.-.,‘A.

”ML“.

 

mwv

.0.
4‘4-
.

,.
ts“.
~o

.: .-
..-~....

. x».
- Nu.”

m: 1. \ ‘
1". mu A54 ‘

'4...

 

 

 

 

 

«I‘d-u.

 

VIN-iv.

 

 

a

S

iii

\iliiiiiilliii

\

This is to certify that the

dissertation entitled

Behavior and development of larval gypsy moth,

Eymantria dispar (L.), on trees of the Upper
Great Lakes forests

presented by

David Berkeley Roden

has been accepted towards fulfillment '
of the requirements for

Ph . D . degree in W

 

  

 
  

Major professor

J.R. Miller
Date_7_Ma3L._1222_

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

> p

"* h-L;~"' :2 8m

lid

 

LiBRARY
Michigan State
i University

*

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

TWKDUE ‘ DATE DUE DATE DUE

 

 

trig 20x99?

 

k

 

 

, (3:533; -

mm 2007
it E in: ~

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
one? WIS-IL 1

BEHAVIOR AND DEVELOPMENT OF LARVAL GYPSY MOTH,

W DISPAB (L.), ON TREES OF THE
UPPER GREAT LAKES FORESTS

By

David Berkeley Roden

A DISSERTATION

Submitted to

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

DOCTOR OF PHILOSOPHY

Department of Entomology
and
Program in Ecology and Evolutionary Biology

1992

or"
4‘
h-

6, u- '7' .20)"

Abstract

BEHAVIOR AND DEVELOPMENT OF LARVAL GYPSY MOTH,
LILMEIBLA DISPAR (L.), ON TREES OF THE
UPPER GREAT LAKES FORESTS

By

David Berkeley Roden

In laboratory and field experiments, all larval instars of gypsy moth, MA
dim (L.), were influenced by the diameter, height and species of a tree. The degree of
larval attraction to an object was positively correlated with the angle presented by the
diameter and height. Larval attraction to red oak (mm grim; L.) was frequently
fifteen-fold greater than to paper birch (Ema mm Marsh.) and trembling aspen
(Mus tremulgides Michx.). Diameter, height and tree species may be important
considerations for standardizing burlap banding and explaining the gypsy moth "wolf

tree" phenomenon.

Pupal weight, developmental time and survival of gypsy moth larvae on three
defoliated (60%) and undefoliated tree species were compared for one season. Host
species and defoliation both affected female pupal weight, whereas only host species
affected male pupal weight. Female and male pupal weights, averaged over defoliation
treatment, were 1.43 and 0.49 g, respectively, on trembling aspen, > the 1.13 and 0.40 g
on paper birch and > the 0.84 and 0.38 g on red oak. Time of development, averaged
over the defoliation treatment, was affected only by tree species: female and male time of

development were 46 and 41 days, respectively, on trembling aspen, < the 48 and 44

days on paper birch and < the 50 and 44 days on red oak. Gypsy moth survival was
unaffected by defoliation or host species. Better gypsy moth performance on trembling
aspen and paper birch was attributed to an imprecise correspondence between host and
herbivore that inhibits outbreaks of gypsy moth on these tree species. It is suggested that
the gypsy moth nuclear polyhedrosis virus and physical features of the host are

responsible for imprecise correspondence.

Gypsy moth larvae were sometimes found to construct silk ladders for climbing
on a smooth vertical surface. The use of silk for climbing was six times more frequent in
less-fit populations and was common on trembling aspen and paper birch, but absent on
red oak. The use of silk for climbing by less-fit larvae in the field may be a result of

wound-induced plant defences.

An incandescent and fluorescent lighting system is described that induced a
change in feeding behavior between small and large gypsy moth larvae in the laboratory.
This change was observed on artificial "tree stems." On the day before pupation began,
85% of the larval population migrated down the artificial tree stems to seek shelter under

felt and cardboard "bark flaps."

Fr

 

COPyrisht by
DAVID BERKELEY RODEN
1992

For Lorne in memory of the years we worked together.
Your thoughtfulness and kind and gentle personality were
instrumental in my development and provided examples I

can only hope to achieve.

