.EQMS EKEQEA‘N

Uwhwim» BstQhOUQNB:

334% Shacziq

6.52%wa 62~QWQUQNBV<

ﬁnnmhctau stick
.Emﬁt SKESSS

kawhrﬁh GNQQhQUQNQS

8325.5 @33onme

GHon

SSW aﬁzmom .x. 05 833

6.3. asticm 22:88 585

in? “Stick Ao\oom-mv 885

EBKQQDDGM .«NDV

830339: 9.530% ARES :3:me

SPEEQQV, 53% A293:
ﬂog. 98 Emma
QEQEEmh wixmol €0va 83.5

838353“ wézmol Agmc -8 833
SQEASVQ £38m
Eggnog kmuw A35 885
.5. @333

@239ch @3886 wxboTOS 685

 

 

P58 4 28¢

 

Criticism of the importance of RIR to herbivores has focused on inconsistency ir
results among experiments (e.g., Fowler and Lawton 1985). However, the plethora Ol
methodologies used in this ﬁeld, especially in early studies, has undoubtedly contributec
to this variation. For example, experiments have used different scales (leaf, branch, tree).
induction methods (scissors, leaf-punches, tearing, insect feeding), timing betweer
induction and bioassays, tree species, herbivore species, types of assays, and levels 01
defoliation. Thus, many of these experiments are not directly comparable and using
variable results as a criterion to dismiss the importance of induced-resistance is contrived.
Given these well-documented sources of variation, it is not surprising that no consensus
on the effects of RIR has been reached.

Three experimental approaches have been used to investigate the effects of RIR on
insect herbivores. One direction of research has focused on behavioral responses by
herbivores to feeding-induced changes in foliar quality, while another area of focus has
been comparison of localized and systemic responses to herbivory. A third area of RIR
research has addressed the effects of whole—tree defoliation on herbivore ﬁtness as
measured by growth, fecundity and survival. While there is overlap among these
approaches, questions are asked at different scales (e.g., Edwards and Wratten 1983) and
from differing perspectives (either plant-oriented or herbivore-oriented).

Dispersion of herbivory over an entire tree may have less impact on the tree than the
same level of herbivory concentrated on individual leaf clusters or branches (Edwards
and Wratten 1983). This suggests that induced-resistance may dissuade herbivores from
feeding in proximity to prior leaf damage. Furthermore, increased herbivore movement

may be correlated with higher mortality, presumably because it increases vulnerability to

natural enemies and dispersal losses (Edwards and Wratten 1983, Schultz 1983, although
see Fowler and Lawton 1985 for alternative hypotheses). Several species of caterpillars
assayed in laboratory and ﬁeld experiments did not respond behaviorally to damaged
leaves in any predictable way (Hartley and Lawton 1987, 1990). In some experiments,
leaves partially damaged early in the season have reduced levels of herbivory later in the
season (Hunter and Schultz 1995, Wold and Marquis 1997), suggesting that herbivorous
insects may avoid previously damaged tissue. However, complete defoliation may have
opposing effects, with some herbivores beneﬁting while populations or guilds of other
species are reduced (Faeth 1988). Interpreting the results and identifying the relevant
mechanisms in behavioral/preference experiments has proven difﬁcult. Furthermore, the
relevance of these processes to herbivore population dynamics remains to be
demonstrated.

In early studies of RIR, a primary objective was to quantify the impact of localized
and systemic responses to herbivory. These experiments used low levels of defoliation
(often by tearing individual leaves) and comparisons were made among undamaged
controls, damaged leaves, and intact leaves immediately adjacent to damaged foliage.
Growth of the autumnal moth, Epirrita autumnata (Lepidoptera: Geometridae) on
mountain birch, Betula pubescens, was reduced when fed either damaged leaves
(localized response) or intact leaves that were adjacent to damaged leaves (systemic
response) (Haukioja and Niemela 1977, Haukioja and Hanhimaki 1985). In an
experiment addressing systemic induction caused by leaf-tearing, autumnal moth pupal
masses were reduced by 0-14% and survival was markedly lower (Haukioja and Niemela

1979)

 

 

 

More recently, Havill and Raffa (1999) compared RIR induced by low levels of
gypsy moth, Lymantria dz'spar (Lepidoptera: Lymantriidae) feeding damage, jasmonic
acid (a well deﬁned biochemical elicitor of induced-responses), and mechanical damage
(scissors) among 12 poplar, Populus spp., cultivars. Not only did the type of damage
inﬂuence induction, the strength of the induced resistance on gypsy moth varied by as
much as 72% between different cultivars. Follow-up studies showed that defoliation
reduced the mass of third instars by 25% when larvae were fed foliage from clones found
to be most inducible (Havill and Raffa 2000). Similar effects on growth have been found
in early instar forest tent caterpillar, Malacosoma disstria (Lepidoptera: Lasiocampidae),
where partial defoliation of poplar reduced growth by three fold (Robison and Raffa
1997). Results of these experiments indicate that systemic induction occurs in poplars as
a response to real or simulated herbivory, and that the impact on herbivores can be
relatively strong. Similar research on other species of trees is needed to determine the
generality of these responses. As with the behavior/preference experiments above,
however, the signiﬁcance of these results to natural populations of insects is not known.

Many investigations of plant-herbivore interactions have used short—temi (one instar
or less) bioassays as surrogates for labor intensive bioassays encompassing the entire
larval stage. While growth rate is generally considered to be correlated with pupal mass
(Ayres et al. 1997), and hence fecundity in outbreak Lepidoptera with non-feeding adults,
this may not always be the case. Some instars are more sensitive to foliar quality than
others (Hanhimaki and Senn 1992, personal observation) and furthermore, critical periods
in foliar development or induction may be missed with short—term bioassays. While these

studies suggest possible consequences of RIR on ﬁtness, caution should be used in

10

 

 

extrapolating population effects from the results unless accompanied by corroborating
full-length larval assays.

Experiments designed to simulate the more intense herbivory associated with
outbreak species have also yielded variable results. Manual defoliation (50%) of
trembling aspen, Populus tremuloides, slightly increased development time and reduced
growth of gypsy moth, whereas the same treatment applied to sugar maple, Acer
saccharum, had no effect (Lindroth and Kinney 1998). In contrast, defoliation did not
affect growth of forest tent caterpillar on aspen but reduced its growth by 20% on sugar
maple (Roth et al. 1998). Defoliation of sugar maple and paper birch, Betula papyrifera,
had no signiﬁcant effect on early instar growth of either forest tent caterpillar or gypsy
moth (Dankert 1996). These results highlight the variability of experiments assessing the
effects of RIR on caterpillar growth.

The effects of RIR on pupal mass and fecundity have been quantiﬁed in only a few
studies. Rossiter et al. (1988) used gypsy moth larvae to generate a 10-58% gradient in
defoliation of red oak, Quercus rubra, and then assessed. the effects of the treatments by
rearing gypsy moth larvae in sleeve cages on the trees from hatch to pupation. Female
pupal mass was negatively correlated with defoliation and was reduced by as much as
20%. Progressive, manual defoliation of black oak, Quercus velutina, and gray birch,
Betula populz'folia, to mimic gypsy moth feeding over the duration of the larval period,
reduced female gypsy moth pupal mass by as much as 18% on oak (Wallner and Walton
1979, Valentine et a1. 1983). On birch, effects of defoliation on pupal mass were more
variable and both increases and decreases were recorded depending on the year. In a

recent study, defoliation (>70%) of red alder, Alnus rubra, using western tent caterpillar,

11

 

 

Malacosoma caliform'cum (Lepidoptera: Lasiocampidae), reduced fecundity by more
than 14% (Rothman 1997). Variability in the effects of RIR among studies may be in part
due to differences in the levels of defoliation used (see Table 1). Roland and Myers
(1987) found that reductions in pupal mass of winter moth, Operophtera brumata
(Lepidoptera: Geometridae) appeared to be correlated linearly with increasing defoliation
in apple, Malus sp. and Garry oak, Quercus garryana.

Data from natural populations of gypsy moth also suggest that RIR can reduce pupal
mass and fecundity while lengthening development times. Relative to an undefoliated
site, pupal masses were reduced by 22% for laboratory reared caterpillars fed foliage
from red oaks with a single year of defoliation and 32% from a site with one year of
previous and one year of current defoliation (Lance et al. 1991). In similar studies of
natural populations, Myers and Williams (1984, 1987) found that RIR had little or no
effect on pupal mass of western tent caterpillar. However, as Neuvonen and Haukioja
(1985) pointed out, this type of experiment may be inherently confounded because the
insects and not the experimenters selected the trees. Furthermore, pathogens or stress
factors such as crowding may also reduce fecundity in outbreak populations, and, the
effects may not easily be separated from those of the host tree. These types of natural
experiment should be interpreted with caution.

The role of RIR in population dynamics of outbreak forest insects has often been
downplayed or ignored. This is unfortunate because, although RIR does not function in a
delayed density-dependent manner and thus can not directly contribute to population
cycles, it still may play a role in population dynamics. For example, RIR can act to slow

population growth, allowing other factors to exert greater negative pressure on population

12

 

 

growth. Temporal variability in RIR could also be important in population processes
because annual ﬂuctuations in its strength may allow populations to escape regulation.
Variability in RIR across landscapes could contribute to spatial heterogeneity in
population densities, contributing to changes in natural enemy activity. For these

reasons, research on the effects of RIR remains valuable and should be continued.

Delayed induced resistance — direct effects on herbivorous insects

In contrast to the variable effects of RIR on caterpillars, DIR has had generally negative
effects on caterpillar ﬁtness (Table 2). Large contributions to the understanding of DIR
have come from research on birches and their defoliators, particularly mountain birch and
autumnal moth in Fennoscandia. Ruohomaki et a1. (1992) summarized 24 separate
experiments conducted in this system over a 15-year period. Although not always
statistically signiﬁcant, their review showed that relative to control trees, pupal mass was
reduced in more than 90% of the experiments following a single year of defoliation.
There was considerable annual variation in the strength of DIR with pupal mass
reductions ranging from 0 to 44% depending on the year. Interestingly, the largest DIR
effects occurred in years where larvae grew the best on control trees. Similar long-term
declines in the quality of foliage following defoliation have been recorded from paper
birch in Alaska. Experiments designed to simulate severe outbreaks of the black-marked
spear moth, Rheumaptera hastata, an important defoliator in this system showed that in
the year following defoliation, survival was reduced by 19% (Werner 1979). In another
experiment, a ca. 28% decrease in pupal mass was recorded for R. hastata caterpillars fed

foliage from birch trees completely defoliated in the previous year (Bryant et al. 1993).

13

 

 

0.0000

 

000020.000 w530=£ @003

3003 EEme 50% E 20:00:80 00 80E Rana :0 300.000 02 5.22% 5%me 28.593 8:303. .03 00005 ~

6.00.0 0.53%

Ammo: 80:. Rana 00800000 QQOmA 20>“: .mmmE Rang @003

$032 % 0030M 0080008 :8» 000305 E foo? £0>0L 0000:0009 305:3 5030.0me 559.5% §E§Q -0: 00003 _
9208:0000 00530000: E @000:

$003 .3 00 Embm o\omv 3 EEK“; 0:« $3.. 3 0002000 08E Guam $5.3: abie§$$~ 505.3ch 038% 12:82 _
00081098 can 000% @002
33: :0 90000000 $310 0002000 0003 8088 Emma .08 #05008

.10 00 30:00:03” .NQQTQE 38050090 00080.50 E 00 3030M 5:52:28 2.55%“ E00800§ 538% .8005 _

 

g0 b00083 00502800
030305
080.08% 200200 8 020200 00005 880% 00005 880% 00; 09? @000;

 

0000658 000 H 0?.0 00000000 8% 8005 000298:
00 A002 530% .08: 08802080 .088 $003 05808 Emmobzm 00 03:00:08 00 @6560: 05808 00006 0 t 038 05 E 000205
300 0003 863m 00808 860% 05 52: 0050 860% 000%: 00 0:0 83 0050:0000 05 880 0800 5 80580000 :23 80b
.00 couﬂgoﬁ 3000:0090 >0. 00 830530 0002: ESE: E .8202 @5000 8 £08000 wEms 3:03:05 0050 00:30 83 003020009
000030000 maﬁa $8805 05 9 Sig 800% 00 S0» wEBEw 05 E 0000:0000 30003800 00 b30000 08>» 80: t 0000200

0003 8EBm .80b 8052000 :0 wcmwoom 000302004 889:8 00 000080803 05 :0 0090:0000 00 $00,006 E00-wcoq .N 0.9;.

14

0.0.000

 

 

53:
:0 00 00500009

383 000003

0009
:0 00 E08003:

203:

202:3 0 E02

68: .0 00 30830

:3: 56530

:8: .0 00 080500

:30 0
:0 00 0000000000052

.32 .0 00 893

:00: >00w 000

00: 0:: 0:00 0:003 :0 80:0 0003 0020000 5 800:000
0300:2050 5 0000080 8000:0000 03080006

0000:0000 $00: 00 0000» 0305000000 m 0000 .N
.: $032000 $00 000 .00 .00 3 0000000 00>0>Sm

5000-0 3 0003000 0003 0000» N 00 000000000
00080: was/0:00 000% E 000380 808:0 00030

800% 0003000 :0 0000:0000
8000 :0 $2-0 03 0003000 80000 0095 000800

0002000 3000000000w00 _0>0>0:m 00000 N 00000 Room
.000» : 00000 808 02030 0020000 :0 00005000 AVON:

.0000:0000 $030200

000» E 0900000000 000 08000 00000 000 0000000 02

.000» wank/0:00 0:0 E 0x02 03 £38m 0000000
000830 0000000 020000000 0000 08000 03 0000:000D

.8000 0000:0000 b32800 000 DEE?
008000000 0008 00050 .8000200000 0000000000:
E 8000 00.05000~ 000 0000:0000 00 000000 02

dw080 00005000000

I 8000 0000200 0000200 .0000:0000 0:030:00
000% E 0008 00020 00000000 :0 00000000 Ax; Tm

.8500 00:20:03

8000600 0003.02300E

0:055:30 3.0.2.5.m

500.000.5530 0500000050

0:850:0Q

00.005000000000000Q

050000005000 050000000050

000.000 00:00:03

0000:3000 80.2.3.0

00000.00 00.030503

SEESQQQ 550%

0200:0000 000000®

000.0.0AQ0Q SEmm

000008000. 350%

0.50: 85%

802000 00000.0

030.0 0.0200.

0.005030% 550%

02008005. 250%

0.53.0 0.00.0000Q

£00:
0202 0. 0% .m .N .0

$02
-03 0352 m 0% .0 .0

00.00
-00 0080 N

 

$5 6080 m

00.00:
102002 N 0000 0

15

00.00
-8 5080 0
@0000 38:0 0
@000
03:02 0
@0000 38:0 0

 

0.08 N 05.0.0

 

 

 

000000 00 300000000
2000000 :03 0000000000 000 000 %00000.0m 000%
Em 00 0000000000 000M000 0200—00000 000000m .o\omn $0000:
0000 ._0 00 0000000 .00 %0 0000000 _0>_>000 000 .00000 00000 003000 0000000000000 0000000000003 0000000000000 000000.000 $0002 m 000 .N .0

0000 000% 000000 00 00000 200000 0:000 00 000000000 0x0:
0:00:03 000 000%: .000% 000 0.0 00000 .0000 0:00:00 00 00.000000 0x03 0000.000000000 000000000000 0,000.0 000; A000 000000 m

000000 0300000000

00 000 :005 %000 00 0000:0000 00000 0000000 00000 0000000000. 000.0%
080 000:0 20800 .08 0003 0 8.0003000 0080 0 000 0.0.0000
00:03 000 000003 0 00000 0005 00000 0000000 00 0000000000 0300000000 00000.00 00.000010 00.00000 0000000® :00002 N 000 0
0000000 000% 000000 00000 00000000 %=00000_0w00 0000020 0000000000000
0000: 000 0000000 0300—00000 38.0: :0000 000003 0.0?
.0 00 0000000030 000 $2-: 0000000 00008 :00000000 00 00000 :0000 000000000 00.2.0030 00000000Q 00000m -va :00002 N 000 0

0.0000 N 030.0

 

16

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Traditionally, experimental tests of DIR have consisted of a single year’s defoliation
followed by phytochemical analyses and/or bioassays in the following year or years. This
approach may underestimate the strength of DIR because outbreaks of many defoliators
are at least two years in duration (e.g., Mattson et a1. 1991). Haukioja et al. (1988)
hypothesized that consecutive defoliations may have cumulative effects on herbivore
performance. Despite this speculation, the effects of cumulative defoliation on trees and
herbivores have received only limited attention in the literature (Kaitaniemi et al. 1999b).

Some studies have shown that there may indeed be cumulative negative effects of
successive years of defoliation on herbivores. Werner (1979) found that each additional
year of defoliation increased the mortality of spear-marked black moth feeding on paper
birch. In similar studies, Wallner and Walton (1979) and Valentine et al. (1983)
defoliated black oak and gray birch for several successive years. Gypsy moth pupal mass
progressively decreased with each additional year of defoliation. Clausen et a1. (1991)
defoliated trembling aspen for three years in succession and found that 18-75% decreases
in pupal mass of large aspen tortrix, Choristoneura conﬂictana (Lepidoptera:
Tortricidae), with the largest losses in the third year. More recently, Kamata et al (1996)
found that pupal mass of beech caterpillar, Quadracalcarifera punctatella (Lepidoptera:
Notodontidae), was reduced by 12% and 26% following one and two years of complete
defoliation, respectively.

Not all studies have shown additive effects of consecutive defoliations. Kaitaniemi et
al. (1999b) found that two years of 75% defoliation on mountain birch did not have
greater effects on winter moth and autumnal moth than a single year of 75% defoliation.

In a unique experiment, Kaitaniemi et al (1999a) maintained undefoliated control trees

17

 

 

 

 

using insecticide throughout a natural outbreak of autumnal moth. Unprotected trees
suffered two consecutive years of defoliation although unlike previous outbreaks, the
amount of foliage removed, 15-25% in 1992 and 30-45% in 1993, was not high.
Bioassays indicated that DIR was weak, reducing pupal mass by only 2—1 1% relative to
control trees following two years of defoliation. These researchers also used regression
to convert the mass of adult males captured in pheromone traps to female equivalents to
determine the reductions in fecundity present in natural populations. They found that
fecundity was decreased maximally by 18% in the year following peak density (two
consecutive years of defoliation). Results from these natural and experimental studies are
well within the range of DIR effects recorded from single year defoliations (Ruohomaki
et al. 1992) suggesting that there were no cumulative effects of defoliation, or
alternatively, that the levels of defoliation were too low to elicit maximal DIR responses.
Delayed induced resistance has been postulated to be a general mechanism driving
the cyclic dynamics of forest insects (e.g., Haukioja 1980, Rhoades 1983). However, a
number of studies have suggested that DIR may be more system speciﬁc than ﬁrst
thought. Roland and Myers (1987) found that moderate levels of defoliation (<50%) in
the previous season actually increased pupal mass of winter moth in the subsequent
season on both apple and oak whereas on apple, higher defoliation (>50%) in the
previous season reduced pupal mass. In some systems, even severe defoliation does not
elicit DIR. Rothman (1997) found that severe defoliation of red alder had no effect on
the mass of western tent caterpillar pupae in the year following defoliation. Similarly,
Harrison (1995) showed that defoliation in the previous year did not affect fecundity of

western tussock moth, Orgyia vetusta (Lepidoptera: Lymantriidae) feeding on bush

18

lupine, Lupinus arboreus. It is interesting to note that both red alder and bush lupine a1
nitrogen-ﬁxing species. Species which ﬁx nitrogen may be buffered from the nitroge
losses that follow defoliation, or may more rapidly recover to predefoliation levels (se
Tuomi et al. 1990).

Since studies of DIR have been confined to a few systems, it is unknown if th
results can be generalized to other species. Among the 10 species with the large:
outbreaks on deciduous trees in North America (Mattson et a1. 1991), research on D12
has been conducted for just three species (gypsy moth, large aSpen tortrix, black-marke
spear moth). In addition, the paucity of studies that examined DIR over time scales the
characterize natural outbreaks suggest that we have only a rudimentary understanding c

the role of DIR in population dynamics.

Mechanisms underlying induced responses in trees

Variability in the effects of RIR and DIR on herbivorous insects may in part stem fror
the effects of underlying environmental heterogeneity on host trees. A number (
mechanisms have been proposed to explain the induction of resistance followin
defoliation. These can be broadly grouped into ’supply-side' hypotheses, focusing on t1“.
partitioning of resources among primary and secondary functions and 'demand-sid
hypotheses, focusing on wound-induced demand for secondary compounds (Lerdau et a
1994, Karban and Baldwin 1997). With respect to deciduous trees, the carbon-nutriei
balance (CNB) hypothesis (Bryant et al. 1983), a subset of predictions under the mor
globally encompassing growth—differentiation balance hypothesis (Herms and Mattsc

1992), has received the most attention of the supply-side hypotheses. The major premi:

19

of the CNB is that growth is more sensitive to changes in nutrients than is photosynthesis.
When growth is limited by nutrients, excess carbon from storage or photosynthesis can be
allocated to production of carbon-based secondary compounds, hence the 'supply-side’
moniker. Rather than a passive allocation of carbon, demand side hypotheses view
resistance as a hormonally mediated response system to various stress factors such as
wounding (Karban and Baldwin 1997).

Baldwin and Karban (1997) suggest that the majority of rapidly induced chemical
responses of plants to herbivory can be explained by the demand-side active response
theories. However, both passive 'supply' hypotheses and active 'demand' hypotheses will
likely be required to explain completely the full temporal range of induced-resistance
responses in trees (e.g., Tuomi et al. 1990, Herms and Mattson 1992, Koricheva et a1.
1998). Integration of different hypotheses across temporal and hierarchical scales has
been hampered by a dichotomy in the research focus and methodology of ecologists
investigating DIR, and plant physiologists and molecular biologists investigating RIR.

Mechanistically, the CNB is predicated on the availability of nitrogen, which is
generally regarded as the growth-limiting nutrient in temperate and boreal forests.
Defoliation, especially early in the growing season, decreases nitrogen stores in
deciduous trees (Harper 1977, Chapin and Kedrowski 1983) and may damage ﬁne roots,
reducing nitrogen acquisition from soil, and exacerbating the nitrogen deﬁcit (Tuomi et
a1. 1990). Thus, until a tree can recover or compensate for the nitrogen lost when
defoliated, production of photosynthate will continue to exceed the carbon needs of

growth processes, maintaining elevated levels of carbon-based secondary compounds.

20

Decreases in foliar nitrogen following extensive defoliation appear to be a genera
phenomenon in deciduous trees (e.g., Tuomi et al. 1991, Bryant et al. 1991). Nitroger
decreases can occur within damaged leaves following herbivory, and if the defoliation i:
extensive, may carry over into the following growing season. Following defoliation
within-season declines in nitrogen have been found in various cultivars of poplars, aspen
sugar maple and a number of other tree species (e.g., Robison and Raffa 1997, Lindrotl
and Kinney 1998, Roth et al. 1998). The mechanisms behind within-season nitroger
decreases in damaged leaves are not well understood. Some nitrogen may be leachec
from the leaves or the tree may withdraw nitrogen from the leaf as has been shown witl
other nutrients (Chapin 1980, Nef 1988). Trees that are severely defoliated produce 2
second set of leaves in the same growing season from dormant buds. Stored nitroger
pools that are available for leaf expansion during bud break in the spring are not available
when trees produce a second set of leaves. Thus, following refoliation, late season leave:
will be signiﬁcantly lower in nitrogen than leaves on trees that were not defoliated.

Long-term, across-year depression of foliar nitrogen following defoliation is more
easily explained than within season changes. Following severe defoliation, recovery 0;
nitrogen may require a year or more in trees growing in nutrient poor soil. Several. years
were required for foliar nitrogen to return to predefoliation levels in defoliated mountair
birch in Fennoscandia (Tuomi et al. 1984). Recovery was shorter for paper birch ir
Alaska (Bryant et al. 1991) and willow, Salix sunnamen, in Finland (Tuomi et al. 1991:
although two years were still required to reach predefoliation levels. Lowe]
concentrations of foliar nitrogen in the year following artiﬁcial or natural defoliation have

also been documented in black oak, gray birch, and red alder (Valentine et al 1983

21

Myers and Williams 1987), suggesting that it may be a general consequence o
defoliation.

The CNB predicts that reductions in nitrogen should lead to increases in carbon-basec
secondary compounds. Congruent with prediction, foliar nitrogen levels following
defoliation have been found repeatedly to be negatively correlated with concentrations 0
phenolics (Tuomi et a1 1990, Bryant et al 1991, 1993). Under the CNB hypothesis
nitrogen fertilization is predicted to attenuate levels of phenolics. In some studies
application of nitrogen fertilizer to defoliated deciduous trees reduced levels of phenolic
while other studies did not show this pattern (e.g., Haukioja and Neuvonen 1985, Bryan
et al. 1993). The actual response of secondary metabolites to fertilization in defoliate<
trees may be dependent on both the ambient soil fertility (Herms and Mattson 1992) am
the speciﬁc phenolic compounds measured (Koricheva et al. 1998).

Following complete defoliation, the phenolic concentrations in reﬂush foliage ma;
revert to levels similar to that found after bud break. Baldwin and Schultz (1982) founi
only low levels of condensed tannins in the reﬂush leaves of red oak following complet
defoliation by gypsy moth, and Faeth (1988) observed a similar pattem in Quercu
emoryi. In contrast, damaged leaves remaining on the tree and sampled at the same tim
as reﬂushed leaves had very high levels of condensed tannins in both studies. Becaus
new leaves are major carbon sinks, growth should take precedence for available carbo:
leaving little for production of condensed tannins.

The CNB hypothesis does not predict some types of rapidly induced changes in tree
including rapid systemic responses induced by low levels of herbivory. In herbaceou

plants, several wound-responsive, signal-transduction pathways have been identiﬁe

22

(e.g., Bergey et al. 1996, Karban and Baldwin 1997, Constabel 1999). In response tr
herbivory, speciﬁc elicitors stimulate secondary metabolism and/or translocation o
secondary compounds from storage sites. Several studies have shown that insect feedin;
damage to the leaves of woody plants elicits different responses than equivalent amount
of tissue removed mechanically (e. g., Haukioja and Neuvonen 1985, Hartley and Lawtor
1987, Krause and Raffa 1992, Havill and Raffa 1999). This suggests that leaves 0
woody plants have insect—speciﬁc wound responses, similar to those demonstrated i1
numerous herbaceous plants. Furthermore, application of the generalized elicito
compound, jasmonic acid, has been shown to induce RIR in hybrid poplars in the absenc:
of tissue damage (Havill and Raffa 1999), indicating that a signal-transduction pathwaj
may be operating. The role of phenotypic variation in rapidly-induced systemic woum

responses in the leaves of trees has not been investigated.

Indirect effects of induced resistance - competition among folivores

Historically, interspeciﬁc competition has not been regarded as an importan
component structuring phytophagous insect communities (e.g., Hairston et al. 196C
Schoener 1983, Strong et al. 1984). Considerable evidence has accumulated suggestin;
that this view was inaccurate. In a comprehensive review, evidence for interspeciﬁ
competition was found in more than 75% of the cases examined (Denno et a1 1995:
Among chewing folivores, plant-mediated, indirect competitive interactions, such a
those driven by induced-resistance, accounted for more than half of reported negativ

interactions between species. Thus, changes in plant quality elicited by caterpillar feedin

23

can have signiﬁcant effects on other members of the phytophagous insect commun
utilizing the same hosts.

Because outbreak species have such a profound impact on the foliage resource, 'tlf
may have particularly large effects on the ﬁtness of competing folivores. T
vulnerability of folivorous insects to indirect competition from outbreak species media
by plant-quality may depend on several different factors. These include, the timing
defoliation in relation to the phenology of competing folivores, life-history attributes sr
as leaf-rolling, leaf-mining, or open feeding, and the degree of dietary specializati
(Hartley and Lawton 1987, Bowers and Stamp 1993, Denno et al. 1995, Dankert et
1997, Agarwal and Karban 1999). Non-outbreak species may be less adapted to lar
changes in nitrogen, water, leaf toughness, or damage-induced secondary compour
than outbreak species.

Experiments have demonstrated that host-plant mediated indirect competition dc
occur and can have signiﬁcant effects. Defoliation of paper birch, sugar maple, and asp
by forest tent caterpillars induced both RIR and DIR, reducing the performance of otl
outbreak species [gypsy moth and white-marked tussock moth, Orgyia leucostigr
Lepidoptera: Lymantriidae)] as well as tiger swallowtail, Papilio canaden
(Lepidoptera: Papilionidae), a non-outbreak species (Dankert et al. 1997). Similar
Harrison and Karban (1986) found that early season feeding by an arctiid caterpillar
bush lupine reduced the quality of foliage for a later feeding tussock moth speci
affecting its growth signiﬁcantly.

Early season feeding damage may induce-effects that continue for the remainder

the growing season. Both manually applied damage and naturally occurring gypsy mc

24

defoliation reduced the probability of subsequent damage by late season feeding ins
(Hunter and Schultz 1995, Wold and Marquis 1997). In forests, gypsy moth outbre
may reduce the population densities and/or species diversity of other lepidopter
(Sample et a1. 1996, Work and McCullough 2000), although the effects documented v
small and restricted to a few taxa. However, not all studies have shown detrime
effects of early season damage. Spring defoliation of Quercus emoryi resulted in hig
densities of some leaf-mining species on late season leaves presumably because w
content was higher in the refoliated leaves (Faeth 1988). The mechanisms underl)
competitive interactions between outbreak folivores and other species are not kno
Using wild radish, Agrawal (2000) demonstrated that the effects of induced-resista
may vary with different species of herbivores and in turn, plants may perceive and r:
to herbivory by different species idiosyncratically. It is not known if trees and ti
associated herbivores respond similarly, particularly to defoliation. In 1
comprehensive evaluation of the competitive effects of outbreak species on o
folivores are lacking in most systems and methodological problems in estimating den

and species abundance confound interpretation of existing studies.

Indirect effects of induced resistance — interaction with the third trophic level.

Twenty years ago, Price et al. (1980) published an inﬂuential paper suggesting that
relationship between herbivorous insects and their host plants need to be considerei
context with higher trOphic levels. The importance of tritrophic interactions in ecol
was not immediately recognized. Indeed, a search of ISI Citation Abstracts revealed (

three papers in the 1980’s using the keyword ‘tritrophic’, although a larger numbe

25

papers were published covering this subject area. In the decade following, there h;
a veritable explosion of papers on tritrophic interactions, 200 papers publishe
1990, with an additional 30 published through July of 2000 alone. The ra
interactions falling under the tritrophic umbrella is extremely broad. More
literature searches suggest that while knowledge on the inﬂuence of plant-1
infochemicals on host seeking behavior in parasitic wasps is growing exponential]:
areas including interactions between plant chemistry and pathogens, as well a
chemistry and entomophagous predators, have lagged behind.

The effects of induced resistance in trees on insect pathogens have not be<
studied. Most research to date has focused on interactions between gypsy moth,
moth nuclear polyhedrosis virus (LdMNPV), and the secondary chemistry of o
outbreaks of gypsy moth, NPV is generally the dominant source of mortality
integral to the collapse of many high density populations. Oaks contain hydro
tannins that may precipitate proteins. Gypsy moth defoliation increases
concentrations in oaks (Schultz and Baldwin 1982, Rossiter et al. 1988), ar
induction is thought to reduce the susceptibility of larvae to mortality from
Laboratory experiments have conﬁrmed that higher tannin levels in defoliated t:
inhibit baculovirus infection (Hunter and Schultz 1993). However, in ﬁeld
D’Amico et a1. (1998) found no evidence that NPV activity was affected by def<
on two species of oak. They attributed their ﬁndings to the lack of RIR during the
time period early in the season when most NPV transmission between larvae occur

Research to date has concentrated primarily on the effects of within season cha

phytochemistry on pathogen interactions. Only two studies have examin

26

interaction between defoliation in the previous year and herbivore susceptibility to
pathogens. Wallner (1983) found that gypsy moth were more susceptible to NPV when
fed leaves from oak trees which had been severely defoliated one year previously than on
undefoliated controls. Conversely, Rothman (1997) found no interactive effect of westem
tent caterpillar NPV with previous defoliation, although there was no direct effect of DIR
on the caterpillars either. Phytochemicals may exhibit multiple modes of action when
interacting with other biochemical processes and pathogens, and these interactions may
be context dependent. Thus, the effects of induced-resistance on mediating insect
susceptibility to pathogens remain to be determined.

Very few studies have speciﬁcally addressed possible relationships between predators
and either RIR or DIR in trees. A recent study examining constitutive resistance in
willows and the inﬂuence of fertility regimes on the impact of predators indicated that
this may be a fruitful area of research. Sipura (1999) found that bird predation of
phytophagous insects was elevated on unfertilized willows of the genotype with highest
constitutive resistance. Invertebrate predators can also interact with induced—responses in
trees. Pear trees infested with psyllids were found to release volatiles attractive to
anthocorid predators (Drukker et al. 1995). Further study indicated that the predators
were not cueing on the insects themselves but rather on volatiles released by the tree in
response to their feeding (Scutareneau et a1. 1997). These recent developments may
herald an expanded interest in predators as a component of tritrophic interactions.

A number of studies have shown that herbivore damaged plants emit signals that are
attractive to parasitoids of caterpillars (e.g., Eller et al. 1988, Turlings et al. 1993, Thaler

1997, Dicke 1999). Thus, increasing emphasis has been placed on the study of plant

27

attributes such as nutritional status, production of secondary compounds, and physi
structure that may enhance or interfere with parasitoid activity (Agrawal 2000). W
the focus of such research has generally been with agricultural plants, there is increasi
evidence that similar tritrophic interactions occur in trees (e. g., Havill and Raffa 2000).

Although hymenopterans have been the focus of research in agricultural systems, ‘
relationship between tachinids and induced-responses have received more attention
trees. It has been recognized for more than 60 years that leaves damaged by feedi
caterpillars are attractive to some tachinids (Bess 1936). In some cases, speci
compounds released by plants and used as cues by tachinids searching for hosts he
been identiﬁed (Roland et al. 1995). Studies addressing the effects of RIR and DIR on .
host location of parasitoids in trees are for the most part lacking. In a recent contributir
Havill and Raffa (2000) found that p0plar leaves damaged by gypsy moth were more tl
three times more attractive to females of the gypsy moth parasitoid Glyptapante
ﬂavicoxz’s (Hymenoptera: Braconidae), than undamaged control leaves. Damaged leai
also signiﬁcantly reduced the growth of gypsy moth caterpillars, indicating that RIR 1r

have a synergistic interaction with this parasitoid.

Induced-resistance and population dynamics

Simulation models have suggested that RIR can regulate herbivore populations and ti
DIR can create cyclical population dynamics in the absence of other density-depend
factors (Edelstein-Keshet and Rausher 1989, Underwood 1999). In Underwood’s (19!
model, outcomes were contingent upon both the strength and timing of induci

resistance. Whether populations were regulated or cycled depended on both the der

28

rate of the resistance and the lag time to maximum strength of the induced-response
date, the effects of RIR on a variety of forest insects have not been large. W
simulation models do not specify absolute values required for regulation, in ft
systems at least, the weak RIR recorded in experiments appears insufﬁcient to regr
insect populations.

Only one study has looked at the effects of induced-resistance on the popula
growth rates of an outbreak folivore. In the absence of natural enemies, Haukioja e
(1987) calculated that for autumnal moth feeding on undefoliated trees, r = 2.9 where
was r = 1.1 on trees defoliated in the previous year, where r is the intrinsic rate of nat
increase for the population. While DIR had a substantive effect on r, acting alone, it
incapable of preventing this population from continuing to grow. In natural systems, i
is not operating independently of other factors and its effects may be sufﬁcient to 5
population growth to the point where other time-lagged factors can drive popula
density down.

Currently, synthetic studies are lacking where the effects of DIR are meaSi
relative to the contributions of other top-down and bottom up regulating factors. U
primarily historical data, Bylund (1995) demonstrated that delayed density-depen
parasitism was the most important factor correlated with ﬂuctuations of autumnal r.
populations in Sweden while other factors including induced-resistance contributed 11
less to population regulation. In another study, Virtanen and Neuvonen (1999) estim
the relative contribution of altitude, climate, host quality, and parasitoids on autur
moth in Finland over a three-year period. Like Byland (1995), they concluded

parasitoids were more important than host plant quality in determining population tre

29

These studies suggest that induced resistance, which has long been thought to drive '

system, may in fact play a lesser role in determining population dynamics.

Overview of dissertation

My dissertation is focused on four general themes: (1) What are the consequences
several consecutive years of defoliation on growth and phytochemistry of trees and
these variables affect ﬁtness parameters of folivorous Lepidoptera using these trees
host plants? (2) Can DIR account for the large reductions in fecundity that characteri
the declining phase of outbreaks in many outbreak Lepidoptera? Population density 01
decreases concurrently with fecundity, suggesting that some delayed-density depend
factor is driving the change. I was interested in whether DIR can account for all or at It
a substantial portion of this fecundity loss. (3) Does defoliation by an outbreak spec
have signiﬁcant indirect, host-tree-mediated effects on other folivorous species belong
to different seasonal guilds and exhibiting different life-history strategies? (4) Can I
and DIR enhance or impair the success of parasitoids attacking outbreak folivores?

To address these questions, I used two poplar systems, one in a natural fo:
environment (Populus tremuloides) and one in replicated blocks of hybrid pol
(Populus x euramericana). The use of poplars offers several advantages for address
questions about plant-herbivore interactions. First, both of my experimental syste
consisted of well-deﬁned clones. Using clones greatly reduces the inter—tree variabi
that has plagued earlier studies of RIR and DIR. Secondly, the phytochemistry of pop:
is reasonably well known and less complex than other species frequently used in sim

experiments such as birches and oaks. Third, trees in the genus Populus are prefer

3O

hosts for two dominant outbreak species, forest tent caterpillar and gypsy moth, all
for experiments in which the experimental trees are defoliated naturally.

In my dissertation, I describe two long-term studies where trees were defoliai
caterpillars to simulate natural outbreak events. In Chapter 2, the results of a study
experimental outbreaks of gypsy moth were established in a hybrid poplar plantati
described. Half of the plots were protected from defoliation. In addition, in
fertilizer was applied to half of the defoliated and undefoliated plots to examii
effects of fertility on tree and herbivore responses to defoliation. The experiment w
for four years and emulated the rise (1 year of light defoliation), outbreak (one y
severe and one year of moderate defoliation), and collapse (one year of ver
defoliation) in population densities that characterize natural gypsy moth outbrea
quantiﬁed the effects of the defoliation and fertilization treatments on pupal
fecundity, and development time of gypsy moth and ﬁve other lepidopterans found
p0plar plots. The species represented a diverse range of life history strategies 2
free-feeding Lepidoptera.

In Chapter 3, I describe an experiment where trees belonging to two different
clones growing in a natural forest environment, were inoculated with high densi‘
forest tent caterpillar. Ten trees in each clone were defoliated for three consecutive
while another ten trees were undefoliated. In the fourth year, ﬁve of the prev
defoliated trees were left undefoliated and ﬁve of the control trees were defoliated :
ﬁrst time. This provided a set of four treatments allowing comparison of the effc
RIR (ﬁrst time defoliated), DIR (three previous years of defoliation), and the con

effects of RIR and DIR (three previous and one current year). I compared fore

31

caterpillar growth, pupal mass, fecundity and survival across these 4 treatments. In
addition, I examined the effects of RIR and DIR on the probability of parasitism by two
species of tachinids, Leschenaultia exul and Patelloa pachypyga, that specialize on forest

tent caterpillar.

LITERATURE CITED

Agrawal, AA. 2000. Speciﬁcity of induced-resistance in wild radish: causes and
consequences for two specialist and two generalist caterpillars. Oikos 89: 493-
500.

Agrawal, AA. and R. Karban. 1999. Why induced defenses may be favored over
constitutive strategies in plants. Pp. 45-61 in R. Tolhian and CD. Harvell, eds,
The Ecology and Evolution of Inducible Defenses. Princeton University Press,
Princeton, New Jersey.

Andrewartha, H.G. and LC. Birch. 1954. The Distribution and Abundance of Animals.
University of Chicago Press, Chicago, Illinois

Ayres, M.P., T.P. Clausen, S.F. MacLean, Jr., A.M. Redman and PB. Reichardt. 1997.
Diversity of structure and antiherbivore activity in condensed tannins. Ecology
78: 1696-1712.

Benz, G. 1974. Negative feedback by competition for food and space, and by cyclic
induced changes in the nutritional base as regulatory principles in the population
dynamics of the larch budmoth, Zieraphera dineana (Guenee) (Lepidoptera:

Tortricidae). Zietschrift fur Angewandte Entomologie 76: 196-228.

32

 

Bergelson, J .M. and J.H. Lawton. 1986. The effects of foliage damage on case-bearing
moth larvae, Coleophora serratella, feeding on birch. Ecological Entomology
11: 241—250.

Bergey, D.R., G.A. Hoi and CA. Ryan. 1996. Polypeptide signaling for plant defensive
genes exhibits analogies to defense signaling in animals. Proceedings of the
National Academy ofScience 93: 12053-12058

Berryman, A.A., N.C. Stenseth and AS. Isaev. 1987. Natural regulation of herbivorous
forest insect populations. 0ecologia 71: 174-184.

Berryman, AA. 1996. What causes population cycles of forest Lepidoptera? Trends in
Ecology and Evolution 11: 28-32.

Bess, HA. 1936. The biology of Leschenaultia exul Townsend, a tachinid parasite of
Malacosoma americana Fabricius and Malacosoma disstria Hiibner. Annals of

the Entomological Society ofAmerica 29: 593-613.

Bowers, MD. and NE. Stamp. 1993. Effects of plant age, genotype, and herbivory on
Plantago performance and chemistry. Ecology 74: 1778-1791.

Bryant, J.P., F.S. Chapin, D.R. Klein. 1983. Carbon/nutrient balance of boreal plants in
relation to vertebrate herbivory. Oilcos 40: 357-368.

Bryant, J.P. J. Tuomi and P. Niemela. 1988. Environmental constraints of constitutive
and long—tenn inducible defenses in woody plants. Pp. 367-389 in K.G. Spencer,
ed., Chemical Mediation of Coevolution. Academic Press, San Diego, CA.

Bryant, J.P., K. Danell, F. Provenza, P.B. Riechardt, T.A. Clausen and RA. Werner.

1991. Effects of mammal browsing on the chemistry of deciduous woody plants.

33

Pp. 135-154 in D.W. Tallamy and M.J. Raupp, eds, Phytochemical Induction by
Herbivores. John Wiley & Sons, Inc., New York.

Bryant, J.P., P.B. Reichardt, T.P. Clausen and R.A.Wemer. 1993. Effects of mineral
nutrition on delayed inducible resistance in Alaska paper birch. Ecology 74:
2072-2084.

Byland, H. 1995. Long-tenn interactions between the autumnal moth and mountain birch:
the role of resources, competitors, natural enemies, and weather. PhD
Dissertation. Swedish University of Agricultural Sciences, Uppsala, Sweden.

Chapin, ES. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and
Sysnnnaﬁcsllz233-260.

Chapin, FS and RA. Kedrowski. 1983. Seasonal changes in nitrogen and phosphorus
fractions and autumn retranslocation in evergreen and deciduous taiga trees.
Ecology 64: 376-391.

Clausen, T.P., P.B. Reichardt, J.P. Bryant and RA. Werner. 1991. Long-tenn and short-
term induction in quaking aspen: related phenomena? Pp. 71-83 in D.W. Tallamy
and M.J. Raupp, eds, Phytochemical Induction by Herbivores. John Wiley &
Sons, Inc., New York.

Constabel, P. 1999. A survey of herbivore inducible proteins and phytochemicals. Pp.
137-166 in AA. Agrawal, S. Tuzin and E. Bent, eds, Induced Plant Defenses
Against Pathogens and Herbivores. APS Press, St. Paul, Minnesota.

D’Amico, V., J .S. Elkinton, G. Dwyer, R.B. Willis and ME. Montgomery. 1998. Foliage
damage does not affect within-season transmission of an insect virus. Ecology

79: 1104-1110.

34

Dankert, BA. 1996. Paper birch and sugar maple resistance to insect folivores as
inﬂuenced by defoliation, drought, and fertilization: MS thesis. Michigan State
University, East Lansing, Michigan.

Dankert, B.A., D.A. Herms, D. Parry, J.M. Scriber and LA. Haas. 1997. Mediation of
interspeciﬁc competition between folivores through defoliation induced changes
in host quality. Pp. 71-88 in A. Raman, ed. Ecology and Evolution of Plant
Feeding Insects in Natural and Man-Made Environments. International Scientiﬁc
Publications. New Delhi, India

Dempster, J .P. 1983. The natural control of butterﬂies and moths. Biological Reviews of
the Cambridge Philosophical Society 58: 461-481.

Demio, R.F., M.S. McClure and J .R. Ott. 1995. Interspeciﬁc interactions in phytophagous
insects: competition revisited and resurrected. Annual Review of Entomology 40:
297-331.

Dicke, M. 1999. Chemical ecology of host-plant selection by herbivorous arthropods: a
multitrophic perspective. Biochemical Systematics and Ecology 28: 601-617.

Drukker, B., P. Scutareanu and M.W. Sabelis. 1995. Do anthocorid predators respond to
synomones from Psylla-infested pear trees under ﬁeld conditions? Entomologia
Experimentalis et Applicata 77: 193-203.

Edelstein-Keshet, L. and MD. Rausher. 1989. The effects of inducible plant defenses on
herbivore populations. I. Mobile herbivores in, continuous time. The American
Naturalist 133: 787—810.

Edwards, P.J. and SD. Wratten. 1983. Wound-induced defences in plants and their

consequences for patterns of insect grazing. 0ecologia 59: 88—93.

35

 

Edwards, P.J., S.D. Wratten and RM. Gibberd. 1991. The impact of inducible
phytochemicals on food selection by insect herbivores and its consequences for
the distribution of grazing damage. Pp. 205-221 in D.W. Tallamy and M.J.
Raupp, eds, Phytochemical Induction by Herbivores. John Wiley & Sons, Inc.,
New York.

Eller, F.J., J .H. Tumlinson and W.J. Lewis.1988. Beneﬁcial arthropod behavior mediated
by airborne semiochemicals: Source of volatiles mediating the host location
ﬂight behavior of Microplitis croceipes (Cresson) (Hymenoptera: Braconidae), a
parasitoid of Heliothis zea (Boddie) (Lepidoptera: Noctuidae). Environmental

Entomology 17: 745-753.

F aeth, SH. 1986. Indirect interactions between temporally separated herbivores mediated

by the host plant. Ecology 67: 479-494.

Faeth, SH. 1988. Plant mediated interactions between seasonal herbivores: Enough for
evolution or coevolution. Pp. 391-414 in D.W. Tallamy and M.J. Raupp, eds,
Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Feeny, PP. 1970. Changes in oak leaf tannins and nutrients as a cause of spring feeding
by winter moth caterpillars. Ecology 51: 565-581.

Fowler, S.V. and J. H. Lawton.l985. Rapidly—induced defenses and talking trees: the

devil’s advocate position. The American Naturalist 126: 181-195.

Fowler, S. and M. MacGarvin. 1986. The effects of leaf damage on the performance of
insect herbivores on birch, Betula pubescens. Journal of Animal Ecology 55:

565-573.

36

 

 

Ginzburg, LR. and DE. Taneyhill. 1994. Population-cycles of forest Lepidoptera —- a
maternal effects hypothesis. Journal of Animal Ecology 63: 79-92.

Hanhimaki, S. and J. Senn. 1992. Sources of variation in rapidly inducible responses to
leaf damage in the mountain birch-insect herbivore system. 0ecologia 91: 318-
331.

Hairston, N.G., F.E. Smith and LB. Slobodkin. 1960. Community structure, population
control and competition. The American Naturalist 115: 354-366.

Harper, J.L. 1977. Population Biology of Plants. Academic Press, San Diego.

Harrison, S. 1995. Lack of strong induced or maternal effects in tussock moths (Orgyia
vetusta) on bush lupine (Lupinus arboreus). 0ecologia 103: 343-348.

Harrison, S. and R. Karban. 1986. Effects of an early season folivorous moth on the
success of a later-season species mediated by a change in the quality of a shared
host, Lupinus arboreus. 0ecologia 69: 354-359.

Hartley, SE. and J.H. Lawton. 1991. Biochemical aspects and signiﬁcance of the rapidly
induced accumulation of phenolics in birch foliage. Pp. 105-132 in D.W.
Tallamy and M.J. Raupp, eds, Phytochemical Induction by Herbivores. John
Wiley & Sons, Inc., New York.

Hassell, M.P., J. Latto and RM. May. 1989. Seeing the wood for the trees: Detecting
density dependence from existing life-table studies. Journal of Animal Ecology
58: 883-892.

Haukioja, E. 1980. On the role of plant defenses in the ﬂuctuation of herbivore

populations. Oikos 35:202-213.

37

Haukioja, E. 1982. Inducible defenses of white birch to a geometrid defoliator, Epirrita
autumnata. Proceedings of the 5th International Symposium on Insect-Plant
Relationships, Wageningen.

Haukioja, E. 1990. Induction of defenses in trees. Annual Review of Entomology 36: 25-
42.

Haukioja, E. and S. Hanhimaki. 1985. Rapid wound resistance in white birch foliage to
the geometrid Epirrita autumnata: A comparison of trees and moths from within
and outside the outbreak range of the moth. 0ecologia 65: 223—228.

Haukioja, E. and P. Niemela. 1977. Retarded growth of a geometrid larvae after
mechanical damage of leaves of its host tree. Annales Zoologici F ennici 14: 48-
52.

Haukioja, E. and P. Niemela 1979. Birch leaves as a resource for herbivores: Seasonal
occurrence of increased resistance in foliage after mechanical .damage of
adjacent leaves. 0ecologia 39:151-159.

Haukioja, E. and S. Neuvonen. 1985. Induced-long term resistance of birch foliage
against defoliators: defensive or incidental. Ecology 66: 1303-1308.

Haukioja, E. and S. Neuvonen. 1987. Insect population dynamics and the induction of
plant resistance: the testing of hypotheses. Pp. 411-432 in P. Barbosa and JG
Schultz, eds, Insect Outbreaks. Academic Press, San Diego.

Haukioja, E., S. Neuvonen, S. Hanhimaki and P. Niemela. 1988. The autumnal moth in
Fennoscandia. Pp. 163-178 in AA. Berryman, ed, Dynamics of Forest Insect

Populations. Patterns, Causes, and Implications. Plenum, New York.

38

Havill, NP. and K.F. Raffa. 1999. Effects of elicitation treatment and genotypic variation
on induced resistance in Populus: impacts on gypsy moth (Lepidoptera:
Lymantriidae) development and feeding behavior. 0ecologia 120: 295-305.

Havill, NP. and K.F. Raffa. 2000. Compound effects of induced plant responses on
insect herbivores and parasitoids: implications for tritrophic interactions.

Ecological Entomology 25: 171-179.

Herms, DA. and W.J. Mattson. 1992. The dilemma of plants — to grow or defend.
Quarterly Review of Biology 67: 283-335.

Hunter, MD. and J .C. Schultz. 1993. Induced plant defenses breached? Phytochemical
induction protects an herbivore from disease. 0ecologia 94: 195-203.

Hunter, MD. and JG Schultz. 1995. Fertilization mitigates chemical induction and
herbivore responses within damaged oak trees. Ecology 76: 1226-1232.

Kaitaniemi, P., K. Ruohomaki, T. Tammaru and E. Haukioja. 1999a. Induced resistance
of host tree foliage during and after natural insect outbreak. Journal of Animal
Ecology 68: 382-389.

Kaitaniemi, P., S. Neuvonen and T. Nyyssonen.l999b. Effects of cumulative defoliations
on growth, reproduction, and insect resistance in mountain birch. Ecology 80:
524-532.

Kamata, N. Y. Igarashi and S. Ohara. 1996. Defoliation of Fagus crenata affects the
population dynamics of the beech caterpillar Quadricalcarifera punctatella. Pp.
68-85 in W.J. Mattson, P. Niemela, M. Rousi, eds, Dynamics of Forest
Herbivory: Quest for Pattem and Principle. USDA Forest Service, North—Central

Forest Experiment Station General Technical Report NC-l 83

39

 

Karban, R. and LT. Baldwin. 1997. Induced Responses to Herbivory. University of
Chicago Press, Chicago, Illinois.

Karban, R. and J .H. Myers. 1989. Induced plant response to herbivory. Annual Review of
Ecology and Systematics 20: 331-348.

Kause A, V. Ossipov, E. Haukioja, K. Lempa, S. Hanhimaki and S. Ossipova. 1999.
Multiplicity of biochemical factors determining quality of growing birch leaves.
0ecologia 120: 102-112.

Koricheva J, S. Larsson, E. Haukioja and M. Keinanen. 1998. Regulation of woody plant
secondary metabolism by resource availability: Hypothesis testing by means of
meta-analysis. Oikos 83: 212-226.

Krause, SC. and K.F. Raffa. 1992. Comparison of insect, fungal, and mechanically
induced defoliation of larch — effects on plant productivity and subsequent host
susceptibility. 0ecologia 90: 411-416.

Lance, D.R., J.S. Elkinton and CF. Schwalbe. 1991. Responses of gypsy moth larvae
(Lepidoptera: Lymantriidae) to foliage of oaks from naturally infested sites on
Cape Cod, Massachusetts. Journal of Entomological Science 26: 214-222.

Lerdau, M., M. Litvak and R. Morison. 1994. Plant chemical defense: Monoterpenes and
the growth-differentiation balance hypothesis. Trends in Ecology and Evolution
9: 58-61.

Lindroth, R.L and K.K. Kinney. 1998. Consequences of enriched atmospheric CO: and
defoliation for foliar chemistry and gypsy moth performance. Journal of

Chemical Ecology 24:1677-1695.

40

 

Mattson, W.J. 1980. Herbivory in relation to plant nitrogen-content. Annual Review of
Ecology and Systematics 11: 119-161.

Mattson, W.J., D.A. Herms, J.A. Witter and DC. Allen. 1991. Woody plant grazing
systems: North American outbreak folivores and their host plants. Pp. 53-84 in
Y.N. Baranchikov, W.J. Mattson, F.P. Hain and TL. Payne, eds, Forest Insect
Guilds: Patterns of Interactions with Host Trees. United States Department of
Agriculture, Forest Service. General Technical Report NE-153.

Mutikainen, P. M. Walls, J. Ovaska, M. Keinanen, R. Julkenen-Tiitto and E. Vapaavuori.
2000. Herbivore resistance in Betula pendula: Effect of fertilization, defoliation,
and plant genotype. EcolOgy 81: 49-65.

Myers, J .H. and KS. Williams. 1984. Does tent caterpillar attack reduce the food quality
of red alder foliage? 0ecologia 62: 74-79.

Myers, J.H. and KS. Williams. 1987. Lack of short or long term inducible defenses in the
red alder- western tent caterpillar system. 0ecologia 48: 73-78.

Murdoch, W.W. 1994. Population regulation in theory and practice. Ecology 75: 271-
287.

Nef, L. 1988. Interactions between the leaf miner, Phyllocnitis suﬂusella, and poplars.
Pp. 239-251 in W.J. Mattson, J. Levieux and C. Bernard-Dagan, eds,
Mechanisms of Woody Plant Defenses against Insects. Springer, New York.

Nicholson, A.J. 1933. The balance of animal populations. Journal of Animal Ecology 2:

132-178.

41

 

Price, P.W., C.E. Bouton, P. Gross, B.A. McPheron, J.N. Thompson and AB. Weis.
1980. Interactions among three trophic levels. Annual Review of Ecology and
Systematics 11: 41-65.

Rhoades, DF. 1983a. Herbivore population dynamics and plant chemistry. Pp. 155-220 in
RF. Denno and MS. McClure, eds, Variable plants and Herbivores in Natural
and Managed Systems. Academic Press, San Diego.

Rhoades, D.F. 1983b. Response of alder and willow to attack by tent caterpillars and fall
webworrns: Evidence for pheromonal sensitivity of willows. Pp. 55-68 in P.
Hedin, ed, Plant Resistance to Insects. Symposium Series 208. American
Chemical Society, Washington, DC.

Robison, DI. and K.F. Raffa.l997. Effects of constitutive and inducible plant traits of
hybrid poplars on forest tent caterpillar feeding and population ecology. Forest
Science 43: 252-267.

Roland, J. and J .H. Myers 1987. Improved insect performance from host-plant defoliation
— winter moth on oak and apple. Ecological Entomology. 12: 409-414.

Roland, J. K.E. Denford and L. Jimenz. 1995. Bomeol as an attractant for Cyzenis
albicans, a tachinid parasitoid of the winter moth, Operophtera brumata L.

(Lepidoptera: Geometridae). The Canadian Entomologist 127: 413-421.

Rossiter, M.-C., J.C. Schultz, I.T. Baldwin. 1988. Relationships among defoliation, red
oak phenolics and gypsy moth growth and reproduction. Ecology 69: 267-277.

Roth, 8., R.L. Lindroth, J .C. Volin and EL. Kruger. 1998. Enriched atmospheric C02
and defoliation: effects on tree chemistry and insect performance. Global

Change Biology 4: 419-430.

42

 

Rothman, L.D. 1997. Immediate and delayed effects of a viral pathogen and density on
tent caterpillar performance. Ecology 78: 1481-1493.

Ruohomaki, K., S. Hanhimaki, E. Haukioja, L. Iso-Iivari, S. Neuvonen, P. Niemela and
S. Suomela. 1992. Variability in the efﬁcacy of delayed—inducible resistance in
mountain birch. Entomologia Experimentalis et Applicata 62: 107-115.

Sample BE, L. Butler, C. Zivkovich, R.C. Whitmore and R. Reardon. 1996. Effects of
Bacillus thuringiensis Berliner var Kurstaki and defoliation by the gypsy moth
[Lymantria dispar (L)(Lepidoptera: Lymantriidae)] on native arthropods in West
Virginia. The Canadian Entomologist 128: 573-592.

Schoener, T.W. 1983. Field experiments on interspeciﬁc competition. The American
Naturalist 122: 240—285.

Schultz, J.C. 1983. Impact of variable plant defensive chemistry on susceptibility of
insects to natural enemies. Pp. 37-54 in P. Hedin, ed, Plant Resistance to Insects.
Symposium Series 208. American Chemical Society, Washington, DC.

Schultz, J .C. 1988. Plant responses induced by herbivores. Trends in Ecology and
Evolution. 3: 45-49.

Schultz, J .C. and LT. Baldwin. 1982. Oak leaf quality declines in response to defoliation
by gypsy moth larvae. Science 217: 149-151.

Scriber, J .M. and F. Slansky. 1981. The nutritional ecology of immature insects. Annual
Review ofEntomology 26: 183-211.

Scutareneau, P., B. Drukker, J. Bruin, M.A. Posthumus and M.W. Sabelis. 1997.
Volatiles from Psylla-infested pear trees and their possible involvement in

attraction of anthocorid predators. Journal of Chemical Ecology 23: 2241-2260.

43

Sipura, M. 1999. Tritrophic interactions: Willows, herbivorous insects, and birds.
Oecologia 121: 537-535.

Strong, D.R., J.H. Lawton and TRF Southwood. 1984. Insects on Plants: Community
Patterns and Mechanisms. Harvard University Press, Cambridge.

Thaler, IS. 1997, Jasmonate-inducible plant defences cause increased parasitism of

herbivores. Nature 399: 686-688

Tuomi, J., P. Niemela, E. Haukioja, S. Siren and S. Neuvonen. 1984. Nutrient stress: an
explanation for plant anti-herbivore responses to defoliation. 0ecologia 61: 208-
210.

Tuomi, J ., P. Niemela and S. Siren.1990. The Panglossian paradigm and delayed-
inducible accumulation of foliar phenolics in mountain birch. Oikos 59:399—410.

Tuomi, J ., T. Fagerstrom and P. Niemela. 1991. Carbon allocation, phenotypic plasticity,
and induced defenses. Pp. 85-104 in D.W. Tallamy and M.J. Raupp, eds,
Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Turchin, P. 1990. Rarity of density-dependence or population regulation with lags?
Nature 344: 660-663.

Turchin, P. 1995. Population regulation: Old arguments and a new synthesis. Pp. 19-40 in
N. Cappuccino and PW. Price, eds, Population Dynamics: New Approaches and
Synthesis. Academic Press, New York.

Turlings, T.C.J., P.J. McCall, H.T. Albom and J.H. Tumlinson. 1993. An elicitor in
caterpillar oral secretions that induces corn seedlings to emit chemical signals

attractive to parasitic wasps. Journal of Chemical Ecology 19: 411-425.

44

 

Underwood, N. 1999. The inﬂuence of plant and herbivore characteristics on the
interaction between induced resistance and herbivore population dynamics. The
American Naturalist 153: 283-294.

Valentine, H.T., W.E. Wallner and PM. Wargo. Nutritional changes in host foliage
during and after defoliation, and their relation to the weight of gypsy moth
pupae. 0ecologia 57: 298-302.

Virtanen, T. and S. Neuvonen (1999). Performance of moth larvae on birch in relation to
altitude, climate, host quality, and parasitoids. 0ecologia 120: 92-101.

Wallner, W.E. 1983. Gypsy moth host interactions: a concept of room and board. Pp 5-8
in R.L. Talerco and M. Montgomery, eds, Proceedings: Forest Defoliator -Host
Interactions: a Comparison between Gypsy Moth and Spruce Budworms. USDA
Forest Service, Northeastern Station General Technical Report NE-85.

Wallner, W.E. and GS. Walton. 1979. Host defoliation: a possible determinant of gypsy
moth population quality. Annals of the Entomological Society of America 72: 62-
67.

Werner, R.A. 1979. Inﬂuence of host foliage on development, survival, fecundity and
oviposition of the spear—marked black moth, Rheumaptera hastata (Lepidoptera:
Geometridae). The Canadian Entomologist 111: 317-322.

Wold, E.N. and RI. Marquis. 1997. Induced defense in white oak: effects on herbivores
and consequences for the plant. Ecology 78: 1356-1369.

Work, T.T. and D.G. McCullough. 2000. Lepidopteran communities in two forest
ecosystems during the ﬁrst gypsy moth outbreaks in northern Michigan.

Environmental Entomology 29: 884-900.

45

CHAPTER 2:

EMULATING INSECT OUTBREAK: EXPERIMENTALLY TESTING THE EFFECTS
OF CUMULATIVE DEFOLIATIONS ON INDUCED-RESISTANCE AND

CATERPILLAR HERBIVORY

INTRODUCTION

Insect outbreaks are a conspicuous component of boreal and temperate forest ecosystems
throughout the world. During outbreaks, high levels of defoliation for one or more years
in succession can signiﬁcantly lower concentrations of leaf nitrogen, water content, and
sugars while elevating ﬁber, toughness and secondary compounds, attributes that
determine the quality of leaves as food for caterpillars (e.g., F eeny 1970, Mattson 1980,
Scriber and Slansky 1981, Schultz 1988, Neuvonen and Haukioja 1991, Kaitaniemi et al.
1998, Kause et al. 1999). Changes in host plant quality elicited by feeding damage that
have negative consequences on herbivores are deﬁned as induced-resistance (Karban and
Baldwin 1997). Two types of induced resistance have been identiﬁed in trees (Haukioja
and Niemela 1977, Haukioja 1980, 1982); rapid—induced resistance (RIR), which occurs
within hours or days of feeding damage, and delayed-induced resistance (DIR), which
has its greatest effects in the year or years subsequent to the defoliation event and usually
requires severe herbivory to be elicited. In simulation models (e.g., Edelstien-Keshet and
Rausher 1989, Underwood 1999), RIR when sufﬁciently strong, can stabilize herbivorous
insect populations because it acts instantaneously on population growth. Conversely, DIR
functions in a delayed-density dependent manner, which tends to destabilize insect

populations (May 1973, Berryman et a1. 1987) and can create cyclic population dynamics

46

 

under some conditions (Underwood 1999). Several factors including maternal effects,
natural enemies, and induced-resistance are capable of producing the delayed-negative
feedbacks (Berryman et al. 1987, 1996, Ginzburg and Taneyhill 1994), identiﬁed in time

series data from many forest insect populations (Turchin 1990, 1995).

Much of our understanding of DIR has come from research on birches (Betula spp.)
and a few other species of deciduous trees. Defoliation of birch elevates levels of foliar
phenolic compounds including hydrolyzable and condensed tannins, and reduces primary
nutrients such as nitrogen and water (Haukioja et al. 1985, Tuomi et al. 1984, 1990,
Bryant et al. 1988, 1993, Kaitaniemi et al. 1998). Caterpillars feeding on foliage from
previously defoliated trees suffer reduced growth and fecundity, and protracted
development times relative to larvae feeding on undefoliated. trees (e.g., Haukioja and
Neuvonen 1985, Ruohomaki et al. 1992, Bryant et al. 1993, Kaitaniemi et al 1999a).
Because changes in foliar biochemistry can persist for several years after the cessation of
defoliation (Bryant et al. 1991), insect ﬁtness may continue to suffer long after an
outbreak has ended (Haukioja and Neuvonen 1987). Long term reductions in foliar
quality following defoliation have also been shown to have signiﬁcant negative effects on
larch bud moth, Zieraphera diniana, feeding on alpine larch in Switzerland (Benz 1974,

Baltensweiler and F ischlin 1988).

Traditionally, experimental tests of DIR have consisted of defoliation in a single year
followed by phytochemical analyses and/or bioassays in the following year or years. This
approach may underestimate the strength of DIR because outbreaks of many defoliators
are at least two years in duration (e.g., Mattson et al. 1991), leading Haukioja et al.

(1988) to hypothesize that consecutive defoliations may have cumulative effects on

47

 

herbivore performance. Despite this speculation, the effects of successive years of
defoliation on trees and their herbivores have received only limited attention in the
literature (Kaitaniemi et al. 1999a). In concordance with the hypothesis of Haukioj a et al.
(1988), several studies have shown increasingly negative effects on insect perfonnance
with each successive year of defoliation (Werner 1979, Wallner and Walton 1979,
Valentine et al. 1983, Clausen et al. 1991). In contrast, successive years of defoliation
were shown to have only minor cumulative effects on phytochemistry and insect
performance in experiments with mountain birch (Betula pubescens) (Kaitaniemi et al.

1999a, 1999b) and trembling aspen (Populus tremuloides) (Parry et al. 2000).

The role of DIR in insect population dynamics may not be as general as ﬁrst thought
(e.g., Haukioja 1980, Rhoades 1983). In some cases defoliation may increase herbivore
performance while in others, there may be no effect. Roland and Myers (1987) compared
the performance of winter moths on apple (Malus sp.) defoliated from 0—100% and Garry
oak (Quercus garryana) defoliated from 0-50% in the previous year. They found that
moderate levels of defoliation (<50%) in the previous season were associated with
increased pupal masses in the subsequent season on both apple and oak. Higher levels of
defoliation (>50%) in the previous season was associated with reduced pupal masses.
Rothman (1997) found no effect of high defoliation of red alder (Alnus rubra) on western
tent caterpillar (Malacosoma californicum) pupal masses in the year following defoliation
that exceeded 70%. Similarly, Harrison (1995) found no effects of previous severe
defoliation of bush lupine (Lupinus arboreus) on fecundity of western tussock moth

(Orgyia vetusta) despite evidence for RIR in the previous year.

48

Defoliation induced changes in plant quality extend beyond the interaction between
trees and the herbivore eliciting the response to include other member of the
phytophagous community. The magnitude and direction of host-plant mediated
competition between an outbreak species and other phytophages depends on several
different factors including the timing of defoliation in relation to the phenology of
competing folivores, the habitat or life history strategy of competitors, and the degree of
feeding specialization by the herbivore (Denno et al. 1995, Dankert et al. 1997). Outbreak
species may be more adapted to defoliation—induced changes in nitrogen, water, or leaf
toughness than species whose populations remain perpetually below the carrying capacity
of the environment. Conversely, generalists may be more susceptible to the effects of
toxic secondary phytochemicals than adapted specialists. In factorial combinations,
Agrawal (2000) showed that induction of wild radish by one species of caterpillar had
variable effects on three other species. The reciprocal effects were interesting in that the
plant perceived damage from each of the four species of caterpillars differently. Such
variation in plant perception of herbivores and variability in herbivore response to
induced-resistance may account for the diversity and sometimes conﬂicting experimental

results within and among tree species (see Karban and Baldwin 1997 for review).

Divergent effects of induced-resistance on herbivores suggest that responses to
defoliation may be system speciﬁc or are mediated by environmental factors,
underscoring the incomplete understanding of mechanisms driving the long-term
consequences of sustained insect outbreak on forest trees. Adding to the complexity,
several mechanisms operating at different hierarchical and temporal scales may be

required to completely explain induced-resistance in trees (e.g., Tuomi et a1. 1990, Herms

49

 

and Mattson 1992, Karban and Baldwin 1997, Koricheva et al. 1998). In deciduous trees,
much focus has been on the carbon-nutrient balance hypothesis (CNB) as a mechanism
explaining the induction of carbon-based secondary compounds. The CNB, a subset of
predictions under the more globally encompassing growth-differentiation balance
hypothesis (GDB), postulates that when nutrients limit growth more than photosynthesis,
excess carbon can be allocated to the production of carbon-based secondary metabolites
(Bryant et al 1983, Tuomi et al. 1984, 1990, Herms and Mattson 1992). Tuomi et al.
(1984, 1990) extended the CNB to encompass changes in nutrient/carbon balance
following defoliation. Defoliation, especially early in the growing season, decreases
nitrogen stores in deciduous trees (Harper 1977, Chapin and Kedrowski 1983) and may
damage ﬁne roots, reducing acquisition from soil, and exacerbating nitrogen deﬁcits
(Tuomi et al. 1990). Thus, until the tree can recover or compensate for the nitrogen lost
when defoliated, photosynthesis will generate carbon in excess of growth requirements;
this carbon is available for allocation to secondary metabolism. The majority of studies
testing predictions of the CNB in trees have focused on the year following defoliation.
Recently, Hunter and Schultz (1995) demonstrated that the CNB could also account for

within year changes of some phenolic compounds.

The domain of the CNB hypothesis does not address rapidly induced systemic
changes in trees. These responses occur over very short-terrn periods of hours or days and
are often invoked by relatively low levels of herbivory, spreading rapidly to undamaged
leaves. Several wound-responsive, signal-transduction pathways have been identiﬁed in
herbaceous plants (e.g., Bergey et al. 1996, Karban and Baldwin 1997, Constabel 1999).

In these pathways, speciﬁc elicitors stimulate secondary metabolism and/or translocation

50

of secondary compounds from storage sites following herbivory. To my knowledge,
similar pathways have not yet been identiﬁed in trees, although there are no compelling
reasons to suggest that they do not occur. Several studies have shown that insect
herbivory elicits different responses in woody plants than equivalent amounts of tissue
removed mechanically (e.g., Haukioja and Neuvonen 1985, Hartley and Lawton 1987,
Krause and Raffa 1992, Havill and Raffa 1999). This suggests that leaves of trees have
insect-speciﬁc wound responses much like those identiﬁed in herbaceous plants.
Furthermore, application of the generalized elicitor compound, jasmonic acid, has been
shown to induce RIR in hybrid p0plars in the absence of tissue damage (Havill and, Raffa

1999), indicating that a signal-transduction pathway may be operating.

Reported here are the results of a four—year study in which I experimentally
established populations of gypsy moth [Lymantria dispar L., (Lepidoptera
Lymantriidae)] in replicated blocks of poplar to emulate the increase, outbreak, and
decline in densities characteristic of natural outbreaks. Several aspects of this study are
unique. The poplar plantation consisted of ontogenetically mature trees (ca. 15 m high, 9
years old) rather than saplings, and was planted on a relatively large spatial scale (four
ha). Furthermore, insects were used to defoliate the trees, control trees had realistic
background levels of herbivory, and the temporal scale of the study was designed to
mimic the duration of a natural outbreak (one year of light defoliation followed by a
severe year, a moderate year, and a year with very low defoliation). I extended my
observations beyond the classically studied relationship between defoliator and host tree
to include ﬁve other generalist and specialist lepidopteran species belonging to three

different seasonal feeding guilds to test for host-plant mediated indirect competition. My

51

main objective was to assess the consequences of long-term defoliation on trees and the
reciprocal effects on their associated herbivores. Speciﬁcally I predicted that: (l)
successive years of defoliation would cause greater changes to tree physiology than a
single defoliation event, and that nitrogen fertilization would mitigate the response as
predicted by the Carbon-Nutrient Balance hypothesis. (2) The quality of leaves would be
reduced for herbivores following defoliation (RIR and/or DIR) and that the magnitude of
the effects would increase with each successive year of defoliation. (3) That
phytochemical changes induced by an outbreak species would have signiﬁcant
detrimental effects on the growth of other members of the leaf-feeding lepidopteran
community, and that the magnitude of the effects would be greater for oligophagous,
non-outbreaking species than for polyphagous species and species characterized by

outbreak pOpulation dynamics.

MATERIALS AND METHODS
Experimental design

My study was conducted in four one—ha stands of a single clone of Populus x
euramericana c.v. "Eugeneii" growing on the Long Term Ecological Research site at
Kellogg Biological Station in southwestern Michigan, USA (see Marino and Gross 1988
for detailed site description). The trees were planted in 1987 as root cuttings in a 1m x 2
m array in ﬁelds that had been under agricultural cultivation for nearly 100 years. The
soil is a Kalamazoo sandy loam (Typic Hapludalf). Poplars were nine-years-old, had

diameters—at—breast-height (DBH) of 7.7 i 0.53 cm (mean i SE), and were 12.1 i 0.49 m

52

(mean i SE) in height when the study was initiated in 1996. These very fast-growing
trees were ontogenetically mature (annual ﬂower production) at the beginning of the

study.

Within each one ha stand of poplar, I placed a 0.25 ha block in the northeast comer,
which was divided into four plots, each separated by a buffer strip of two rows of trees,
which were included in the treatments but not used in the experiment. Each of the four
plots within a block was randomly assigned one of four treatments: (1) undefoliated and
unfertilized, (2) undefoliated and fertilized (100 kg ha”1 N as NH4NO3), (3) defoliated and
unfertilized, or (4) defoliated and fertilized. Fertilizer was applied in spring shortly after
bud break in each year of the study. Undergrowth vegetation was suppressed from the
time of planting and throughout my experiment by twice—annual applications of
glyphosate (2% v/v, Roundup, Monsanto Corporation, St. Louis, MO). Before
randomization of plots, a minor restriction was placed on the location of defoliation
treatments within a block. Because I was relying solely on physical barriers (insecticides
were not used) to keep caterpillars from colonizing the control plots, the possibility of the
random assignment of defoliation plots diagonally from each other was eliminated.
Locating the defoliation plots diagonally would have doubled the edge exposure of the
control plots increasing substantially the difﬁculty of maintaining low herbivory levels in
controls. Initially all plots had the same probability of being assigned to a defoliation
treatment. Once the ﬁrst plot assignment had been made, the second defoliation treatment
could be assigned to only two of the three remaining plots because of this restriction. I
felt that the practical beneﬁts of employing a restricted design far outweighed the loss of

fully random deployment of treatments within each block.

53

Establishing the gypsy moth outbreak — A number of studies have shown that artiﬁcial
removal of leaves by tearing or using scissors elicits different physiological responses
than does leaf removal by caterpillars (e.g., Haukioja and Neuvonen 1985, Hartley and
Lawtonil987, Krause and Raffa 1992, Havill and Raffa, 1999). To more closely emulate
processes experienced by trees in natural outbreaks, defoliation was applied by
inoculating the plots with large numbers of caterpillars. I chose gypsy moth because
many trees in the genus Populus are a preferred hosts (Liebhold et al. 1995, Havill and
Raffa 1999), egg masses and larvae were already present in some of the plots, and I could
obtain large numbers of eggs from several high-density populations elsewhere in
southern Michigan. Prior to the initiation of the study, gypsy moth populations in my
plots and. in the surrounding oak/hickory/maple forests were at low but detectable levels
and had been present in the area since at least 1991 (S. Gage, personal communication).
In the spring of 1996, I collected large numbers of gypsy moth eggs from high-density
populations in Ottawa County, MI, and in 1997, from Wayne and Kalamazoo Counties,

MI for innoculative release in the experimental plots.

In the laboratory, egg masses were separated manually to release the eggs from the
hair matrix in which they are embedded. I submerged eggs for 1 hour in a 10% fonnalin
solution (Bell et al. 1981), which was agitated periodically. Fonnalin removes nuclear
polyhedrosis virus, critical because this pathogen is the major source of mortality in high-
density gypsy moth populations and is spread from contaminated egg surfaces to
emerging neonates (Woods and Elkinton 1987). After the fomialin treatment, eggs were
placed on a mesh screen and rinsed under running water for 1 h, then dried overnight at

room temperature under a fume hood. When dry, eggs were weighed and 100g allotments

54

were placed in cheesecloth packages and secured with twist ties. Eggs were kept at 4°C in
an environmental chamber until bud break approached. To synchronize hatch of
collected eggs with the natural phenology of gypsy moth, in situ egg masses in the poplar
stands were monitored for hatch as bud break approached. Once natural hatching was
observed in the ﬁeld, egg packages were placed in an environmental chamber at 27°C
until the ﬁrst eggs began hatching within a few days. The cheesecloth packages were then
stapled to the bark of 10—12 trees in the center of each defoliation treatment plot.

Emerging caterpillars climbed from the cloth bags and ascended the trees.

In 1996, I increased gypsy moth densities from background levels in the defoliation
treatment plots by adding ca. 248,000 eggs to each of the 8 defoliation plots (150 grams
of eggs/plot or 8 x 103 egg masses/ha assuming 500 eggs/mass). A further ca. 3.1 million
eggs (2.25 kg /plot or the equivalent of 1 x 105 egg masses/ha assuming 500 eggs/mass)
were added to each defoliation plot in 1997 to bolster egg masses from the previous
year’s population. Egg density in the release packages was estimated from the total mass
of washed eggs divided by the average weight of individual gypsy moth eggs, 0.65 mg
(from Diss et al. 1996, and DP, unpublished). I reduced the estimated amount by 10%
to account for eggs parasitized by Ooencyrtis kuwaniae (Hymenoptera: Encyrtidae), and
those that were infertile, calculated from randomly drawn subsamples of eggs. Densities
of egg masses in the poplars surrounding the plots increased very rapidly following my
introductions in 1996, likely driven by large numbers of caterpillars wandering out of the
defoliation treatment plots. By 1997, the entire one ha stands of poplar surrounding my
research blocks contained densities of gypsy moth sufﬁcient to severely defoliate most of

the trees that were not protected. Because these larvae were used only to defoliate the

55

 

trees and were not part of any experiment, I made no attempt to control them, nor did I
distinguish between those originating within or outside the defoliation plots. It was not
necessary to add any gypsy moth egg masses in 1998 because egg mass densities
remained very high within the plots and surrounding poplar stands. Progeny of these egg
masses were sufﬁcient to cause moderate to severe defoliation in 1998. To my
knowledge, this is the ﬁrst instance of a self—sustaining, experimental outbreak population

of gypsy moth.

Following the 1998 season, gypsy moth populations crashed in my plots apparently
due to a variety of contemporaneous mortality sources. These included larval parasitoids
(primarily the tachinid Compsilura concinnata), pupal parasitism [likely Brachymeria
intermedia (Hymenoptera: Chalcididae)], extensive predation by ﬂocks of starlings
(Sturnus vulgaris), black-billed cuckoos (Coccyzus erythropthalmus), as well as moderate
levels of NPV. I did not quantify sources of mortality but combined, they were very
effective and reduced populations to low levels. No mortality from the fungus
Entomophaga maimaiga, responsible for devastating epizootics in gypsy moth
populations elsewhere in North America, was detected in my plots. Few egg masses
were laid in 1998 and as a result, larval populations were very sparse in 1999. Because I
was interested in the recovery of trees after the collapse of an outbreak, gypsy moth eggs

were not added to the existing low densities in 1999.

Gypsy moth neonates are highly mobile and may wander extensively before settling
and beginning to feed (Mason and McManus, 1981). Following release, neonates readily
dispersed throughout the canopy of my plots by ballooning on silk threads or by crawling

from tree to tree. To minimize colonization of control trees, tree wrap coated with

Tangletrap was placed on the trunk of each control tree ca. 0.5 m above the ground. I
trimmed back all branches and other vegetation that could provide a conduit for
caterpillar incursion from defoliation plots to control plots. To remove caterpillars that
circumvented my barriers or ballooned onto the control trees, I used a combination of
manual and mechanical methods. I accessed the canopy using ladders and a truck
mounted ‘cherry picker’ and used a pressure water sprayer to dislodge small caterpillars
from the treetops. Third and later instar gypsy moth caterpillars descend the trees in the
morning to seek daytime hiding places in low density populations so I placed burlap
bands on each control plot tree above the sticky barriers. Burlap bands were checked
daily and any caterpillars found were returned to the defoliation plots. With the
exception of one plot in 1997, I was very successful in maintaining very low densities of
gypsy moth in the control plots throughout the course of the experiment. I did not
actively attempt to remove any other herbivorous insects from the control trees and other
than gypsy moth, no other species of insects were particularly abundant in the plots
although there was considerable diversity of other herbivorous insects present. In natural
forests, estimates indicate that between 5—20% of the foliage is removed by non-outbreak
populations of herbivorous insects (e. g., Mattson and Addy 1975, Reynolds and Crossley
1997). Thus, to assess the effects of defoliation induced changes in host plant quality on
herbivores, I felt that trees subjected to background herbivory were a more appropriate

control than trees maintained free of insect damage.

To estimate defoliation in 1997 and 1998, I relied upon visual estimates recorded
independently by three observers for each tree in the plots. These estimates were

averaged to obtain mean plot defoliation levels. Defoliation was partitioned into four

57

classes: 1. 0-25%, 2. 25-50%, 3. 50-75%, and 4. 75-100%. In 1996 I collected insect frass
in 50 cm diameter traps. Two traps were placed in each plot. Each trap consisted of a
bucket with a screen funnel on the open end that directed falling frass into collection
cups. Frass was dried and weighed. I used the same frass traps in 1997 and discerned that
frass capture was highly correlated with visual estimations of defoliation (r2 = 0.75). This
regression was used to estimate defoliation in 1996 because the relatively low level of
defoliation was difﬁcult to estimate visually. These data showed that in the treatment
plots, light to moderate levels of defoliation were achieved in 1996, uniformly severe
levels of defoliation in 1997, and moderate to severe levels in 1998 (Figure 1). I did not
quantify defoliation in 1999 because caterpillar density was so low that all trees were
well below the threshold where any defoliation could be detected visually (overall, less
than 1 egg mass per plot was found in the spring of 1999 as opposed to several hundred

per plot in the spring of 1998).
Tree responses to defoliation

Tree growth — To ascertain the effects of treatments on whole tree processes, diameter
was recorded annually from each living tree in the plots. Two measurements of diameter
were recorded 90° apart at 1.4 m above the ground using digital calipers. I permanently
marked the measurement points on each tree so that diameter could be recorded from the

same point each year.

Phytochemistry — Several primary and secondary biochemical constituents of leaves were
recorded each year. I based the timing of leaf samples on the phenology of gypsy moth so
as to standardize collections among different years. From 1996-1999, samples of leaves

were collected in June, which corresponded to the presence of ﬁnal instar caterpillars in

58

 

  
 
  

 

 

 

 

 

4
—U— Def-, N—
—I— Def-, N+
" —O— Def+,N-
—O— Def+,N+
8 3 -
L“
0
c
.9 -
as
.9!
~§
o 2 -
1 . .
1996 1997 1998 1999
Year

Figure 1. Defoliation levels (mean i SE) in treatment and control plots
1996-1999. Means (i SE) were estimated from the four plots in each
treatment. Defoliation was estimated visually using four classes: Class 1 —
0-25%, Class 2 — 25-50%, Class 3 — 50-75%, Class 4- 75-100%. Treatment
codes are undefoliated = Def -, defoliated = Def+, unfertilized = N-, and
fertilized = N+.