ACKNOWLEDGMENTS

I would like to express my gratitude to many people who made this research
possible. First and foremost, an acknowledgment must be extended to my wife Pat.
Without her continued encouragement, support, personal sacrifice and love, completion
of this study would not have been possible. When there was no one else, or when
technical support was insufficient, she was always available and was essential for

completion of much of the laboratory work.

A special thank you is extended to Gary A. Simmons. As my major professor, he
helped me through many difficulties that should never have arisen. His understanding of

and compassion for people and dedication to the classroom and his students was

exemplary.

Grateful acknowledgment is also extend to the other members of my graduate
committee, Stuart H. Gage, Gordon M. Howse, William J. Mattson, James R. Miller and
David R. Smitley for their suggestions and consultation throughout this study. However,
I am particular indebted to Bill Mattson and Jim Miller from Michigan State University
and Gary Grant from Forestry Canada in Sault Ste. Marie. Their helpful suggestions,
perceptive comments and critical reviews of my manuscripts and research contributed

immensely to my growth and development as a scientist.

A special debt of gratitude is owed to Jim Miller. The assumption of
responsibilities as my major professor after Gary Simmons' death was profoundly
appreciated. I am also indebted to him for his classroom lectures and philosophies which
were instrumental in molding a research theme that was not anticipated by either of us
and one that helped fill a vacuum when initial funding collapsed. Jim became not only

my major professor, but before I was through, a personal friend and respected colleague.

I also gratefully acknowledge Forestry Canada and their commitment to
employees, training and education. The study was conducted while on educational leave
and was made possible by financial resources provided by Forestry Canada, Ontario

Region, Sault Ste. Marie, Ontario.

vii

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................ xi

LIST OF FIGURES .................................................................................................... xiii

INTRODUCTION ......................................................................................................... 1
CHAPTER ONE. Visual Stimuli Influencing Orientation by Larval

Gypsy Moth, Ma dim (L.) .............................................. 3

Introduction ......................................................................................................... 3

Materials and Methods ......................................................................................... 4

Rearing .................................................................................................... 4

Larval Preference for Black versus White Artificial Trees ........................ 5

Diameter ................................................................................................ 10

Height .................................................................................................... 12

Species ofTree 12

Statistical Procedures ............................................................................. 12

Results ............................................................................................................... 13

Larval Preference for Black versus White Artificial Trees ...................... 13

Diameter ................................................................................................ 17

Height .................................................................................................... 27

Species of Tree ....................................................................................... 3O

viii

Discussion ......................................................................................................... 3O
Larval Preference for Black versus White Artificial Trees ...................... 30
Diameter, Height and Species of Tree .................................................... 35

CHAPTER TWO. Influence of Current Defoliation and Different Tree
Species on Gypsy Moth, Emeritus discs; (L),

Growth and Development ........................................................... 40

Introduction ....................................................................................................... 40
Materials and Methods ....................................................................................... 41
Results ............................................................................................................... 43
Pupal Weight .......................................................................................... 43
Development Time and Survival ............................................................ 45
Discussion ......................................................................................................... 50

CHAPTER THREE. Laddering: A climbing behavior of the

Gypsy Moth, mm dim (L.) ............................................ 5 8
Introduction ....................................................................................................... 58
Materials and Methods ....................................................................................... 58
Results and Discussion ....................................................................................... 62

CHAPTER FOUR. A Laboratory Technique to Study a Change in Feeding
Behavior between Small and Large Larvae

of the Gypsy Moth, mm 51131231 (L.) .................................. 72

Introduction ....................................................................................................... 72
Material and Methods ........................................................................................ 73
Lighting System ..................................................................................... 73
Artificial Trees ....................................................................................... 80
Results and Discussion ....................................................................................... 86

ix

 

SUE

REC (

APPE

LIST

SUMNIARY AND CONCLUSIONS ........................................................................... 90

RECOMNIENDATIONS ............................................................................................. 93
APPENDIX A. Instructions for Wiring the Incandescent Light Circuit ......................... 96
LIST OF REFERENCES ............................................................................................ 98

 

 

Tab

Tab

Tab

Tat

Tat

Tat

Tat

Tat

Tab

Tab

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

LIST OF TABLES

Attraction of gypsy moth larvae on a horizontal surface to black versus
white columns .......................................................................................

Numbers of gypsy moth larvae of various instars that were attracted to
adjacent versus non-adjacent black or white columns in the laboratory

Influence of instar on larval gypsy moth choice of the black versus the
white arm of a Y after climbing 25 em up a 5-cm-diameter column ......