59

 

the ﬁeld. In 1997 and 1998, I also sampled foliage from the plots one month after gypsy
moth feeding was complete, and defoliated trees had completely refoliated, in early to
mid July. Additionally, in 1998, I collected an early season sample two weeks after bud

break and the hatch of gypsy moth eggs.

To sample leaves, I randomly selected healthy trees (no dead tops or other signiﬁcant
dieback) in each plot. June leaf samples were taken from 10-14 trees/plot in 1996-1998
and from 4 trees per plot in 1999. July samples in 1997 and May and July samples in
1998 were collected from four trees per plot. Leaf sampling was very labor intensive,
therefore the number of trees sampled in any given collection was dictated by the
availability of help. To sample, I used extendable pole-pruners to cut two branches with
approximately equal light exposure from the upper canopy of each tree. Poplar leaves
exhibit considerable variability in phytochemistry that is dependent on the position of the
leaf on a shoot (Meyer and Montgomery 1987, Robison and Raffa 1997). To standardize
sampling, the ﬁrst six expanded leaves (corresponding to a plastochron index of 4—1 1, see
Meyer and Montgomery 1987) on each branch, were detached, placed in small paper
envelopes and immediately ﬂash frozen in liquid nitrogen. Frozen leaves were packed in
dry ice for transport to the laboratory and were held at —40°C until lyophilized. Following
lyophylization, leaves were ground in a miniature Wiley mill through an #40 mesh screen

and stored at —20°C over silica until used in analyses.

Levels of total phenolics and condensed tannins were quantiﬁed in 1996. In 1997 and
1998, I quantiﬁed foliar nitrogen, water, total phenolics, condensed tannins, and six other

secondary phenolic compounds (phenolic glycosides, cinnarnic acid, catechins, luteolin,

60

quercetin-glycoside, and rnyricetin-glycoside). In 1999, levels of nitrogen, water, total

phenolics, condensed tannins were determined.

To quantify levels of foliar nitrogen, 9-11 mg of leaf powder was weighed into 8 x 5
mm tin capsules, the edges were folded over and the capsule compressed to remove air.
These tin pellets, along with acetanilide standards, were then placed in an autosampler of
a Carlo-Erba model NA-1500 NC analyzer. Water content of leaves was determined
gravimetrically. At each sampling date, leaves were collected from the trees, weighed on
an electronic balance and placed in paper envelopes in a drying oven at 45°C for several

days. Samples were removed and reweighed to determine water content.

To determine concentrations of foliar total phenolics and condensed tannins, leaf
powder was weighed on an electronic balance (70-80 mg) and placed in 15-ml
polypropylene screw—top centrifuge tubes. Aqueous methanol (50%) was added (100:1
ml/ g). Tubes were tightly capped, and extractions were conducted in the dark at 250 C on
a shaker table. Samples were then centrifuged at 2500 rpm for 10 min, and the
supernatant ﬁltered through disposable ﬁlter columns (200—300 rm pore size; Fisher

Scientiﬁc) into 3.5 ml polystyrene sample cups.

The Folin-Denis method for measuring total phenolics (Swain and Hillis 1959) was
modiﬁed for use in an air-segmented, continuous ﬂow analyzer (RFA-300 Rapid Flow
Analyzer and 301 autosampler; Astoria-Paciﬁc International, Clackamas, Oregon) (Nitao
et al., in review). The analyzer was conﬁgured to run at room temperature, and to dilute
the phenolic samples by a factor of 78.3 before mixing with the Folin-Denis reagent in a
ratio of 1213.8 (F-D reagent : diluted sample) for 0.3 min at room temperature. Sodium

carbonate solution was added in a ratio of 1:148 (sodium carbonate: sample stream) and

61

 

 

mixed for 0.6 min before being measured at 750 nm using a standard curve based on
tannic acid (Fisher Scientiﬁc). The autosampler, standard curve construction (polynomial

ﬁt), quantiﬁcation and data capture were all controlled through RF A-PC software.

The same sample preparation and extraction procedure was used for condensed
tannins. The sulfuric acid method for measuring condensed tannins (Bate-Smith and
Rasper, 1969; Bae et al.1993) was adapted for use with the autosampler above (Nitao et
al., in review). The analyzer was conﬁgured to mix 43% sulfuric acid diluted with
methanol (v/v) with tannin samples in a ratio of 1:6.3 (samplezsulfuric acid) for 0.4 min
at room temperature, heated to 500 C for 5.0 min, and allowed to mix for an additional
1.2 min at room temperature. Condensed tamrins were measured at 580 nm using a
standard curve based on quaking aspen tarmin standard (provided by Karl Kleiner,

University of Wisconsin, Madison, Wisconsin).

To quantify other phenolic components of poplar foliage, lyophilized leaf powder (8-
10 mg) was placed in an Eppendorf vial and 0.7 ml of methanol added. Vials were left
standing for 15 min in a ice-bath. Samples were removed from ice and homogenized with
an Ultra-Turrax homogenizer for 2 min and then centrifuged at 10000 rpm for 3 min.
Solvent was taken into 6 ml glass-vial (in a cold-bath) and evaporated under nitrogen.
The extraction was repeated three times with washing. Prior to HPLC, the sample was
dissolved into HZOzmethanol (1:1, 600 pl :600 pl). The HPLC-runs followed the
procedure of Julkunen-Tiitto et al. (1996). Despite my best efforts, HPLC revealed
considerable quantities of salicin and salicin derivatives, generally regarded as hydrolytic
breakdown products of unstable phenolic glycosides such as salicortin and tremulacin

(Lindroth and Pajutee 1987, Lindroth and K055, 1996). Because of this degradation, I

62

 

pooled salicin, tremulacin, and salicortin (appreciable levels of tremuloiden were not
detected) to get an estimate of total phenolic glycosides as was done in Lindroth and

Pujutee (1987).

In addition to the direct measures of foliar chemistry, I estimated the mean area and
toughness of individual leaves. For these measures, leaves were sampled from two
branches on each tree and standardized for plastochron position. Leaves were
photocopied shortly after being clipped to provide a permanent record. The photocopies
were placed on a ﬂatbed scanner and the area calculated by a digital image analysis
program (C1400 Computer Image Analysis System, Version 2.0, Jandel Scientiﬁc, San
Rafael, CA). Leaf toughness was determined using a digital force meter attached to a
penetrometer (Chatillon DFM2, Chatillon Inc., Greensboro, NC). Two measurements
were taken on each leaf, one on the right and one on the left of the midrib. To standardize
my measurements, force data were collected from leaf lamina between the ﬁrst and

second lateral veins.
Effects of defoliation on caterpillars

Bioassay insects — To assay for changes in foliage quality following treatments, I utilized
six species of Lepidoptera. The caterpillars used, gypsy moth, forest tent caterpillar,
Malacosoma disstria Hiibner (Lasiocampidae), white-marked tussock moth, Orgyia
leucostigma (J .E. Smith) (Lymantriidae), poplar tent maker Clostera inclusa (Hilbner)
(Notodontidae), big-poplar sphinx, Pachysphinx modesta (Harris) (Sphingidae), and fall
webworm, Hyphantria cunea (Drury) (Arctiidae), were selected because they occurred
naturally in the poplar plantation. These species represent a broad cross section of life

history strategies for exposed leaf-feeding caterpillars. Three species are highly

63

polyphagous generalists (gypsy moth, fall webwomr, white-marked tussock moth), one
species is moderately polyphagous (forest tent caterpillar), two are poplar—feeding
oligophages (poplar tent maker, big poplar Sphinx), and one species (white-marked
tussock moth) is bivoltine in Michigan although I assayed only ﬁrst generation larvae.
The four polyphagous species are characterized by outbreak pOpulation dynamics
whereas the two oligophagous species exhibit stable dynamics in natural systems. Three
species (forest tent caterpillar, poplar tent maker, and fall webworm) are highly
gregarious and live in colonies whereas the other three species are solitary. Using feeding
phenology (e.g., Hunter 1995), these lepidopterans can be broadly partitioned into three
seasonal guilds (Figure 2). Gypsy moth, forest tent caterpillar, and ﬁrst generation white-
marked tussock moth are spring feeders initiating feeding at or shortly after bud break,
poplar tent maker and big poplar Sphinx are mid-season feeders, and in Michigan, the fall

webworm is a late season feeder, hatching in mid to late July.

Long term bioassays — Life-long bioassays (egg to adult) were conducted for gypsy moth
each year from 1996 to 1999. The height of the trees and the propensity of poplars to
maintain branches only in the upper third of the tree precluded on-tree bioassays so all
larvae were reared in the laboratory. Caterpillars were obtained from eggs collected in a
hybrid poplar stand near the plots but not used as part of the experiment. Gypsy moth
caterpillars were reared in pairs in 150 x 25 mm petri dishes in 1996 and 1997.
Thereafter, bioassays for all species were done with groups of 10 in 5.6 L clear plastic
boxes with ventilation holes in the lids. Four trees in each plot were used to supply
foliage in 1996 and 6-8 trees were used from 1997—1999. Trees were randomly selected

on each foliage collection date with the stipulation that the same tree could not be used on

64

 

it

Figure 2. Approximate larval feeding periods of the six species of Lepidoptera used in
the study in relation to important phenological events during the study. * White-marked
tussock moth has a spring and a summer generation in southern Michigan, I used only
spring generation larvae in the bioassays.

65

>mo_o:o;n_ mEuwou _m>._m._

 

 

 

 

 

 

 

33 25. >23.
. . .
. . _ . .
Show 93. 22. 95.. >22). Ego.
_ _ _ _
s s s
{3:828 E2 598 T
i IE9: >83
e£oE x0083
5me E? :2me
xcEQw Econ 05
8538; __ms.
new”. 53.3.88 v.81 aerosm

 

 

saroeds

66

consecutive dates to avoid changing tree physiology through excessive sampling. Tw
with attached leaves were clipped from the selected trees using pole-pruners and returr
to the laboratory in plastic bags on ice. The stem of each twig was out under water 2
inserted into water-ﬁlled aqua-picks to maintain leaf turgor and placed in the caterpil
rearing containers. This method allowed caterpillars to choose from ~15-30 leaves
which to feed. Foliage was never allowed to become depleted and fresh leaves w
provided every second or third day as needed through to pupation. In 1996 and 19'
caterpillars were reared in environmental chambers at 25:18°C and 16:8 (lightzda
photoperiod. Thereafter, a naturally lit and ventilated room was used and temperatu

were allowed to ﬂuctuate daily with ambient conditions.

In 1998, in addition to gypsy moth, 20 forest tent caterpillar, 20 poplar tent maker
big poplar sphinx, and 20 fall webwomr were reared on foliage from each plot. For
tent caterpillars were obtained from egg masses collected from the hybrid poplar plc
Big poplar sphinx eggs were obtained from adults collected in the plots and mated
captivity. Poplar tent maker and fall webworm were collected from poplars as larvae
the year prior to the bioassay, reared to maturity in the laboratory, and mated in captiv
to produce the eggs used in the 1998 bioassay. Groups of ten insects were reared in
plastic boxes described above with the exception of big poplar sphinx. Caterpillars
this species were very aggressive to conspeciﬁcs, necessitating individual rearing in p
diShes. Overall, caterpillar phenology in the bioassays was similar to that of Ian
Observed in the field in 1996-1998 and slightly faster in 1999. I again reared 20 for
tent caterpillar and 20 poplar tent maker larvae in 1999 and also used 20 ﬁrst generat

white-marked tussock moth.

67

Following pupation, the pupae were weighed on an electronic balance, and pla-
individually in 25-ml plastic cups. The species used in my study have non-feeding adi
and female mass and fecundity are generally highly correlated for species exhibiting 1
life-history strategy (Tammaru and Haukioja 1996). Pupae were checked daily

emergence, which was recorded along with the sex of each individual adult.

To estimate fecundity, gypsy moth, forest tent caterpillar, and poplar tent maker w
reared in cages on poplar foliage collected from outside of the experimental plots. Pu;
were weighed and at emergence, females were dissected and the number of eggs count
allowing me to develop a regression equation that describes the relationship betwl
pupal mass and fecundity for each species. I derived the following equations for
gypsy moth, y = 475.56}: — 106.23, r2 = 0.90, n = 18, (2) forest tent caterpillar, ;
376.74x + 41.92, r2 = 0.74, n = 44 and (3) poplar tent maker, y = 1147.0x + 56.77, r
0.65, n = 29, where y = fecundity and x = pupal mass. Sufﬁcient numbers of wh
marked tussock moth, big poplar sphinx, and fall webworm were available only for
bioassays, so it was not possible to generate fecundity regression equations independer

of the experimental larvae.

Short-term bioassays - I conducted short term, within instar bioassays using gypsy m
in 1997 and 1998. Late fourth instar caterpillars that were preparing to molt w
collected from poplars surrounding my plots and allowed to molt in the laboratc
Within 24 hours of molting, I placed ﬁfth instar caterpillars individually in petri dis
and assigned them randomly to each plot. Leaves collected from each plot were t1
allocated to the appropriate petri dish. Seven caterpillars per plot were reared in 1997 a

10 caterpillars per plot were reared in 1998. In 1997 I used a leaf area meter to determ

68

the amount of foliage consumed by subtracting the ﬁnal area from the initial. Lea
were then placed in a drying oven and the wet/dry conversion ratio used to estin‘.
initial weight. After 48 hours in 1997 and 72 hours in 1998, the caterpillars w
removed, frozen, and placed in a drying oven at 40°C for 5 days along with the frass :
remaining leaf material. To obtain initial dry weights, a regression equation \
calculated from the mass of newly molted caterpillars before and after drying. Beca
there was no association with any of the treatments, caterpillars that died or did not f
were not included in any analyses. Relative growth rate was calculated as RGF
ln(weightf) - ln(weight,-)/T where In is the natural logarithm, weightf = ﬁnal weig
weight,- = initial weight, and T is the elapsed time in days (Gordon 1968). Measures
efﬁciency of conversion of ingested food (ECI), approximate digestibility (AD), :
efﬁciency of conversion of digested food (ECD) were calculated using the stand
formulas of Waldbauer (1968). In 1998, rather than calculate total consumption from

area of leaf consumed, gravimetric techniques were used.
Statistical analyses

I analyzed the effects of my treatments as a randomized block design. I treated
‘block’, ‘defoliation’, and ‘fertilization’ terms as ﬁxed effects and F -tests were made 0
the mean square error. The poplar plots at Kellogg Biological Station are unique :
were not chosen at random from some greater pool of poplar stands, leading me to ass
them as a ﬁxed term, recognizing the limitations that this assignation has on
generality of my conclusions. Both defoliation and fertilization represent spec
treatment levels and were also assigned as ﬁxed terms. All data collection was done us

subsamples from within each plot. These were averaged and statistics performed us

69

the plot as the sampling unit. To test the effects of leaf quality variables (nitrogen, wz
allelochemicals, physical properties) on insect performance (growth, pupal mass,
development time), I used Pearson product-moment correlation. As above, indivit
plots were considered the sampling unit and all analyses were performed using the

means.

RESULTS
Effects of defoliation on trees

Growth — Defoliation signiﬁcantly reduced the diameter growth of trees in 1997
1998 (Figure 3, Table 3). In 1996 there was no effect of defoliation on growth am
1999, previous defoliation did not effect the growth of trees. Fertilizer positively affe<
growth in every year although the effects were not signiﬁcant in 1996. Previ
fertilization continued to enhance growth in 1999 although the effect was of c

marginal signiﬁcance.

Phytochemistry - I found that the defoliation levels achieved in 1996, the ﬁrst year of
study, were sufﬁcient to signiﬁcantly increase concentrations of total phenolics

condensed tannins (Figure 4, Table 4). There was no main effect of fertilizer treatn
on either condensed tannins or total phenolics but defoliation and fertilizer exhib
marginally signiﬁcant, interactive effects on condensed tannin concentration. Tani
increased in response to fertilization in undefoliated plots but decreased in defolia

fertilized plots.

70

 

 

 

 

 

—O—Def—,N-

A ‘ +Def-,N+

E 6 _ —E1—Def+,N-

v —I—Def+,N+

g _

9 4

or _

h

o -

4.0

‘e’

.9 2 —

D

O I l I l I l l
1996 1997 1998 1999
Year

Figure 3. Annual diameter growth (mean i SE) for poplars subjected to
defoliation and fertilization treatments over the course of the study. Means
(i SE) were calculated from the four plots within each treatment category.
Treatment codes are undefoliated = Def -, defoliated = Def+, unfertilized =
N-, and fertilized = N+.

71

 

 

 

 

mo; a 34.0 a omd a med 0 Hobm

 

2.0 02.0 a 0.0 _0.0 a 0.2 00.0 a N0 :0 _ Em x Essa
.2 3. 0:. _ ._. 0.0 00.0 a ._. 0.2 20 a 0.0 N0; 2 aeaggm
M3 00.0 _ r... 0.00 00.: _ r... 3.0 00.0 a 0.2 00.0 a 803203
2 a: m 0.0 3.0 m 0.0 00.0 m a. 0.0 03 0 v6ch
[.4 02 :0 m 02 .6 m 02 a0 a 02 a0 add

0\
O\
H
00
O\
v—4
[\
O\
‘I—d
\0
ON
0—4

 

 

.36
V H 04 60.0 V m H * .56 v R H .1. 650.0 V an .13.. 80 EOQEm .895 m2 546 @202 new eoxm 00 @8005 380$ :0 ﬁts 202: BE
2: maﬁa osoe was» 8:098» mo 0630904 38% 32>on E @2390 mcmoacacb 80¢ 80 32 E 38me $93-39 E 32%? 803

35:53:. 033-033 3030 So» #4 2: $>o 526% SEENE out so 8:258: coumﬁztom was coca—£30 mo ﬂooﬁo BE. .m eight

72

 

15 _
Total phenolics

12 _ UTannins
6 .
'. ‘a a b ab
3 a
x x x
0 I i i i y i y

Def-, N- Def-, N+ Def+, N- Def+, N+

Treatment

% Dry Weight
(0

 

 

 

Figure 4. The effects of defoliation and fertilization treatments on concentrations
(mean i SE) of total phenolics and condensed tannins in June 1996. Means (i
SE) were calculated from the four plots within each treatment category. Different
letters indicate signiﬁcant pairwise differences among means (p < 0.05) after
signiﬁcant main defoliation or fertilization effects in ANOVA. Treatment codes
are undefoliated = Def -, defoliated = Def +, unfertilized = N-, and fertilized =
N+.

73

F1-

 

00.00 0 . 00.00

 

 

0 Sum

0 0.0 00.0 a 0.0 00.0 a 80 x 605
0.0 00.0 a 0.0 :0 0 50580

r. 0.0_ 00.00 _ r. 0.: 00.00 0 0000:0000.

0. 0.0 00.0 0 a. 0.0 00.0 0 002m 0000
.0 02 00 ,0 02 00 8500 29:00
00:00:31 305280 0020503 38H

 

 

.oﬁo V H .0 60.0 V a.“ 0. .56 V at” *0. .506 V 00” 03.3.. cobm m2 BB @203 was 30¢ 00 0300b 08800

:0 :23 0588 BE 0% maﬁa once 003 880900, .00 20%—0:400 .355 08302820 0003 8809 40:507. 83800 09:02 80¢ 35303

e08 3800.08 295 0953800 .003 5 02:59:00 02022.3 8.38 so 35:58.5 mocmmﬁtﬁ ES cobmmﬁmwxo 030%.? «is .V QENH

74

In June 1997, concentrations of foliar nitrogen, total phenolics, condensed tannins,
and six other phenolic compounds were measured shortly before peak defoliation. I found
that defoliation signiﬁcantly reduced the concentration of nitrogen by 10% and 15% in
unfertilized and fertilized plots, respectively. Defoliation increased levels of total
phenolics by 15% in unfertilized plots and by l 1% in fertilized plots (Figure 5, Table 5).
Tannins also were increased by defoliation (24% in unfertilized plots, 15% in fertilized
plots). Fertilizer had no signiﬁcant effects on the concentration of foliar nitrogen nor did
it inﬂuence total phenolics and condensed tannins. One control plot was removed from
the analysis because large numbers of gypsy moth larvae moved into it from the
adjoining defoliation plot. Although I removed these caterpillars, analysis indicated that
concentrations of total phenolics and condensed tannins increased dramatically and were

as high or higher than in the defoliation plots.

Reﬂush leaves sampled from defoliation plots in July 1997 had marginally lower
nitrogen levels (-9% and -10% in unfertilized and fertilized plots, respectively), whereas
nitrogen was signiﬁcantly higher in fertilized trees (11% in undefoliated plots, 12% in
defoliated plots; Figure 5, Table 5). Total phenolics (-11% in both unfertilized and
fertilized plots) and condensed tannins (-20% and —42% in unfertilized and fertilized
plots, respectively) were signiﬁcantly lower in the July foliage of trees that had been
completely defoliated earlier in the season (Fig 5, Table 5). There was no main or
interactive effect of fertilizer on levels of total phenolics or condensed tannins in July
leaves. Interestingly, the condensed tannin and total phenolic concentrations in the
control plot that had experienced signiﬁcant gypsy moth herbivory had returned to levels

found in the other control plots by late season.

75

 

 

     

 

 

 

 

 

     

 

 

 

 

10 U T t I h l' 4
0a p enoncs -
8 A JuneD Tannins 3 June Nitrogen
6 ‘ 2 ‘
4.0
.C 4 - l
.9 1 4 ab
a: 2 - ,
g o 0
E10 5 T t I h I 4
. oa p enorcs -
D 8 7 JUIyEI Tannins 3 . July Nitrogen
c\° 6 4 l
2 4
4 - .
2 a 1 ' ab
0 0
Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+
Treatment

Figure 5. The effects of 1997 defoliation and fertilization treatments on
concentrations (mean t SE) of total phenolics, condensed tannins, and foliar
nitrogen in June and July. Means (i SE) were calculated from the four plots within
each treatment category. Different letters indicate signiﬁcant pairwise differences
among means (p < 0.05) after signiﬁcant main defoliation or fertilization effects in
ANOVA. Treatment codes are undefoliated = Def -, defoliated = Def +, unfertilized
= N-, and fertilized = N+

76

 

 

 

 

one 0 Had a 03. 0 Sum
90 8.0 H No w H .o H HS 8.0 H Em x Heoo
* as 3.0 H no 2.0 H 0.0 8.0 H Swansea
,H 3 N2 H 0.0 me H * 3 08 H Sansone
NH 8.0 m do 8.0 m 9H 93 m €on 33
#2 w 8.0 m 3H w 55
90 No.0 H no 8.0 H S 8.0 H Em x Heom
2 H H .o H 3 8.0 H No 8.0 H HHOHHEHHHEVH
* SH 30 H t w: NNH H t a: a: H 8:20am
2 2.0 m t we HMS m t 3H H2 m vHoonH 2:;
m m: e m m: E m m: (He 85% 295m
amolele god 8:823 Heoe

 

 

.03 v n Ht woe v mu * .50 v a n I Hood v an 1* as 286er .85, m: a: 832 Has SE a

@335 mHoHoﬁ =H~ 5:5 388 SE 2: mam: 38 was» BENCH?» .«o $9223 .933 Home 5:08 use 28 8:58 cotmzomow Mama 8&3

btoﬁ 3383 803 8284 .32 E cowobE USN 3:823 8.38 so mEoEEob :oHHmNHHEﬂ 98 2232839 mo muooﬁo BEL .m «3.3.

77

The treatment effects on minor phenolic constituents were variable (Table 6, Table 7).
Defoliation signiﬁcantly increased the levels of phenolic glycosides but had no main
effects on any of the other compounds measured. Fertilizer signiﬁcantly reduced levels
of quercetin and marginally reduced myricetin concentration. With the exception of
luteolin, concentrations of all minor phenolic compounds were altered by defoliation
earlier in the season. In refoliated leaves, phenolic glycosides, cinnamic acid, and
quercetin increased whereas levels of catechins and myricetin were decreased. Fertilizer
had marginally signiﬁcant effects on phenolic glycosides and there was a marginally

signiﬁcant interactive effect of defoliation and fertilizer on cinnamic acid.

Leaves were sampled at three intervals in 1998. Despite signiﬁcantly lower nitrogen
in the late season leaves in 1997, there was no effect of either defoliation or fertilization
treatments on foliar nitrogen concentrations following bud break in May 1998 (Figure 6,
Table 8). Herbivory was minimal at this time because caterpillars were still very small
indicating that the effects of severe defoliation and fertilization treatments in the previous
season did not carry over to 1998. Conversely, the previous year’s treatments did alter
concentrations of secondary compounds in the following spring. May 1998 samples had
signiﬁcantly higher levels of total. phenolics (10% in unfertilized plots, 23% in fertilized
plots) and condensed tannins (21% and 31% in unfertilized and fertilized plots,
respectively) were marginally higher in plots defoliated in 1997 (Figure 6, Table 8).

Fertilizer did not have an effect on concentrations of secondary compounds in spring.

Defoliation and fertilization treatments had signiﬁcant effects on nitrogen in leaves
sampled in June 1998 shortly before peak defoliation. Nitrogen was lower in defoliated

plots (-3% in unfertilized plots, -10% in fertilized plots) and elevated in plots that had

78

PEOU

 

 

 

79

000.0 H 02.0 80.0 H 000.0 000.0 H RH.0 000.0 H H000 03. 28056
03.0 H mmm.m 21.0 H 60H 09.0 H com; wood H mmvg 8208wa 83:25 hi
H000 H 20.0 000.0 H $0.0 000.0 H 20.0 30.0 H $0.0 0060030-:H08E2
30.0 H 000.0 000.0 H 090 0000 H 000.0 20.0 H H30 oeHmoobwaeHoeeHo
H000 H 30.0 0000 H 0000 0000 H 000.0 20.0 H 03.0 H5023
000.0 H 000.0 0000 H H000 03.0 H 000.0 000.0 H H50 23098
$00 H 02.0 000.0 H 3.0.0 300 HHOH.0 000.0 H 02.0 0H8 25966
08.0 H 000.0 02.0 H 0.00.0 :00 H 08.0 20.0 H 52.0 828030 00825 2:;
£0 H 9320 mm H 5520 50 H $020 50 H 08020

+2 #me .2 £me 2 memo -z :me 0550600 206me

“GOEHNOHH.

 

 

.+Z H coNEtow EH0 r Z H woNEtotHSH .+ (SD M 882800 r mom H 3020.325 80 02000
0.3820on .315 2030232 Howe 0::po _ H050 A230 coumzomoc Mama €809 btoﬁ 39:3 0962 Sod womssasc 803 09300380

.SQH E 33800 Amm H 5095 09:69:00 23:23 SEE no 358880. :oHHmmEtom 98 20502820 mo gumbo 2:. .9 29¢.

 

wood H mHod

wmod H Elmd

:dd H cmmd

mvod H mind

wood H ASod

Nmod H whmd

oHod H mHNd

mmod H Humvd

wood H Edd

mmod H Rand

omod H mad

nmod H wmmd

mood H ded

mmod H mmmd

NNod H Emd

mwod H mm©d

oEmoobwécooaz
0208398008030
:28qu

2.50800

 

0008 0 2000.

80

 

81

 

H00.0 0 000.0 0 000.0 0 H00 0 00.0 0 3.0 0 8:0
H0 000.0 H 0.0 000.0 H 0.0 H000 H 0.0 00.0 H :0 00.0 H 0.0 H00 H 503000
0.0 000.0 H 0.0 000.0 H 0.0 000.0 H 0H 00.0 H 0.0 00.0 H 20.0 00.0 H 000000200

.00 H000 H *00 000.0 H 0.0 H000 H ...0.0 00.0 H ._.0.0 00.0 .H .20.: 00H H 00000000
0.0 000.0 0 :0 000.0 0 0.0 000.0 0 0.0 H00 0 0.0 00.0 0 0.0 00.0 0 H020

0%
:00 0 000.0 0 000.0 0 00.0 0 0H.0 0 00.0 0 5:0
0.0 000.0 H 0.0 H000 H 0.H 000.0 H :0 00.0 H 0.H H00 H H0 00 H 5010000

H3 000.0 H :0 £00 H 0.0 000.0 H 0.0 00.0 H S 00.0 H 0.0 0.0 H 000000050
00 000.0 H H.H 000.0 H 0.0 H00.0 H S 00.0 H H.H 50 H .1. 0.0H 00.0 H 000000000
00 000.0 0 0.0 000.0 0 0H 000.0 0 0.0 H00 0 0H 50 0 .00 00.0 0 00on
{.0 0H2 00 0 000 00 2 0H2 00 2 02H 00 DH 02 00 2 03H 00 202

30. 090450.
008202 008000 2180.3 01% 231020.300 00208 H 00002: 3%

 

 

 

d2d v H H .mod v an * Jod V mu .1. .2ood v an is.“ 020 02020203 .<>OZ< 2: 22 0002 0203 00002 00 22002 028002 20
223 020022 H03 .3220 0003 202020202 w22$0202 2020 A023 80: 202020200 20002 020.202 32020 00222200 00>002 0202 002220220 0203

0020020200 .39 22 0022002200 02020202 20222 20 020202200200 20 0020022002 2020022202 020 202020200 aH0 0200222 .5 030B

 

   

 

 

   

 

 

 

 

   
    

 

 

 

 

 

 

10
‘ 5 Total phenolics May Nitrogen
8 ‘ Ma El Tannins
.C < Total phenolics
.9 8 - JuneEl Tannins
o _
3 6 '
2‘ 4 '
n 2 -
0° 0 g
10 I
‘ Total phenolics
8 ‘ JUIyEl Tannins
6 _ .
4 _
2 — .
0 — o T ,

 

 

 

 

 

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+

Treatment

Figure 6. The effects of 1998 defoliation and fertilization treatments on foliar
concentrations (mean i SE) of total phenolics, condensed tannins, and nitrogen in
May, June, and July. Means (i SE) were calculated from the four plots within each
treatment category. Different letters indicate signiﬁcant pairwise differences among
means (p < 0.05) after signiﬁcant main defoliation or fertilization effects in ANOVA.
Treatment codes are undefoliated = Def -, defoliated = Def +, unfertilized = N—, and
fertilized = N+.

82

 

 

 

 

0.250
00.0 0 00.0 0 H00 0 8:0
._. 00.0 00.0 H 0H 00.0 H S 00.0 H :02 x H0000
.1. 0.0H 0H0 H H.0 00.0 H 0.0 0H0 H 202022202
3. 0.0H 0H0 H r... 0.00 00.00 H r... 0.00 $.0H H 000000002
.2. 0.0H 0H0 0 .3. H.HH £0 0 0.0 H00 0 00on 2:2
00.0 0 H00 0 00.0 0 Sum
00H 00.0 H 0 00.0 H 3 00.0 H EHH x H0000
~02 mod 2 m2 mvd 2 0d 3 .o 2 20200220002 &
00.0 00.0 H .2 0.0 0H.H H ._. 0.0 0H.0 H 000000000
._. 00.0 00.0 0 0.0 00.0 0 0H 00.0 0 00on 002
2 02 00 0 000 00 .0 02 00 8050 20800
20w022 Z 0222200 000200200 002202022. 20002.

 

 

d2d V H 0. .mod V 00.1. * Jod V an ** .2ood V 0‘” is.” 020 0202225 .2022 m2 20>0 000002 020 00002 00 000002 0000002 20 223

020002 0022 w2202 0200 003 002020> 20 0200020240N .925 020 202002 020 020 A0223 202020.200 2000 020.202 0002020 60006 @2020 002200

22 00222200 0003 00000012

0002 22 20w0222 020 002202022 00.05 20 002022002 2020022202 020 202020.200 .20 000020 022. .w 030E

 

Nd

*NR

**N.N2

0.0

mod

20.0

mmd

020.0

8.0

0.0

** N:

**m.om

0.2

$6

mmd

omd

mm.m2

mod

20.0
"33.“th
“will

:0

0.010

2w.o

no.0

voN

mud

202m

2002 x 20.2002
20220022222002
202202202002

v200:2

2::

 

0.008 0 0300

 

84

been fertilized (Figure 6, Table 8). There was a signiﬁcant interaction between
defoliation and fertilizer. Defoliation increased total phenolics (24% in unfertilized plots,
26% in fertilized plots) and condensed tamiins (31% in unfertilized plots, 37% in
fertilized plots) but fertilizer itself did not change levels of either compound signiﬁcantly

(Figure 6, Table 8).

Samples collected from refoliated trees in defoliation plots in July 1998,
approximately one month after peak defoliation had signiﬁcantly lower nitrogen (-12% in
unfertilized plots, -14% in fertilized plots) than trees in control plots (Figure 6, Table 8).
Fertilizer increased nitrogen levels in both undefoliated and defoliated plots. In contrast
to 1997, leaves sampled from plots defoliated earlier in the season had elevated levels of
total phenolics (7% in unfertilized plots, 20% in fertilized plots) and condensed tannins
(20% and 35% in unfertilized and fertilized plots, respectively; Figure 6, Table 8). There
was a signiﬁcant reduction in phenolics and tannins in fertilized plots and a marginally

signiﬁcant interaction between defoliation and fertilization on total phenolic levels.

Of the minor phenolics quantiﬁed in May 1998, quercetin concentrations were
reduced in plots defoliated in the previous year while fertilizer had no measurable effect
on any compound (Table 9, Table 10). Concentrations of catechin, luteolin, and myricetin
were all increased by defoliation in June leaf samples (Table 9, Table 10) whereas
fertilizer signiﬁcantly decreased luteolin and marginally increased catechins, but had no
effects on other minor phenolics. In July leaves, defoliation had signiﬁcant effects on
levels of luteolin and in addition, there was an interaction between fertilization and
defoliation. Myricetin-glycosides were signiﬁcantly higher in the plots defoliated earlier

in the season.

85

20.2200

 

 

 

86

000.0 0 000.0 000.0 0 000.0 H000 0 000.0 000.0 0 0000 008.022
000.0 0 000.0 000.0 0 000.0 0H0.0 0 H000 0H00 0. 000.0 008500
30.0 H 03.0 20.0 H 03.0 000.0 0 22.0 000.0 0 02.0 00023
000.0 .0 :00 00H.0 0 000.0 300 0 000.0 0H H0 0 H020 020200
$00 0 000.0 000.0 0 000.0 000.0 0 000.0 000.0 0. 000.0 00% 0200006
000.0 0 000.0 0H 3 .0 000.0 000.0 0 000.0 000.0 .0 0000 0000820 0000002 202
200 H 000:0 200 H 0025 200 H 0025 200 H 000:0
+2 £202 .2 0:202 +2 .0222 -z .0222 000000000 20600
2202200;

 

 

.20 v n 2 .000 v 0.” ... .50 v 2n .2. 2000 v 2n .2... 80 200220
.<>O72< 022 22 20002 0203 20002.2 00 200200.: 0202002 :0 H2223 020002 20202 .0223 02022 2020220202 w22>>02202 2020 A0222 02022 202202202020
02002 0.20202 22222020 .3006 220202 30 H2208 2202230 2020 020022 20202 20220 002003 N 2002022200 00>002 0202 20222220220 0203 0202220022200

.2302 22 0222022200 02202022 202222 20 020202200200 20202 20 022022002 2022002222202 2020 20220220200 .20 0200.22m2 .2 030%

 

 

000.0 H NN0.0

wm0.0 H 000.0

2N0.0 H 002.0

mm0.0 H 000.0

mm0.0 H 0NN.0

00N0 H 200.2

w20.0 H 0N0.0

w2m.0 H 0.00.0

mmod H 002.0

220.0 H 000.0

0m0.0 H w0N.0

m2m.0 H w0m.0

000.0 H NN0.0

mm0.0 H 220.0

020.0 H m02.0

30.0 H 000.0

0m0.0 H 022.0

w2N.0 H .0022

000.0 H 20.0

002.0 H M000

08.0 H N020

220.0 H 000.0

N000 H m2m.0

002.0H 0200

000.0 H N200

08.0 H 000.0

0N0.0 H 2Nm.0

000.0 H 200.0

020.0 H 002.0

002.0H 000.2

000.0 H 000.0

002.0H 020.0

wmod H 0mm.0

220.0 H 220.0

2m0.0 H 22m.0

02N.0 H 2mm.0

2N0.0 H 000.0

200.0 H N000

000.0 H 202.0

202.0 H 000.0

0m0.0 H 002.0

00.00 H mmm2

000.0 H 2N0.0

202.0H 0N2.0

020.0 H 0mm.0

000.0 H Nm0.0

«.80 H 22.0

$2.0 H wmmd

2220002232
222002020
222002212
220200200

2020< 02802220
00202000020 0220202202
222002.222
22002020
222002212
222200200

020< 022202220

0020200002w 022020022

202

0222

 

0008 0 2000

 

 

20.2200

 

 

 

0.0 200.0 m .3. 2.22 000.0 m 0 m.m 000.0 m .0 m2 22m.0 m 2.2 N200 m.N 0N.0 m 2005

a

M0000 0 N200 0 200.0 0 022.0 0 2m0.0 00.0 0 202m2

m0 0000.0 2 2.N M000 2 mN 000.0 2 N0 000.0 2 2.0 N000 2.0 00.0 2 20.2 x 20.2002

2.0 2000.0 2 0.0 000.0 2 0.0 000.0 2 0.0 200.0 2 0.0 200.0 2.0 00.0 2 202200222202

0.0 0000.0 2 ._. m0 000.0 2 0.2 000.0 2 0.0 200.0 2 0.0 000.0 2.0 00.0 2 202202202002

mN 2000.0 m ._. 0.0 000.0 m ._c_. 0.0 200.0 m m0 020.0 m ._. m2 020.0 .0... 2.02 20.0 M 020022

.2 000 20 .2 000 00 .2 000 20 .2 020 00 .2 000 20 .2 000 .00 2000

02420.0 .1230...

220.002.2222 22208020 @2242. 02220|0200 203404224023 0020200020 02202022 3

 

 

020vu2020v2u*220v2n**

.2000 v 2. H .12.. 020 02222me .<>O72< 0H: 22 2000: 0203 20002.2 00 002002 0202000 220 22223 020002 20222 005,2 00 2002002 0202002 220 02223

020022 2022 0022 02202 02020 003 0020220> 00 020222020V .0223 02020 002002 2020220002 2.0 202020220 022302202 2020 A0220 02020 202202202020

02002 020202 022020 6202.6 2302 22022 009% 2020 0200202 20202 000 002003 N 20022800 00>002 22022 2000220220 0203 02022022200

.0002 22 0022022200 02202022 202202 20 0202202200200 0202 20 0220022002 2022002222202 2020 20202202020 00 0200220 0020 .02 030,—.

88

 

0200.0

N.0 0000.0
0.N m000.0

._. 0.02 N200.0

m2 N000.0

00m0.0
0.0 2000.0
2.0 2000.0

.2 0.0 0000.0

.2

2

2220.0

2.m 020.0

0.0 M000

0N N200

0N 220.0

00m.0

2.0 200.0

0N NN0.0

m.m 0N0.0

.H. 2.0

0.2

... 2.0

.H. 2.0

0.2

.0 m0

220.0

000.0

N000

020.0

000.0

000.0

200.0

N200

030* 2.2mm 020.0

wN
0.0
0N

m.N

0.0
.2 0.0

* 2.0

2N2.0

wm0.0

000.0

020.0

2m0.0

mm0.0

000.0

02N.0

m0m.0

m2

0.2

0.0

02

N0

0.0

m.N

200.0

000.0

000.0

000.0

20.0

000.0

200.0

000.0

No.0

2.0

m0

N.0

0.2

N.0

0.0

m0

20.N

No.0

00.0

00.0

N20

002

No.0

00.0

M00

v-I'l

v—It

202m

2.2002 x 20.2002
202200222202

0002000002

20on
2.2.2

202m

2002 x 20200
0000000002

000000000

 

0.208 S 2002

 

89

Water concentration was signiﬁcantly reduced in the defoliation plots in June 1998
(Figure 7, Table 11), but was not affected by fertilizer. There was no signiﬁcant effect of
either defoliation or fertilizer on July water content although the among-plot variability
was much higher than earlier in the season (Figure 7, Table 11). Water content was 5-8%
higher in plots that had been severely defoliated and produced a second set of leaves.
Conversely, it was lower than controls in moderately defoliated plots where trees retained
their damaged leaves. As well, the area of intact leaves was reduced in defoliation plots
but was not affected by fertilizer (Figure 7, Table 11). Leaf toughness increased in

defoliation treatment plots but was not altered by fertilizer (Figure 7, Table 11).

Leaves were sampled only once in 1999, corresponding with the presence of late
instar gypsy moth. Because gypsy moth develops more slowly in low-density populations
due primarily to shifts in feeding behavior (Lance et al. 1986), the samples were taken at
a point phenologically later in the season than in the previous three years. No defoliation
was evident and fertilizer was not applied in 1999. Although there was trend toward
lower nitrogen in plots defoliated over the three previous years, the difference was not
signiﬁcant (Figure 8, Table 12). Conversely, nitrogen was elevated in plots that had been
fertilized in the three preceding years (8% in undefoliated plots, 5% in defoliated plots).
Concentrations of total phenolics (19% in both unfertilized and fertilized plots) and
condensed tannins (15% in unfertilized plots, 27% in fertilized plots) were higher in
previously defoliated plots (Figure 8, Table 12). Prior fertilization marginally reduced
total phenolics and signiﬁcantly reduced condensed tannins. Treatment history was also

reﬂected in foliar water content with reduced levels in the defoliation plots, although

90

 

 

June Water July Water

         

60—
.\°40— 240—
20— 20—

 

 

June Leaf Area . June Leaf Toughness

 

 

 

 

   

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+

Treatment

Figure 7. The effects of 1998 defoliation and fertilization treatments on measures (mean
i SE) of water content in June and July, and leaf area and leaf toughness in June. Means
(i SE) were calculated from the four plots within each treatment category. Different
letters indicate signiﬁcant pairwise differences among means (p < 0.05) after signiﬁcant
main defoliation or fertilization effects in ANOVA. Treatment codes are undefoliated =
Def -, defoliated = Def +, unfertilized = N—, and fertilized = N+.

91

 

 

 

00.22 0 000.0 0 00.0 0 w0.2 0 202m

 

 

0.0 m0.0 2 2. 0.m 200.0 2 N0 N2 2 0.0 No.0 2 2022 x 2020a

0.0 2m.0 2 2.2 000.0 2 0.0 02.0 2 2.0 22.0 2 2022022222002

.2... 0.0N 2m.N0m 2 ._. 0.0 m00.0 2 0.0 20.0 2 0...... 0.0N 20.0N 2 202202202022

02 00.0N m .H. 2.2 N000 m 0.N 00.22 m 0.N w2N m V2002m2

n2 m2>2 .220 r2 £2 .220 22 m2 220 .02 92 .220 00220m
0222 - 0020320012 0222 - 000222m202. 2222 2 - 202025 3

 

 

020vn2020v2n.
.200 V 2. H .0... .2000 V 2. H .23. 020 02002223 2005 m2>2 20>0 2002002 2020 20002.2 00 2002002 0202002 :0 H2223 020002 2022 0222 M2202 02020 003

00202022 .20 0202220202 .0002 22 0000229202 .2002 020 .0020 2002 .2020? 20 022022002 2022002222202 2020 20202202020 20 0200220 022. .22 02220.2.

92

 

    
  

      

 

 

   

 

4-0 12 El . 4
g, June D $gt1anliﬁgenollcs , , ' June Nitrogen
._ 3 *
OJ 8 .
3 2 —
a 4 * b
n 1 a
o\° 0 . ' ' . . , 0 . . . ~
Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+
Treatment

Figure 8. The effects of previous defoliation and fertilization treatments on foliar
concentrations (mean t SE) of total phenolics, condensed tannins, and nitrogen in
June 1999. Means (i SE) were calculated from the four plots within each treatment
category. Different letters indicate signiﬁcant pairwise differences among means (p
< 0.05) after signiﬁcant main defoliation or fertilization effects in ANOVA.
Treatment codes are undefoliated = Def -, defoliated = Def +, unfertilized = N-, and

fertilized = N+.

93

 

 

 

 

220.20 22 0220 a 2020 a 2820
2.2 220.20 2 2.20 020.20 2 20.2 20.20 2 :02 02 2202022
.3. 20.22 22.20 2 .2 0.2 2.2% 2 .3. 20.202 2.0 2 2220222222202
22:2. 8.20 2 .3. 2.2 2002.2 2 .3... 2.202. 3.22 2 28202220202
2 220.20 m 22.2 2.22 m .. 22 32 m v2820 22:2
2 £2 220 .2 £2 220 .2 m: 220 8280 0222250
gleam 332224200. 822202202222 25202

 

 

.02.0 v H 2. .m0.0 v 2.“ .2. .200 v RH .23.. .2000 v 2.“ .13.. 020 0200225 20.2.5 m2 20>o 2002002 2020 2000222 020 20020022 0202002 2222

22223 020022 20222 022 w22ms 02020 903 0020220.» .20 0202220272.. 20202.2 020 22 00.2202 222022 209$ 202022 2022.2 20 002000222 523 m22202022oo 02222

0202 22 2002222200 0203 00200.2 .0002 22 20w02222 2020 0022020222 20.022 20 0220220022 202202222202 20220 2022202202020 .220 0200,2020 0222. .2 02220.2.

94

there was no effect of previous fertilizer application (Figure 9, Table 13). In 1999, leaf

area was not affected by previous defoliation or fertilization regimes (Figure 9, Table 13).

Effects of defoliation and fertilization on caterpillar performance

Gypsy moth lifelong bioassays - Neither defoliation nor fertilization had any effects on
gypsy moth pupal mass in 1996 despite signiﬁcant treatment effects on foliar
biochemistry (Figure 10, Table 14). The development time of males was signiﬁcantly
increased in the defoliation plots, although the increase was small (less than one full day)
and unlikely to have biological relevance. There was no impact of either treatment on

female development time.

I increased the density of gypsy moth eggs in the defoliation plots by more than 40
fold in 1997 resulting in near complete defoliation by the end of the larval feeding period.
Caterpillars in the plots switched from feeding only at night to feeding both diurnally and
nocturnally, a typical density related response of gypsy moth larvae seen in natural
outbreak populations (Lance et al. 1986). Surprisingly, only a marginal decrease in
female pupal mass was detected, and there was no effect of defoliation on males (Figure
10, Table 14). There was no direct or interactive effect of fertilization on the pupal mass
of either sex, nor were there any effects of defoliation or fertilizer on the development
times. From the regression equation derived for female gypsy moth pupal mass and egg
number (see methods), I estimated that fecundity was reduced from 717 to 632 eggs, a

difference of 11.9% for caterpillars reared on foliage from the defoliated plots.

95

 

 

80 40
June Water ~ June Leaf Area

 

 

 

 

   

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+

Treatment

Figure 9. The effects of previous defoliation and fertilization treatments on water
content (mean i SE) and leaf area (mean i SE) in June 1999. Means (i SE) were
calculated from the four plots within each treatment category. Different letters
indicate signiﬁcant pairwise differences among means (p < 0.05) after signiﬁcant
main defoliation or fertilization effects in AN OVA. Treatment codes are undefoliated
= Def -, defoliated = Def +, unfertilized = N-, and fertilized = N+.

96

 

 

 

 

 

00.0 0 00.0 0 202m
2.0 0N2 2 N2 00.20 2 220m x 220.20D
m0 2.0.N 2 m. 2 0: 2 2022002222202
Wm N228 2 0.0 00.0». 2 220220220200
m0 0210 m 0.0 0N.m M 020222
.02 £2 220 02 $2 220 005cm
022 2. - 00222 00012 0222 - 202025
.02.0 v H

2‘ 60.0 v 2. H .2. .200 v 2. H 2.... .2000 v 2. H. 232.... 020 022022202m .2otm2 m2 20>o 2002002 2020 20002.2 00 2002002 02202002 220 22223 02005 2022 022

M00220: 02020 003 00202202» 20 0202220522 .0002 22 00.20 0002 2020 20203 20 02202220022 2023222202 2020 20220220200 20 0200220 0222. .2 022208

97

 

Pupal Mass Development Time

1996 - 1996

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

     

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+
Treatment
FemaleCl Male

 

Figure 10. The effects of defoliation and fertilization treatments on male and female
gypsy moth pupal mass and development time for 1996-1999. Means (1 SE) were
calculated from the four plots within each treatment category. Different letters
indicate signiﬁcant pairwise differences among means (p < 0.05) after signiﬁcant
main defoliation or fertilization effects in AN OVA. Treatment codes are undefoliated
= Def -, defoliated = Def +, unfertilized = N-, and fertilized = N+.

98

daovn2wqo
v 2. H .2. .200 v 22 H .232. .2000 v 2. H ._.._..2. 020 022022202m 02202220022 0200.2 02030222 m 2002002202 0202222220200 2222 20002202200 0w022o.2 0202020222
.0002 22 200222220 0203 02202220022 02 0200222 m w222202202 0222 .2220 20002 0202220220 00 2022022022 2200222 202 0220 2002220220 003 02200222 0220 .0002

22 20222 02072 germ m2>2 20>o 2002002 2020 20002.2 00 20020022 02202002 220 22223 020022 2022 0222 w2202 02020 003 00202202» .220 02022202 00020002

.2002 222000 22 02222 2202222020520 2020 0008 202222 222022 220223 02080.2 2020 02022 20 2202202222200 2020 202202200020 00 0200.200 022. .3 0220.2.

 

 

99

 

 

 

2.20 a 02.20 0 02020.20 0 202.2020 0 2220
2.20 00.20 2 0.0 02.20 2 2.20 202020.20 2 20.20 202020.20 2 220.2 x 22002
20.2 3.20 2 2.0 00.2 2 2.20 22020.20 2 2.20 02020.20 2 02.0.2
2.2 022.20 2 20.20 2020.20 2 2.0 02020.20 2 20.2 220.20 2 2200.02
20 02.20 0 20 02.20 m 0.20 22020.20 0 20.2 220.20 0 022020222 082
R2 0 00.2 0 20220.20 0 820.20 0 02.20
0.2 2222 2 2.20 020.2 2 22.20 2020.20 2 22 2220.20 2 220.2 x 220.202
2.20 2.2.20 2 0.20 00.2 2 2.20 22020.20 2 0.2 2220.20 2 020.2
02 2.0.2 2 20.20 2020.20 2 2.20 22020.20 2 .2. 0.: 022.20 2 2.02022
2.0 20.0 m 2.20 202.20 2.. 00 0020.20 m .2. 2.02 2202.20 2.. 022020222 082
02.2 a 02.2 0 0020.20 0 22020.20 0 02.00
2.20 020.2 2 20.20 2020.20 2 0.20 202020.20 2 0.20 0220.20 2 0.5.2 x 220.202
20.20 020.20 2 2.20 22.20 2 0.20 202020.20 2 20.20 202020.20 2 220.2
0.20 02.20 2 0.2 00.2 2 20.0 02020.20 2 2. 2.2. 002.20 2 3002
0.20 2.2.20 0 0.20 :20 m 0.20 22020.20 0 2.2 2.02020 0 02020222 302
02.20 a 200.20 a $20.20 0 02020.20 0 0820
2. 3. N20 2 2.0 202.2 2 2.20 22020.20 2 2.20 0220.20 2 220.220.2022
20 202.20 2 2.20 020.20 2 2.2 02020.20 2 0.2 2.20.20 2 25202002222202
.. 0.0 20.2.0 2 0.0 00.2 2 2.20 202020.20 2 0.20 0020.20 2 2520202220002
22.0 20.2 m 2.0 20.2 m 2.20 02020.20 0 2 0220.20 m 02820 0002
.2 022,2 020 .2 022 020 m 022 020 {.2 02 020 8080 08>
@202. 04% 30,2. 0022.42.00
022222 22022222020322 00.292 220.2222

 

 

100

I relied on in situ egg masses and did not supplement gypsy moth densities in 1998.
As a result, defoliation levels among plots were more variable than in the previous year,
ranging from moderately to severely defoliated. In lifelong bioassays, gypsy moth
females suffered a signiﬁcant decline in pupal mass in defoliation treatments (-15.8%
unfertilized plots, -8.3% fertilized plots) whereas males were not affected in 1998 (Figure
10, Table 14). There were no effects of defoliation or fertilization on development time of
either males or females. Because of high mortality in my gypsy moth rearing stock (33%
survival) due to an unidentiﬁed, presumably bacterial disease, data was available for only
three of the four blocks (12 of 16 plots) and sample sizes within cells were reduced
relative to previous years. Nonetheless, my data show that the combination of previous
(1996 and 1997) and current year defoliation reduced fecundity by 14%, or 95 eggs per

female, similar to the reduction seen in 1997.

In 1999, defoliation levels were very low and I did not apply fertilizer, thus any
treatment effects reﬂect the history of the plots over the previous year or years. There was
no signiﬁcant effect of defoliation or fertilization on male or female gypsy moth pupal
masses or male and female development time (Figure 10, Table 14). The lack of effects
on gypsy moth following three consecutive years of defoliation indicate that induced-
resistance in these poplars was driven primarily by within season changes in leaf quality,

and not by DIR.

Gypsy moth short-term bioassays - In a short-term bioassay of 48 hours, relative growth
rates (RGR) for 5th instars did not vary among treatments in 1997 (Figure 11, Table 15).
Relative consumption rate (RCR) increased by 9% in response to defoliation in fertilized

plots and by 24% in unfertilized, defoliated plots. Defoliation and fertilization did not

101

hr

 

Figure 11. The effects of previous defoliation and fertilization treatments on mean (i SE)
relative growth rates (RGR), relative consumption rates (RCR), approximate digestibility
of consumed foliage (AD), efﬁciency of conversion of ingested foliage (ECI), and
efﬁciency of conversion of digested foliage (ECD), for ﬁfth instar gypsy moth in 1997
and 1998. Means (i SE) were calculated from the four plots within each treatment
category. Different letters indicate signiﬁcant pairwise differences among means (p <
0.05) after significant main defoliation or fertilization effects in ANOVA. Treatment
codes are undefoliated = Def —, defoliated = Def +, unfenilized = N-, and fertilized = N+.

102

 

 

0.8 1997 0.8 1998
0.6 - 0.6 -

0.4 r 0.4 ~

 

 

A0
420

 

 

 

mg-mg'1- day'1
: El
: -E

-El

 

 

%

 

 

 

 

ECI

 

 

 

{if ' ., 53 ‘5?

 

 

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+
Treatment

103

    

 

 

104

 

0.02 0 022.20 00.0N 200.0 000.0 0 2202.22

m2 0N2 2 m0 20.2 2.2 mv.0m Wm wmmd ._.... M22 m00.0 2 2.202 x 20202

0.0 0.0 2 2.0 $0 N2 00.2m 2.N 002.0 0.2 200.0 2 2022322222022

Wm Wow 2 .2. 0.m wwhm m0 000 .232. 0.02 0N2: 0.2 200.0 2 202202220202
02 0.02 m .2 2.m 22.02 0.0 Qumm 0.0 000.0 .2. 0.0 2000.0 m 02002m2 0002

000 w m0.m N00m 202.0 m00.0 0 2202.22

N0 N02 2 2.0 202.0 2.0 000 0.2 0mN.0 m2 000.0 2 2202 x 20.202

0.0 2.0m 2 ._. 0.0 N002 2.0 0N2“ 2.0 020.0 .2. Wm 220.0 2 2022022222002

.2 w.m 0.0mm 2 .232. 0.02 02.2% 2 0.0 00.2 2 .2. 0.0 2000.0 2 0.0 000.0 2 2202202202002
0. 0.0 WEN m ._.._.._. NMN N220. m 0.2 02.2% m .2 w.m m2m.0 m 0.2 m00.0 m v2002m2 0002
.22 m2 .220 r2 m2>2 220 K m2>2 .220 n2 m2 0220 .22 m2 .220 0022om 200%

.QIUIM ﬁlm. ml... Mud. ma

 

 

.02.0 V H .2 60.0 V 02” .2. .200 V mu ._.... .2000 V 2.” 232.... 020 020222225 .2002 22 220222 020 202 2002020002 82 2222200202022 003 02020
20222222202200 20002.2 00 20020022 02202002 220 22223 020022 2022 0222 20002 0020220> .20 02022202».2 .0002 2020 0002 22 22022 209$ 202022 em 2202
22020 20200.2 200200m220 .20 220202022200 20 2020202220 2020 A2020 200202 2002000022 .220 2020202200 20 2220202202220 ADA: 222222222200w220 0202220822220

.8020 0202 20222220200 0320202 .920”: 0202 2223202w 0220202 20 0220222002 2022002222202 2020 20202202020 .20 0200.22.02 .3 030B

alter the approximate digestibility of foliage. However, defoliation signiﬁcantly
decreased the efﬁciency of caterpillars in converting ingested (ECI) and digested (ECD)
foliage to biomass by 19% and 21%, respectively. Fertilizer did not affect either of these
measures of feeding efﬁciency. The 1998 short-temi bioassay yielded similar results to
those seen in 1997 (Figure 1 1, Table 15). There was no treatment effect on RGR but RCR
increased by 16% and 29% in caterpillars feeding on foliage from fertilized and
unfertilized defoliation plots respectively. ECD was reduced by 23% and ECI by 22% for
larvae fed foliage from defoliated plots. Fertilizer did not signiﬁcantly affect measures of

efﬁciency in the short-term bioassay.

I used Pearson correlation coefﬁcients to assess the associations among
phytochemical variables measured in this experiment, and pupal mass and development
times of gypsy moth (Table 16). Nitrogen was positively correlated with the pupal mass
of gypsy moth females in 1997 and 1998. In 1999, development time of both male and
female gypsy moth was negatively correlated with nitrogen although for females the
effect was only of marginal signiﬁcance. Water content was not measured in 1997 but
was positively correlated with female gypsy moth pupal mass in 1998. Of the secondary
compounds determined, total phenolics were negatively correlated with pupal masses of
both sexes of gypsy moth in 1997 and with female mass in 1998. Total phenolics were
signiﬁcantly correlated with longer development times for both males and females in
1996. Condensed tannins make up a large portion of the total phenolics in these poplars
and thus had similar correlations with pupal mass and development times of gypsy moth.
Male gypsy moth development time was marginally correlated with condensed tannins in

1996, 1997, and 1998. On the other hand condensed tannins were correlated with female

105

.36 v R v mod 8m moﬁgﬁ BEBE:

60.0 v m 2m m03m>i 09$ Eom .moEmoobw 55238 H 7:1ng 98 .moEmoobm cubocosc H 2.5530 6:082 H AOMHDA

106

.msEoBmo H momﬂzo 66m 28986 H 527:0 .moEmoobm 0:05an wﬂoom H >ADZmEm £2683 womcowcoo H Z<HQZU
“monocoam :88 H ZEQHOH £395 H ONE hammer: H Z 8m mcoumgoﬁﬁx .303 05 8on 835mm 32*: E mﬁomEonogﬁm

mo 32855800 98 0E: anmoﬂgow was mmmE Essa £08 39% 222:8 98 23: comics mnosﬁoboo :83me .3 03.31

.a...
.

 

 

 

 

 

 

 

. . . . . . . . . . . . 00.0 3.0 00.0 00.0 000 00.0. 00.0 $0. 0
. . . . . . . . . . . . 000 2.0 2.0 0.0.0 00.0 v0.0- 000 00.0- 2 002
50 00.0- :0 00.0 00.0 :0- 000 3.0 3.0 :0 000 8.0 00.0 3.0 00 20- R0 000. 00.0 00.0 m
:0 8.0- 00.0- 20 00.0 00.0 00.0 00.0 00.0 8.0 $0 2.0. 050 000 S0 S0 000 $0- 20 N00- 2 0000
$0 20 30 00.0- N00 20. 20 8.0. 2.0 00.0 £0 £0 £0 2.0 0.0.0 000 . . £0 £0. 0
~00 000- 20 00.0- 000 00.0- 000 E0- N00 20 00.0 N00 00 00.0 :0 ~00 . . $0 .20- 2 L02
. . . . . . . . . . . . 00.0 00.0 000 $0 . . . . m
. . . . . . . . . . . . 00.0 000 20 $0 . . . . 2 002
l L Q L Q L l L Q L n& L K L AN L Q L AN L va uwo>
2.5602 70.5350 400:3 mum—05 0,2226 >5sz z<Ezo 7:308 g m «Baa—0.00.3530
. . . . . . . . . . . . 000 2.0. 00.0 $0. 00.0 :0- 00.0 00.0- m
. . . . . . . . . . . . 00.0 2.0- 00.0 2.0- 00.0 00.0 00.0 2.0 E 002
000 00.0- 2.0 :0 00.0 00.0- 00.0 $0 2.0 00.0- 00.0 000 S0 000. 00.0 00.0- 00.0 000 N00 80 m
20 000 .00 ~00 :0 000 20 $0 2.0 30 80 E0 000 0.00. 00.0 2.0. R0 000- 3.0 3.0 2 0000
$0 2.0- 00.0 30 N00 000 00.0 00.0 :0 00.0. 00.0 000. N00 000- .00 ~00- . . ~00 00.0 m
8.0 00.0 $0 :0 :0 20 00.0 2.0 00.0 2.0 00.0 00.0 3.0 $0- 8.0 00.0- . . 2.0 2.0 E 82
. . . . . . . . . . . . 00.0 00.0- .80 8.0 . . . . m
. . . . . . . . . . . . 00.0 3.0 E0 E0 . . . . 2 0000
Q L Q L Q L Q L Q L l L Q L Q L Q L Q L 03% .8an
a a and 0550.9 24225 5020:: 25:20 79.3.50 % a “:2ng

 

 

 

7
O
1

gypsy moth development time in 1996 only. Phenolic glycosides had only a marginal
negative association with female gypsy moth in 1997, and had a signiﬁcant positive
relationship to male and female pupal mass in 1998. Because concentrations were low
(>1%) in this pOplar clone, other intercorrelated phytochemical constituents may be
determining the variable associations of phenolic glycosides with gypsy moth pupal
mass. A similar process may explain why male pupal mass was positively correlated with
myricetin- and quercetin—glycosides in 1998 but not in 1997 and had no effect in either

year on females.

Long-term bioassays on other species of Lepidoptera - In addition to gypsy moth, I
determined the effects of fertilization and gypsy moth defoliation on four other species of
poplar-feeding lepidopterans in 1998 and three species in 1999. A large portion of the
feeding period of forest tent caterpillar overlaps with gypsy moth (see Figure 2). In
lifelong bioassays, pupal mass of female forest tent caterpillar was signiﬁcantly reduced
by defoliation whereas male pupal mass was unaffected in 1998 (Figure 12, Table 17).
Using the regression equation for fecundity and pupal mass, I calculated that defoliation
reduced fecundity by 9%, or 21 eggs/female. Pupal mass was not different among plots in
1999 suggesting that the reductions in 1998 were due to rapid-induced responses elicited
by gypsy moth feeding damage. Fertilization did not affect forest tent caterpillar pupal
mass in 1998 but signiﬁcantly increased female mass in 1999 (Figure 12, 13, Table 17,
18). Development time of forest tent caterpillar was not affected by treatments in either

year (Figure 12, 13, Table 17, 18).

The poplar tent maker initiates feeding in mid-season when gypsy moth defoliation is

at its peak. Both female and male pupal masses were signiﬁcantly reduced in the

108

 

Pupal Mass 80 Development Time
, M. disstn'a

0.6

 

M. disstn'a

”917.14
A El

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+

Treatment
I FemaleEl Male

Figure 12. The effects of defoliation and fertilization treatments in 1998 on mean (i SE)
pupal mass and development time of forest tent caterpillar (M. disstria), poplar tent
maker (C. inclusa), big poplar sphinx (P. modesta), and fall webworm (H. cunea). Means
(i SE) were calculated from the four plots within each treatment category. Different
letters indicate signiﬁcant pairwise differences among means (0 < 0.05) after signiﬁcant
main defoliation or fertilization effects in ANOVA. Treatment codes are undefoliated =
Def -, defoliated = Def +, unfertilized = N-, and fertilized = N+.

109

00.0vn L .000va ._. .000vmu ._.... .2000 v m

H ._.._.._. 20 ﬂonﬁxm 020:5,“ was 8:2: cﬂoom :o 0:20 003 203050 om 3:80:50 3338 Ho: €03 30$th Q :2: 807:» .Hotm m2

110

8>o 8me USN coxm mm @8080 £80090 :0 5:3 £538 83 05 wEm: meow £03 88033, mo membmsxx .033 E Aeﬁwﬁooxﬂ .00 £08

06033 0338-853 was Aeozxo 5% 0503903 :& AEwmnoE .3 08:30 SEOQ m5 Aegean .03 5me EB 830m .3328 V5

85838 ES 688 com 2:: EoEQBoSw 38 $008 Ema 038$ use 208 so :ocﬁﬁtﬂ 98 08302820 mo ﬁooﬁo Bit .: oEnH

 

 

 

 

 

M00 0 R0 0 ~0000 0 0000.0 0 H0cm
0.0 ~00 0 0.0 :0 0 0.0 0000.0 0 0.0 800.0 0 Em x 308
0.0 00.0 _ 2 00.0 0 0._ 800.0 0 ~.0 ~0000 0 Em
0. L0 0m~ _ 0.0 00.0 _ 2 A0000.0 _ 00 ~0000 0 608
._. 3 R~ m 0.0 $0 0 0.0 ~0000 m ~.0 00000 0 e285 858 .m
. . 000. 0 . . 000.0 0 Sum
. . . m0 0: 0 . . . E 2 _.0 _ Em x :03
. . . m0 0: 0 . . . m0 0~00 0 Em
. . . ._. 0.0 0~.0_ 0 . . . ._. E 03.0 0 600a
. . . 0.0 30 m . . . ~.0 20.0 0 seem ”sense m
0&0 0 0~._ 0 000.0 0 000.0 0 Sum
00 2.0 0 m0 000 0 2 000.0 0 ~.0 000.0 0 Eu x 303
2 :0 0 0.0 ~00 _ ~0 000.0 _ 0.0 000.0 0 Em
._.... 0.3 00.0 _ ._. 0.0 00.2 _ ._.... 0.2 000.0 0 ._. 0.0 000.0 0 605
0.~ 00.0 m w.~ 00.0 m E 0000 m S 0000 m seem Sees .o
00.0 0 $0 0 S00 0 ~000 0 Sum
00 m _ .0 _ 0.0 00.0 0 0.~ 0000 0 0.0 ~00.0 _ Em x 305
0.~ 00.2 0 00 00.0 0 0.0 000.0 0 0.0 000.0 0 souﬁzeom
0.0 R0 0 50 90 0 0.0 000.0 0 ._. 00 ~50 0 aoséeom
._.... 0.~_ 000 m ._.... 00 ~00 m ._. 0.0 80.0 m ._.... 0.0 20.0 0 seem 35% .3 002
u m: 00 m m: 00 L m: 00 L 02 00 08mm seam 5;
3234 galﬁ g 3%.
m2? Ewe/$03500 332 40:50

 

 

111

0 8 Pupal Mass 40 Development Time

4 M. disstria J M. disstn'a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 -‘
10 -
0 j
Def-, N- Def-, N+ Def+, N- Def+, N+ Def-, N- Def-, N+ Def+, N- Def+, N+
Treatment

FemaleEl Male

Figure 13. The effects of previous defoliation and fertilization treatments on 1999
pupal mass and development time of forest tent caterpillar (M. disstria), white-marked
tussock moth (O. leucostigma) and poplar tent maker (C. inclusa). Means (i SE) were
calculated from the four plots within each treatment category. Different letters indicate
signiﬁcant pairwise differences among means (p < 0.05) after signiﬁcant main
defoliation or fertilization effects in ANOVA. Treatment codes are undefoliated = Def
-, defoliated = Def +, unfertilized = N—, and fertilized = N+.

112

.36 v H + .36 v
o. H * 40.0 v R H in Jocd v o. H *** 80 Eonﬁzm .058» 085 "BESS 2: E 8:QO 3:258: vogue—m8 E22883 >9 coEsmcoo
owmzom 0:5 .32 E @2390 803 3580020 oz gobm m2 S>o @808 was woxm mm @2080 086$ :0 :23 0:02: 83 2t wEms

got 003 meant? we Ewing dog E A532: My 5me EB $33 98 AeEmEoome .Qv £08 #0033 woﬁmEoHEB Aegean $5

50:58:00 EB 0088 com 2:: EoEmoEBw ES 308 Rose 208mm 98 205 :o cEENEtE can nonsmomow mo ﬂooto 2:1 .2 03.;

113

 

 

 

0.0 0 00.0 0 0000.0 0 80.0 0 Sum
m 0.0 02 H 0.0 00.0 m 0.0 500.0 _ 0.0 000.0 m EmiEo
._. E 00.0 m 0.0 00.0 m 0.0 500.0 m 0.0 000.0 m Em
S 00.0 0 .m 0.0 00.0 m E 3 300.0 m ... 0.0 000.0 0 BED
2 00.0 0 0.0 000 0 0.0 0000.0 0 0.0 000.0 0 EOE SE3 .9
00.0 0 00; 0 500.0 0 000.0 0 Sum
00 00.0 m 0.0 00.0 E 2 500.0 m 0.0 000.0 E Em 0 ES
0.0 00.0 m 0.0 00.0 m 0; 500.0 E m 0.0 20.0 m Em
S 00.0 0 0.0 00.0 m 0.0 00000 H 5.0 000.0 _ SEQ
0.0 00.0 0 0; 00; 0 00 500.0 0 0.0 000.0 0 EOE sameness .o
00.0 0 50.0 0 0000.0 0 30.0 0 85
0.0 00.0 _ 00 00.0 m 0.0 0000.0 H 0; 50.0 m Em x :Eo
0.0 00.0 m 0.0 00.0 H 0.0 0000.0 0 ... 0.0 000.0 E Em
2 8.0 _ 0; 5.0 m 0.0 0000.0 _ E 80.0 0 BED
0; 00.0 0 0.0 3.0 0 m 0.0 :000 0 ._. 3 000.0 0 EOE 032% e0 0000
m 02 00 m 02 00 m 02 00 m 02 m0 scum 02800 Em
.Elaﬂ 04.0% 00002 gum.
m2; mZmEmoqm>mo mm<2 imam

 

 

114

defoliated treatments in 1998 (Fig 12, Table 17). Using regression, I estimated that
defoliation reduced fecundity by 9% (30 eggs/ female). There were no main or interactive
effects of fertilizer on pupal mass. Both male and female poplar tent maker took
signiﬁcantly longer to complete the larval period on trees which had been defoliated. The
performance of poplar tent maker in 1999 was surprising because female pupal masses
were signiﬁcantly heavier and males marginally heavier on previously defoliated trees
than on control trees (Figure 13, Table 18). Female development time was marginally
faster on control trees. Observations suggested that this species may prefer vigorous
shoots. The proliferation of rapidly growing epicormic shoots on surviving trees in
previously defoliated plots may account for the apparent positive effects of the treatment

history on this species. There was no effect of fertilizer on either sex in 1999.

Big poplar sphinx initiates feeding in July after previously defoliated trees have
refoliated. In this study, big poplar sphinx did not exhibit sexual dimorphism in pupal
mass. Males and females were pooled for analysis. I found that defoliation earlier in the
season had signiﬁcant, negative effects on big poplar sphinx pupal mass in 1998 (Figure
12, Table 17). Although not estimated in this study, the declines in pupal mass would
likely translate to signiﬁcant reductions in fecundity because the adults do not feed. In
addition, development time of big poplar sphinx was lengthened signiﬁcantly on trees
defoliated earlier in the season (Figure 12, Table 17. I was unable to obtain sufﬁcient

numbers of big poplar sphinx to use in bioassays in 1999.

Fall webworm had the latest feeding phenology of any species used in my bioassays,
beginning near the end of July. I found no effect of either defoliation or fertilizer

treatment on male or female pupal mass for this species nor was there an effect on female

115

development time (Figure 12, Table 17). Male development time was marginally faster
on previously defoliated trees although the difference was less than one day. Fall

webworm was not assayed in 1999.

White-marked tussock moth was used in bioassays for the ﬁrst time in 1999. In
Michigan this species is bivoltine having a spring and summer generation. I used ﬁrst
generation larvae only and found that previous defoliation did not signiﬁcantly affect
pupal masses of either males or females (Figure 13, Table 18). Female pupal mass was
marginally heavier on trees that had received fertilizer in previous years. There were no

effects of defoliation and fertilizer on development time of white-marked tussock moth.

As with gypsy moth above, correlation analysis was used to assess the association of
individual phytochemical components with pupal mass and development time of the
insects (Table 19). Pupal mass of female forest tent caterpillars was positively correlated
with foliar nitrogen in 1998 and 1999. Male forest tent caterpillar pupal mass was
marginally correlated with nitrogen in both years. Pupal masses of male and female
poplar tent makers as well as big poplar sphinx were correlated with nitrogen levels in
1998 although the signiﬁcance was only marginal. Pupal mass of fall webworm was not
correlated with foliar nitrogen levels. In 1998, development times were longer when
nitrogen was low for both sexes of forest tent caterpillar and for female poplar tent
makers. Nitrogen was also negatively correlated with development time of male poplar
tent makers although the effects were only marginal. There was no correlation between

development time of any other species and foliar nitrogen in 1999.

Water content was correlated with pupal masses of female forest tent caterpillar, and

male poplar tent makers in 1998 (Table 19). Water was also correlated with male poplar

116

 

.26 v 00V mod 80 moE0>i 806:0: 60.0 v Q80 SEE/i
0&0 200m 800830 008:? M 250002 000 0000830 :EEE u Emmbo 00022 n momma 050808 u momzu .060
28056 H $2225 .moEmooﬁm 0:823 Boom H >q©mem .mEEmS 80:06:00 H Z<EQZU .8205me E009 H ZEmHOF £203
H om: doonc H Z 20 00530305900 .wo>mo>E 028% 600E oE mom 2E0 E0528 bm0oEwomoeosa 0008 Rt 00 @8828 003 00E
29:00 m02 2: 500300. 0:00 803 00832880 .083 2: 80¢ 33500 00202 E £0EEoEoo§Ea mo 0503055280 900 @Ewumoosﬂ
.OV EoE x8003 $0338-03?» ES 3038 iv E83063 20% 5:03.09: .03 xEEQm 0an En A0032: UV EMS: E8 SEOQ

Animate .05 359200 E8 0028 m8, 2E0 005508636 30 0008 00:3 208$ E0 208 59,309 mcosﬂoboo :00m0om .3 0308

 

117

 

 

 

 

 

 

 

 

. . . . . . . . . . . . 00.0 00.0 00.0 00.0 00.0 00.0- 3.0 00.0- m

. . . . . . . . . . . . 00.0 00.0 00.0 00.0- 00.0 00.0. 000 00.0 2 0320.9

. . . . . . . . . . . . 00.0 8.0 00.0 00.0 00.0 00.0 00.0 00.0 m

. . . . . . . . . . . . 00.0 00.0 2.0 00.0 00.0 00.0 8.0 00.0 2 00.00080006

. . . . . . . . . . . . 00.0 00.0 00.0 00.0 00.0 8.0 00.0 00.0 m

. . . . . . . . . . . . 00.0 00.0- 000 000. 00.0 8.0- 00.0 00.0- 2 5.00008 0.00:
00.0 00.0- 00.0 00.0- :0 00.0- 00.0 00.0- 00.0 00.0 00.0 00.0 00.0 00.0 00.0- :0 00.0 00.0 00.0 00.0 m
00.0 00.0- 8.0 00.0- 00.0 00.0- 000 00.0- 00.0 8.0- 00.0 00.0- 00.0 00.0- 00.0 00.0- :.0 00.0- 00.0 2.0 2 $03.:
00.0 00.0 00.0 00.0 00.0 000 00.0 00.0 $0 2.0 00.0 00.0 00.0 2.0 00.0 :0- 8.0 000 0.0.0 00.0- mi 300005.00
00.0 20. 8.0 :0 :0 0:0 :0 00.0 00.0 00.0 00.0 00.0- 8.0 00.0 8.0 00.0 8.0 00.0- 00.0 8.0- m
8.0 2.0 00.0 80 0:0 000 8.0 00.0 8.0 00.0 00.0 00.0 8.0 8.0 8.0 00.0 8.0 00.0- 00.0 00.0. 2 0305.9
8.0 $0. 000 00.0- :0 8.0- 8.0 00.0- 00.0 :0 8.0 00.0- 00.0 00.0 00.0 00.0 00.0 00.0- 00.0 8.0- m
00.0 $0- 000 $0- 80 00.0- 00.0 00.0- 00.0 00.0 :0 00.0- 000 00.0 00.0 000 00.0 :0- 80 00.0- 2 0.50008 000:
AN & Ax x K k 0.x k l L AN & 0% k n-N x 0* k AN .0 Xom momoomm 53>
a 2.5.3.0 4883 mummku 80.725 50205.. 72820 zmmmbm a a

. . . . . . . . . . . . 000 00.0 8.0 00.0 00.0 00.0- 00.0 00.0 m

. . . . . . . . . . . . :0 00.0 00.0 00.0 00.0 00.0- 00.0 3.0- 2 03.050

. . . . . . . . . . . . 00.0 00.0- 00.0 00.0- 00.0 8.0- 00.0 000 m

. . . . . . . . . . . . 00.0 00.0- 00.0 00.0- 00.0 3.0 8.0 00.0 2 0500080050

. . . . . . . . . . . . 2.0 00.0- 00.0 00.0- 00.0 00.0- 00.0 00.0 m

. . . . . . . . . . . . 00.0 00.0- 00.0 00.0- 000 00.0 00.0 $0 2 5300.8 000:
8.0 8.0- 000 00.0 00.0 00.0 000 000. 00.0 00.0 000 0:0 :0 00.0- 00.0 00.0- 00.0 000 00 00.0 m
8.0 00.0- 00.0 00.0 00.0 :0- 00.0 00.0- 00.0 000. 00.0 00.0- 00.0 00.0- :00 00.0- 00.0 :0 00.0 :0 2 02:8,:
00.0 00.0- 00.0 8.0- 000 00.0- 8.0 00.0- 00.0 $0. 000 00.0- 8.0 00.0- 00.0 00.0. :.0 00.0 00.0 2.0 m3 8.000500
00.0 00.0- 00.0 8.0 00.0 :0- 000 00.0 00.0 00.0- P0 00.0 00.0 00.0- 2.0 00.0- 00.0 00.0 00.0 :0 m
00.0 00.0- 8.0 00.0 00.0 :0- 000 00.0- :0 00.0- 00.0 00.0 8.0 00.0- 00.0 00.0- 8.0 00.0 00.0 00.0 2 02020.9
00.0 0.0- 00.0 00.0 00.0 00.0- 00.0 00.0 00.0 00.0- 00.0 00.0 8.0 00.0- :0 00.0. 80 00.0 8.0 8.0 m
00.0 00.0 3.0 00.0 00.0 3.0 00.0 00.0- 00.0 000- 00.0 00.0 8.0 00.0- 8.0 00.0- 00.0 00.0 00.0 00.0 2 0.520.: 000:
m m l L l m m g m m 0‘ L Q x Q m l m l L 5% 86on 80>
awe-a 78550 gem-03 mom—$6 E4725 0,-5sz 722.020 zmmeoe INN m Egg-=43

 

 

 

 

 

 

tent maker pupal mass in 1999. Pupal masses of the late-feeding species, big poplar
sphinx and fall webworm, were not affected by water content. Correlation analysis
suggested that development time of big poplar sphinx was faster when feeding on leaves
with higher water content, as was development time of both sexes of poplar tent maker in
1998. Only the development of male poplar tent makers varied with water content in

1999.

Male poplar tent maker and big poplar sphinx pupal mass were negatively correlated
with total phenolics in 1998, although only marginally in the case of big poplar sphinx
(Table 19). In fact, the relationship between total phenolics and pupal mass had a
negative sign for all species in both years except for poplar tent maker in 1999. The
development times of both male and female poplar tent maker increased with total
phenolic concentration in 1998. The pupal masses of female forest tent caterpillar and big
poplar sphinx were negatively correlated with condensed tannins in 1998. There was a
marginal association between, female forest tent caterpillar development time and
condensed tannins and a signiﬁcant relationship for condensed tannins and development
time for both sexes of poplar tent maker in 1998. However, there was no correlation
between condensed tannins and pupal masses of fall webworm in 1998 or for any species

in 1999.

Other phenolic constituents in poplar leaves had variable and species speciﬁc effects
on pupal mass and development time. Of the species assayed, only pupal mass of big
poplar sphinx was consistently associated with concentrations of these compounds (Table
19). Cinnamic acid, catechins, luteolin, quercetin, and myricetin were all negatively

correlated with pupal mass of this herbivore, although the relationship was of only

119

marginal signiﬁcance in some cases. Of the minor phenolics, myricetin showed the
strongest correlation with herbivore performance, having a negative association with
pupal mass of male poplar tent maker, big poplar sphinx, and both sexes of fall
webworm. This was the only compound I measured that correlated either positively or.
negatively with pupal mass of fall webworm. Forest tent caterpillar development time
was sensitive to phenolic glycosides, myricetin-glycosides, and quercetin-glycosides,
which all negatively correlated with female development time. Male development time
was also negatively correlated with myricetin and quercetin although only marginally
whereas cinnamic acid. exhibiting a positive correlation. Decreases in the development
time of male and female fall webworm were associated with increasing myricetin

concentrations.

DISCUSSION
Effects ofdefoliation-induced changes in phytochemistry on insect herbivores

I conducted laboratory bioassays using gypsy moth in each of the four years of the study,
and in addition to gypsy moth, used ﬁve other species over the last two years of the
study. In 1996 gypsy moth pupal mass was unchanged among treatments despite
defoliation—induced increases in total phenolics and condensed tannins. In contrast, gypsy
moth fecundity was reduced by 12% by defoliation in 1997 and by 14% in 1998. The
fecundity of forest tent caterpillar, poplar tent maker, and big poplar sphinx was also
reduced in 1998 in the defoliation plots by 9%, 9%, and 10%, respectively. The
performance of gypsy moth and three other species assayed in 1999 were unaffected by

three consecutive years of previous defoliation.

120

The complete lack of signiﬁcant negative effects on herbivores in 1999 suggests that
RIR rather than DIR was determining foliar quality in these poplars in 1997 and 1998.
However there was considerable variation in the strength of RIR on individual herbivore
species both within a season and among years. In the ﬁrst year of the study, large
increases in total phenolics and condensed tannins had no effect on gypsy moth fecundity
or female development time and only minimal effects on male development time. In
contrast, partial defoliation has had signiﬁcant effects on gypsy moth in other
experiments using poplars. Havill and Raffa (1999) compared RIR induced by gypsy
moth feeding, the elicitor jasmonic acid, and mechanical damage to 12 poplar cultivars.
Not only did the type of damage inﬂuence induction, the strength of the induced
resistance on gypsy moth varied by as much as 72% among different cultivars. Similarly,
growth rates of early instar forest tent caterpillar were three times higher on undefoliated
poplars than on trees partially defoliated (Robison and Raffa 1997). Studies using other
tree species have also shown signiﬁcant RIR effects on gypsy moth. On red oak with
current year defoliation of 10-58%, Rossiter et a1. ( 1988) found that female gypsy moth
pupal masses were reduced by as much as 20%. Furthermore, the largest portion of the
variance in gypsy moth fecundity was explained by the concentration of phenolics which
in turn, increased with defoliation. Data from natural populations of gypsy moth also
suggest that RIR can reduce the pupal mass and fecundity of gypsy moths while
lengthening development times. Relative to an undefoliated site, pupal mass was reduced
by 22% for caterpillars fed foliage from a site with a single year of defoliation and 32%
from a site with one year of previous and one year of current defoliation (Lance et al.