Orientation time and accuracy in response to columns as influenced
by insect instar. .....................................................................................

Larval gypsy moth response, by instar, to 13.8 i 3.1 cm wide by 1 m
long, vertically positioned bolts from three naturally available tree
species .................................................................................................

Analysis of variance of gypsy moth pupal weights as influenced
by defoliation and tree species ...............................................................

Mean female and male pupal weight (gm) for gypsy moth larvae
reared on undefoliated and defoliated trees averaged over host
species. Mean standard error = 0.02 and 0.005, respectively .................

Mean female and male pupal weight (gm) for gypsy moth larvae

as influenced by host species averaged over defoliated and
undefoliated treatments. Mean standard error = 8.86 and

7.13, respectively ..................................................................................

Analysis of variance of gypsy moth time of development as influenced

by defoliation and tree species ...............................................................

Mean time of development (days) for female and male gypsy moth
larvae as influenced by host species averaged over defoliated and
undefoliated treatments. Mean standard error = 8.86 and

7.13, respectively ..................................................................................

xi

....... l4

....... 15

....... 16

....... 22

....... 31

....... 44

....... 46

....... 46

....... 47

....... 48

 

Table

Table

 

Table ‘

Table

Table

Table

Table

Table

Table 11.

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17 .

Table 18.

Analysis of variance of gypsy moth survival as influenced by defoliation
and tree species ............................................................................................ 49

Source of experimental mortality and number of larvae surviving from
second-instar to adult eclosion for female and male gypsy moth larvae
reared on defoliated and undefoliated host species ....................................... 49

Mean number of larvae per replicate, by instar, that demonstrated
laddering behavior on smooth and rough columns of ABS plastic
pipe. All larvae were reared on ICN diet ...................................................... 63

Mean time for sy moth larvae with and without the spinneret sealed
with Crazy Glue to climb 25 cm on a smooth or rough 50-cm column
of 5-cm-diameter ABS plastic pipe. Standard error = 0.55 ........................... 67

Mean time for gypsy moth larvae reared on ICN and Bell's artificial

diets (averaged over low and high population levels) climbing 25 cm

in 30 min on a smooth column of ABS pipe. Standard error of the

mean = 0.10 ................................................................................................. 68

Measures of fitness for 15 larval gypsy moth reared on Bell's
and ICN diets ............................................................................................... 68

Mean number of gypsy moth larvae per replicate that exhibited laddering
behavior on 30-cm bolts of red oak, paper birch and trembling aspen

(MSE = 0.86) and the mean time per replicate for larvae to climb

25 cm (MSE = 0.38) .................................................................................... 70

Variance and maximum/minimum number of larval gypsy moth per

tree (n=12) on 5 subsequent days before (days 5-9) and after (days 16-20)

a change in feeding behavior was observed. During photophase on

days 16-20, 80-85% of the population (n=30) changed

feeding behavior .......................................................................................... 88

 

Tigu

Tigt

Fig!

Flg|

Figl

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

LIST OF FIGURES

The arrangement of black (B) and white (W) artificial tree trunks used
to evaluate the attractiveness of vertical objects to gypsy moth
larvae. (ri = points at which larvae were released) ......................................... 7

The ranked position, by instar, of the number of gypsy moth larvae
attracted to columns of different diameters. Columns A to 1,

respectively, represent the smallest to the largest-diameter columns;
X and Z, respectively, represent larvae that exceeded the 30-min
test period and larvae that wandered out of the arena. Numbers

for columns above the same horizontal line are not significantly
different (Waller-Duncan k-ratio test, k=100). Numbers below each
column are the mean ranked numbers of the first 20 larvae attracted.