1991).

121

There are a number of possibilities why I did not see any effect of elevated levels of
phenolics on gypsy moth in 1996. Densities of gypsy moth were relatively low in 1996
compared to 1997 and 1998. It is possible that the negligible feeding damage by early
instars may have been insufﬁcient to induce any changes. Subsequent feeding by large
larvae may have triggered the induced response, but too late to have any signiﬁcant
impact on growth. The data of Schultz and Baldwin (1982) indicates that tannin levels in
oaks reach a peak after gypsy moth has ﬁnished feeding. Similarly, D’Amico et al. 1998
found no difference in tannin levels between defoliated and undefoliated oaks in May and
June when gypsy moth are actively feeding, although signiﬁcant differences were
detected later in the season. My short—tem1 bioassays in 1997 and 1998 indicate that
growth rates of late instars were not affected by the defoliation treatments because they
were able to compensate for reduced leaf quality by increasing consumption rates. Havill
and Raffa (1999) showed that feeding damage to P. X euramericana “Eugeneii”, the
same cultivar used in my experiments, reduced growth rates of laboratory reared second
instar gypsy moth by 44%, suggesting that larval ontogeny may inﬂuence susceptibility

to secondary compounds.

Similar variability in RIR response to defoliation is apparent in other Populus species.
On aspen, Lindroth and Kinney, (1998) showed that herbivory by forest tent caterpillar
elicited a signiﬁcant increase in condensed tannins, but that most of the increase came
late in the larval period when caterpillars had nearly completed their development. As a
result, there was no signiﬁcant effect on the growth of the caterpillars. In contrast, Parry
et al. 2000, found that current year defoliation of aspen signiﬁcantly reduced growth,

pupal mass of both sexes, and fecundity of forest tent caterpillar. They used much higher

122

defoliation levels (ca. 80% vs. 30%) than did Lindroth and Kinney (1998) suggesting that
induction of RIR may be density dependent and that the timing of induction mediates the
outcome of experiments. If the timing of induction also determines the effects on gypsy
moth, then the signiﬁcant reductions in gypsy moth fecundity in 1997 and 1998 may be
due to increased levels of defoliation achieved in those years relative to 1996. Because
densities of larvae were much higher and signiﬁcant damage occurred earlier, induction
could have occurred at an earlier point in the season thus affecting the more sensitive

early instars.

A more parsimonious explanation for the lack of effects on gypsy moth in 1996 is
that by themselves, total phenolics and condensed tannins play little role in determining
gypsy moth performance. This explanation seems contradictory given the strong negative
correlations between gypsy moth pupal mass and the concentrations of total phenolics
and condensed tannins in 1997 and 1998. In those two years however, nitrogen levels
were signiﬁcantly lower in the leaves of defoliated trees. Thus, the performance of gypsy
moth in 1997 and gypsy moth and the other species assayed in 1998 may reﬂect a
response to foliar nitrogen concentration. This suggests that the negative correlations
between caterpillar performance and total phenolics or condensed tamiins are spurious. A
number of studies have shown that nitrogen and condensed tannins are themselves
negatively correlated (see Tuomi et al. 1990). In 1999, levels of foliar nitrogen in the
previously defoliated plots were not different from controls, and there were no negative
effects of treatments on the four species assayed. The lack of any effects on larval
performance in 1999 Occurred despite total phenolic and condensed tannin levels in trees

from previously defoliated plots that were as high as the concentrations in 1997 and

123

1998. This strongly suggests that induced secondary phenolics were not substantially

inﬂuencing the performance of the caterpillars in this experiment.

The lack of a direct effect of condensed tannins on herbivore performance does not
necessarily mean that they had no biological activity. Negative correlations between
pupal mass and condensed tannins were higher than were positive correlations between
foliar nitrogen and pupal mass for all species assayed in 1998. This suggests that when
nitrogen levels vary as in 1997 and 1998, condensed tannins have greater negative effects
on herbivore performance than when nitrogen levels were more uniform among
treatments, as in 1999. Bryant et al. (1993) describe a similar relationship between
nitrogen and condensed tannins on the performance of the lepidopteran Rheumaptera
hastata feeding on paper birch. Condensed tannins were highly negatively correlated with
pupal masses of R. hastata whereas nitrogen had the opposite effect. They concluded that
condensed tannins were the most important determinant of the induced-resistance they
found in birch because in laboratory bioassays, artiﬁcially increasing tannin levels
reduced performance of the larvae. However, they did not vary the nitrogen content
simultaneously, which may have better elucidated the relationship between condensed
tannins, nitrogen, and insect growth. Bryant et al. (1987) varied the concentration of
condensed tannins and nitrogen in artiﬁcial diets fed to large aspen tortrix. The
magnitude of the tannin effects was greater on the low nitrogen diet suggesting that
tannins may interact with dietary nitrogen. Low nitrogen levels have been found to
exacerbate the activity of other phenolic compounds (Lindroth and Bloomer 1990). In the

study, phenolic glycosides were more detrimental to forest tent caterpillars when they

124

were reared on diets low in nitrogen. Interactions between secondary phenolics and foliar

nitrogen may be a fruitful area of future research.

My study does suggest that the principal induced-response in poplars to herbivory has
little direct effect on several of its herbivores. This raises interesting evolutionary
questions about the actual function of condensed tannins. In other Populus spp. such as
aspen, condensed tannins have also been shown to have little or no effect on herbivorous
insects. Constitutive levels of condensed tannins do not affect forest tent caterpillar and
gypsy moth growth (Hemming and Lindroth 1995, Hwang and Lindroth 1997).
Condensed tannins exhibit major differences in structure among tree species and also
vary considerably in their antiherbivore activity (Ayres et a1. 1997). An assay of six
insect herbivores against condensed tannins from 16 woody plants found that condensed
tannins had signiﬁcant effects in only eight of the 45 insect-tannin combinations that
were run (Ayres et a1. 1997), suggesting that as a resistance mechanism, tannins have
limited utility.

Not all studies have shown effects of DIR following defoliation. The performance of
winter moth feeding on oak and apple was found to decrease with increasing levels of
current year defoliation. However, performance on oak was positively correlated with
previous years defoliation, at least across a range of 10-50% (Roland and Myers 1987). In
the same study, higher defoliation of apple in the previous year signiﬁcantly decreased
the pupal mass of winter moth. In natural gypsy moth populations, Lance et a1. (1991)
found that 70% defoliation of red oak had little effect on gypsy moth pupal mass or
development time in the following year. Current year defoliation was much more

important in reducing pupal mass, although the site with the lowest pupal masses had

125

experienced both current and previous defoliation. On red alder, severe (>70%)
defoliation by western tent caterpillar had no effect on larval performance in the
following year. Similarly, Harrison (1995) found that pupal mass or fecundity of western
tussock moths was not reduced on bush lupines that had been completely defoliated in the

previous year.

Most experimental tests of DIR have consisted of a single year’s defoliation followed
by assays in the following year or years. However, it has been suggested that the
cumulative effects of successive defoliation may have larger effects than a single event
(Haukioja et al. 1988). Many natural insect outbreaks partially or completely defoliate
trees for at least two years in succession, indicating that longer studies may be more
realistic. Some studies have shown that there may indeed be cumulative effects of
sustained defoliation on herbivores. Werner (1979) found that additional years of
defoliation had additive effects on mortality of spear-marked black moth. Similarly
defoliation of black oak for three consecutive years reduced gypsy moth pupal masses by
increasing amounts in each successive year presumably because of long term depletions

in foliar nitrogen (Walton and Wallner 1979, Valentine et al. 1983).

A few other studies have also experimentally determined the effects of cumulative
defoliation on insect performance. The results of Clausen et al. (1991) are difﬁcult to
interpret. Treatments of one, two, or three consecutive years of defoliation were applied
to aspen, and the performance of large aspen tortrix fed foliage from these trees
compared with insects fed foliage from undefoliated controls. Unfortunately, the
researchers reared tortix larvae on the refoliated leaves of the aspen, which is not

phenologically relevant because the caterpillars normally initiate feeding at bud break

126

and complete their development on the spring leaves (Parry et al. 1997). Thus, the instars
(2-5) used in their bioassay would never encounter reﬂush foliage in nature. The large
negative effects on large aspen tortix pupal mass are interesting from a physiological
standpoint but do not reveal anything about the effects of defoliation on populations of
this insect. Research with mountain birch has shown that the effects of two consecutive
years of defoliation were not signiﬁcantly greater than a single defoliation (Kaitaniemi et
al. 1999b). They found no difference in the fecundity of two defoliators, autumnal moth
and winter moth, Opherophtera brumata, reared on foliage from plots suffering 25% and
75% defoliation, or two years of 75% defoliation. Similarly, four consecutive years of
high defoliation (>70%) did not reduce forest tent caterpillar fecundity more than a single
year of high defoliation in two clones of trembling aspen (Parry et al. 2000), although
there was a trend towards lower performance on trees with the highest cumulative
defoliation. My study also showed that the effects of repeated defoliation events did not

elicit any additional response in the poplars beyond that occurring in a single season.
Effects of defoliation on tree growth and phytochemistry

The CNB hypothesis predicts that under conditions where growth but not photosynthesis
is limited by nutrient deﬁciencies, excess carbon will be allocated to carbon-based
secondary compounds such as phenolics (e.g., Bryant et a1. 1983, Tuomi et al. 1984,
1990). The ability of trees to recover nitrogen after defoliation is contingent upon both
species-speciﬁc compensatory responses and on environmental constraints such as
resource availability (Tuomi 1990). Carbon-based secondary compounds are predicted to
remain elevated until a tree’s nitrogen pool has recovered to predefoliation levels, a

process that may take several years in nutritionally poor environments (Tuomi et al.

127

1984, 1990, 1991, Bryant et al. 1991). Under these conditions, the application of
nitrogen fertilizer to trees suffering defoliation should mitigate the production of

secondary phenolics because carbon will be preferentially allocated to growth.

The effects of defoliation on foliar nitrogen within a growing season have received
less attention than processes occurring in the subsequent year. I found lower nitrogen
levels in poplar leaves from defoliated plots in both mid and late season samples in 1997
and 1998. Since early season nitrogen levels did not differ among treatments, current
year herbivory was likely responsible for the change. Declines in nitrogen concurrent
with defoliation, or later within the same growing season, have been shown in several
other tree species (e. g., F aeth 1986, Robison and Raffa 1997, Lindroth and Kinney 1998,
Roth et al. 1998). The mechanism behind nitrogen decreases in damaged leaves is not
well understood. Some nitrogen may be leached from the leaves, or the tree may
withdraw nitrogen from the damaged leaf as has been shown for other nutrients (Chapin
1980, Nef 1988). Reasons for nitrogen reductions in refoliated leaves are more obvious.
When trees produce a second set of leaves from dormant buds, they must do so without
the stored nitrogen pools available for leaf expansion during the spring, leading to lower
levels of foliar nitrogen. The poplars in my study exhibited this pattern in 1997.
However, in 1998, many of the trees were not defoliated to the extent that extensive
reﬂush occurred and instead, retained damaged leaves for the duration of the season.
Foliar nitrogen was signiﬁcantly lower in these damaged leaves than in control foliage.
This suggests that feeding damage and the physiological constraints of producing new

leaves may both account for within season declines in nitrogen following defoliation.

128

Increases in secondary phenolic compounds are often correlated with decreases in
foliar nitrogen (Tuomi 1990, Bryant et al. 1993). The two main secondary phenolic
classes found in poplars are condensed tannins and phenolic glycosides, both products of
the shikimic acid pathway (Clausen et al. 1991, Lindroth and Hwang 1996). I found that
gypsy moth feeding induced signiﬁcant within season changes in total phenolics and
condensed tannins in each of the three years that defoliation treatments were applied. In
1996, the ﬁrst year of the study, I recorded signiﬁcant increases in total phenolics and
condensed tannins. Within season increases in condensed tannins in response to
defoliation have been shown repeatedly (e.g., Schultz and Baldwin 1982, Rossiter et al.
1988, Faeth 1988, Hunter and Schultz 1995, Roth et al. 1998). In my study, foliage
samples collected in 1999, one year after the cessation of defoliation also had elevated
levels of secondary phenolics in previously defoliated plots showing that induced effects
on condensed tannins occur both within and across years. Given these results, the
elevated levels of condensed tannins in 1997 and 1998 likely reﬂect defoliation effects

from both current and past years, a result that has not been documented previously.

The addition of N-fertilizer reduced levels of both total phenolics and condensed
tannins in defoliated plots in 1996 and 1998 as predicted by the CNB. In 1997, the year
with most severe defoliation, there was no effect of fertilizer on condensed tannins in the
defoliated plots. Why fertilizer did not decrease phenolic concentrations in 1997 as it did
in other years is not clear. One possibility is that very high levels of defoliation and/or
input of large amounts of nitrogen-rich frass in 1997 may have obscured any effects of
fertilization. Alternatively, the near complete loss of leaf tissue could have reduced

photosynthesis so much that carbon also became a limiting factor. There was no

129

difference in tree growth between unfertilized and fertilized trees in the defoliated plots
in 1997 suggesting that nitrogen was not limiting growth in contrast to the growth

differences in 1998 when defoliation was less severe.

In 1997, levels of condensed tannins in the refoliated leaves were lower than foliage
from the control plots. During refoliation, defoliated trees must allocate substantial
amounts of carbon to producing a new set of leaves, which are major carbon sinks.
Furthermore, trees must also replenish the depleted carbohydrate stores prior to autumn
leaf fall. Both of these processes are likely to take precedence over the production of
carbon-demanding phenolics leading to lower concentrations of carbon-based secondary
compounds such as condensed tannins. In marked contrast, total phenolics and
condensed tannins were signiﬁcantly higher in late season leaf samples taken from plots
defoliated earlier in the season in 1998. This pattern may be due to lower defoliation
levels and the retention of damaged leaves rather than a complete reﬂush as was seen in
1997. Thus the levels of condensed tannins in the late season samples in 1998 were
similar to those taken at mid-season. Schultz and Baldwin (1982) and Faeth (1988) also
found lower levels of condensed tannins in reﬂush leaves following complete defoliation
of two different species of oak. As in my study, partially damaged oak leaves had higher
levels of condensed tannins than either undamaged controls or refoliated leaves in late

season samples.

The six minor phenolic compounds measured in my study had variable responses to
defoliation and fertilization. In 1997, defoliation resulted in a signiﬁcant increase in
phenolic glycoside concentrations but this did not occur in 1998. In aspen, phenolic

glycoside response to defoliation has also been less predictable than condensed tannins

130

with some studies showing short-term increases in phenolic glycosides while others did
not (Clausen et al. 1991, Roth et al. 1998, Lindroth and Kinney 1998). Phenolic
glycosides play a role in very rapid responses and may rapidly turnover (Clausen et al.
1989, 1991), making detection of changes highly dependent on the timing of sampling.
Low levels of damage in hybrid poplar has been shown to induce rapid increases in total
phenolics but the effects disappeared within ﬁve days (Baldwin and Schultz 1983),
indicating that unless there is subsequent damage to reinforce the response, the induction
is ephemeral. This was particularly evident in one control plot in 1997 where signiﬁcant
numbers of gypsy moth circumvented the sticky trap barriers. Despite relatively low
herbivory levels, at least compared to the defoliation plots, total phenolics increased
dramatically in foliage from this plot and were as high or higher than in the defoliation
plots. Yet after the removal of the caterpillars, the levels of total phenolics had returned
to control levels in samples taken one month later. At least with respect to fertilizer and
light availability, phenolic glycosides do not respond as predicted by the CNB (Hemming
and Lindroth 1999). In my study, fertilization resulted in increased phenolic glycoside
levels in July leaves, opposing predictions of the CNB. On the other hand, I found that
fertilization reduced levels of quercetin and myricetin in June 1997, as predicted by CNB
although concentrations of the other minor phenolic compounds did not vary among

treatments.

Processes driving DIR have received much more attention in trees than those
responsible for RIR. Reduced nitrogen levels in the year or years following defoliation
are thought to underlie DIR in deciduous trees (Tuomi et al. 1984, 1990). In forests

growing in nutrient poor soil, the recovery of foliar nitrogen to predefoliation levels may

131

take several years as shown for mountain birch (Tuomi et al. 1984). Shorter recovery
times have been recorded for defoliated birches in Alaska (Bryant et al. 1991) and
willows in Finland (Tuomi et al. 1991), although two years were still required for
nitrogen to return to control levels. Reductions in nitrogen in the year following
defoliation have also been shown in black oak (Quercus velutina Lambert), gray birch
(Betula populifolia Marshall), and red alder (Alnus rubra Bong) (Valentine et al 1983,
Myers and Williams 1987). In contrast, I found that foliar nitrogen in previously
defoliated poplars, while slightly depressed was not statistically different from levels in
control leaves. This pattern was evident in May 1998 and June 1999 even though late
season leaf samples from the previous year had markedly reduced nitrogen levels relative
to control foliage. These data indicate that the poplars were able to rapidly replenish
nitrogen pools prior to the start of the following growing season, which has not

previously been shown for any tree species.

The rapid recovery of foliar nitrogen levels in the defoliated poplars is somewhat
puzzling given that their rate of ammonium and nitrate uptake was only 33% of that in
control trees even though there was little evidence of ﬁne root mortality (Kosola et al.
2000). The management practices in this poplar plantation may provide a possible
explanation for this pattern. Application of herbicide in the spring and summer to control
understory vegetation resulted in essentially bare soil beneath the trees, in marked
contrast to many natural forest systems that have extensive woody and herbaceous
undergrowth. Thus, following decomposition, nitrogen in frass, greenfall, insect exuviae,
and cadavers is available for uptake by trees free of competition from other vegetation.

In contrast, nitrogen lost through defoliation may not be recovered by trees in natural

132

forests because competing herbaceous and woody plants in the understory intercept and
utilize it. Furthermore, the poplars were growing in ﬁelds formerly used in a corn,
alfalfa, and soybean rotation, and were undoubtedly richer in nitrogen than soils in boreal
forests and subarctic soils where the strongest DIR responses have been observed. The
tree growth data mirror the changes in foliar nitrogen levels. Despite reductions in
diameter growth of ca. 11%, 50%, and 50% in 1996, 1997, and 1998 respectively, there
was no difference in growth among previously defoliated and control trees in 1999. My
results are congruent with the prediction of Tuomi et al. (1990) that recovery of foliar
nitrogen to predefoliation levels would be most rapid for trees growing in nutritionally

rich environments.

In contrast to the ephemeral effects on foliar nitrogen, strong across-year effects of
defoliation on levels of total phenolics and condensed tannin levels were recorded. This
was most pronounced in 1999, where condensed tannins remained 16% and 28% higher
than control leaves in the previously unfertilized and fertilized defoliated trees,
respectively. Similar differences in condensed tannin levels (10% and 22%) were evident
in foliage sampled from the defoliated plots in May 1998. Since herbivory was minimal
at the time of sampling, these differences reﬂect effects of treatments from the previous
year. These results are not unexpected given that the return of phenolic concentration to
levels found in control trees generally lags behind nitrogen by at least a year (Tuomi et al.

1991).

Variability in the induced responses of phenolic compounds to defoliation and
fertilization both within and between years suggests that more than one mechanism may

be governing the response. While condensed. tannins responded as predicted to

133

defoliation and fertilization, other phenolic compounds did not. Outside of the phenolic
glycosides, little is known about the majority of the other minor phenolic compounds
identiﬁed from poplar foliage. It is possible that some of these biochemicals are involved
in wound-speciﬁc responses and thus are unaffected by alterations in the nutritional
environment of the tree. Tuomi et al. (1991) suggested that non-speciﬁc chemical
changes caused by alterations of the carbon/nutrient balance and speciﬁc wound-induced
resistance mechanisms may both be components of plant resistance to herbivory.
Characterization of wound-speciﬁc responses in trees lags behind analogous research
with herbaceous plants. A full understanding of induced-responses in trees will not be
possible until the relationship between both nutritionally driven and wound speciﬁc

responses has been identiﬁed.

In a seeming paradox, I view the results of my experiment as complimentary rather
than contradictory to other studies on DIR. My study suggests that rather than a general
response, the outcome of experiments is system speciﬁc and highly contingent upon
environment. Tuomi et al. (1984, 1990) suggested that expression of DIR would be
greatest in marginal habitats deﬁcient in nitrogen. Because my experimental system is at
the other end of a fertility continuum, a signiﬁcant DIR response should not be expected.
I note that in two well-designed studies where DIR responses were not detected (Harrison
1995, Rothman 1997), both were conducted on nitrogen ﬁxing species. In these trees,
nitrogen depletion is unlikely to occur unless defoliation occurred repeatedly for many

years; thus any expression of DIR should be negligible.

Indirect effects of induced resistance on herbivores - Indirect interactions between

induced resistance and the natural enemies of caterpillars are increasingly recognized as

134

important. For example, the induction of secondary compounds such as phenolics may
interact with pathogens to either enhance or reduce mortality of the insect host. In a
laboratory study, gypsy moth mortality from nuclear polyhedrosis virus (NPV) was
negatively correlated with hydrolyzable tannin concentrations in oak leaves induced by
defoliation (Hunter and Schultz 1993), presumably because tamiins precipitate viral
proteins, interfering with the infection process (Keating et al. 1990). Hunter and Schultz
(1993) suggested that the interaction between secondary phenolics and pathogens could
prolong gypsy moth outbreaks in oak-dominated forests. In ﬁeld studies however,
D’Amico et al. (1998) found no evidence that mortality of gypsy moth from NPV was
reduced by defoliation on two species of oak. They attributed their ﬁndings to the lack of
induction of tannins during the critical time period early in the season when most viral

transmission between larvae occurs.

An aspect of the relationship between pathogens and tannins that has been
hypothesized (e.g., Herms and Mattson 1992), but not explored in detail, is the effect of
increased consumption rate on susceptibility to pathogen infection. Elevated consumption
rates by gypsy moth and other species such as forest tent caterpillar are likely a
generalized response to low nutritive levels in leaves (e.g., Slansky and Feeny 1977,
Stockhoff 1992, Lindroth et al. 1997, Williams et al. 1998, Parry et al. 2000). If tannins
interact negatively with NPV, decreases in susceptibility may be mitigated if the
caterpillars simultaneously increase consumption rates on foliage with elevated
phenolics, thus ingesting greater numbers of viral propagules and resulting in no net
change in mortality. Such effects are unlikely in poplars as their condensed tannins only

weakly reduce NPV mortality (Lindroth et al. 1999). The effects of DIR on pathogens has

135

received little attention despite Wallner’s (1983) study showing increased gypsy moth
susceptibility to NPV when reared on foliage from previously defoliated oak trees.
Disparate effects of tannins on insect pathogens documented in previous studies suggest
that the relationship between primary nutrients, inducible secondary allelochemicals, and

pathogens is not yet understood and should be explored further.

Until recently, it was generally thought that herbivorous insect communities were not
structured by competition (e.g., Hairston et al. 1960, Schoener 1983, Strong et al. 1984).
Evidence has now accumulated suggesting that these earlier views were not entirely
accurate. A comprehensive review found that interspeciﬁc competition between
herbivorous insects occurred in more than 75% of the cases examined (Demio et al.
1995). Among chewing folivores, plant-mediated competitive interactions such as those
driven by induced-resistance accounted for more than half of reported negative
interactions between species. Early season defoliation may impact folivorous insect
communities in two ways. First, the outbreak species competes directly with other species
for foliage. Secondly, the feeding activity of an outbreak species may trigger induced-
resistance that reduces the quality of remaining foliage for co-occurring species. Changes
in host plant quality due to defoliation may persist thus affecting herbivores feeding later
in the season. For example, defoliation of paper birch, sugar maple, and aspen by forest
tent caterpillars induced both RIR and DIR, reducing the performance of gypsy moth
larvae early in the season as well. as the tiger swallowtail and second generation white-
marked tussock moth feeding later in the year (Dankert et al. 1997). Similarly, Harrison
and Karban (1986) found that early season feeding by an arctiid caterpillar on bush lupine

reduced the quality of foliage for a tussock moth that initiated feeding later in the

136

growing season. Early season damage to oak leaves reduces the probability of herbivory
later in the season (Hunter and Schultz 1995, Wold and Marquis 1997) although Faeth
(1988) found that the effects on late season species depended on both the severity of the

previous herbivory and the species examined.

Effects of induction may be perceived differently among different species of
herbivores. There are two sides to the speciﬁcity of induction. Plants may perceive and
react to herbivory caused by different species differently, and induced responses elicited
by a single species may have variable effects on different species of herbivore (Agrawal
1999, 2000). Dietary specialization can mediate the outcome of the effects of induced
responses on herbivores (Hartley and Lawton 1987, Bowers and Stamp 1993, Agrawal
and Karban 1999). I predicted that the polyphagous and outbreak species in the study
would be buffered to changes in poplar quality after defoliation, because substantive
variation in the nutritional quality and biochemical proﬁle of host plants is incumbent to
these life-history strategies. Contrary to expectation, the pupal mass and/or fecundity of
gypsy moth, forest tent caterpillar, poplar tent maker, and big poplar sphinx were reduced
by similar magnitudes in 1998. Thus, uniform responses to changes in foliage occurred
despite the disparate life histories represented among these species. One explanation is
that all species were responding to changes in nitrogen rather than an induced secondary
compound. Similar interspeciﬁc responses to a primary nutrient such as nitrogen is more
probable than uniform responses to secondary compounds that may require the evolution
of speciﬁc detoxiﬁcation mechanisms. In contrast to my study, Agrawal (2000) found
considerable variation in the effects of induced physical and biochemical resistance in

wild radish (Raphanus saliva) among a guild of foliage-feeding caterpillars. The

137

differences in performance among species could not predicted by the degree of dietary

specialization.

The only species not affected by my defoliation treatments in 1998 was fall
webworm. This species is extremely polyphagous suggesting that it has evolved
mechanisms to process leaves with wide variations in secondary compounds, toughness,
water, and nutrient concentrations. Williams and Myers (1984) concluded that fall
webworm beneﬁts from early season defoliation, although they found no differences in
pupal mass of larvae reared on undefoliated and previously defoliated trees at one site
and only a 12 mg increase at a second site. Their interpretation is even more problematic
because sexes in this dimorphic species were not analyzed separately and the insects
rather than the experimenters selected the trees. Rather than improving foliar quality, I
suggest that their results concur with mine: early season defoliation has no biologically

meaningful effect on fall webwomi performance.

I measured only the indirect effects of gypsy moth defoliation on other lepidopterans
in the poplar-feeding community. In natural outbreaks, direct competition for foliage
undoubtedly compounds the indirect effects on. the folivorous community. Displacement
from preferred hosts and habitats by defoliation may be associated with higher mortality,
especially if the species relies on crypsis to avoid predation or parasitism (Heinrich
1993). The protective structures of leaf rolling and leaf mining species can be
compromised by the feeding activities of defoliators. Hunter and Willmer (1989) showed
that the free-feeding winter moth had a competitive advantage over T ortrix viridana
when they co-occurred on oak because winter moth damaged the leaf-rolls essential for

osmoregulation of tortrix. In several. studies, leafminer survival was negatively correlated

138

with free-feeding caterpillar damage (Faeth 1988, West 1985). Gypsy moth defoliation
has been shown to reduce abundance and/or species diversity of other lepidopterans in
oak forests (Sample et al. 1996, Work and McCullough 2000), although the effects are
relatively minor. It is unclear if indirect or direct competitive interactions, or some other

process drive these changes.

Simulation models have suggested that induced resistance can regulate herbivore
populations and drive cyclic dynamics without the inﬂuence of other density-dependent
factors (Edelstein-Keshet and Rausher 1989, Underwood 1999). In Underwood’s (1999)
model, whether populations were regulated or cycled depended on both the decay rate of
the resistance and the lag time to maximum strength of the induced-response. In forest
systems at least, induced-responses appear unlikely to be capable of regulating insect
populations by themselves. Only one study has looked at the effects of induced-resistance
on the population growth rate of a defoliator. In the absence of natural enemies, Haukioja
et a1. (1988) calculated the per capita growth rate, r, for autumnal moth feeding on
undefoliated trees to be 2.9 compared with 1.1 on trees defoliated the previous year.
Although DIR decreased population growth rate signiﬁcantly, by itself, it was not capable
of preventing this population from growing. In natural systems, DIR interacts with other
biotic and abiotic factors and its effects may be sufﬁcient to slow population growth to
the point where other time-lagged density-dependent agents such as pathogens or

parasitoids can drive populations down.

Considerable knowledge has accumulated on the effects of DIR on herbivorous
insects, albeit in relatively few systems. Lacking, however, are synthetic studies where

the effects of DIR are measured relative to the contributions of other top-down and

139

 

bottom up regulating factors. Bylund (1995) attempted such a study using primarily
historical data and demonstrated that delayed density-dependent parasitism was the most
important factor correlated with ﬂuctuations of autumnal moth populations in Sweden.
Other factors including depletion of food, low winter temperatures, and induced-
resistance contributed to population regulation. Over a three-year period, Virtanen and
Neuvonen (1999) estimated the relative contribution of altitude, climate, host quality, and
parasitoids on autumnal moth in Finland, and like Bylund (1995), concluded that
parasitoids were more important than host plant quality in determining population
dynamics. Although not looking explicitly at induced-resistance, Hunter et al. (1997)
used time series analysis and data from supplementary experiments to determine the
relative contributions of top-down and bottom-up forces to the population dynamics of
two herbivores on English oak. This approach could be adapted for use with long—term
data sets from other forest insects to look at the relative contribution of induced-
responses and natural enemies to population dynamics. Unfortunately, in many of the
systems where induced-resistance has been intensively studied, there is a dearth of
information on other potential regulating factors. For example, the paper birch-black-
marked spear moth system and the large aspen tortrix-trembling aspen system have been
subjected to detailed research on induced-resistance mechanisms in Alaska, yet lacking
entirely from these systems is any quantitative estimates of the impact of natural enemies.
Given the intractability of studying outbreak species over the decades required to collect
data on natural cycles, manipulative studies such as mine may be a useful tool to address
these questions, especially if interactions with the third trophic level are incorporated into

the experimental design.

140

 

 

 

ACKNOWLEDGEMENTS

I am extremely grateful for the invaluable help with the chemical analyses provided by
Bruce Birr and Bill Mattson (USDA Forest Service, Rhinelander, WI). Muralee Nair and
Russell Ramsewak (Natural Products, Michigan State University) endured my constant
requests for the use of their lyophilizer, and Phil Robertson’s laboratory (KBS) donated
equipment, people, and vehicles to assist with the project. Many LTER and Kellogg
Biological Station employees helped with ‘bug-wrangling’, sampling, equipment loans,
and rearing of insects. Thanks to Sandy Halstead (KBS) for organizing crews and
providing help on short notice. I are especially grateful to C. Miller, E. Duncan, C.
Easley, M. Fisher, J. Edwards, S. Sanders, L. White, and C. Schoen who put in long and
sometimes nearly thankless hours on various aspects this project. This research was
funded by NSF grant DEB 95-10044, and the following grants to D.P.: Graduate
Research Training Group at the Kellogg Biological Station (NSF grant DIR-9602252), a
Hutson Grant, Department of Entomology, Michigan State University, Sigma Xi Grant-
in-Aid of Research, and a Dissertation. Completion Fellowship, The Graduate School,
Michigan State University. Summer salary support in 2000 was provided by the

Michigan Agricultural Experiment Station (Project # MICL 01644 to J. Mark Scriber)

LITERATURE CITED
Agrawal, AA. 2000. Speciﬁcity of induced-resistance in wild radish: causes and
consequences for two specialist and two generalist caterpillars. Oikos 89: 493-

500.

141

Agrawal, AA. 1999. Induced responses to herbivory in wild radish: effects on several
herbivores and plant ﬁtness. Ecology 80: 1713-1723.

Agrawal, AA. and R. Karban. 1999. Why induced defenses may be favored over
constitutive strategies in plants. Pp. 45-61 in R. Tollrian and CD. Harvell, eds,
The Ecology and Evolution of Inducible Defenses. Princeton University Press,
Princeton, New Jersey.

Ayres, M.P., T.P. Clausen, S.F. MacLean, Jr., A.M. Redman and PB. Reichardt. 1997.
Diversity of structure and. antiherbivore activity in condensed tannins. Ecology
78: 1696-1712.

Baldwin, LT. and J .C. Schultz. 1983. Rapid—changes in tree leaf chemistry induced by
damage: evidence for communication between plants. Science 221: 277-279.

Baltensweiler, W. and A. Fischlin. 1988. The larch budmoth in the Alps. Pp. 331-351 in
AA. Berryman, ed, Dynamics of Forest Insect Populations. Patterns, Causes, and
Implications. Plenum Press, New York.

Bate-Smith, EC. and V. Rasper. 1969. Tannins of grain sorghum: luteforol
(leucoluteolinidin) 1’,4,4’5,7-pentahydroxyﬂavan. Journal of Food Science 34:
203-209.

Bae, H.-E., T.A. McAllister, A.D. Muir, L.J. Yanke, K.A. Bassendowski and K.J.
Cheng.1993. Selection of a method of condensed tannin analysis for studies with
rumen bacteria. Journal of Agricultural and Food Chemistry. 41: 1256-1260.

Bell, R.A., C.D. Owens, M. Shapiro and J .R. Tardif. 1981. Development of mass-rearing

technology. Pp. 599-633 in CC. Doane and ML. McManus, eds, The Gypsy

142

Moth: Research Toward Integrated Pest Management. United States Department
of Agriculture Technical Bulletin 1584.

Benz, G. 1974. Negative feedback by competition for food and space, and by cyclic
induced changes in the nutritional base as regulatory principles in the population
dynamics of the larch budmoth, Zieraphera dineana (Guenee) (Lepidoptera:
Tortricidae). Zietschrift fur Angewandte Entomologie 76: 196-228.

Bergey, D.R., G.A. Hoi and CA. Ryan. 1996. Polypeptide signaling for plant defensive
genes exhibits analogies to defense signaling in animals. Proceedings of the
National Academy of Science 93: 12053-12058

Berryman, A.A., N.C. Stenseth and AS. Isaev. 1987. Natural regulation of herbivorous
forest insect populations. 0ecologia 71: 174-184.

Berryman, AA. 1996. What causes population cycles of forest Lepidoptera? Trends in
Ecology and Evolution 11: 28-32.

Bowers, MD. and NE. Stamp. 1993. Effects of plant age, genotype, and herbivory on
Plantago performance and chemistry. Ecology 74: 1778-1791.

Bryant, J.P., F.S. Chapin, D.R. Klein. 1983. Carbon/nutrient balance of boreal plants in
realtion to vertebrate herbivory. Oikos 40: 357-368.

Bryant, J.P. J. Tuomi and P. Niemela. 1988. Environmental constraints of constitutive
and long-term inducible defenses in woody plants. Pp. 367-389 in K.G. Spencer,
ed, Chemical Mediation of Coevolution. Academic Press, San Diego, CA.

Bryant, J.P., K. Danell, F. Provenza, P.B. Riechardt, T.A. Clausen and RA. Werner.

1991. Effects of mammal browsing on the chemistry of deciduous woody plants.

143

Pp. 135-154 in D.W. Tallamy and M.J. Raupp, eds, Phytochemical Induction by
Herbivores. John Wiley & Sons, Inc., New York.

Bryant, J.P., P.B. Reichardt, T.P. Clausen and R.A.Wemer. 1993. Effects of mineral
nutrition on delayed inducible resistance in Alaska paper birch. Ecology 74: 2072-
2084.

Bryant J .P., T.P. Clausen, P.B. Reichardt, M.C. McCarthy and R.A.Wemer 1987. Effects
of nitrogen fertilization upon the secondary chemistry and nutritional value of
quaking aspen (Populus tremuloides Michx.) leaves for the large aspen tortrix
(Choristoneura conflictana (Walker)). 0ecologia 73: 513-517.

Bylund, H. 1995. Long-term interactions between the autumnal moth and mountain birch:
the role of resources, competitors, natural enemies, and weather. PhD
Dissertation. Swedish University of Agricultural Sciences, Uppsala, Sweden.

Chapin, F .S. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and
Systematics 11: 233-260.

Chapin, FS and RA. Kedrowski. 1983. Seasonal changes in nitrogen and phosphorus
fractions and autumn retranslocation in evergreen and deciduous taiga trees.
Ecology 64: 376-391.

Clausen, T.P., P.B. Reichardt, J .P. Bryant and RA. Werner. 1991. Long-term and short-
term induction in quaking aspen: related phenomena? Pp. 71—83 in D.W. Tallamy
and M.J. Raupp, eds, Phytochemical Induction by Herbivores. John Wiley &

Sons, Inc., New York.

144

Constabel, P. 1999. A survey of herbivore inducible proteins and phytochemicals. Pp.
137-166 in AA. Agrawal, S. Tuzin and E. Bent, eds, Induced plant defenses
against Pathogens and herbivores. APS Press, St. Paul, Minnesota.

D’Amico, V., J .S. Elkinton, G. Dwyer, R.B. Willis and ME. Montgomery. 1998. Foliage
damage does not affect within-season transmission of an insect virus. Ecology 79:
1 104-1 1 10.

Dankert, B.A., D.A. Herms, D. Parry, J.M. Scriber and LA. Haas. 1997. Mediation of
interspeciﬁc competition between folivores through defoliation induced changes
in host quality. Pp. 71-88 in A. Raman, ed, Ecology and Evolution of Plant
Feeding Insects in Natural and Man-Made Environments. International Scientiﬁc
Publications. New Delhi, India

Denno, R.F., MS. McClure and JR. Ott. 1995. Interspeciﬁc interactions in phytophagous
insects: competition revisited and resurrected. Annual Review of Entomology 40:
297-331.

Diss A.L, J.G. Kunkel, ME. Montgomery and DE. Leonard. 1996. Effects of maternal
nutrition and egg provisioning on parameters of larval hatch, survival and
dispersal in the gypsy moth, Lymantria dispar L. 0ecologia 106: 470-477.

Edelstein-Keshet, L. and MD. Rausher. 1989. The effects of inducible plant defenses on
herbivore populations. 1. Mobile herbivores in continuous time. The American
Naturalist 133: 787-810.

Faeth, SH. 1986. Indirect interactions between temporally-separated herbivores mediated

by the host plant. Ecology 67 :479-494.

145

Faeth SH. (1988). Plant mediated interactions between seasonal herbivores: Enough for
evolution or coevolution. Pp. 391—414 in D.W. Tallamy and M.J. Raupp, eds,
Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Feeny, PP. 1970. Changes in oak leaf tannins and nutrients as a cause of spring feeding
by winter moth caterpillars. Ecology 51: 565-581.

Ginzburg, LR. and DE. Taneyhill. 1994. Population-cycles of forest Lepidoptera — a
maternal effects hypothesis. Journal of Animal Ecology 63: 79-92.

Gordon, HT. 1968. Quantitative aspects of insect nutrition. Zoology 8: 131-138.

Hairston, N.G., FE. Smith and LB. Slobodkin. 1960. Community structure, population
control and competition. The American Naturalist 115: 354-366.

Harper, J.L. 1977. Population Biology of Plants. Academic Press, San Diego.

Harrison, S. 1995. Lack of strong induced or maternal effects in tussock moths (Orgyia
vetusta) on bush lupine (Lupinus arboreus). Oecologia 103: 343-348.

Harrison, S. and R. Karban. 1986. Effects of an early season folivorous moth on the
success of a later-season species mediated by a change in the quality of a shared
host, Lupinus arboreus. Oecologia 69: 354-359.

Hartley, SE. and J.H. Lawton. 1987. Effects of different types of damage on the
chemistry of birch foliage and the response of birch feeding insects. 0ecologia
74: 423-437.

Haukioja, E. 1980. On the role of plant defenses in the ﬂuctuation of herbivore

populations. Oikos 35: 202-213.

146

Haukioja, E. 1982. Inducible defenses of white birch to a geometrid defoliator, Epirrita
autumnata. Proceedings of the 5th International Symposium on Insect-Plant
Relationships, Wageningen.

Haukioja, E. and P. Niemela. 1977. Retarded growth of a geometrid larvae after
mechanical damage of leaves of its host tree. Annales Zoologici Fennici 14: 48-
52.

Haukioja, E. and S. Neuvonen. 1985. Induced-long term resistance of birch foliage
against defoliators: defensive or incidental. Ecology 66: 1303-1308.

Haukioja, E. and S. Neuvonen. 1987. Insect population dynamics and the induction of
plant resistance: the testing of hypotheses. Pp. 411-432 in P. Barbosa and J .C.
Schultz, eds, Insect Outbreaks. Academic Press, San Diego.

Haukioja, H., P. Niemela and S. Sirena. 1985. Foliage phenols and nitrogen in relation to
growth, insect damage, and the ability to recover after defoliation, in the mountain
birch Betula pubescens ssp. tortuosa. 0ecologia 65:214-222.

Haukioja, E., S. Neuvonen, S. Hanhimaki and P. Niemela. 1988. The autumnal moth in
Fennoscandia. Pp. 163-178 in AA. Berryman, ed, Dynamics of Forest Insect
Populations. Patterns, Causes, and Implications. Plenum, New York.

Havill, NP. and K.F. Raffa. 1999. Effects of elicitation treatment and genotypic variation
on induced resistance in Populus: impacts on gypsy moth (Lepidoptera:
Lymantriidae) development and feeding behavior. 0ecologia 120: 295-305.

Heinrich, B. 1993. How avian predators constrain caterpillar foraging. Pp. 224-247 in N.
E. Stamp and T. M. Casey, eds, Caterpillars: Ecological and Evolutionary

Constraints on Foraging. Chapman and Hall, New York.

147

 

 

Herms, DA. and W.J. Mattson. 1992. The dilemma of plants — to grow or defend.
Quarterly Review of Biology 67: 283-335.

Hemming J .D.C., R.L. Lindroth. 1995 . Intraspeciﬁc variation in aspen phytochemistry —
effects on performance of gypsy moths and forest tent caterpillars. 0ecologia 103:
79-88.

Hunter, AF. 1995. Ecology, life—history, and phylogeny of outbreak and non-outbreak
species. Pp. 41-64 in N. Cappuccino and PW. Price, eds, Population Dynamics:
New Approaches and Synthesis. Academic Press, New York.

Hunter, M.D., and PG. Willmer. 1989. The potential for interspeciﬁc competition
between 2 abundant defoliators on oak — leaf damage and habitat quality.
Ecological Entomology 14: 267-277.

Hunter, MD. and J .C. Schultz. 1993. Induced plant defenses breached? Phytochemical
induction protects an herbivore from disease. 0ecologia 94: 195—203.

Hunter, MD. and J.C. Schultz. 1995. Fertilization mitigates chemical induction and
herbivore responses within damaged. oak trees. Ecology 76: 1226-1232.

Hunter MD, G.C. Varley and GR. Gradwell. 1997. Estimating the relative roles of top-
down and bottom-up forces on insect herbivore populations: A classic study
revisited. Proceedings of the National Academy of Science 94: 9176-9181.

Hwang, S.Y. and R.L. Lindroth. 1997. Clonal variation in foliar chemistry of aspen:
effects on gypsy moths and forest tent caterpillars. 0ecologia 111: 99-108.

Julkunen-Tiitto, R, M. Rousi, J.P. Bryant, S. Sorsa, M. Keinanen and H. Sikanen. 1996.
Chemical diversity of several Betulaceae species: Comparison of phenolics and

terpenoids in northern birch stems. T recs-Structure and Function 11:16-22.

148

Kaitaniemi, P. K. Ruohomaki, V. Ossipov, E. Haukioja and K. Pihlaja. 1998. Delayed
induced changes in the biochemical composition of host plant leaves during an
insect outbreak. 0ecologia 116: 182-190.

Kaitaniemi, P., K. Ruohomaki, T. Tammaru and E. Haukioja. 1999a. Induced resistance
of host tree foliage during and after natural insect outbreak. Journal of Animal
Ecology 68: 382-389.

Kaitaniemi, P., S. Neuvonen and T. Nyyssonen.l999b. Effects of cumulative defoliations
on growth, reproduction, and insect resistance in mountain birch. Ecology 80:
524-532.

Karban, R. and LT. Baldwin. 1997. Induced Responses to Herbivory. University of
Chicago Press, Chicago, Illinois.

Kause A, V. Ossipov, E. Haukioja, K. Lempa, S. Hanhimaki and S. Ossipova. 1999.
Multiplicity of biochemical factors determining quality of growing birch leaves.
0ecologia 120:102-112.

Keating, S.T., M.D. Hunter and J.C. Schultz. 1990. Leaf phenolic inhibition of gypsy
moth nuclear polyhedrosis virus: role of polyhedral inclusion body aggregation.
Journal ofChemical Ecology 16: 1445-1457.

Koricheva J, S. Larsson, E. Haukioja and M. Keinanen. 1998. Regulation of woody plant
secondary metabolism by resource availability: hypothesis testing by means of
meta-analysis. Oikos 83: 212—226.

Kosola, K.R., D.I. Dickmann, E.A. Paul and D. Parry, 2000. Repeated insect defoliation
effects on growth, nitrogen acquisition, carbohydrates and root demography of

poplars. 0ecologia. In Press.

149

 

Krause, SC. and K.F. Raffa. 1992. Comparison of insect, fungal, and mechanically
induced defoliation of larch — effects on plant productivity and subsequent host
susceptibility. 0ecologia 90: 411-416.

Lance, D.R., J.S. Elkinton and GP. Schwalbe. 1986. Feeding rhythms of gypsy moth
larvae: effect of food quality during outbreaks. Ecology 67: 1650-1654.

Lance, D.R., J .S. Elkinton and GP. Schwalbe. 1991. Responses of gypsy moth larvae
(Lepidoptera: Lymantriidae) to foliage of oaks from naturally infested sites on
Cape Cod, Massachusetts. Journal of Entomological Science 26: 214-222.

Liebhold, A.M., K.W. Gottschalk. R.-M. Muzika, ME. Montgomery, R. Young, K.
O’Day and B. Kelley. 1995. Suitability of North American tree species to the
gypsy moth: A summary of ﬁeld and laboratory tests. United States Department
of Agriculture, Forest Service. Northeast Forest Experimental Station General
Technical Report NE-211.

Lindroth, R.L. and MS. Bloomer. 1991. The chemical ecology of the forest tent
caterpillar: responses to dietary protein and phenolic glycosides. 0ecologia 86:
408-413.

Lindroth, R.L. and S.-Y. Hwang. 1996. Diversity, redundancy, and multiplicity in
chemical defense systems of aspen. ), Pp. 25-56 J .T. Romeo, J .A. Saunders and P.
Barbosa, eds, Phytochemical Diversity and Redundancy in Ecological
Interactions. Plenum Press, New York.

Lindroth, R.L. and RA. Koss. 1996. Preservation of Salicaceae leaves for phytochemical

analyses: Further assessment. Journal of Chemical Ecology 22:765-772.

150

Lindroth, R.L. and MS. Pajutee. 1987. Chemical analysis of phenolic glycosides: art,
facts, and artifacts. 0ecologia 74: 144-148.

Lindroth, R.L and K.K. Kinney. 1998. Consequences of enriched atmospheric CO2 and
defoliation for foliar chemistry and gypsy moth performance. Journal of
Chemical Ecology 24: 1677-1695.

Lindroth, R.L., K.A. Klein, J.D.C. Hemming and A.M. Feuker. 1997. Variation in
temperature and dietary nitrogen affect performance of gypsy moth (Lymantria
dispar L.). Physiological Entomology 22: 55-64.

Lindroth, R.L., S. Hwang and TL. Osier. 1999. Phytochemical variation in quaking
aspen: effects on gypsy moth susceptibility to nuclear polyhedrosis virus. Journal
ofChemical Ecology 25: 1331-1341.

Marino, PC. and KL. Gross. 1988. Competitive effects of conspeciﬁc and herbaceous
plants (weeds) on growth and branch architecture of Populus x euramericana cv.
Eugenei. Canadian Journal of Forest Research 28: 359-367.

Mason, OJ. and ML. McManus, 1981. Larval dispersal of the gypsy moth. Pp. 161-202
in CC. Doane and ML. McManus, eds, The Gypsy Moth: Research Toward
Integrated Pest Management. United States Department of Agriculture Technical
Bulletin 1584.

Mattson, W.J. 1980. Herbivory in relation to plant nitrogen-content. Annual Review of
Ecology and Systematics 11: 119-161.

Mattson, W.J. and ND. Addy. 1975. Phytophagous insects as regulators of forest primary

production. Science 190: 515-522.

151

Mattson, W.J., D.A. Herms, J.A. Witter and DC. Allen. 1991. Woody plant grazing
systems: North American outbreak folivores and their host plants. Pp. 53-84 in
Y.N. Baranchikov, W.J. Mattson, F.P. Hain and TL. Payne, eds, Forest Insect
Guilds: Patterns of Interactions with Host Trees. United States Department of
Agriculture, Forest Service. General Technical Report NE-153.

May, RM. 1973. Stability and Complexity in Model Ecosystems. Princeton University
Press, Princeton, New Jersey.

Meyer, GA. and ME. Montgomery (1987). Relationships between leaf age and the food
quality of cottonwood foliage for the gypsy moth, Lymantria dispar. 0ecologia
72: 527-532.

Myers, J .H. and KS. Williams. 1987. Lack of short or long term inducible defenses in the
red alder- western tent caterpillar system. 0ecologia 48: 73-78.

Nef, L. 1988. Interactions between the leaf miner, Phyllocnitis suﬂusella, and poplars.
Pp. 239-251 in W.J. Mattson, J. Levieux and C. Bemard—Dagan, eds, Mechanisms
of Woody Plant Defenses against Insects. Springer, New York.

Neuvonen, S. and E. Haukioja. 1991. The effects of inducible resistance in host foliage
on birch-feeding herbivores. Pp. 277-291 in D.W. Tallamy and M.J. Raupp, eds,
Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Nitao, J.K, B.A. Birr, M.G. Nair, D.A. Herms, and W.J. Mattson. 2000. Rapid
quantiﬁcation of condensed tannins with a continuous ﬂow analyzer. Journal of
Agricultural and Food Chemistry. In review.

Parry, D. W.J.A. Volney and CR. Currie. 1997. The relationship between trembling

aspen phenology and larval development of the large aspen tortrix. Natural

152

Resources Canada, Canadian Forest Service, Northwest Region, Northern
Forestry Centre NOR-X—350. 14 pp.

Parry, D., D.A. Herms and W.J. Mattson. 2000. Induced-responses of aspen to multiple,
consecutive defoliations: their effects on an outbreak folivore and its natural
enemies. Journal of Animal Ecology. Submitted.

Reynolds, BC. and DA. Crossley. 1997. Spatial variation in herbivory by forest canopy
arthropods along an elevation gradient. Environmental Entomology 26: 1232—
1239.

Rhoades, DF. 1983. Herbivore population dynamics and plant chemistry. Pp. 155-220 in .
R.F. Denno and MS. McClure, eds, Variable plants and Herbivores in Natural
and Managed Systems. Academic Press, San Diego.

Robison, DJ. and K.F. Raffa.l997. Effects of constitutive and inducible plant traits of
hybrid poplars on forest tent caterpillar feeding and population ecology. Forest
Science 43: 252-267.

Roland, J. and J .H. Myers 1987. Improved insect performance from host-plant defoliation
— winter moth on oak and apple. Ecological Entomology 12: 409-414.

Rossiter, M.-C., J .C. Schultz, I.T. Baldwin. 1988. Relationships among defoliation, red
oak phenolics and gypsy moth growth and reproduction. Ecology 69: 267-277.

Roth, S., R.L. Lindroth, J.C. Volin and EL. Kruger. 1998. Enriched atmospheric CO2
and defoliation: effects on tree chemistry and insect performance. Global Change
Biology 4: 419-430.

Rothman, L.D. 1997. Immediate and delayed effects of a viral pathogen and density on

tent caterpillar performance. Ecology 7 8: 1481-1493.

153

Ruohomaki, K., S. Hanhimaki, E. Haukioja, L. Iso-Iivari, S. Neuvonen, P. Niemela and
S. Suomela. 1992. Variability in the efﬁcacy of delayed-inducible resistance in
mountain birch. Entomologia Experimentalis et Applicata 62: 107-115.

Sample BE, L. Butler, C. Zivkovich, R.C. Whitmore and R. Reardon. 1996. Effects of
Bacillus thuringiensis Berliner var Kurstaki and defoliation by the gypsy moth
[Lymantria dispar (L)(Lepidoptera: Lymantriidae)] on native arthropods in West
Virginia. The Canadian Entomologist 128: 573-592.

Schoener, T.W. 1983. Field experiments on interspeciﬁc competition. The American
Naturalist 122: 240-285.

Schultz, J .C. 1988. Plant responses induced by herbivores. Trends in Ecology and
Evolution 3: 45-49.

Schultz, J .C. and LT. Baldwin. 1982. Oak leaf quality declines in response to defoliation
by gypsy moth larvae. Science 217 : 149-151.

Scriber, J .M. and F. Slansky. 1981. The nutritional ecology of immature insects. Annual
Review of Entomology 26: 183-21 1.

Slansky, F. and P. Feeny. 1977. Stabilization of the rate of nitrogen accumulation by the
larvae of the cabbage butterﬂy on wild and cultivated food plants. Ecological
Monographs 47: 209-228.

Stockhoff, B. A. 1992. Diet-switching by gypsy moth: effects of diet nitrogen history vs.
switching on growth, consumption and food utilization. Entomologia
Experimentalis et Applicata 64: 225-238.

Strong, D.R., J .H. Lawton and T.R.E. Southwood. 1984. Insects on Plants: Community

Patterns and Mechanisms. Harvard University Press, Cambridge.

154

Swain, T. and WE. Hillis. 1959. The phenolic constituents of Prunus domestica. I. The
quantitative analysis of phenolic compounds. Journal of the Science of Food
Agriculture 10: 63-68.

Tammaru, T. & Haukioja, E. (1996) Capitol breeders and income breeders among
Lepidoptera — consequences to population dynamics. Oikos 77: 561-564.

Tuomi, J ., P. Niemela, E. Haukioja, S. Siren and S. Neuvonen. 1984. Nutrient stress: an
explanation for plant anti-herbivore responses to defoliation. 0ecologia 61: 208-
210.

Tuomi, J., P. Niemela and S. Siren.1990. The Panglossian paradigm and delayed-
inducible accumulation of foliar phenolics in mountain birch. Oikos 59:399-410.

Tuomi, J ., T. Fagerstrom and P. Niemela. 1991. Carbon allocation, phenotypic plasticity,
and induced defenses. Pp. 85-104 in D.W. Tallamy and M.J. Raupp, eds,
Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Turchin, P. 1990. Rarity of density-dependence or population regulation with lags?
Nature 344: 660-663.

Turchin, P. 1995. Population regulation: Old arguments and a new synthesis. Pp. 19-40 in
N. Cappuccino and PW. Price, eds, Population Dynamics: New Approaches and
Synthesis. Academic Press, New York.

Underwood, N. 1999. The inﬂuence of plant and herbivore characteristics on the
interaction between induced resistance and herbivore population dynamics. The

American Naturalist 153: 283-294.

155

Valentine, H.T., WE. Wallner and PM. Wargo. Nutritional changes in host foliage
during and after defoliation, and their relation to the weight of gypsy moth pupae.
0ecologia 57:298-302.

Virtanen, T. and S. Neuvonen (1999). Perfomiance of moth larvae on birch in relation to
altitude, climate, host quality, and parasitoids. 0ecologia. 120: 92-101.

Waldbauer, GP. 1968. The consumption and utilization of food by insects. Advances in
Insect Physiology 5: 229—289.

Wallner, WE. 1983. Gypsy moth host interactions: a concept of room and board. Pp. 5-8
in R.L. Talerco and M. Montgomery, eds, Proceedings: Forest Defoliator —Host
Interactions: a Comparison between Gypsy Moth and Spruce Budworms. USDA
Forest Service, Northeastern Station General Technical Report NE-85.

Wallner, WE. and GS. Walton. 1979. Host defoliation: a possible determinant of gypsy
moth population quality. Annals of the Entomological Society of America 72: 62-
67.

Werner, RA. 1979. Inﬂuence of host foliage on development, survival, fecundity and
oviposition of the spear-marked black moth, Rheumaptera hastata (Lepidoptera:
Geometridae). The Canadian Entomologist 111: 317-322.

West, C. 1985. Factors underlying the late seasonal appearance of the lepidopterous leaf-
mining guild on oak. Ecological Entomology 10: 111-120.

Williams, KS. and J.H. Myers.1984. Previous attack of red alder may improve food

quality for fall webworm larvae. 0ecologia 62: 74-79.

156

Williams, R.S., D. E. Lincoln and R.J. Norby. 1998. Leaf age effects of elevated CO2-
grown white oak leaves on spring-feeding lepidopterans. Global Change Biology
4: 235-246.

Wold, E.N. and R.J. Marquis. 1997. Induced defense in white oak: effects on herbivores
and consequences for the plant. Ecology 78: 1356-1369.

Woods, SA. and J.S. Elkinton. 1987. Bimodal patterns of mortality from nuclear
polyhedrosis virus in gypsy moth (Lymantria dispar) populations. Journal of
Invertebrate Pathology 50: 151—157.

Work, T.T. and D.G. McCullough. 2000. Lepidopteran communities in two forest
ecosystems during the ﬁrst gypsy moth outbreaks in northern Michigan.

Environmental Entomology 29: 884-900.

157

CHAPTER 3:
INDUCED-RESPONSES OF ASPEN TO SUCCESSIVE YEARS OF DEFOLIATION:

THE EFFECTS ON AN OUTBREAK F OLIVORE AND ITS NATURAL ENEMIES

INTRODUCTION

Decline in foliage quality following defoliation of trees is one of several mechanisms
potentially responsible for driving the delayed-density dependent feedback necessary to
produce the cyclic population dynamics characteristic of some outbreak forest
Lepidoptera (Berryman et al. 1987). In response to defoliation, two types of induced
resistance have been identiﬁed, rapid-induced resistance (RIR) and delayed-induced
resistance (DIR) (Neuvonen and Haukioja 1991). Induction of RIR occurs within hours
or days of herbivory, affecting the herbivore generation responsible for the damage, and
is thereby expected to exert a stabilizing inﬂuence on population density. In contrast,
DIR responses are not manifested until the following growing season, and therefore
affect future generations of herbivores (Haukioja 1982). A number of authors have
speculated that long-term reductions in host plant quality and the time-lag in the recovery
of trees after defoliation, act to destabilize populations thus contributing to population
cycles (e.g., Benz 1974, Haukioja 1980, Rhoades 1983). A recent model has veriﬁed that
DIR can produce population cycles if there is a sufﬁcient time-lag before maximum
induction, or if following induction, the relaxation of DIR is sufﬁciently protracted

(Underwood 1999).

Research on mountain birch, Betula pubescens ssp. czerepanovii (Orlova) Hamet-

Ahti, and its primary herbivore, the autumnal moth, Epirrita autumanta (Lepidoptera:

158

Geometridae) in F ennoscandia, has generated substantive contributions to the
understanding of the role of induced-resistance in insect population dynamics.
Defoliation of mountain birch induces phytochemical changes in the leaves characterized
by elevated levels of secondary compounds such as hydrolyzable and condensed tannins,
and reductions in primary nutrients including nitrogen and water (e. g., Tuomi et al. 1984,
1990, Kaitaniemi et al. 1998). These changes in host quality persist for at least the
subsequent growing season, and in some cases do not return to predefoliation levels for
several years (Tuomi 1984). Caterpillars feeding on foliage from previously defoliated
mountain birch suffer signiﬁcantly reduced growth and fecundity relative to undefoliated
control trees. In some studies, reductions in fecundity as high as 60-80% were found,
although lesser effects are more common (Haukioj a 1982, Haukioja and Neuvonen 1987,
Ruohomaki et a1. 1992, Kaitaniemi et al. 1999a, 1999b). In other systems, artiﬁcial
defoliation of paper birch, Betula papyrifera Marshall, and trembling aspen, Populus
tremuloides Michaux, simulating outbreaks of the black-marked spear moth,
Rheumaptera hastata (Lepidoptera: Geometridae), and large aspen tortrix, Choristoneura
conﬂictana (Lepidoptera: Tortricidae), respectively, induced phytochemical changes and
reductions in herbivore performance (Clausen et al. 1991, Bryant et a1. 1993). These
studies suggest that defoliation driven declines in host plant quality may be a general

phenomenon in deciduous trees.

During outbreaks of forest insects, host trees are often partially or completely
defoliated for at least two successive years (e.g., Mattson et al. 1991). However, the
majority of studies documenting DIR have been based on only a single defoliation event,

followed by phytochemical analyses and larval bioassays in the following year

159

(Kaitaniemi et al. 1999b). Single year defoliation studies have been used repeatedly
despite the speculation of Haukioja et al. (1988) that consecutive years of defoliation
might have cumulative negative effects on the quality of trees for herbivores. Kaitaniemi
et al. (1999b) suggested that the effects of multiple years of defoliation on trees and their
reciprocal effects on herbivores may be one of the least understood aspects of induced
resistance to herbivory in trees. Increasingly negative effects on insect performance with
each successive year of defoliation have been shown in gray birch, Betula populifolia
Marsh, paper birch, and black oak, Quercus velutina Lamarck (Werner 1979, Wallner
and Walton, 1979, Valentine et al. 1983). More recently, successive years of defoliation
were shown to have only minor cumulative effects on phytochemistry and insect
performance in mountain birch and hybrid poplar, Populus x euramericana c.v.
‘Eugeneii’ (Kaitaniemi et al. 1999b, Parry et al. 2000a). Reasons for the disparity among
the studies above have not been identiﬁed. The magnitude of the effects could be system
speciﬁc or could result from differences in experimental methods such as insect-caused
versus manually applied defoliation, background soil fertility, nutrient addition to
compensate for the lack of frass in manual defoliation experiments, and the timing or
severity of the defoliation. The lack of congruence among the studies also highlights the

gaps in our knowledge on the effects of long term defoliation on trees.

Outbreaks of the forest tent caterpillar, Malacosoma disstria (Lepidoptera:
Lasiocampidae) on trembling aspen rank among the largest in the world with contiguous
areas of defoliation exceeding 13 million ha in a single season (e.g., Mattson et al. 1991).
The majority of this defoliation is concentrated in the vast trembling aspen forests found

across much of Canada and in the Great Lake States of Michigan, Wisconsin, and

160

Minnesota. In these regions of North America, forest tent caterpillar exhibits cyclical
population dynamics with outbreaks occurring every 9-12 years, although at smaller
scales, there is considerable variability around the mean cycle length (Hildahl and Reeks
1960, Sippell 1962, Hodson 1977). The length of outbreaks can be protracted with aspen
trees often defoliated for three consecutive years although longer outbreaks are not
uncommon (Hildahl and Reeks 1960, Sippell 1962, Ives 1971, Witter et al. 1975, Hodson
1977). Surprisingly, the effects of such sustained, severe defoliation on phytochemistry,
and the potential effects that any changes in host quality may have on forest tent

caterpillar population dynamics have received little attention.

Trembling aspen is typical of early succesional trees, quickly colonizing recently
disturbed landscapes and relying on rapid growth to out-compete other vegetation
(Barnes 1969, Peterson and Peterson 1992). Although aspen produces copious quantities
of seed, reproduction is primarily by clonal suckering from parental rootstock (Barnes
1969, Peterson and Peterson 1992). This reproductive strategy may result in vast areas of
even-aged stands with relatively little genetic diversity, especially following ﬁre, clear
cut logging, or other large disturbances. The low inter- and intra-speciﬁc diversity in
aspen dominated environments led Mattson et a1. (1991) to suggest that herbivore
outbreaks are inevitable in these forests. From theory, fast growing, shade intolerant early
succesional tree species such as aspen, should invest relatively little in constitutive
resistance mechanisms, favoring instead resistance mechanisms that are inducible (Herms
& Mattson 1992). Given the frequency of large—scale defoliation events in environments
where most competition is intraspeciﬁc, trees such as aspen are also predicted to have

evolved strategies allowing tolerance of high herbivory (Mattson et a1. 1991).

161

Factors underlying the outbreak and collapse of forest tent caterpillar populations
have been the subject of much speculation. Weather, parasitoids, and pathogens may all
play a role, although linkage of these factors to changes in population density has not
been shown. A component of population dynamics that has received relatively little
attention is the marked declines in forest tent caterpillar fecundity that characterize the
latter years of an outbreak. This decline can be large with egg masses being as much as
50% smaller at the end of an outbreak (Ives 1971, Witter et al. 197 5, Parry, unpublished).
In western tent caterpillar, Malacosoma californicum Dyar, fecundity remained high
through the early years of outbreaks when densities were highest before declining and
remaining low even as populations collapsed (Myers and Kukan 1995). These data
suggest that rather than a simple shortage of food, a delayed-density dependent factor is
operating on fecundity. Some authors have suggested that sublethal pathogens maybe
primarily responsible for declines in fecundity (Rothman and Myers 1994, Myers and
Kukan 1995, Rothman 1997), although the evidence is equivocal. The possibility that
reductions in forest tent caterpillar fecundity over the course of an outbreak could be
driven by DIR, manifested through changes in aspen primary or secondary
phytochemistry, as has been suggested from studies of the large aspen tortrix (Clausen et

al. 1991), has not been investigated.

Increasing attention has been focused on the effects of induced-resistance on the
natural enemies of herbivorous insects. A number of studies have shown that plants
damaged by herbivores emit chemical signals that are attractive to natural enemies of
caterpillars (e.g., Eller et al. 1988, Turlings et al. 1993, Thaler 1997). Thus, increased

emphasis has been placed on the study of plant attributes such as nutritional quality,

162

production of secondary compounds, and physical structure that may enhance or interfere
with the activity of natural enemies (Agrawal 2000). While the focus of such research
has generally been with agricultural plants, there is increasing evidence that similar

tritrophic interactions occur in trees (e. g., Havill and Raffa 2000).

Conceptually, the idea that induction of volatiles attractive to natural enemies
following herbivory is not new. More than 60 years ago, it was suspected that some
tachinid ﬂies (Diptera: Tachinidae) attacking caterpillars were attracted to volatiles from
host trees because they would only oviposit in the presence of caterpillar damaged leaves
(Bess 1936). This has now been documented for tachinids in both agricultural and
forested environments (e.g., Monteith 1964, Roland et al. 1995). Two species of leaf-
ovipositing tachinids, Leschenaultia exul (Townsend) and Patelloa pachypyga (Aldrich
and Weber) dominate the larval parasitoid complex across the northern range of forest
tent caterpillar (Sippell 1957, Witter and Kulrnan 1979, Parry 1995, Parry et al. 1997).
Experiments have shown that both species are attracted to volatiles released when tent
caterpillars feed on leaves (Mondor and Roland 1997, 1998) conﬁrming the hypothesis
proposed by Bess (1936). To my knowledge, the effect of induced responses in trees
(RIR or DIR) on the activity of parasitoids has not been investigated. A number of studies
have shown that the ﬁtness of dipteran and hymenopteran parasitoids can be reduced
when utilizing hosts that have been feeding on diets high in secondary metabolites
(Bourchier 1991, Roth et al. 1997, Havill and Raffa 2000). Thus parasitoids may make
tradeoffs between attacking larger numbers of poor quality hosts on plants with induced-

resistance, or searching for higher quality hosts which may be at lower density.

163

The objective of this study was to emulate the interactions that occur between forest
tent caterpillar and aspen during outbreak conditions. I incorporated as much realism as
is possible under the constraints imposed by the nature of a manipulative experimental
ﬁeld study. Defoliation treatments were applied by inoculating trees with outbreak
densities of caterpillars over a three—year period. Trees used in the experiment were
ontogenetically mature belonging to two different clones growing in a forest
environment. In 2000, the fourth year of the study, half of the previously defoliated trees
were left undefoliated and half were defoliated. again. In addition, another set of trees was
defoliated for the ﬁrst time in 2000. A fourth set of undefoliated trees served as the
control. These treatments allowed the determination of the relative effects of RIR, DIR,
and the combined effects of both, on growth, pupal mass, fecundity, and survival of forest
tent caterpillars. I also examined the effects of the defoliation treatments on parasitism by
two species of tachinid ﬂies that are attracted by volatiles released by aspen leaves

damaged by forest tent caterpillar feeding.

MATERIALS AND METHODS

EXPERIMENTAL SYSTEM

The study was conducted in an early successional forest (ca. 15 yr old) growing in an
abandoned ﬁeld at the edge of a mature wood-lot on the Michigan State University
campus, East Lansing. I identiﬁed six aspen clones at this site based on leaf morphology,
timing of leaf drop in autumn, and phenology of bud break. I selected two of the clones

for the experiment because of the six, they contained the most ramets which afforded me

164

the greatest opportunity to standardize the experimental trees with respect to height,
aspect, exposure, and foliar volume. These clones were separated by ca. 20 m of mixed
white ash, Fraxinus americanus L., red maple, Acer rubrum L. and red oak, Quercus
rubra L., saplings and were growing on similar sites with respect to drainage and
insolation. The timing of bud break between the clones differed by approximately ﬁve
calendar days each year. At the onset of the experiment in 1997, I selected 20 trees in
each clone and permanently identiﬁed each tree with a metal identiﬁcation tag. In the
spring of 1997, trunks of the trees averaged 3.4 cm in diameter at 45 cm above ground

(below ﬁrst major branching point) and ranged from 3.6 to 4.5 m in height.

To my knowledge, no signiﬁcant defoliation had occurred on these trees prior to this
study. The Baker Woodlot, adjoining the plots on one edge, is a heavily utilized
recreational and instructional area on campus and signiﬁcant defoliation would likely
have not gone unnoticed. Herbivory occurred at background levels (<10%) in the aspen
stands in 1996, the year before the study began. Although I can not rule out signiﬁcant
herbivory occurring prior to the experiments, given the long time frame of the study,
control trees would have been free of herbivory for at least 5 years previous to the

bioassays in 2000.

IMPLEMENTATION OF TREATMENTS

In 1997, the initial year of the study, I randomly assigned half of the trees in each clone to
the forest tent caterpillar defoliation treatment. These trees were defoliated in 1997, 1998,
and 1999. In 2000, half of the trees defoliated from 1997-1999 were not defoliated and

half were defoliated again (Table 20). In addition, ﬁve trees in each clone not previously

165

Table 20. History of defoliation treatments implemented between 1997-2000 to assess
the effects of RIR and DIR on forest tent caterpillar performance. Short-term bioassays
using second and ﬁnal (ﬁfth) instars and a life-long egg-adult bioassay were conducted in

2000. The same 40 trees ﬁrst selected in 1997 were used throughout the study.

 

 

 

 

Defoliation

Treatment Clone N (trees) 1997 1998 1999 2000

CTRL 1 5 No No No No
2 5

RIR 1 5 No No No Yes
2 5

DIR 1 5 Yes Yes Yes No
2 5

RIR/DIR l 5 Yes Yes Yes Yes
2 5

 

166

defoliated, were defoliated for the ﬁrst time (Table 20). This experimental design allowed
me to simultaneously test the effects of RIR (trees defoliated for the ﬁrst time concurrent
with insect bioassays = RIR), DIR (trees defoliated for the previous 3 years but with no
defoliation in the year of the insect bioassays = DIR), and the combined effects of DIR
and RIR (trees with 3 years of defoliation as well as defoliation concurrent with the insect
bioassays = RIR/DIR). Control trees (CTRL) were always maintained free of herbivory

from tent caterpillars.

To inoculate trees with caterpillar densities capable of creating high levels of
defoliation, large numbers of forest tent caterpillar eggs were obtained from outbreak
populations. I used eggs collected near Cochrane, Ontario, Canada in 1997 and 1998,
Flin Flon, Manitoba, Canada in 1999, and Ontonagon, Michigan in 2000. Nuclear
polyhedrosis virus WPV) is a major concern with ﬁeld collected eggs. To remove viable
NPV from egg surfaces, spumaline, a frothy secretion from the female accessory glands
used to cover the eggs, was scraped ’off using a razor blade. Egg bands were then
immersed in household bleach, diluted to approximately 4%, for 3 minutes. In 1999 and
2000, I added a small amount of household dish detergent to the bleach and agitated the
eggs using a toothbrush which appeared to give better results than the bleach soak alone.
This method was very effective in eliminating pathogen problems and despite many
thousands of tent caterpillars on each defoliation treatment tree, only a few cases of NPV

mortality were observed. No NPV deaths were recorded in. any of the bioassay larvae.

After surface sterilization, packages of eggs were held at ca. 4°C over the winter, and
misted occasionally with water to prevent desiccation. Trees were monitored daily as

spring approached and once buds began to swell, I placed 5—10 egg bands in cheesecloth

167

packages and warmed them at room temperature in the laboratory until approximately
one day prior to hatch. Eggs were then returned to the refrigerator until bud break. Forest
tent caterpillar cannot initiate feeding until buds soften and the scales begin to separate
(D. Parry, unpublished data). Once bud scales began to separate, I removed the eggs from
the refrigerator and attached them to treatment trees with plastic twist ties. Depending on
the size of the tree, 3-6 packages were placed throughout the canopy. Eggs began to hatch
within two days of being place on the trees, closely approximating the natural synchrony

between aspen and forest tent caterpillar (see Parry et al. 1998).

Upon eclosion, groups of caterpillars moved to the buds and initiated feeding on the
newly expanding leaves. Because forest tent caterpillars is a highly gregarious species
and exhibits considerable ﬁdelity to their ‘maternal tree’ until at least the middle of the
fourth instar (Batzer et al. 1995, D. Parry, personal observation), larvae do not need to be
restrained. Unlike other Malacosoma species, forest tent caterpillar does not make a tent,
and exhibits a nomadic feeding behavior where larval groups wander throughout the
canopy feeding at different locations. The density of caterpillars placed on the trees was
sufﬁcient to cause 50-70% defoliation by the end of the fourth instar, levels that are
similar to those observed in many natural outbreaks (D. Parry, personal observation).
After the fourth molt, forest tent caterpillars generally leave the maternal tree and wander
extensively, irrespective of population density or defoliation level. In natural outbreaks,
densities of these wandering ﬁnal instar larvae are often high enough that any remaining
aspen foliage is completely defoliated, along with. many other tree and shrub species.
However, in my plots, unrestrained larvae dispersed into the surrounding forest and

disappeared. To achieve the higher levels of defoliation characteristic of outbreaks, I

168

enclosed groups of 30 to 60 ﬁnal instars in ﬁne mesh sleeve cages and moved the cages
throughout the canopy of each treatment tree on nearly a daily basis to simulate the
wandering behavior of larvae. Caterpillars in these bags were used solely to defoliate the
trees and were not part of any bioassay. This method also allowed me to equalize the
defoliation among trees by removing or adding cages of larvae to achieve the desired
levels of herbivory. Unrestrained larvae were prevented from ascending control trees by

a band of tangle trap placed over paper tree wrap at the base of the trunk.

An essential element of realism in this study is that comparisons of treatment effects
were made with control trees that had background levels of herbivory in addition to any
effects caused by the 30 caterpillars used in the on-tree bioassays. The low level
herbivory on the control trees is important because if induced-responses are to have any
role in the population dynamics of forest tent caterpillar, they must increase beyond
background levels caused by groups of this gregarious species, as well as herbivory from
other species. Unlike on-tree bioassays done with solitary species where density effects
can confound results, forest tent caterpillar forage in clusters, thus rearing in groups is
consistent with the species natural behavior. A single-family group consists of 100-500
caterpillars depending on latitude and population density (Parry et al. 2000b). In many
previous studies, researchers compared the effects of defoliation treatments against
control trees with no herbivory. This is unrealistic, with respect to gregarious species at
least, because even in low-density populations, larvae will experience the induction
effects of their own feeding, as well as feeding by the family group. Thus, the correct

comparison for this and other gregarious species is between trees experiencing severe

169

defoliation and those affected by background herbivory and the feeding of individual

family groups, a design element that I have utilized throughout this experiment.

I estimated defoliation by assigning trees to 10% classes based on the amount and
degree of damage to the remaining foliage. In a study using hybrid poplar (Parry et al.
2000a), similar visual estimates of defoliation were highly correlated with damage
assessments done on individual leaves. Because of its severity, defoliation in the three
induction treatments was relatively easy to estimate. However, herbivory on control trees
was very low and difﬁcult to estimate. To further quantify defoliation on control trees, I
estimated feeding damage on two dominant lateral branches in the mid-canopy, the
maximum height accessible from a 3 m ladder. The ﬁrst 50 leaves from the short shoots
beginning at the branch tip were classiﬁed as undamaged if they were free from
herbivory, suffered minor blemishes, or occasional small shot holes. Leaves with more
than ca. 20% of the surface removed were classiﬁed as damaged. This showed that 18%
of the leaves on undefoliated trees in Clone 1 (controls and DIR trees) had some damage.
In Clone 2, the damage on the undefoliated trees (controls and DIR) was 20%. The
majority of the damaged leaves had more than 70% of their total area remaining. In
contrast, 100% of the leaves on all trees in the defoliated treatments were classiﬁed as

damaged and most had lost 60% or more of the total leaf area.

INSECT BIOASSAYS

Lifelong Bioassay —— To estimate the effects of the defoliation treatments on ﬁtness, I
reared forest tent caterpillars from egg hatch through adult on each of the experimental

trees. Experimental larvae were obtained from egg bands collected the previous fall from

170

a moderately high, expanding population in Ontonagon County, Michigan. The collected
egg bands were large, containing 150-200 eggs, indicating little previous stress and the
population was growing (ratio of current year egg bands to previous year egg bands
128:27 for three sampled trees). Eggs were kept in an environmental chamber in plastic
bags at ca. 4°C and 70% RH for the duration of the winter. Shortly prior to bud break, 45
egg bands were removed from cold. storage, surface sterilized as above, and placed at
room temperature to hatch. The progeny of each egg band, representing the full
reproductive compliment of a single female, were allowed to mingle in a large plastic
container for 24 hours. I used a ﬁne paintbrush to randomly allocate 30 neonates to petri
dishes assigned to each of the 40 trees. Neonate larvae are diminutive and could pass
through the mesh in the screen cages. Thus, each group of 30 larvae was reared in the
laboratory on foliage clipped from the assigned tree until the end of the ﬁrst instar. Twigs
were placed in ﬂorist’s aquapicks to maintain leaf turgor. I reared the ﬁrst instars at 18°C
with a 16:8 photoperiod. At this temperature, growth of larvae kept pace with that of
larvae in the ﬁeld. At the time of molt to second instar, each group of 30 was transferred
to the ﬁeld and placed within a large (60 x 30 cm) mesh sleeve cage which enclosed
foliage and was attached to the appropriate experimental tree. While the larvae were
small (L2-L4), sleeves were moved to new branches every 2-3 days to simulate the
movement of the nomadic feeding groups through the trees. Extensive ﬁeld observations
on the natural foraging patterns of forest tent caterpillars suggest that larvae often utilize
a single branch for several days before moving elsewhere on the tree, particularly during
early instars. Once the larvae entered the fourth instar, sleeves were moved more

frequently, and nearly daily through the ﬁnal (L5). Foliage within the bioassay sleeves

171

was never allowed to become depleted. I did not protect any foliage from non-bioassay
caterpillars defoliating the trees although I did remove caterpillars during the ﬁnal instar
to prevent trees from becoming completely defoliated prior to the bioassay larvae
completing their development. Therefore, the bioassay caterpillars were feeding on
foliage representative of the damage levels on each tree, which added an essential degree
of realism to the study, and also highlights the critical importance of using disease-free
larvae to defoliate the trees. When moving the sleeve cages, an effort was made to
provide the bioassay caterpillars with the best foliage remaining on a tree under the
assumption that foraging caterpillars discriminate between severely damaged and
partially consumed leaves. Even so, caterpillars in the defoliated treatments were

completing development on leaves that had suffered considerable damage.

When caterpillars had ﬁnished feeding and were spinning cocoons in preparation for
pupation, the sleeve cages were removed from the trees and returned to the laboratory.
Within 24-48 hours of pupation, the silk cocoons were carefully cut away with scissors
and the pupae were weighed on an electronic balance (to 0.1 mg) and then placed
individually on squares of paper towel in plastic snap cap vials. Pupae were held at 22°C
and 16:8 and checked daily for emergence of adults, which were sexed and the

emergence date recorded.

To estimate the effects of the treatments on fecundity, caterpillars from the same
population used in the bioassay were reared from egg hatch to pupation on aspen trees
near the experimental plots. At pupation, pupae were weighed and individually placed in
vials for emergence. Using the methods of Parry et al. (2000b), adult females were

dissected, the number of eggs counted and a least-squares regression equation ﬁtted to the

172

relationship between fecundity and pupal mass. This regression equation (y = 390.85x —
19.11, r2 = 0.88, p < 0.0001, df = 24) was used to estimate fecundity changes based on

the pupal mass data from the lifelong bioassay.

Short-term bioassay — Short-term bioassays were conducted using second (L2) and ﬁfth
(L5) instars, respectively in 2000. The progeny of 40 egg bands collected the previous
year in Ontonagon County, Michigan were allowed to mingle for 24 hours and then
several hundred were placed in the ﬁeld on aspen trees outside of the plots. These larvae
served as the stock for both the L2 and L5 short-term bioassay and were from the same
population as those used in the long-term bioassay above. Because hatch was
synchronous and caterpillars highly gregarious through the ﬁrst four instars, sufﬁcient
numbers of similar aged larvae could easily be collected for bioassays. For the L2
bioassay, I retrieved caterpillars from the ﬁeld when the ﬁrst instars stopped feeding and
congregated on silk pads in preparation for molt. Forest tent caterpillar is sensitive to
phenological changes in foliar quality (Parry et al. 1998). Thus, using caterpillars reared
in the ﬁeld in close proximity to the plots ensured that they were synchronized with the
host plant and larvae used in a life-long bioassay below. Following collection,
caterpillars were returned to the laboratory and held in petri dishes on moist paper towels
without food. Within 12 hours of collection, most of the collected caterpillars had molted
to the second instar. Foliage was then obtained by clipping representative samples of
leaves at the base of the petiole throughout the mid-canopy of each experimental tree.
Clipping leaves in this manner does not induce phytochemical changes in aspen (Mattson
and Palmer 1988). Aspen is heterophyllous, producing both determinate shoots from leaf

primordia at bud break and indeterminate shoots that continue to develop new leaves

173

throughout the growing season. To standardize the collections, only leaves from

determinate shoots were used in the short-term bioassays.

Collected leaves were placed in zip-lock bags on ice, returned to the laboratory,
weighed on an electronic balance and placed in 125 x 50 mm plastic petri dishes. The
bottom third of each dish had been ﬁlled with plaster saturated with water, which
maintained humidity and. prevented leaf turgor loss over the 2-3 day duration of the
bioassays. I randomly allocated groups of second instars (n = 12) to foliage from each of
the 40 trees. Fifteen larvae were individually weighed fresh and then dried to provide an
initial dry weight estimate. The bioassay was conducted at in an environmental chamber
at 23°C with 16:8 photoperiod until the most rapidly developing groups ﬁnished feeding
in preparation for the second molt (51 hours for Clone 1 and 42 hours for Clone 2). At the
conclusion of the bioassay, caterpillars were placed individually in vials and the
remaining portions of leaves and frass were collected then dried at 40°C in a drying oven
for ﬁve days prior to weighing. Relative growth and consumption rates were calculated
for the L2 caterpillars from the weights of insects and foliage consumed as described

below.