Numbers within brackets are the total number of larvae attracted

to each column. (TLT = total larvae tested for each instar until

20 larvae orientated to a column; instars tested = Neonate, I, 111, V
and VtField) .................................................................................................. 19

Tracing of gypsy moth larval tracks from the diameter experiment for

(a) neonate, (b) third-instar and (c) field-collected fifth- and sixth-instar

larvae. Ticks perpendicular to the trace show the posterior position of

a larva every 15 seconds ................................................................................. 21

Numbers of each instar attracted to columns of (a) different diameters

and the same height and (b) different heights and the same diameter,

as a function of the angles 0 and B, respectively, as subtended by

a column from the arena center where larvae were released ....................... 24

An illustration of the angles 0 and 2» used to calculate the expected
proportion of larvae attracted to the different columns or open spaces
between columns ........................................................................................ 26

xiii

 

Fl

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

The ranked position, by instar, of the ranked number of gypsy moth
larvae attracted to columns with different heights. Columns A

to 1, respectively, represent the shortest to tallest columns; X

and Z, respectively, represent larvae that exceeded the 30-min

test period and larvae that wandered out of the arena. Numbers

for columns above the same horizontal line are not significantly
different (Waller-Duncan k-ratio test, k=100). Numbers below each
column are the mean ranked number of the first 20 larvae attracted.
Numbers within brackets are the total number of larvae attracted

to each column. (TLT = total larvae tested for each instar until 20
larvae orientated to a column; instars tested = Neonate, I, 111, V”
and Vtmd) .................................................................................................. 29

Silk ladder used by a sixth-instar gypsy moth larva to climb on a
smooth column of ABS plastic pipe ............................................................ 65

Mean number of larvae (j; SE) descending below 1.5 m on 3 white oak

trees, Om alba L. (mean dbh = 38.1 i 4.59 cm; mean height = 17.2 i
0.61 m) during subsequent 15 min intervals between 0200 and 0800 h

EST on 30 June 1989 (Kaladar, Ont) .......................................................... 75

Schematic of the brass union that coupled the drive motor to the handle
of the rheostat: a, drive motor; b, rheostat handle; c, brass union; (1, 5 mm
pin; e, brass rod; f, rheostat; g, set screws; h, drive-motor mounting

bracket ........................................................................................................ 78
Wiring diagram for the incandescent light circuit: DM, drive-motor;

VR, variable rheostat; CP, 0.25 MFD capacitor .......................................... 82
Artificial tree used to observe a change in feeding behavior ........................ 84

xiv

 

 

bclo

mai

bag-
DOT
and
of 1
him

CS.

9‘11

INTRODUCTION

The gypsy moth, Lymantda dispar (L.), is a univoltine defoliator of hardwood
trees and is indigenous to Eurasia and the Mediterranean (Mills 1983). It was first
introduced to North America by Leopold Trouvelot at Medford, Massachusetts in 1868
or 1869 (Forbush and Femald 1896). Passive dispersal of life stages, particularly egg
masses, via human interaction (McManus and McIntyre 1981), and active dispersal by
ballooning of first-instar larvae (Capinera and Barbosa 1976; Cameron gt, al. 1979;
Mason and McManus 1981) have enabled gypsy moth to colonize most of the New
England states, as well as areas as far away as California, Michigan and British Columbia

(Prittchett 1975; Campbell 1979; Sterner and Davidson 1981).

Many studies in Europe and the United States show that species of trees
belonging to the genus mm are the preferred gypsy moth host most capable of
maintaining populations (Campbell and Sloan 1977; Hough and Pimentel 1978; Barbosa
1978; Barbosa gt, al. 1979; Houston 1979). However, infestation of the upper Great Lakes
basin, which has a tree species composition substantially different from that in the
northeastern United States, has raised concern about the utilization of secondary hosts
and their potential to support outbreaks of gypsy moth similar to those in the oak forests
of New England. Trees such as trembling aspen, 11mm tremulgides Michx. and paper
birch, Begun mm Marsh, hosts classified as secondary in the forests of the eastern
USA , are a major component of the forest in the upper Great Lakes basin and an
important sector of the resource base utilized by many forest industries. Studies by Witter

gt a1. (1990) and Roden and Surgeoner (1991) have indicated that gypsy moth larval

l

 

 

 

' 1
mm

min
Lal
def

2

development on these hosts may produce more fecund adults in a shorter developmental
period than gypsy moth larvae reared on a "traditional" host such as red oak, gum
{ulna (L.). Control of an herbivore and the potential subsequent effect on population
dynamics is based on the assumption that the additon or deletion of a particular phase of
host-colonization behavior will result in interruption of the host-selection process (Harris
and Miller 1991). The objective of the present research was to evaluate the potential of
trembling aspen and paper birch in the development of gypsy moth in comparison with
that of red oak and investigate the possible influence that the physical features of the host

may have on larval behavior.