For the ﬁfth instar bioassay, the same methods as for the L2 bioassay were used.
Fourth instars were collected from silk molting mats in the ﬁeld when their head capsules
showed signs of slippage, and were returned to the laboratory to complete the molt. I
reared four caterpillars individually on leaves collected from each of the trees for 72
hours at 24°C. Observations from previous experiments suggest that this time period
brackets the period of maximum growth in this instar. Leaves were weighed prior to

presentation to caterpillars. Sub-samples of 15 larvae were dried prior to the bioassay and

174

linear regression was used to estimate the initial dry weight of the experimental insects.
In this Michigan population, newly molted ﬁfth instar females were larger (ca.190-250
mg) than males (ca.125-190 mg) so I allocated two females and two males (as identiﬁed
from initial weight) to foliage from each tree. After the termination of the bioassay,
larvae were frozen for a few hours and then examined microscopically to conﬁrm their
sex using diagnostic characters described in Stehr and Cook (1968). Caterpillars were
then placed individually in vials and dried. Frass was collected into 1g plastic vials, and

uneaten leaf portions were placed in small paper envelopes prior to drying.

For L5 caterpillars, I calculated relative growth and consumption rates, and standard
estimates of efﬁciency of digestion of ingested food (ECI), approximate digestibility of
ingested food (AD), and efﬁciency of conversion of digested food to biomass (ECD)
using the gravimetric methods of Waldbauer (1968). Caterpillars that died or did not feed
were not included in any analyses. Relative growth rate was calculated for individual
caterpillars as RGR = ln(weightf) - ln(weight,-)/ T where In is the natural logarithm, weightf
= ﬁnal weight, weight,- = initial weight, and T is the elapsed time in days (Gordon 1968).
Estimates of initial dry weight of larvae were obtained from a subset of insects weighed
wet and dried at the start of the experiment. To estimate initial leaf dry weight, a portion

of the leaves collected from each tree was dried before the experiment started.

Survival and parasitism — Overall survival ﬁ'om hatch to pupation was assessed by
counting the number of larvae from the original 30 that successfully pupated. To assess
treatment effects independently of parasitism, I included in this analysis all caterpillars
that were parasitized. The effects of parasitism on survival were addressed in a separate

analysis (described below). Two sleeve cages ripped during a windstonn and some early

175

instars were lost. For these groups, the remaining larvae were transferred to a new sleeve,
and survival to pupation estimated as a proportion of the larvae that were transferred,

recognizing that this introduces a small amount of error to the estimate.

1 was interested in the contribution that induced resistance could make to population
growth, relative to the effects of other contemporaneous factors such as natural enemies.
For each tree, sources of mortality were partitioned among parasitoids and unknown
causes. The two dominant parasitoids of forest tent caterpillar larvae place microtype
eggs on foliage and thus are not greatly affected by the presence of sleeve cages. Other
parasitoid species can attack the larvae through the sides of the screen cages although
rates of parasitism are often lower than in unrestrained larvae (D. Parry, personal
observation). When sleeve cages were removed from the trees at pupation, a careful
search was made of each bag for ﬂy puparia and hymenopteran cocoons that had emerged
from bioassay caterpillars. These were retumed to the laboratory and placed in vials until
adult emergence. Parasitoids emerging from tent caterpillar pupae or collected from
inside the sleeve cages were identiﬁed using the keys of Sippell (1961) and Williams et

al. (1996).

To address the relative contribution of induced resistance and parasitism to
population dynamics, 1 estimated the approximate net reproductive rate (R0) of forest tent
caterpillar in each treatment. To simplify the calculations, 1 ﬁrst converted males to
female equivalents using the regression equation y = 2.0798x — 0.1484 (r2 = 0.67, p <
0.001, df = 40) determined from the mean male pupal mass for each experimental tree.
The number insects dying from unknown causes were subtracted from the original 30

insects in each cohort to estimate survival, and then multiplied by the mean fecundity for

176

 

each treatment to determine R0 (Southwood 1991). Fecundity was determined from the
regression equation calculated above (y = 390.85x — 19.11, df = 24, r2 = 0.88, p <
0.0001). Because parasitized individuals were included in the survival calculation, their
effect could be determined from R0 with and without including parasitized individuals in
the calculations. In addition to estimates obtained from my experiment, I calculated
parasitism rates from ﬁrst, third, and fourth year outbreaks of forest tent caterpillar in
Alberta, Canada (Parry 1995), which are similar to levels recorded elsewhere across the
northern range of forest tent caterpillar (e.g., Sippell 1957, Witter and Kulman 1979). I
calculated R0 using these values in place of those estimated in my experiment. The
relative impact of parasitoids in natural populations could then be compared under the
assumption that the induced-resistance levels found in my experiment are representative

of aspen forests elsewhere.

STATISTICAL ANALYSES

In all statistical analyses, I used a mixed-model ANOVA where Yijk = u + C,- + Dj +
Tk(C-) + (C: x Dj) + eU-k, where C,- = aspen clone, Dj = defoliation treatment, and Tk =
tree. In this model, aspen clone was a random effect and defoliation treatment was a
ﬁxed effect. F-tests for defoliation treatment, clone x defoliation treatment, and
tree(clone) were over the mean square error. Clone was tested over the C,- x D,-
interaction. I used the PDIFF option following the LSMEANS statement (PROC GLM,
SAS Institute, 1997) to make preplanned a priori pairwise comparisons of treatment
means within a clone and between same treatments among clones. Prior to analysis, data

were checked for normality and homoscedasticity and appropriately transfonned if

177

 

necessary. To improve the normality, proportional and percentage data (nutritional

indices, survival, parasitism) were arcsine-square root transformed prior to analysis.

Some have cautioned that each ramet in an aspen clone (‘Tree’ in my analysis) may
not be entirely independent (Osier et al. 2000), presumably because of concerns about
common root connections. This may be more of a problem with young trees than with
older trees which have well developed root systems and may sever root connections to
parental and adjoining trees entirely (Peterson and Peterson 1992). To my knowledge,
only Mattson and Palmer (1988) have looked at the effect of defoliation of individual
aspen trees on surrounding ramets. They found that potassium concentration varied with
distance, but that there was no effect on any other primary or secondary compound.

Individual ramets are used as the experimental unit throughout my study.

RESULTS

EFFECTS OF DEFOLIATION ON CATERPILLARS

Life-long bioassay — In egg to adult bioassays, both aspen clone and defoliation
treatments had signiﬁcant effects on forest tent caterpillar pupal mass. Female pupae
were heavier when reared on trees in Clone 1, whereas there was no effect of clone on
male pupal mass (Fig 14, Table 21). Pupal mass of both sexes was reduced relative to
controls in all three defoliation treatments. However, there was no difference in pupal
masses among the three defoliation treatments for females or males. Among trees, male

and female pupal masses were highly correlated (r = 0.82, p < 0.001, n = 40) indicating

178

 

  

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

F m Ies DCTRL
- e a 1:] RIR
[:1 DIR
0-8 r l RIR/DIR
0.6 e
J E
0.4 e
a ab bc x y y
0.2 A
g _
u) 0
0’ 1
‘2“ Males ElCTRL
‘ I] RIR
, r1611 DIR
0'8 l RIR/DIR
0.6 ,
0.4 - L T
0.2 , a a a b x y xy
0 , -
Clone 1 Clone 2

Figure 14. Pupal mass (mean i SE) for female and male forest tent
caterpillars reared from egg hatch to pupation on trees in each of the
treatments. Different letters indicate signiﬁcant pairwise differences between
means (p < 0.05) following a signiﬁcant treatment effect in ANOVA.
Treatments are CTRL = Control (undefoliated), RIR = one year of defoliation
concurrent with the bioassay, DIR = defoliation for three consecutive years
previous to the bioassay, and RIR/DIR = defoliation for the three previous
years and one year of defoliation concurrent with the bioassay.

179

Table 21. Results of ANOVA for an egg to adult, on tree bioassay. Thirty caterpillars
were reared on each experimental tree. Two variables were measured, pupal mass and the
duration of development from egg to adult. The F -test for ‘Clone’, a random effect, was
over the mean-square term for ‘Tree(Clone)’. F—tests for the ﬁxed effects ‘Treatment’,
‘Clone >< Treatment’, and ‘Tree (Clone)’ were over the mean-square error term. Analyses
were done separately for males and females. Symbols are: *** p < 0.001, ** p < 0.01, * p
< 0.5, 'l p < 0.10.

 

 

 

Females Males

Variable Source df Mean Square F Mean Square F

Pupal Mass Clone 1 0.035 7.2 * 0.000 0.6
Treatment 3 0.036 9.2 ** 0.005 8.5 **
Clone >< Treatment 3 0.004 0.9 0.001 0.1
Tree (Clone) 8 0.005 1.2 0.001 2.0 ‘1'
Error 24 0.004 0.001

Duration Clone ’ 1 53.20 62.3 *** 63.06 68.2 ***
Treatment 3 1.90 1.4 1.78 1.7
Clone >< Treatment 3 1.38 1.0 0.55 0.5
Tree (Clone) 8 0.85 0.6 0.92 0.9
Error 24 1.34 1.03

 

180

that defoliation had similar effects on the performance of both sexes. Relative to controls,
females were 8 and 18% smaller after a single concurrent defoliation of Clones l and 2,
respectively, reﬂecting effects of RIR. In both clones, female pupae were 15% smaller on
trees with three previous years of defoliation, reﬂecting the effects of DIR. Three years of
previous defoliation, coupled with an additional concurrent year of defoliation, manifests
the combined effects of RIR and DIR. In this treatment, female pupal mass was reduced
by 17% and 20% in Clone 1 and 2, respectively. Although pupal masses of both male and
female forest tent caterpillars were reduced by defoliation, the duration of development
from egg to adult was not affected for either males or females (Fig. 15, Table 21). There
were signiﬁcant clonal effects however, with the deve10pment times of both males and

females longer on Clone 2.

Using the regression equation derived for this forest tent caterpillar population, the
effects of the treatments on female pupal mass can be estimated as decreases in fecundity.
In the single year defoliation treatment (RIR), fecundity was reduced by 23.0 eggs (8.5%)
and 56.3 eggs (21.7%) in Clones 1 and 2, respectively. Three years of previous
defoliation (DIR) decreased egg production in Clone 1 females by 37 eggs (13.5%) and
by 49 eggs (18.7%) in Clone 2 females. The combined effects of DIR and RIR reduced
fecundity by 59 eggs (21.5%) and 55 eggs (21.3%) in females feeding on trees in Clones
1 and 2, respectively. While reductions in pupal mass of similar magnitude to that of

females occurred in males, no direct effect on male ﬁtness was assessed in our study.

Short term bioassays - I found no signiﬁcant effect of the defoliation treatments on
relative growth (RGR) of second instars (Fig. 16, Table 22). Caterpillars fed foliage from

Clone 2 had signiﬁcantly higher RGR than those feeding on Clone 1. There was no

181

 

8O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Females 1:1 CTRL
: [:1 RIR
D DIR
60 g I RIR/DIR
4O ,
20 -
(D
>. 0 -
(U 80
D Males El CTRL.
- El RIR
I:I DIR
60 , I RIR/DIR
40 -
20 e
0 -
Clone1 Clone 2

Figure 15. Development time (mean t SE) for female and male forest tent
caterpillars reared from hatch to pupation on trees from each of the
treatments. Means comparisons done only following signiﬁcant treatment
effect (p < 0.05) in overall ANOVA. Treatments are CTRL = Control
(undefoliated), RIR = one year of defoliation concurrent with the bioassay,
DIR = defoliation for the three years previous to the bioassay, and
RIIUDIR = defoliation for the three previous years and one year of
defoliation concurrent with the bioassay.

182

mgomg'1 0 (la)r1

1.5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I:ICTRL
RGR DRIR
I:ID|R
.RIR/DIR
1 - i
0.5 .
0
5
- RCR DCTRL
EIRIR
4 2 I:IDIR
.RIR/DIR
3 ,
2 ,
1
o -

Clone 1 Clone 2

Figure 16. Mean (i SE) relative growth rates (RGR) and relative
consumption rates (RCR) for second instar forest tent caterpillar.
Treatments are CTRL = Control (undefoliated), RIR = one year of
defoliation concurrent with the bioassay, DIR = defoliation for three
consecutive years previous to the bioassay, and RIR/DIR = defoliation
for the three previous years and one year of defoliation concurrent with
the bioassay.

183

r“ 1‘

Table 22. Results of ANOVA for short-term bioassay using second instar forest tent
caterpillar larvae. The F -test for ‘Clone’, a random effect, was over the mean-square for
‘Tree(Clone)’. F-tests for the ﬁxed effects ‘Treatment’, ‘Clone X Treatment’, and ‘Tree
(Clone)’ were over the Mean-Square Error. RGR = relative grth rate, RCR = relative
consumption rate. One control tree in Clone 2 was inadvertently excluded during set up,
resulting in 38 degrees of freedom. Symbols are: *** p < 0.001, ** p < 0.01, * p < 0.5, T
p < 0.10.

 

 

 

Variable Source df Mean Square F

RGR Clone 1 0.696 288.09 ***
Treatment 3 0.002 1 .32
Clone >< Treatment 3 0.001 0.37
Tree (Clone) 8 0.001 1.63
Error 23 0.001

RCR Clone 1 4.11 40.30 ***
Treatment 3 0.39 2.40 T
Clone X Treatment 3 0.18 1.03
Tree (Clone) 8 0.10 0.63
Error 23 10.08

 

184

interaction between treatment and clone, or any effect of trees within clones, on second
instar relative growth rates. As with relative growth rate, relative consumption (RCR)
was signiﬁcantly affected by aspen clone (Fig 16, Table 22). Relative consumption was
higher for larvae feeding on foliage from trees in the defoliated treatments although the
difference was only marginally signiﬁcant. The elevation in RCR suggests that even
early instars may compensate for defoliation-induced changes in host quality, since RGR
remained constant across treatments. Alternatively, any changes in the young leaves
induced by previous and current year defoliation by early instars were relatively slight,

and had only small effects on the feeding behavior and growth of second instars.

The effects of the defoliation treatments on caterpillars were more pronounced on
ﬁfth than on second instars. Females were more sensitive to treatments than males.
Female RGR in the ﬁfth instar was signiﬁcantly affected by clone and in contrast to the
second instars was higher on Clone 1 foliage (Fig. 17, Table 23). Relative growth rate of
females was signiﬁcantly reduced when fed foliage from defoliation treatment trees.
Conversely, RGR of males was unchanged among clones and was not signiﬁcantly
affected by defoliation treatments. Relative consumption rates did not differ among
clones for either males or females. Female RCR was signiﬁcantly lower on control tree
foliage than on leaves from treatment trees. There was no signiﬁcant effect of clone or
defoliation treatments on approximate digestibility (AD) of either females or males. The
efﬁciency of conversion of ingested food (ECI) varied signiﬁcantly among clones for
males and marginally so for females. Both sexes had lower ECI when reared on foliage

from all of the defoliated treatment when compared to undefoliated controls. There was

185

Figure 17. Mean (i SE) relative growth rate (RGR), relative consumption rate (RCR),
approximate digestibility (AD), efﬁciency of conversion of ingested foliage to biomass
(ECI), and efﬁciency of conversion of digested foliage to biomass (ECD) for ﬁnal (ﬁﬁh)
instar female and male forest tent caterpillars. Treatments are CTRL = Control, RIR =
one year of defoliation concurrent with insect bioassay, DIR = defoliation for the three
years previous to insect bioassay, and RIR/DIR = defoliation for the three previous years
and in the year concurrent with the insect bioassay.

186

1 Females 1 Males

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

— RGR RGR
0.8 - 0.8 .
0.6 - 0.6 .
3 0.4 — 0.4
q 0
g 0.2 d a b b x xy xy 0.2
. - .
‘7 O I I I O - r .
g 2.5 2.5
('3': 2 : RCR 2 7 RCR
E 1.5 - 1.5 -
l r a ab b x x y xy‘ 1 i
0.5 - I 0.5
O - . - O - -
80 80
AD ' AD
60 . 60 .
% 4o - 40 -
20 - 20 .
O - - 0 ,
60 60
— ECI ECI
4o — 40
°/- « ~
0 I I r 0 I I I
100 100
E
80 . CD 80 7 ECD
60 < 60
%
40 A a b b x x x 40
20 : 20 ,
0 , - . 0 , , - . .
Clone 1 Clone 2 Clone 1 Clone 2

187

Table 23. Results of AN OVA for short-term bioassay of ﬁfth instar forest tent caterpillar
larvae. The F -test for ‘Clone’, a random effect, was over the mean-square for
‘Tree(Clone)’. F-tests for the ﬁxed effects ‘Treatment’, ‘Clone >< Treatment’, and ‘Tree
(Clone)’ were over the Mean-Square Error. Analysis was done separately for males and
females. RGR = relative growth rate, RCR = relative consumption rate, AD =
approximate digestibility, ECI = efﬁciency of conversion of ingested food, and ECD =
efﬁciency of conversion of digested food. Symbols are: *** p < 0.001, ** p < 0.01, * p <
0.5, T p < 0.10.

 

 

Females Males

df Mean Square F

 

Variable Source Mean Square F

 

 

RGR Clone 1 0.029 8.2 ** 0.007 2.3
Treatment 3 0.011 5.1 ** 0.006 1.4
Clone >< Treatment 3 0.003 1.5 0.000 0.1
Tree (Clone) 8 0.004 1.7 0.003 0.8
Error 24 0.002 0.004

RCR Clone 1 0.021 0.5 0.001 0.0
Treatment 3 0.045 6.4 ** 0.069 1.4
Clone >< Treatment 3 0.009 1.3 0.044 0.9
Tree (Clone) 8 0.043 6.2 ** 0.057 1.1
Error 24 0.007 0.048

AD Clone 1 118.89 3.2 46.09 2.5
Treatment 3 5.97 0.3 53.38 1.6
Clone X Treatment 3 16.39 0.8 1.26 0.0
Tree (Clone) 8 37.29 1.9 18.44 0.5
Error 24 20.19 34.02

ECI Clone 1 182.91 4.5 ‘1 46.70 10.7 *

Cont’d

188

 

Table 23 cont’d

 

ECD

Treatment
Clone >< Treatment

Tree (Clone)

Error

Clone

Treatment
Clone >< Treatment

Tree (Clone)

Error

3

3

8

24

1

3

3

8

24

54.38

5.06

40.51

8.78

228.88

340.71

65.53

101.85

47.67

6.2 **
0.6

4.6 **

2.3
7.2 **
1.4

2.11

13.75

12.60

4.38

19.83

27.57

121.69

62.96

19.24

78.22

0.7

0.6

0.2

1.4

1.6

0.8

0.3

 

189

 

no effect of clone on efﬁciency of conversion of ingested food (ECD) for either sex but

females had signiﬁcantly lower ECD when consuming foliage from defoliated trees.

There was no correlation between the mean RGR for second instars on each tree and
pupal mass of males (r = -0.l4, p < 0.41, n = 39) but second instar RGR was negatively
correlated with pupal masses of females (r = -0.34, p< 0.04, n =39). Male pupal mass and
ﬁfth instar RGR was not correlated (r = -0.l3, p < 0.42, n = 40) but ﬁfth instar RGR was
signiﬁcantly correlated with female pupal mass (r = 0.40, p < 0.02, n =40). The low
proportion of the variance in pupal mass that was explained by RGR over the ﬁrst three
days of the ﬁnal instar suggests that the treatment effects on pupal mass may be due

primarily to foliage quality experienced in the latter part of the ﬁnal instar.

EFFECTS OF TREATMENTS ON SURVIVAL AND PARASITISM

Survival of larvae from hatch to pupation was generally high (937 out of 1200
successfully pupated) and did not vary among the treatments although there was a trend
toward lower survival on defoliated trees (Fig. 16, Table 24). Mortality in the sleeve
cages was classiﬁed as unknown if the cocoons or puparia of parasitoids were not found.
Observations suggested that the primary cause of the unknown mortality were pentatomid
bugs that preyed on caterpillars resting on the sides of the sleeve cages. Due to the
obvious liquefaction of caterpillars that succumb to NPV, I am conﬁdent that this

pathogen killed none of the bioassay caterpillars.

Larvae were attacked by several species of parasitoids. Low levels of parasitism
(<1%) were caused by the braconid parasitoids Meteorus sp. and Hypositor fugivitus. I

was concerned that the sleeve cages might interfere with these species, but similar levels

190

%

 

- ++ IBiS/DIR
++

50*

25*

 

 

 

 

 

 

 

 

 

 

Clone 1 Clone 2

Figure 18. Percent survival (mean i SE) of forest tent caterpillar reared
from hatch to pupation on trees from each of the treatments. Analysis were
done on arcsine—square root transformed data. Survival data includes
caterpillars that were parasitized. Means comparisons done only following
a signiﬁcant (p < 0.05) treatment effect in overall ANOVA. Treatments
are CTRL = Control, RIR = one year of defoliation concurrent with the
bioassay, DIR = defoliation for the three years previous to the bioassay,
and RIR/DIR: defoliation for the three previous years and one year of
defoliation concurrent with the bioassay.

191

Table 24. Results of ANOVA for survival and parasitism of forest tent caterpillar in the
egg to adult bioassay. To determine if tree-mediated treatment effects inﬂuenced
survival, individuals that were parasitized were counted as surviving and the treatment
effects on parasitism were analyzed separately. Analyses were done using arcsine-square
root transformed data to meet assumptions of normality. The F -test for ‘Clone’, a random
effect, was over the mean-square for ‘Tree(Clone)’. F —tests for the ﬁxed effects
‘Treatrnent’, ‘Clone X Treatment’, and ‘Tree (Clone)’ were over the Mean-Square Error.
Symbols are: *** p < 0.001, ** p < 0.01, * p < 0.5, T p < 0.10.

 

 

 

Parameter Source df Mean Square F

Survival Clone 1 292.4 7.9 *
Treatment 3 1 72.5 1 .5
Clone X Treatment 3 161.2 1.4
Tree (Clone) 8 36.9 0.3
Error 24 1 15.9

Parasitism Clone 1 482.6 4.5 '1
Treatment 3 681.2 9.2 ***
Clone >< Treatment 3 96.1 1.3
Tree (Clone) 8 107.9 1. 5
Error 24 74.29

 

192

were recorded from samples of caterpillars from outside of the sleeve cages. A larger
source of mortality in the experiment was due to the multivoltine and highly polyphagous
tachind Lespesz'a frenchii. This ﬂy attacks caterpillars directly by cementing eggs to
larval setae. Parasitism rates were much greater in caterpillars collected from outside of
the bags (30%) than in the bioassay larvae (2%). Of the mortality attributable to
parasitoids in bioassay larvae, 78% was due to the tachinid ﬂies Leschenaultia exul and
Patelloa pachypyga, which are specialists on forest tent caterpillar (see Parry 1995, Parry
et al. 1997 for the biology of these species). Although they emerge from pupae, both
species are actually parasitoids of fourth and ﬁfth stage larvae. Adults deposit microtype
eggs on foliage, which must be ingested for parasitism to occur. In my experiment,
virtually all of the mature maggots emerged from pupae, although L. exul maggots may
also emerge from prepupational larvae (Parry 1995). A proportion of the unknown
mortality may be due to this ﬂy, as the maggots are capable of escaping by burrowing
through the mesh of the sleeve cages. Since these two species have nearly identical

lifecycles, their parasitism was pooled for analysis.

There were signiﬁcant differences in parasitism rates among treatments and also a
marginal signiﬁcant difference among clones (Fig. 17, Table 24). The highest parasitism
rates occurred on trees in the RIR and RIR/DIR treatments, where large numbers of
larvae were defoliating the trees concurrent with estimates of parasitism rates. In the DIR
treatment, caterpillar densities were equivalent to controls because only the bioassay
larvae were feeding on the foliage in the year that parasitism was assessed. There was no

difference in the rate of parasitism among control trees and the DIR treatment. This

193

 

 

 

 

 

 

 

50
ElCTRL
1:1 RIR
40 _ .DIR
I RIR/DIR
30 - i
%
20 - a b ab b x y X
10 I I
0 -- n Y ,
Clone 1 Clone 2

Figure 19. Percent parasitism rate (mean i SE) of forest tent caterpillar
reared from hatch to pupation on trees in each treatment. Analysis was done
on arcsine-square root transformed data. Different letters indicate signiﬁcant
pairwise differences (p < 0.05) among means following signiﬁcant treatment
effects in ANOVA. Treatments are CTRL = Control, RIR = one year of
defoliation concurrent with the bioassay, DIR = defoliation for the three
years previous to the bioassay, and RIIUDIR= defoliation for the three
previous years and one year of defoliation concurrent with the bioassay.

194

suggests that the parasitoids were responding in a spatially-density dependent manner and

were not inﬂuenced 'by the treatment history of the trees.

Relative to the effects of induced-resistance, the effect of larval parasitism on
approximate reproductive rates was variable (Table 25). In Clone 1, the additional
negative effects of parasitism on R0 ranged from 9-14% and values remained very high.
In Clone 2, parasitism reduced R0 below the effects of induced-resistance by an additional
40% in RIR and additional 30% in RIR/DIR. The data from natural populations yielded
reductions in R0 similar to the values obtained from the experimental populations. The
very high R0 values in all treatments suggest that large additional reduction in fecundity
and/or increases in mortality are required to slow population growth to rates observed in

natural populations.

DISCUSSION

This study was the ﬁrst to use an experimentally created outbreak to emulate natural
defoliation processes and examine the effects of short and long term induction on a North
American defoliator. The experiment was run for four years, an appropriate time span for
emulating the effects of natural forest tent caterpillar outbreaks on trees. Furthermore,
the unique design allowed me to isolate the relative contributions of RIR and DIR to
herbivore performance. 1 found that forest tent caterpillar defoliation elicited both short-
terrn and long-term induced responses in aspen. All three of the defoliation treatments
reduced the growth, pupal mass, and fecundity of forest tent caterpillars relative to

control trees. Contrary to my hypothesis however, the effects of three and four

195

Table 25. Approximate reproductive rates (R0) for each treatment. Reproductive rates were
calculated for surviving females with and without parasitism. Initial cohorts had 30 female
equivalents. Numbers in parentheses are percentage decreases from control values in each row.
Larval parasitism rates in natural outbreak populations were obtained from Parry (1995) (see
text). I Data on parasitism in natural low-density p0pulations corresponding to the controls are
not available so values from my control plots were substituted.

 

 

 

Clone Comparison CTRL RIR DIR RIR/DIR

1 N0 parasitism 114.7 95.0 (-17.2) 78.4 (-31.6) 86.4 (-24.7)
Parasitism 112.7 77.4 (-31.3) 66.8 (40.7) 71.3 (-36.7)

2 NO parasitism 112.3 88.8 (20.9) 93.7 (-16.5) 78.7 (-29.9)
Parasitism 105.5 53.7 (-49.1) 86.7 (-17.8) 55.1 (-47.8)

1 Parasitism rates in 112.7”r 69.7 (-38.1) 66.8 (-40.1) 38.3 (660)
natural populations

2 Parasitism rates in 10551 79.5 (-24.6) 82.9 (-21.4) 44.8 (575)

natural populations

 

196

consecutive years of defoliation were not signiﬁcantly greater than the reduction in

caterpillar performance caused by a single year of defoliation.

Rapid—induced resistance had surprisingly strong effects on forest tent caterpillar
growth. In previous studies, the effects of RIR on the forest tent caterpillar have been
variable. For example, Parry et al. 2000 found that gypsy moth defoliation of Populus
reduced pupal masses of forest tent caterpillar feeding on the same trees by 10%.
Similarly, Robison and Raffa (1997) found that partial defoliation of two different poplar
cultivars signiﬁcantly reduced the growth rates of second instar tent caterpillars. In
contrast, Roth et a1. (1998) found no signiﬁcant effect of defoliation on forest tent
caterpillar feeding on aspen. Cappuccino et al. (1995) found no deleterious effects of
herbivory on forest tent caterpillar, but rather observed increased pupal mass of males
feeding on damaged birch leaves. Differences in defoliation levels among these studies
could account for the variable results. The data of Roth et al. (1998) indicated that most
of the changes in aspen phytochemistry induced by forest tent caterpillar defoliation
occurred in response to feeding by ﬁnal instar (ﬁfth) caterpillars which account for
approximately 80% of the total larval consumption (Hodson 1941). At the time of their
fourth instar bioassays, little phytochemical change was evident between control and
defoliated trees, and by the end of the experiment, trees were only ca. 50% defoliated. In
my study, no treatment differences in relative growth rates of second instars were found.
Growth rates of female ﬁfth instars were only modestly correlated with female pupal
mass while male growth rate and. male pupal mass were not correlated. In addition,
development time was not signiﬁcantly lengthened by the defoliation treatments for

either sex. This suggests that RIR is primarily a response to feeding damage in the latter

197

part of the ﬁnal instar where, in my experiment, defoliation exceeded 70% on all
treatment trees and was higher than 90% on some trees. As Roth et al.’s (1998) data also
suggested, lighter defoliation levels (30-60%) have little effect on forest tent caterpillar
growth, at least in one-year studies (D. Parry and DA. Herms unpublished). Removal of
some aspen leaves has been shown to increase nitrogen in the remaining leaves (Mattson
and Palmer 1988) which could account for the lack of negative effects at lower levels of
defoliation. Conversely, the late onset of RIR may account for the reduced performance
of gypsy moth, white-marked tussock moth, Orgyia leucostigma, and tiger swallowtail,
Papilio canadensis, larvae feeding on birch, sugar maple, and aspen trees moderately

defoliated by forest tent caterpillar earlier in the season (Dankert et al. 1997).

Forest tent caterpillar may compensate for reductions in foliar quality by increasing
relative consumption rates (RCR). Williams et a1. (1998) found that FTC increased
consumption rates by 8% on white oak foliage with lower nitrogen levels. I found that on
aspen, both early and late instars were capable of increasing consumption rates. In second
instars, elevation of RCR on defoliated treatments was only of marginal statistical
signiﬁcance although it may have served to maintain equal RGR across all treatments. In
ﬁnal instars, female consumption rates were 6% higher in response to RIR in Clone l and
were unchanged in Clone 2, while male consumption rates were 15% higher in Clone 1
and 5% higher in Clone 2. Although females increased consumption by 9-12% and 5-
12% in Clone 1 and Clone 2, respectively, in response to DIR and the combined effects .
of RIR and DIR, the changes in RCR by ﬁnal instars were insufﬁcient to prevent a

signiﬁcant reduction in pupal mass.

198

Change in nitrogen utilization is an aspect of compensatory feeding behavior in forest
tent caterpillar not investigated in my study. Williams et al. (1998) fOund that although
RCR increased on low nitrogen foliage, most of the forest tent caterpillar’s ability to
compensate for poor quality food was through a large increase in nitrogen utilization
efﬁciency. Final instars were able to increase nitrogen utilization efﬁciency by as much
as 20% when consuming low nitrogen leaves. The mechanism underlying this
physiological behavior is not understood, although it would undoubtedly be adaptive for
an outbreak species that frequently must process damaged foliage with low nitrogen

levels.

The effects of DIR on herbivore performance have been best studied in the mountain
birch systems of northern Fennoscandia. In the year following defoliation, the quality of
leaves of mountain birch are reduced for its primary herbivore, the autumnal moth, as
well as other species and these effects may linger for several years (Tuomi et al. 1990).
Mechanisms underlying the reduction in performance of caterpillars on previously
defoliated mountain birch are not fully understood. Elevation of secondary phenolics
including gallotannins and proanthocyanidins as well as decreases in foliar nitrogen may
contribute to observed reductions in pupal mass and fecundity (Haukioja et al. 1985,
Kaitaniemi et al. 1998). Some research has suggested that a suite of traits may interact
and undergo phenological shifts so that no single measure of leaf quality had signiﬁcant
effects on the growth of both early and late instars (Kause et al. 1999). Instead,
gallotannins, proanthocyanidins, foliar nitrogen, and water all contributed at different
times to the performance of larvae. Reductions in autumnal moth fecundity exhibit large

variation depending on the year and the outbreak, although more than 90% of the studies

199

reviewed showed some DIR related reduction in fecundity (Ruohomaki et al. 1992,

Kaitaniemi et al. 1999a).

Only one study in the mountain birch system has examined the effects of multiple
seasons of defoliation on autumnal moth performance, a species with outbreaks that
persist from 1-3 years at any one locality (Bylund 1995). Kaitaniemi et al. (1999b) found
that two consecutive years of defoliation did not have greater effects than a single season
of 75% defoliation on the growth and fecundity of either autumnal moth or the winter
moth, Operophtera brumata (Lepidoptera: Geometridae). In contrast, defoliation of trees
for two or more years in succession in other systems has had additive deleterious effects
on the herbivores assayed. With each successive year of defoliation of black oak,
Valentine et al. (1983) found increasingly negative effects on gypsy moth pupal mass.
Similarly, Werner (1979) found that survival of black-marked spear moth decreased with
consecutive years of defoliation of Alaska paper birch, while Clausen et al. (1991) found
that defoliation of aspen from 1-3 years had additive negative effects on pupal mass of
the large aspen tortrix. I found that three consecutive years of defoliation reduced
fecundity by 13.5% and 18.7% in Clones 1 and 2, respectively, translating into 37 and 49
fewer eggs per female. However, these effects were not signiﬁcantly lower than those
caused by RIR resulting from a single defoliation. Furthermore, although the reduction in
fecundity was greatest in the four-year defoliation treatment (ca. 22% in both clones), the
combined effects of RIR and DIR did not differ from the single-year defoliation
treatment. This suggests that there may be an upper threshold for induced resistance in

aspen beyond which additional response is not physiologically possible.

200

The basis for induced-resistance in aspen is not well understood. Levels of primary
compounds, including water and nitrogen, decrease in damaged leaves (Roth et al. 1998).
Clausen et al. (1991) suggested that leaf size was one of the best predictors of large aspen
tortrix performance on defoliated trees. Reductions in leaf size are a common response of
trees in the year following severe defoliation and are thought to be a function of the
decrease in available nitrogen (Tuomi et al. 1984, 1990). Smaller leaf sizes are often
accompanied by increase in secondary phenolics and in some cases leaf toughness.
Studies have shown that the constitutive level of nitrogen is the best predictor of forest
tent caterpillar performance among aspen clones (Hwang and Lindroth, 1995, 1997), thus
changes in foliar nitrogen levels may also determine performance on defoliated trees. In
other Populus species, defoliation-induced reductions in nitrogen were associated with
reduced performance of forest tent caterpillar (Robison and Raffa 1997, Parry et al.

2000a)

Phenolic glycosides and condensed tannins, products of the shikimic acid pathway,
dominate the secondary chemistry of aspen. Clausen et al. (1989) suggested that RIR in
aspen was due to increased production of phenolic glycosides and developed a pathway
to describe the phytochemical changes involved in the RIR response following herbivory
(see also Clausen et a1. 1991, Lindroth and Hwang 1996). They suggested that the
phenolic glycosides salicortin and tremulacin increase signiﬁcantly in the ﬁrst 24 hours
following feeding damage. When ingested by caterpillars, these unstable compounds are
rapidly converted to salicin, tremuloiden and the byproduct 6-HCH. When salicortin,
tremulacin, and the by-product 6-hydroxy-2-cyclohexenone (6-HCH) were incorporated

in to artiﬁcial diets, they had deleterious effects on the growth of large aspen tortix.

201

However, in subsequent experiments, inducibility has been inconsistent. Roth et al.
(1998) showed that forest tent caterpillar herbivory had no signiﬁcant on levels of
salicortin, while tremulacin increased initially but then decreased to levels below that of
controls. In other Populus species, the response of phenolic glycosides to defoliation has
also been unpredictable with signiﬁcant increases in one year and no change in the
following year (Parry et al. 2000a). In contrast to the instability of phenolic glycosides,
condensed tannins concentrations were elevated by defoliation in all of these
experiments. However, condensed tannins in aspen and other poplars while readily
inducible, appear to have little effect on a variety of herbivorous insects (Roth et al. 1998,
Ayres et al. 1997, Lindroth and Kinney 1998, Parry et al. 2000a). Clausen et al. (1991)
suggested that RIR and DIR in aspen are related processes. My experiment agrees with
this premise because RIR and DIR individually had effects on larval performance of

similar magnitude to that of RIR and DIR combined.

The costs of induction relative to any accrued beneﬁts is an aspect of induced-
resistance in aspen and other trees that has received little attention. In large-scale
outbreaks of the forest tent caterpillar in boreal forests, virtually all woody deciduous
species are defoliated, and few if any aspen trees are spared even in areas where hundreds
of aspen clones/hectare occur. This suggests that despite wide variability in constitutive
and induced resistance in aspen clones, the end result may still be complete defoliation
during outbreaks (Mattson et a1. 1991). The proposition that variability in defoliation
intensity is due to clonal differences in phenolic glycosides (e.g., Lindroth and Hwang
1996), is speculative and has not been conﬁrmed experimentally. In extensive ground

and aerial surveys of forest tent caterpillar outbreaks in Canada, variation in defoliation

202

intensity appeared to be driven more by features of the landscape than any distinct clonal
attributes (D. Parry, personal observations). Tree phenology, oviposition preferences by
females, and spatial heterogeneity in mortality from natural enemies may all contribute to
variation in defoliation levels among trees, and should be investigated before differences
in herbivory are ascribed to any phytochemical attribute. Thus, while there are
phytochemically based differences in the performance of forest tent caterpillar among

clones, at outbreak densities, trees are defoliated irrespective of biochemical components.

In aspen, chemical defenses are costly in terms of their trade-off with growth (Hwang
and Lindroth (1997). If defoliation is inevitable for 2-5 years in succession, trees with the
greatest investment in resistance may actually incur the highest ﬁtness costs since they
will be defoliated anyway, but will grow more slowly and be less competitive in the
absence of defoliation (i.e., Herms and Mattson 1992). Mattson et a1. (1991) suggested
that in aspen forests susceptible to frequent forest tent caterpillar outbreak, tolerance to
defoliation might represent the optimal life-history strategy. Assessing the beneﬁts and
costs of resource allocation to resistance during widespread outbreaks is likely to be a

productive area of future research.

Parasitoids undoubtedly play an important role in the population dynamics of forest
tent caterpillar (Sippell 1957, Hodson 1977, Witter and Kulman 1979, Parry 1995, Parry
et al. 1997). While parasitoids probably do not initiate population decline, the action of
larval and pupal parasitoids in older outbreaks may hasten the demise of populations and
drive them to very low levels. I found no evidence of synergistic or antagonistic effects of
either RIR or DIR and the dominant larval parasitoids. Highest parasitism levels occurred

in the two treatments where, in addition to the bioassay larvae, large numbers of

203

caterpillars were used to defoliate the trees. In the year that parasitism rates were
assessed, the control and DIR treatment had only the bioassay larvae feeding on the trees.
Conversely, the RIR and COMB treatments had high densities of larvae used to defoliate
the trees in addition to the bioassay larvae. Parasitism rates were equivalent among the
two treatments with low larval densities (CNTL and DIR) and were signiﬁcantly higher
but equivalent between the two high-density treatments (RIR and COMB). This result
indicates that the response of the tachinids was driven by spatial responses to host density
rather than an effect of the induction treatments. Spatial density dependent responses

have been shown previously for both P. pachypyga and L. exul (Parry et al. 1997).

Although only marginally signiﬁcant, there were intriguing differences in parasitism
rates among the two clones. Since the aspen clones were separated by only a short
distance, were similar in foliar area, tree size, proximity to clearings, and insolation, and
caterpillar densities were approximately equal, the markedly higher rates of parasitism in
Clone 2 may reﬂect an interaction between phytochemistry and tachinid olfaction. Both
L. exul and P. pachypyga locate tent caterpillars using volatiles given off by leaves
damaged by larval feeding (Bess 1936, Mondor and Roland 1997, 1998). Thus, there may
be subtle differences in the attractiveness of clone speciﬁc combinations of the
phytochemicals important in host-detection. To my knowledge, mediation of interactions
between parasitoids and hosts through intraspeciﬁc variation in chemistry of trees has not
been explored in a ﬁeld environment. An interaction between natural enemies and
phytochemical variation in host plants could be an important component of spatial

heterogeneity in herbivore population dynamics.

204

A longstanding criticism of the role of induced-responses in plants is their relevance
for insect population dynamics. Fowler and Lawton (1985) suggested that many of the
observed effects of induced resistance, while statistically signiﬁcant, are small, easily
overwhelmed, and rendered unimportant by other processes acting on insect populations.
While their point is well taken, the criticism might be most valid for studies looking at
RIR elicited by low-level herbivory and at the scale of individual leaves or branches.
Such research has revealed large variability in the magnitude and direction of the effects
(see Karban and Baldwin 1997 for review). In the context of outbreak species, however,
few would argue that reductions in fecundity of 40% or more caused by DIR,
documented in some experiments with autumnal moth on mountain birch (e.g.,

Ruohomaki et al. 1992), do not have measurable impact on population level processes.

For DIR to play an important role in population dynamics, signiﬁcant decreases in
fecundity must be manifested at the population level. It is well established that large
reductions in fecundity are characteristic of declining populations of many outbreak
folivores (Mason et al. 1977, Baltensweiler and Fischlin 1988, Carter et a1. 1991). Similar
trends are evident in forest tent caterpillar populations. Following several years of
defoliation, forest tent caterpillar egg masses can be 50% smaller than early in an
outbreak (Ives 1971, Witter et al. 1975, D. Parry, unpublished). However, DIR is not the
only mechanism operating in outbreak population that can reduce fecundity, thus it is

important to consider alternative hypotheses.

Undoubtedly, a lack of food contributes to reductions in the average egg mass size of
defoliating Lepidoptera, especially in years of peak densities when virtually all edible

foliage in a forest is consumed. Hodson (1941) found that partial starvation of ﬁnal instar

205

forest tent caterpillar reduced fecundity by 45%. Field data also indicate that female
pupal mass can be dramatically reduced in stands with complete defoliation (Hodson
1941, Parry et al. 2000b). In a study of forest tent caterpillar population dynamics in
Minnesota, Witter et al. (1975) showed that fecundity was lowest in the year of highest
density suggesting that intraspeciﬁc competition for foliage was a primary determinant of
clutch size. In high-density populations of forest tent caterpillar, most of the aspen foliage
is removed prior to the ﬁnal larval molt. Final instars must then complete development on
secondary host trees and shrubs after they completely consume all remaining aspen
foliage. The quality of many secondary hosts is considerably lower than aspen. When
ﬁnal instars were switched from aspen to foliage from the ﬁve most abundant secondary
host species in boreal aspen forests, fecundity was reduced by 5-31% (D. Parry & J.R.
Spence, unpublished). Thus, even in forests where densities are not sufﬁcient to cause
starvation, fecundity can be reduced when aspen foliage is depleted prior to pupation. In
outbreak species where adults of both sexes are capable of efﬁcient dispersal, such as the
con‘iferophagous budworms, Choristoneura spp., large aspen tortrix, and tent caterpillars,
emigration of moths from large outbreaks may swamp processes operating in local
populations. Thus low or high fecundity may be a property of the dispersing individuals

and not of processes operating in the forest in which oviposition occurs.

While factors such as RIR and competition for resources can account for reductions
in fecundity within a season, the effect is directly density dependent and fecundity should
immediately rebound following population decrease. However, in some defoliator
populations, fecundity does not appear to follow this pattern. Myers and Kukan (1995)

showed that for western tent caterpillars, fecundity appeared to increase, or at least

206

remained stable, even as densities were increasing. F ecundity decreased in later outbreak
years, but remained low even as populations collapsed. There is some evidence that
similar patterns in fecundity may occur in forest tent caterpillar. A long-term study in
Minnesota suggested that fecundity ﬂuctuated considerably before declining during the
last few years of the outbreak (Witter et al. 1975). Unfortunately, the study was initiated
several years into the outbreak and values from the ﬁrst few years were not recorded.
Batzer et al. (1994) monitored populations for ﬁve years and found that fecundity
decreased in three stands as outbreaks declined, but fecundity changed little in one stand
while an increase was observed in another stand. During an outbreak in Alberta, Canada,
fecundity was highest in the ﬁrst two years of outbreak and was lower in later years (Ives
1971). These studies suggest that some time-lagged process or processes contribute to

changes in forest tent caterpillar fecundity over the course of an outbreak.

A number of authors have suggested that reduced fecundity and survival caused by
induced-resistance may be responsible for driving cycles of forest defoliators (e.g., Benz
1974, Haukioja 1980, Rhoades 1983). Others have been critical of an important role for
DIR in the forest insect populations (e.g., Myers 1988a, b, 1993). The main criticism of
hypotheses based on plant quality is that outbreak populations may collapse
simultaneously across large regions irrespective of their defoliation histories.
Furthermore, manipulative studies where populations were cropped or moved to

undefoliated areas failed to prevent declines (Myers 1981, 1990).

Myers (1993) suggested that instead of induced-resistance, sublethal infection by
baculoviruses such as NPV is a more parsimonious explanation for fecundity declines.

This hypothesis proposes that virus loads increase in the environment when populations

207

are high. Increasing numbers of females with sublethal infection lowers average
fecundity. Since NPV may persist for some years in the environment, there may be a time
lag in the recovery of fecundity levels to preoutbreak levels. While NPV treatments are
associated with modest reductions of western tent caterpillar in laboratory studies
(Rothman and Myers 1994), there were no differences in fecundity among infected and
uninfected high-density populations in a ﬁeld study (Rothman 1997). Moreover,
extensive research with autumnal moth and larch budmoth, two systems which show the
most regular cycles, have failed to demonstrate that pathogens are of much importance
(Baltensweiler and F ischlin 1988, Bylund 1995). There is also no evidence to suggest that
sublethal NPV infections impact fecundity of gypsy moth (Murray and Elkinton 1989).
Thus by themselves, sublethal levels of baculovirus does not appear to provide a

satisfactory general mechanism for explaining cycles in forest defoliator populations.

Considerable research over the past three decades has focused on identifying a single
causal agent that drives the population dynamics of cyclically outbreaking forest
defoliators. I suggest that this effort for the most part has been fruitless. While most
researchers now agree that delayed-density dependent factors must be involved (Turchin
1990, 1995), several processes are thought to be capable of producing cycles (e.g.,
Berryman et al. 1987, Ginzburg and Taneyhill 1994, Berryman 1996, Underwood 1999).
In fact, recent studies have suggested that even within a species, the relative role of
different agents may vary in different population peaks. For example, a granulosis virus
infection thought responsible for the collapse of larch budmoth populations in
Switzerland was not found to have any impact on previous or subsequent outbreaks in the

same area (Baltensweiler and Fischlin 1988). Similarly, Bylund (1995) suggested that

208

different agents were responsible for the collapse of outbreak autumnal moth populations
during three separate population peaks in 1955, 1965, and 1987. Clearly, induced-
resistance plays a role in the population dynamics of some outbreak species but rather
than operating independently, it should be viewed in context with contemporaneously
acting natural enemies. In addition, the interactive effects of plant-quality and natural
enemies need to be coupled with spatially synchronizing density-dependent and density-
independent processes like dispersal and weather. I echo the call of other authors (e.g.,
Hunter and Price 1992, Hunter et al. 1997, Karban and Baldwin 1997) that plurality is
needed and that viewing plant-quality driven and natural enemy driven hypotheses as

mutually exclusive is artiﬁcial and counter productive.

I clearly showed that that delayed-density dependent declines in fecundity could be
attributed solely to changes in tree quality in this study. Yet DIR induced by defoliation
at most decreased fecundity by 22% which is insufﬁcient to account for the 50% or more
declines in fecundity that have been recorded in natural forest tent caterpillar outbreaks.
Furthermore, parasitoids had a much greater impact on reproductive rates than the effects
of induced-resistance. Neither factor was sufﬁcient to slow the approximate reproductive
rates of my experimental populations to levels found in nature. My results suggest that
the integrating the effects of agents acting contemporaneously on populations is the next

major hurdle in understanding the mechanisms underlying cycles of forest Lepidoptera.

ACKNOWLEDGEMENTS
I gratefully acknowledge the exemplary ﬁeld and laboratory assistance of Sara Sanders,

Christina Schoen, and Linsey White. I thank Frank Telewski, Department of Botany,

209

Michigan State University for providing the necessary permits to conduct research in the
study area. This study was partially funded through a NASA-TECO grant to D. Parry and
W.J. Mattson, and the following grants to D. Parry: Michigan State University Hutson
Fund, Sigma Xi Grant-in-Aid of Research, and a Dissertation Completion Fellowship,

The Graduate School, Michigan State University.

LITERATURE CITED
Agrawal, AA. 2000. Mechanisms, ecological consequences and agricultural implications

of tri-trophic interactions. Current Opinion in Plant Biology 3: 329-335.

Ayres, M.P., T.P. Clausen, S.F. MacLean, Jr., A.M. Redman and PB. Reichardt. 1997.
Diversity of structure and antiherbivore activity in condensed tannins. Ecology 78:

1696-1712.

Baltensweiler, W. and A. Fischlin. 1988. The larch budmoth in the Alps. Pp. 331-351 in
AA. Berryman, ed, Dynamics of Forest Insect Populations. Patterns, Causes, and

Implications. Plenum Press, New York.

Barnes, B.V. 1969. Natural variation and delineation of clones of Populus tremuloides

and P. grandidentata in northern lower Michigan. Silvae Genetic 18: 130-142.

Batzer, H., MP. Martin, W.J. Mattson and WE. Miller. 1994. The forest tent caterpillar
in aspen stands: Distribution and density estimation of four life stages in four

vegetation strata. Forest Science 41: 99-121.

Benz, G. 1974. Negative feedback by competition for food and space, and by cyclic

induced changes in the nutritional base as regulatory principles in the population

210

dynamics of the larch budmoth, Zieraplzera dineana (Guenee) (Lepidoptera:

Tortricidae). Zietschrift fur Angewandte Entomologie 76: 196-228.

Berryman, A.A., N.C. Stenseth and AS. Isaev. 1987. Natural regulation of herbivorous

forest insect populations. 0ecologia 71: 174-184.

Berryman, AA. 1996. What causes population cycles of forest Lepidoptera? Trends in

Ecology and Evolution 11: 28-32.

Bess, HA. 1936. The biology of Leschenaultia exul Townsend, a tachinid parasite of
Malacosoma americana Fabricius and Malacosoma disstria Hiibner. Annals of

the Entomological Society ofAmerica 29: 593-613.

Bryant, J.P., P.B. Riechardt, T.P. Clausen and R.A.Wemer. 1993. Effects of mineral
nutrition on delayed inducible resistance in Alaska paper birch. Ecology 74: 2072-

2084.

Bourchier, RS. 1991. Growth and development of Compsilura concinnata (Miegan)

(Diptera: Tachinidae) parasitizing gypsy moth larvae feeding on tannin diets. The

Canadian Entomologist 123: 1047-1055.

Byland, H. 1995. Long-term interactions between the autumnal moth and mountain birch:
the role of resources, competitors, natural enemies, and weather. PhD. thesis.

Swedish University of Agricultural Sciences, Uppsala, Sweden.

Cappuccino, N., H. Damman and IF Dubuc. 1995. Spatial behavior and temporal
dynamics of outbreak and nonoutbreak species. Pp. 65-82 in N. Cappuccino and
PW. Price, eds,. Population Dynamics: New Approaches and Syntheses.

Academic Press, San Diego.

211

Carter, M.R., F.W. Ravlin and ML. McManus. 1991. Changes in gypsy moth
(Lepidoptera: Lymantriidae) fecundity and male wing length resulting from

defoliation. Environmental Entomology 20: 1042-1047.

Clausen, T.P., P.B. Reichardt, J.P. Bryant, R.A. Werner, K. Post and K. Frisbee. 1989. A
chemical model for short-term induction in quaking aspen (Populus tremuloides)

foliage against herbivores. Journal of Chemical Ecology 15: 2335-2346.

Clausen, T.P., P.B. Reichardt, JP. Bryant and RA. Werner. 1991. Long-term and short-
term induction in quaking aspen: related phenomena? Pp.71-83 in D.W. Tallamy
and M.J. Raupp, eds, Phytochemical Induction by Herbivores. John Wiley &

Sons, Inc., New York.

Dankert, B.A., D.A. Herms, D. Parry, J.M. Scriber and LA. Haas. 1997. Mediation of
interspeciﬁc competition between folivores through defoliation induced changes
in host quality. Pp. 71-88 in A. Ramen, eds, Ecology and Evolution of Plant
Feeding Insects in Natural and Man-Made Environments. International Scientiﬁc

Publications. New Delhi, India

Eller, F.J., J .H. Tumlinson and W.J. Lewis.1988. Beneﬁcial arthropod behavior mediated
by airborne semiochemicals: Source of volatiles mediating the host location ﬂight
behavior of Microplitis croceipes (Cresson) (Hymenoptera: Braconidae), a
parasitoid of Heliothis zea (Boddie) (Lepidoptera: Noctuidae). Environmental

Entomology 17: 745-753.

Fowler, S.V. and J. H. Lawton.1985. Rapidly—induced defenses and talking trees: the

devil’s advocate position. The American Naturalist 126: 181-195.

212

 

Ginzburg, LR. and DE. Taneyhill. 1994. Population cycles of forest Lepidoptera — a

maternal effects hypothesis. Journal of Animal Ecology 63: 79-92.
Gordon, HT. 1968. Quantitative aspects of insect nutrition. Zoology 8: 131-138.

Haukioja, E. 1980. On the role of plant defenses in the ﬂuctuation of herbivore

populations. Oikos 35:202-213.

Haukioja, E. 1982. Inducible defenses of white birch to a geometrid defoliator, Epirrita
autumnata. Proceedings of the 5th International Symposium on Insect-Plant

Relationships, Wageningen.

Haukioja, E. J. Suomela and S. Neuvonen. 1985. Long-term inducible resistance in birch

foliage: triggering cues and efﬁcacy on a defoliator. 0ecologia 65: 363-369.

Havill, NP. and K.F. Raffa. 2000. Compound effects of induced plant responses on
insect herbivores and parasitoids: implications for tritrophic interactions.

Ecological Entomology 25: 171-179.

Hemming, J.D.C. and R.L.Lindroth. 1995. Intraspeciﬁc variation in aspen
phytochemistry: effects on performance of gypsy moths and forest tent

caterpillars. 0ecologia 103: 79-88.

Herms, DA. and W.J. Mattson. 1992. The dilemma of plants — to grow or defend.

Quarterly Review of Biology 67: 283-335.

Hildahl, V. and WA. Reeks. 1960. Outbreaks of the forest tent caterpillar and their
effects on stands of trembling aspen in Manitoba and Saskatchewan. The

Canadian Entomologist 92: 199-209.

213

 

 

 

Hodson, AC. 1941. An ecological study of the forest tent caterpillar, Malacosoma

disstria an. Minnesota Agricultural Experiment Station Technical Bulletin 148.

Hodson, AC. 1977. Some aspects of forest tent caterpillar population dynamics.

Minnesota Agricultural Experiment Station Technical Bulletin 310: 5-16.

Hunter, MD. and PW. Price. 1992. Playing chutes and ladders — heterogeneity and the
relative role of bottom-up and top-down forces in natural communities. Ecology

73: 724-732.

Hunter, M.D., G.C. Varley and GR. Gradwell. 1997. Estimating the relative roles of top-
down and bottom-up forces on insect herbivore populations: A classic study

revisited. Proceedings of the National Academy of Science. 94: 9176-9181.

Hwang, S.Y. and R.L. Lindroth. 1997. Clonal variation in foliar chemistry of aspen:

effects on gypsy moths and forest tent caterpillars. 0ecologia 111: 99-108.

Ives, W.G.H. 1971. The forest tent caterpillar in Alberta. Environment Canada, Canadian

Forest Service, Northern Forestry Centre, Internal Report NOR-4.

Kaitaniemi, P. K. Ruohomaki, V. Ossipov, E. Haukioja and K. Pihlaja. 1998. Delayed
induced changes in the biochemical composition of host plant leaves during an

insect outbreak. 0ecologia 116: 182-190.

Kaitaniemi, P., K. Ruohomaki, T. Tammaru and E. Haukioja. 1999a. Induced resistance
of host tree foliage during and after natural insect outbreak. Journal of Animal

Ecology 68: 382-389.

214

Kaitaniemi, P., S. Neuvonen and T. Nyyssonen.l999b. Effects of cumulative defoliations

on growth, reproduction, and insect resistance in mountain birch. Ecology 80:

524-532.

Karban, R. and LT. Baldwin. 1997. Induced Responses to Herbivory. University of

Chicago Press, Chicago.

Kause A, V. Ossipov, E. Haukioja, K. Lempa, S. Hanhimaki and S. Ossipova. 1999.
Multiplicity of biochemical factors determining quality of growing birch leaves.

Oecologia 120:102-112.

Lindroth, R.L. and S.-Y. Hwang. 1996. Diversity, redundancy, and multiplicity in
chemical defense systems of aspen. ), Pp. 25-56 in T. Romeo, J.A. Saunders and
P. Barbosa, eds, Phytochemical Diversity and Redundancy in Ecological

Interactions. Plenum Press, New York.

Lindroth, R.L and K.K. Kinney. 1998. Consequences of enriched atmospheric CO2 and
defoliation for foliar chemistry and gypsy moth performance. Journal of

Chemical Ecology 24:1677-1695.

Mason, R.R., R.C. Beckwith and HG. Paul. 1977. Fecundity reduction in a collapse of a
Douglas-ﬁr tussock moth outbreak in Northeast Oregon. Environmental

Entomology 6: 623-626.

Mattson, W.J. and SR. Palmer. 1988. Changes in levels of foliar minerals and phenolics
in trembling aspen, Populus tremuloides, in response to artiﬁcial defoliation. Pp.
157-169 in W.J. Mattson, J. Levieux and C. Bernard-Dagan, eds, Mechanisms of

Woody Plant Defenses against Insects, Springer, New York.

215

Mattson, W.J., D.A. Herms, J.A. Witter and DC. Allen. 1991. Woody plant grazing
systems: North American outbreak folivores and their host plants. Pp. 53-84 in
Y.N. Baranchikov, W.J. Mattson, F .P. Hain and TL. Payne, eds, Forest Insect
Guilds: Patterns of Interactions with Host Trees (eds),. United States Department

of Agriculture, Forest Service. General Technical Report NE-153.

Mondor, EB. and J. Roland. 1997. Host locating behaviour of Leschenaultia exul and
Patelloa pachypyga: two tachinid parasitoids of the forest tent caterpillar,

Malacosoma disstria. Entomologia Experimentalis et Applicata 85: 161-168.

Mondor, EB. and J. Roland. 1998. Host searching and oviposition by Leschenaultia exul,
a tachinid parasitoid of the forest tent caterpillar, Malacosoma disstria. Journal of

Insect Behavior 11: 583-592.

Monteith, LG. 1964. Inﬂuence of the health of the food plant of the host on host-ﬁnding

by tachinid parasites. The Canadian Entomologist 96: 1477-1482.

Murray, K.D. and J .S. Elkinton. 1989. Environmental contamination of egg masses as a
major component of transgenerational transmission of gypsy moth nuclear

polyhedrosis virus (LdMNPV). Journal of Invertebrate Pathology 53: 324-334.

Myers, J .H. 1981. Interactions between western tent caterpillars and wild rose: A test of

some general plant-herbivore hypotheses. Journal of Animal Ecology 50: 11-25.

Myers, J.H. 1988a. Can a general hypothesis explain population cycles of forest

Lepidoptera? Advances in Ecological Research 18: 179-242.

216

Myers, J.H. 1988b. The induced-defense hypothesis: does it apply to the population
dynamics of insects. Pp. 345-366 in KC. Spencer, ed, Chemical Mediation of

Coevolution. Academic Press, San Diego.

Myers, J.H. 1990. Population cycles of western tent caterpillars: Experimental

introductions and synchrony of ﬂuctuations. Ecology 71: 986-995.

Myers, J.H. 1993. P0pulation outbreaks in forest LepidOptera. American Scientist 81:

240-251.

Myers, J.H. and B. Kukan. 1995. Changes in the fecundity of tent caterpillars — a
correlated character of disease resistance or sublethal effect of disease? 0ecologia

103: 475-480.

Neuvonen, S. and E. Haukioja. 1991. The effects of inducible resistance in host foliage
on birch-feeding herbivores. Pp. 277-291 in D.W. Tallamy and M.J. Raupp, eds,

Phytochemical Induction by Herbivores. John Wiley & Sons, Inc., New York.

Osier, T.L., S.Y. Hwang and R.L. Lindroth. 2000. Effects of phytochemical variation in
quaking aspen Populus tremuloides clones on gypsy moth, Lymantria dispar,

performance in the ﬁeld and laboratory. Ecological Entomology 25: 197-207.

Parry, D. 1995. Larval and pupal parasitism of the forest tent caterpillar in Alberta. The

Canadian Entomologist 127: 877-893.

Parry, D., J.R. Spence and W.J.A. Volney. 1997. Responses of natural enemies to
experimentally increased forest tent caterpillar populations. Ecological

Entomology 22: 97-108.

217

Parry, D., J .R. Spence & W.J.A Volney. 1998. Budbreak phenology and natural enemies
mediate survival of early instar forest tent caterpillar (Lepidoptera:

Lasiocampidae). Environmental Entomology 27: 1368-1374.

Parry, D., D.A. Herms, K.R. Kosola, J. M. Scriber and R. Julkenen-Tiitto. 2000a.
Emulating Insect Outbreak: Experimentally Testing the Effects of Cumulative
Defoliations on Induced—Resistance and Caterpillar Herbivory. Ecology. In

preparation.

Parry, D., R.A. Goyer & G.A. Lenhard. 2000b. Macrogeographic variation in egg and
neonate size of the forest tent caterpillar, Malacosoma disstria. Ecological

Entomology. In press.

Peterson, EB. and NM Peterson. 1992. Ecology, management, and use of aspen and
balsam poplar in the Prairie Provinces. Forestry Canada, Northwest Region,

Northern Forestry Centre Special Report 1.

Rhoades, DF. 1983. Herbivore population dynamics and plant chemistry. Pp. 155-220 in
RF. Denno and MS. McClure, eds, Variable Plants and Herbivores in Natural

and Managed Systems, Academic Press, San Diego.

Robison, DJ. and K.F. Raffa.l997. Effects of constitutive and inducible plant traits of
hybrid poplars on forest tent caterpillar feeding and population ecology. Forest

Science 43: 252-267.

Roland, J. K.E. Denford and L. Jimenz. 1995. Bomeol as an attractant for Cyzenis
albicans, a tachinid parasitoid of the winter moth, Operophtera brumata L.

(Lepidoptera: Geometridae). The Canadian Entomologist 127: 413-421.

218

Roth S, C. Knorr R.L. Lindroth. 1997. Dietary phenolics affects performance of the
gypsy moth (Lepidoptera: Lymantriidae) and its parasitoid Cotesia melanoscela

(Hymenoptera: Braconidae). Environmental Entomology 26: 668-671.

Roth, 8., R.L. Lindroth, J.C. Volin and EL. Kruger. 1998. Enriched atmospheric CO2
and defoliation: effects on tree chemistry and insect performance. Global Change

Biology 4: 419-430.

Rothman, L.D. 1997. Immediate and delayed effects of a viral pathogen and density on

tent caterpillar performance. Ecology 78: 1481-1493.

Rothman, L.D. and J .H. Myers. 1994. Effect of nuclear polyhedrosis virus treatment on
reproductive potential of western tent caterpillar (Lepidoptera: Lasiocampidae).

Environmental Entomology 23: 864-869.

Ruohomaki, K., S. Hanirnaki, E. Haukioja, L. Iso—Iivari, S. Neuvonen, P. Niemela and S.
Suomela. 1992. Variability in the efﬁcacy of delayed-inducible resistance in

mountain birch. Entomologia Experimentalis et Applicata 62: 107-115.

Sippell, W.L. 1957. A study of the forest tent caterpillar Malacosoma disstria an. and
its parasite complex in Ontario. PhD. thesis. University of Michigan, Ann Arbor,

Michigan.

Sippell, W.L. 1961. Key to the puparia of common dipterous parasites of the forest tent
caterpillar, Malacosoma disstria an. in Ontario. The Canadian Entomologist 93:

433-436.

219

 

 

 

Sippell, W.L. 1962. Outbreaks of the forest tent caterpillar Malacosoma disstria an., a
periodic defoliator of broad-leafed trees in Ontario. The Canadian Entomologist

94: 408-416.
Southwood, T.R.E. 1991. Ecological Methods, 3rd Edition. Chapman and Hall.

Stehr, F.W. and BF. Cook.1968. A revision of the genus Malacosoma Hiibner in North
America (Lepidoptera: Lasiocampidae): Systematics, biology, immatures, and

parasites. United States National Museum Bulletin 276.

Thaler, J.S. 1997, Jasmonate-inducible plant defences cause increased parasitism of

herbivores. Nature 399: 686-688

Tuomi, J ., P. Niemela, E. Haukioja, S. Siren and S. Neuvonen. 1984. Nutrient stress: an
explanation for plant anti-herbivore responses to defoliation. 0ecologia 61: 208-

210

Tuomi, J., P. Niemela and S. Siren.l990. The Panglossian paradigm and delayed-

inducible accumulation of foliar phenolics in mountain birch. Oikos 59: 399-410.

Turchin, P. 1990. Rarity of density-dependence or pOpulation regulation with lags?

Nature 344: 660-663.

Turchin, P. 1995. Population regulation: Old arguments and a new synthesis. Pp. 19-40 in
N. Cappuccino and PW. Price, eds, Population Dynamics: New Approaches and

Syntheses. Academic Press, San Diego.

Turlings, T.C.J., P.J. McCall, H.T. Alborn and J.H. Tumlinson. 1993. An elicitor in
caterpillar oral secretions that induces corn seedlings to emit chemical signals

attractive to parasitic wasps. Journal of Chemical Ecology 19: 411-425.

220

Underwood, N. 1999. The inﬂuence of plant and herbivore characteristics on the
interaction between induced resistance and herbivore population dynamics. The

American Naturalist 153: 283-294.

Valentine, H.T., W.E. Wallner and PM. Wargo. Nutritional changes in host foliage
during and after defoliation, and their relation to the weight of gypsy moth pupae.

0ecologia 57: 298-302.

Waldbauer, GP. 1968. The consumption and utilization of food by insects. Advances in

Insect Physiology 5: 229-289.

Wallner, W.E. and GS. Walton. 1979. Host defoliation: a possible determinant of gypsy
moth population quality. Annals of the Entomological Society of America 72: 62-67.

Werner, RA. 1979. Inﬂuence of host foliage on development, survival, fecundity and
oviposition of the spear-marked black moth, Rheumaptera hastata (Lepidoptera:

Geometridae). The Canadian Entomologist 111: 317-322.

Williams, D. J. M., D. Parry and D.W. Langor. 1996. Sampling and identiﬁcation of
forest tent caterpillar parasitoids in the Prairie Provinces. Natural Resources

Canada, Canadian Forest Service, Northwest Region, Northern Forestry Centre

NOR-X-345.

Williams, R.S., D. E. Lincoln and R.J. Norby. 1998. Leaf age effects of elevated C02-
grown white oak leaves on spring-feeding lepidopterans. Global Change Biology

4: 235-246.

Witter, J .A. and HM. Kulman. 1979. The parasite complex of the forest tent caterpillar in

northern Minnesota. Environmental Entomology 8: 723-731.

221

Witter, J .A., W.J. Mattson and HM. Kulman. 1975. Numerical analysis of a forest tent
caterpillar (Lepidoptera: Lasiocampidae) outbreak in Northern Minnesota. The

Canadian Entomologist 107: 837-854.

222

 

 

CHAPTER 4:
CONCLUSIONS

My dissertation research focused on four main themes outlined in the introductory
chapter. The primary objectives were to determine if (1) consecutive years of defoliation
have cumulative effects on tree physiology and if this in turn successively reduces
herbivorous insect performance; (2) induced-resistance has strength sufﬁcient to account
for the large (ca. 50%) reductions in fecundity found in many declining populations of
outbreak Lepidoptera; (3) what are the indirect competitive effects of defoliation caused
by a dominant outbreak species on other members of the folivorous lepidopteran
community; (4) can RIR or DIR interact to enhance or reduce the susceptibility of
caterpillar to mortality from higher trophic levels such as parasitoids?

Three consecutive years of defoliation did not have cumulative effects on the growth,
development time, and fecundity of lepidopterans in the hybrid poplar system. A single
year of severe defoliation had the same magnitude of effect on gypsy moth growth rate,
pupal mass, and fecundity as two consecutive years. This was evident from the
comparison of short and long-term bioassays in 1997, which were preceded by one year
of mostly light defoliation, and the 1998 assays that were preceded by severe defoliation
in 1997 in addition to light defoliation in 1996. Similarly in the aspen system, reductions
in growth, pupal mass, and fecundity experienced by forest tent caterpillar was no greater
following four consecutive years of defoliation than after a single year of defoliation.
These results concur with those of Kaitaniemi et al. (1999) who found that the effects of
two years of 75% defoliation of mountain birch on two lepidopteran folivores was greater

than a single year of defoliation. Conversely, cumulative defoliations of black oak did

223

 

 

 

 

have cumulative negative impacts on gypsy moth pupal mass (Wallner and Walton 1979,
Valentine et al. 1983) and on survival of black-marked spear moth (Werner 1979).
Further experiments on other species of trees will be required to determine if there are
general response of trees to successive years of defoliation or if the variability in results
to date reﬂect system speciﬁc responses.

In outbreaks of several different species of Lepidoptera, reductions in fecundity in
declining populations can approach 50% (e.g., Carter et al. 1991, Witter et al. 1975,
Myers and Kukan 1995). Since DIR has been shown to reduce pupal mass and fecundity,
it has been proposed as mechanism to account for changes in fecundity. In both of my
studies, the effects of induced resistance could not account for more than a portion of the
declines observed in natural populations. For example, in declining gypsy moth
populations, declines in fecundity of 52% relative to low-density populations have been
recorded (Carter et al. 1991). Similar changes in fecundity have been documented in
forest tent caterpillar populations (Ives 1971, Witter et al. 1975). In my study, the effects
on gypsy moth fecundity were maximally 14% and furthermore, the effects were directly
dependent on density. In the forest tent caterpillar study, there were delayed effects of
l3-19%, which account for approximately a third of the changes seen in natural outbreak
populations. This indicates that factors in addition to DIR are driving fecundity changes.

An interesting aspect of the hybrid poplar study was that, even though the
physiology of the trees responded as theory predicted, the insect herbivores did not. No
residual effects of previous defoliation were detected on four species of Lepidoptera
assayed in 1999, one year after the cessation of treatments. Elevated levels of condensed

tannins as well as some minor phenolics appeared to have little impact on the

224

 

 

 

have cumulative negative impacts on gypsy moth pupal mass (Wallner and Walton 1979,
Valentine et a1. 1983) and on survival of black-marked spear moth (Werner 1979).
Further experiments on other species of trees will be required to determine if there are
general response of trees to successive years of defoliation or if the variability in results
to date reﬂect system speciﬁc responses.

In outbreaks of several different species of Lepidoptera, reductions in fecundity in
declining populations can approach 50% (e.g., Carter et a1. 1991, Witter et al. 1975,
Myers and Kukan 1995). Since DIR has been shown to reduce pupal mass and fecundity,
it has been proposed as mechanism to account for changes in fecundity. In both of my
studies, the effects of induced resistance could not account for more than a portion of the
declines observed in natural populations. For example, in declining gypsy moth
populations, declines in fecundity of 52% relative to low-density populations have been
recorded (Carter et al. 1991). Similar changes in fecundity have been documented. in
forest tent caterpillar populations (Ives 1971, Witter et a1. 1975). In my study, the effects
on gypsy moth fecundity were maximally 14% and furthermore, the effects were directly
dependent on density. In the forest tent caterpillar study, there were delayed effects of
l3-19%, which account for approximately a third of the changes seen in natural outbreak
populations. This indicates that factors in addition to DIR are driving fecundity changes.

An interesting aspect of the hybrid poplar study was that, even though the
physiology of the trees responded as theory predicted, the insect herbivores did not. No
residual effects of previous defoliation were detected on four species of Lepidoptera
assayed in 1999, one year after the cessation of treatments. Elevated levels of condensed

tannins as well as some minor phenolics appeared to have little impact on the

224

 

 

 

performance of the herbivores. This was most evident in 1999 where the trees continued
to exhibit an induced-response to defoliation with elevated levels of total phenolics and
condensed tannins. However, in 1999 there was no effect of three previous years of
defoliation on the growth of gypsy moth, forest tent caterpillar and white-marked tussock
moth, and even an apparent increase in performance of the mid-season feeding poplar
tent maker. This result suggests that simplistic cause and effect statements about
secondary compounds and herbivore performance need to be reevaluated. It also lends
credence to the results of Ayres et al. (1997), which showed that tannins had little effect
on many herbivorous insects, leading the authors to suggest that the function of
condensed tannins is enigmatic and may not necessarily be related to insect resistance.
Trees have physiological limits to their responses to defoliation. In the hybrid poplar
system, even though there was no cumulative effect on herbivores, there was a lagged
response on tree physiology. Kosola et al. (2000) found that there was no difference in
tree mortality among the treatments through the ﬁrst three years of the experiment.
However, increases in mortality became evident in 1999 with defoliated, unfertilized
treatments suffering the highest losses. Furthermore, after coppicing in the winter of
1999, height and mortality measurements on the stump sprouts in 2000 showed continued
residual effects of the fertilizer and defoliation treatments applied last in 1998 (AA.
Agrawal, K.R. Kosola and D. Parry, unpublished manuscript). In the aspen system, the
response of the trees to repeated defoliation was different from the patterns seen in the
hybrid poplar experiment. Even four successive years of defoliation did not increase tree

mortality. Over the course of the study, none of the experimental trees died. While not

225

 

 

 

quantiﬁed, there appeared to be many more dead branches on the repeatedly defoliated
aspen trees, especially in the lower third of the canopy.

The response of hybrid poplars to defoliation and fertilization generally ﬁt the
predictions of the CNB hypothesis, at least with respect to condensed tannins, the most
prevalent phenolic compound. With the exception of 1997, N—fertilizer mitigated the
production of condensed tannins in defoliated trees. The predicted allocation of carbon to
growth in the presence of nitrogen fertilizer appeared to be validated because tree grth
was higher in defoliated plots with added fertilizer than in defoliated unfertilized plots.
The effects of fertilizer on induced-responses of poplars to defoliation corroborate many
previous experiments addressing the constitutive responses of condensed tannins to
fertilization (e. g., Koricheva et al. 1998).

My study showed that the effects of defoliation caused by an outbreak species has
signiﬁcant effects on the performance of other species in the folivorous lepidopteran
community. The magnitude of the effects was similar for each of the species reared in
1998 with the exception of fall webworm. This result was unexpected because of the
variety of life histories represented among the herbivores. While Agrawal (2000) found
that induced-responses of wild radish to defoliation had variable effects on four
caterpillars assayed, the results from my poplar experiment suggest that performance of
these caterpillars may be driven by similar determinants of foliar quality. This result also
indicates that the effects of defoliation on competitors in natural outbreaks may be driven
more by direct competition for foliage or displacement from preferred habitat than by

phytochemical induction.

226

 

 

 

Parasitoids have often been identiﬁed as one of the most important mortality factors
in populations of outbreak Lepidoptera (e. g., Bylund 1995, Parry et a1 1997, Gould et al.
1990). Despite current interest in tritrophic relationships in agricultural crops, little
research has examined the interaction between trees, parasitoids, and caterpillars. In
agricultural plants such as corn and tomatoes, speciﬁc compounds are elicited in response
to caterpillar feeding and these are in turn used by parasitic Hymenoptera to locate
potential hosts (Turlings et al. 1993, Thaler 1997). In laboratory tests, Havill and Raffa
(2000) found that a parasitoid of gypsy moth responded to speciﬁc wound-related
volatiles released following larval feeding on poplar leaves. Some tachinids appear to
respond similarly to caterpillar feeding on the leaves of forest trees (Roland et al. 1995,
Mondor and Roland 1997).

The response of the tachinids Leschenaultia exul and Patelloa pachypyga to forest
tent caterpillar was mediated by the aspen clone on which they were feeding.
Intraspeciﬁc differences in tree physiology have not previously been shown to alter
susceptibility of caterpillars to parasitism. While preliminary in nature, this result could
have signiﬁcant bearing on the nature of host-parasitoid relationships. For example, it
could contribute to spatial heterogeneity in susceptibility to parasitism, creating, if not
enemy-free space, at least enemy reduced space. Considering that one of the tachinid
species, L. exul, is an important source of mortality in low-density populations of forest
tent caterpillar (Parry 1995, Parry et al. 1997), variance in susceptibility to parasitoids
among clones over large forest landscapes could have important implications to

population dynamics.

227

 

 

 

Future Research

My dissertation research has suggested a number of potentially interesting directions
for future research. The lack of DIR in the hybrid poplar system and the presence of DIR
in the aspen experiments is interesting because the trees are similar in phytochemistry
and life-history strategy. This indicates that as Tuomi et al. (1990) hypothesized, DIR
varies along environmental gradients in soil fertility. Experiments utilizing clones
varying in inducible traits coupled with gradients in fertility regimes would be make a
nice contribution to our understanding of induced-resistance. Aspen would be a
particularly good experimental species in this regard because of its extremely wide
distribution and occurrence over a range of site productivity.

Competition between outbreak species and other phytophages has not been well
researched. Although biodiversity surveys suggest that some species do decrease
following gypsy moth defoliation (e.g., Sample et al. 1996), the mechanisms underlying
the changes in populations is not known. The scale and duration of forest tent caterpillar
outbreaks (e.g., Mattson et al. 1991) suggest that populations of some competitors,
particularly species such as contemporaneous leaf-rollers are likely to decline. My study
has suggested that reductions in competitors could be due to changes in host plant
chemistry although other mechanisms likely contribute signiﬁcantly. Identifying the
mechanisms responsible for determining the outcome of competition under experimental
conditions in natural outbreaks might be a productive avenue of research. Such effects
have the potential to cascade through ecosystems because several of the dominant

defoliators in North America (gypsy moth, forest tent caterpillar) are not preferred prey

228

 

 

 

species of many forest dwelling birds that may regulate populations of other forest
insects.

Finally, the integration of top-down processes with bottom-up is critical. My studies
have shown that while the effects of induced-resistance can reduce fecundity by
appreciable amounts, these reductions only account for a fraction of the total change
observed in natural populations. Thus, full understanding of complex ecological
processes such as the population dynamics of outbreak species can not be achieved until

all of the relevant processes are included.

LITERATURE CITED

Agrawal, AA. 2000. Speciﬁcity of induced-resistance in wild radish: causes and
consequences for two specialist and two generalist caterpillars. Oikos 89: 493-
500.

Ayres, M.P., T.P. Clausen, S.F. MacLean, Jr., A.M. Redman and PB. Reichardt. 1997.
Diversity of structure and antiherbivore activity in condensed tannins. Ecology
78: 1696-1712.

Byland, H. 1995. Long-term interactions between the autumnal moth and mountain birch:
the role of resources, competitors, natural enemies, and weather. PhD
Dissertation. Swedish University of Agricultural Sciences, Uppsala, Sweden.

Carter, M.R., F.W. Ravlin and M.L. McManus. 1991. Changes in gypsy moth
(Lepidoptera: Lymantriidae) fecundity and male wing length resulting from

defoliation. Environmental Entomology 20: 1042-1047.

229

 

 

 

 

Gould, J.R., J.S. Elkinton and WE. Wallner. 1990. Density dependent suprresion of
experimentally created gypsy moth, Lymantria dispar (Lepidoptera:
Lymantriidae) populations by natural enemies. Journal of Animal Ecology 59:
213-233.

Havill, NP. and K.F. Raffa. 2000. Compound effects of induced plant responses on
insect herbivores and parasitoids: implications for tritrophic interactions.

Ecological Entomology 25: 171-179.

Ives, W.G.H. 1971. The forest tent caterpillar in Alberta. Environment Canada, Canadian

Forest Service, Northern Forestry Centre, Internal Report NOR-4.

Kaitaniemi, P., S. Neuvonen and T. Nyyssonen.l999. Effects of cumulative defoliations
on growth, reproduction, and insect resistance in mountain birch. Ecology 80:
524-532.

Koricheva J, S. Larsson, E. Haukioja and M. Keinanen. 1998. Regulation of woody plant
secondary metabolism by resource availability: hypothesis testing by means of
meta-analysis. Oikos 83: 212-226.

Kosola, K.R., D.I. Dickmann, E.A. Paul and D. Parry, 2000. Repeated insect defoliation
effects on growth, nitrogen acquisition, carbohydrates and root demography of
poplars. 0ecologia. In press.

Mattson, W.J., D.A. Herms, J.A. Witter and DC. Allen. 1991. Woody plant grazing
systems: North American outbreak folivores and their host plants. Pp. 53-84 in
eds Y.N. Baranchikov, W.J. Mattson, F.P. Hain and TL. Payne, eds, Forest Insect
Guilds: Patterns of Interactions with Host Trees. United States Department of

Agriculture, Forest Service. General Technical Report NE-153.

230

Mondor, EB. and J. Roland. 1998. Host searching and oviposition by Leschenaultia exul,
a tachinid parasitoid of the forest tent caterpillar, Malacosoma disstria. Journal of

Insect Behavior 11: 583-592.

Myers, J.H. and B. Kukan. 1995. Changes in the fecundity of tent caterpillars — a
correlated character of disease resistance or sublethal effect of disease? 0ecologia

103: 475-480.

Parry, D. 1995. Larval and pupal parasitism of the forest tent caterpillar in Alberta. The

Canadian Entomologist 127: 877-893.

Parry, D., J.R. Spence and W.J.A. Volney. 1997. Responses of natural enemies to
experimentally increased forest tent caterpillar populations. Ecological

Entomology 22: 97-108.

Roland, J. K.E. Denford and L. Jimenz. 1995. Bomeol as an attractant for Cyzenis
albicans, a tachinid parasitoid of the winter moth, Operophtera brumata L.

(Lepidoptera: Geometridae). The Canadian Entomologist 127: 413-421.

Sample BE, L. Butler, C. Zivkovich, R.C. Whitmore and R. Reardon. 1996. Effects of
Bacillus thuringiensis Berliner var Kurstaki and defoliation by the gypsy moth
[Lymantria dispar (L)(Lepidoptera: Lymantriidae)] on native arthropods in West
Virginia. The Canadian Entomologist 128: 573-592.

Thaler, J.S. 1997, Jasmonate-inducible plant defences cause increased parasitism of

herbivores. Nature 399: 686-688.

231

 

 

 

 

 

Turlings, T.C.J., P.J. McCall, H.T. Albom and J.H. Tumlinson. 1993. An elicitor in
caterpillar oral secretions that induces corn seedlings to emit chemical signals
attractive to parasitic wasps. Journal of Chemical Ecology 19: 411-425.

Tuomi, J., P. Niemela and S. Siren.l990. The Panglossian paradigm and delayed-
inducible accumulation of foliar phenolics in mountain birch. Oikos 59:399-410.

Valentine, H.T., W.E. Wallner and PM. Wargo. Nutritional changes in host foliage
during and after defoliation, and their relation to the weight of gypsy moth pupae.

0ecologia 57: 298-302.

Wallner, W.E. and GS. Walton. 1979. Host defoliation: a possible determinant of gypsy
moth population quality. Annals of the Entomological Society of America 72: 62—

67.

Witter, J .A., W.J. Mattson and HM. Kulman. 1975. Numerical analysis of a forest tent
caterpillar (Lepidoptera: Lasiocampidae) outbreak in Northern Minnesota. The

Canadian Entomologist 107: 837-854.

232

 

 

 

 

 

 

 

 

 

 

$.47“.qu 1%....7

..
my. «cur—.5: .s...

25L. .
—Q A 4...» u.
T H.H...“
.92..
z r
,

.r ..

 

as .

£5.54...

 

 

 

 

 

: 2 . 5.. ..

a: ”U: 7
5:. 2.. a. :

.._
7: s

... r .9

. 2...:
“erasaeuﬁ
6?... a a».
...2. a 2
z\.