Chapter one identifies visual stimuli that influence the orientation of gypsy moth
larvae and how physical features of a host such as its diameter, height and species
influence the larval host-selection process. Chapter two evaluates the potential that
trembling aspen and paper birch provide for the development of gypsy moth in the Great
Lakes basin compared with red oak and how this potential may be affected when plant
defences are evoked by artificial defoliation. Chapter three reports a previously
unidentified behavior ("laddering") of gypsy moth larvae that is used for climbing on a
smooth vertical surface and the influence of diet on this behavior. Chapter four describes
a lighting system for the laboratory that can be used with artificial trees to produce a
change in the feeding behavior between small and large larvae and suggests that such a
system could be used to further evaluate physical features of a host and how they

influence larval behavior.

 

 

 

by f:
emp
Bari
sub:
Sch
ins:

Be:

Che
SCV'
lab
(D:-

lieu

8115

C011

CHAPTER ONE

Visual Stimuli Influencing Orientation by Larval
Gypsy Moth. Lang's slim (L)

Introduction

Host colonization by the gypsy moth, Ma dim (L.), may be influenced
by feeding behavior, foraging behavior, or both. Studies of feeding behavior have
emphasized either the nutritional suitability of the hosts (Hough and Pimentel 1978;
Barbosa and Greenblatt 1979), or the debilitating effects of ingested secondary
substances (Keating and Yendol 1987; Meyer and Montgomery 1987 ; Rossiter and
Schultz 1988). Chemical ecologists have suggested that the behavioral response of an
insect to a host is mainly a series of simple reactions to chemicals (Thomsteinson 1960;
Beck 1965; Shorey 1977). This explanation of insect behavior subscribes to the theory of
"labeled lines" (Perkel and Bullock 1968; Dethier 1971, 1982) and implies that a
chemical molecule, when it impacts on a chemoreceptor, transmits nonambiguous but
severely restricted information that initiates an appropriate behavioral response. The
labeled-line model is gradually being replaced by the theory of across-fiber patterning
(Dethier 1971, 1982) and across-modality stimulus summation (Miller and Strickler
1984; Miller and Harris 1985). According to these latter concepts, the relative patterns of
neuronal action potentials are integrated with sensory patterns from other modalities (i.e.,
gustatory, olfactory, visual and mechanical) to trigger decision-making intemeurons that

control behavior. Although there is no doubt that chemical cues play an important
3

 

 

 

Rearin
years b

obtaine

(44°39

Surface:
experim
results fi
storage c
Used for
1981) in
humidit;
behavior
with Ian
condUCte
by lan'ae
fiel‘L “'6:
1983) am

minimize

4

regulatory role in the colonization of hosts by the gypsy moth, the foraging behavior of
folivorous insects has received little attention (Schultz 1983; Dethier 1987). Because
there is evidence to suggest that the orientation of gypsy moth larvae to hosts in the field
is influenced by stimuli from sensory modalities other than chemoreception (Doane and
Leonard 1975; Lance and Barbosa 1982), the present study posed the question of how the

physical features of a tree influence the host-selection process of larval gypsy moths.
Materials and Methods

Rearing. Behavioral experiments were conducted in the laboratory over a period of 3
years between January and September from 1986 to 1989. Gypsy moth larvae were
obtained annually from eg masses collected the preceding October near Kaladar, Ont.
(44°39'N, 77°07'W) and held at 5°C until diapause was completed. Eggs were then
surface sterilized (Shapiro 1977) and incubated, as required, at 22°C :1: 2°C. No
experimental measurements of larval behavior were conducted afier 15 September;
results from the black-white artificial tree experiment described below indicated that the
storage of egg masses each year until 15 September did not affect larval behavior. Larvae
used for laboratory experiments were reared individually on artificial diet (Bell gt a1-

1 981) in screened Lab-Tek® 150- by 25-mm plastic Petrie plates at 70% relative

humidity with a 16L:8D photoperiod. In 1988, to validate the laboratory results,
behavioral studies of fifth- and sixth-instar larvae were repeated in the field at Kaladar
with larvae collected from the base of red oak (Om min-a L.) trees. These tests were
conducted under an overstory of red oak to simulate the lighting conditions experienced
by larvae crawling on the forest floor. All larvae tested, whether in the laboratory or the
field, were held without food overnight to increase their activity (Leonard 1967; Dethier
1982) and were gently transferred from a Petrie plate with a camel-hair brush to

minimize the effects of handling. Each larva was tested only once during an instar.

Larval Preference for Black versus White Artificial Trees. Deane and Leonard (1975)
reported that larval gypsy moth appeared to orient themselves with respect to vertical
objects such as tree trunks. To test these observations and determine if larvae are
influenced by contrasting colors, I constructed 24 artificial tree trunks l m in height from
7.5-cm ABS plastic pipe. A circular plate, cut from 5-mm Plexiglass® and glued in place
with ABS cement, sealed the top of each artificial trunk. The 24 trunks, randomly

divided into two groups of 12, were then painted with a mixture of 150 mL of sand
(sifted through a 20-mesh screen) and 250 mL of either black or white latex paint (Color
Your World, Toronto, Ont. M8W 3R5; paints #5900 and #5920, respectively). Sand was
added to the paint to roughen the surface of the tree trunks because previous trials had
revealed that larval gypsy moth had difficulty climbing on a smooth surface (Roden,
unpublished data). The 24 artificial tree trunks were then arranged in a 4- by 5-m rearing
room in a pattern of seven rows and seven columns offset from each other by 0.5 m; trees
of the same color were separated by 1 m (Figure 1). Thus, there were four tree trunks,
respectively, in each alternating row and column. This arrangement provided a total of 13
release points, one in the center of each square formed by the rows and columns, which
were equidistant from the two nearest black (B) and white (W) tnmks (Figure 1). These
points were used to release 13 larvae for each replication and to compare selection of the
tree trunks. The populations tested were classified as neonate larvae. These were unfed,
first-instar larvae that had been eclosed from the eg for at least 24 h, but that were less
than 48 h old; first-instar larvae (I) that had fed, but that were older than 48 h; separately,
second- (II), third- (III) and fourth-instar larvae (IV); collectively, fifth- and sixth-instar
larvae (V); and, collectively field-collected fifth- and sixth-instar larvae (V+F‘°“). In
addition to the above, three Other tests included first-instar larvae from eggs randomly
39130th and held until 15 September (I-SEPT); first-instar larvae tested on 122-cm

double-density 40-weight kraft paper (Lewis Paper, Toronto, Ont. M9N 2Y8) placed on

Figure 1. The arrangement of black (B) and white (W) artificial tree trunks used to
evaluate the attractiveness of vertical objects to gypsy moth larvae. (ri = points at which
larvae were released)

 

 

Fluorescent Lighting

I|'||'l ul||'||||l-

 

8

the floor between replications (I-PAP); and third-instar larvae (III-OAK) reared from
eclosion on leaves of red oak. Full water content was maintained in excised leaves by ‘
placing individual leaves in "water piks" and changing them every other day (Scriber

1977).

A complete randomized block design, blocked by time and with six replications,
was used to compare three treatments for each larval population: larvae that oriented
toward black artificial tree trunks, white artificial tree trunks and larvae that did not
orient to any of the artificial tree trunks within 1 h. In the laboratory, the color of the
rows and columns was reversed each time the experiment was replicated to remove the
possibility that the arrangement of the fluorescent lights (General Electric F40/C7 5 lamps
with a standard ballast operating at 110/115 V) biased larval choice. Field tests of larvae
were conducted with the artificial trees arranged in rows and columns running north-
south and east-west on the forest floor; the rows and columns were also reversed between
replications. Replication in the field was conducted at 0800, 1200 and 1600 h EST on
each of 2 days (7 and 9 July 1988) to allow for a possible effect caused by the change in
the position of the sun on larval orientation (Doane and Leonard 1975). In both the
laboratory and the field, each larva was released with its head pointed in one of the four
cardinal compass directions at the time of release. The order of larval release at the 13
release points and the direction in which they were released were both randomized. One
hour after release, the tree trunk to which larvae were attracted was recorded. A Waller-
Duncan Bayesian k-ratio Test with k = 100 was used for separation of means. A k-ratio
of 100 is 5 equivalent to 0t=0.05 (Waller and Duncan 1969; Duncan 1975). Observations
of whether or not larvae were attracted to the closest column were made for one larva
selected randomly from each replication for neonates and first, second, third, fourth, fifth

and sixth instars.

9

Larval response to contrasting black and white colors while climbing was
investigated using a Y-shaped column constructed from 5-cm-diameter ABS plastic pipe.
Arms of the "Y", 15 cm in length, were angled at 900 and then glued together with ABS
cement before being fastened with a hot-air welder (Leister-Kombi, Type: Triac,
purchased from Johnson Industrial Plastics, Toronto, Ont.) to a 5-cm-diameter, 60-cm-
tall vertical stem of ABS plastic pipe so that each arm of the Y was at 1350 to the vertical
stem. Additional support was provided by inserting the bottom of the Y into a 5- by 7-cm
coupling. The black paint-sand mixture described previously was then applied to one
arm of the Y and the corresponding half of the stem; the other arm and the other half of
the stem were painted with a similar mixture of white paint and sand. The larval instars
tested were: neonate, first, second, third, fourth, fifth and sixth and, collectively, field-
collected fifth- and sixth-instar larvae. When testing larvae, the Y was placed on a table
between overhead fluorescent lights so that each color was illuminated equally. Release
of larvae onto the column was accomplished by allowing each larva to climb a piece of
wooden dowelling positioned at an angle of ca. 30° against the column's black-white
interface 25 cm below the Y. Initially, 100 neonate larvae were released; however, the
number released for older instars decreased because of mortality. After testing each larva,
the column was brushed to remove trails of silk (Leonard 1967) and was wiped with.
hexane to remove possible pheromone or other chemical substances associated with the
silk. Between odd- and even-numbered larva, the Y was also rotated 180° to remove the
effect of room lighting, left-right larval preferences, or both. A test was considered
completed when a larva had climbed to the top of one arm of the Y; larvae that did not
climb to the top within 10 min were not included in the analysis. Sex was determined by
rearing larvae until they pupated. Statistical significance for either a black or white
surface preference (p=0.5) was tested for each instar with a table of values for a two-

tailed binomial test (Zar 1984).

10

Diameter. The influence of the diameter of an artificial tree on larval orientation was
investigated using nine l-m-tall columns with diameters of 3, 6, 12, 25, 50, 100, 150,
200 and 300 mm. Materials used for the 3-mm (steel rod) and the 6- and 12-m (wood
doweling) columns differed from those used for the others (cardboard: Design Tubes,
Markham, Ont.) because it was too difficult to obtain a broad range of diameters in one
material; however, so that the image and the light reflected by each material was the
same, the surface of each column was painted with the black paint-sand mixture
described earlier. A randomized complete block design, blocked by time and with five
replications, was used to test neonate larvae, first-instar larvae at least 48 h old that had
fed, third-instar larvae and, collectively, fifih- and sixth-instar larvae and field-collected
fifth- and sixth-instar larvae. The nine test columns, randomly arranged for each
replication, were placed along the circumference of a l-m-diameter circle at 40°
intervals; the area within the l-m circle was the arena used to test larval response. The
three smallest columns were supported by nylon fishing line suspended from the ceiling;
nylon line was also suspended from the ceiling above each of the larger columns, but was
not used for support. For each replication, larvae were released individually with a
camel-hair brush from the center of the circle so that the head of each larva was pointed
in one of the four cardinal compass directions; the order and direction of each release
were both randomized. Each replication was continued until a total of 20 larvae had
climbed on a column. Larvae that did not climb on a column within 30 min, or that
wandered outside the arena, were also recorded. To remove the possibility that larvae
followed a pheromone or silk trail, a piece of 1.2-m-square krafi paper with 15- by 15-
cm pencil-drawn grids was placed on the floor for each larva tested. The track of the
larval path on a parallel grid and the time required for each larva to reach a column was
recorded. Measurements of tortuosity (the number of turns that exceeded 90°) and the
time required for larvae to select a column were measured for 20 larvae selected

randomly from each instar.

ll

Larval testing in the field was similar to that in the laboratory; however, columns
were placed on a 1.2-m-square plywood platform raised ca. 9.0 cm off the forest floor.
The smaller-diameter columns (3, 6, 12 and 25 mm) were supported by forcing them
through the lcraft paper into slightly smaller holes drilled in the plywood; the lengths of
these columns were increased to compensate for the distance they were inserted. Three
randomly chosen times during the day (0930, 1200 and 1500 h EST) were used to test
larval response in the field. Replications 1 and 3 were tested at 0930 and 2 and 4 were
tested at 1500 h on separate days; the fifth replication, initiated on a 3rd day, began at
1200 hours. Larval preference for diameter within each respective instar, including the
number of larvae that wandered out of the arena and those that did not select a column
within 30 min, was ranked for the first 20 larvae tested (Conover and Iman 1981) and
then subjected to AN OVA and Waller-Duncan's Bayesian k-ratio Test with k = 100 for
separation of means. The perceived angular width of each column (0) from the center of
the arena was also calculated; these were 0.18, 0.36, 0.72, 1.44, 2.84, 5.52, 8.09, 10.45
and 14.82 degrees. The transformed values (loge) for larvae of each instar (100) that
selected columns were tested for normality and regressed against 0. The slopes of each

regression equation for each population were compared.

I also investigated the probability that larvae would simply randomly encounter
columns when they wandered away from the center of the circle. This was tested with
chi-square values by summing and comparing, across instars (because there was no
difference between the slopes for each population tested), the number of larvae attracted
to columns versus the number of larvae that wandered out of the arena. The expected
distribution of larvae attracted to columns was the proportion for 0 versus the sum of
angles (A) between columns. A chi-square test was also used to compare the 20 larvae
attracted to columns based on an expected proportional distribution of 0. It is conceivable

that, although the number of larvae attracted to a column might be significantly

12

correlated with 0, the proportion of the population attracted to columns with large or

small values of 0 might be different from what was expected.

Height. Based on the results of the diameter experiment, the effect of height on the
choice of a host by L, dim was investigated by using 150-mm-diameter cardboard
columns for the following heights: 20, 40, 80, 160, 320, 640, 1280, 1600 and 2000 mm.
Columns were also painted with the black paint-sand mixture. Experimental design and
analysis was similar to that for the diameter experiment. The angles ([3) from the center
of the arena to the top of each column were, respectively, 1.15, 2.29, 4.57, 9.09, 17.75,
32.62, 52.00, 57.99 and 63.44 degrees.

Tree Species. The influence of the species of a tree on gypsy moth larval orientation was
investigated using red oak, trembling aspen (Regulus Weider, Michx.) and paper
birch (BS3113 papyflfera Marsh). The experimental design and the larval instars
examined were the same as those in the diameter and height experiments, except that the
nine cardboard columns were replaced with three l-m-tall equal-diameter bolts of each
species. These were randomly arranged in a 1 -m circle and replaced each time an instar
was tested. The number of larvae attracted to each species was analyzed with one-way
ANOVA; means were compared with a Waller-Duncan Bayesian k-ratio Test with k =

100.

Statistical Procedures. The Waller-Duncan Bayes test (BLSD) was specifically chosen
for means comparisons in this chapter because its ability to detect real differences
between means does not depend on the number of means being compared. Consequently,
the use of this test to compare the diameter and height means avoids problems assocaited
with disagreement in the literature about the use of comparisonwise and experimentwise

error rate approaches for means comparisons (Jones 1984; Perry 1986; Day and Quinn

13

1989). When the BLSD F-value is low (indicating a set of homogeneous means), the test
has a conservative characteristic that is typical of an experimentwise error rate approaCh
whereas, when the F -value is high (indicating heterogeneous means), it assumes a

comparisonwise error rate approach (Peterson 1985).

Results

Larval Preference for Black versus White Artificial Trees. The color of an artificial
tree clearly influenced the orientation of larval gypsy moth crawling on a horizontal
plane. All instars (Table 1) significantly preferred black to white artificial trunks
(0t=0.05); the number of larvae attracted to white trunks and those that left the perimeter
of the arena were not significantly different from each other for all instars examined.

Most larvae were attracted to the nearest column (Table 2).

Larval preference for a black or a white surface while climbing vertically varied
with instar (Table 3). Neonate larvae preferred white (or=0. 1) whereas first-, second-,
third- and fourth-instar larvae did not discriminate (or=0.1). However, laboratory-reared
fifth- and sixth-instar larvae and field-collected fifth- and sixth-instar larvae preferred
black (0t=0.05). These results were not influenced by sex, except for neonate larvae
(Ot=0.1); neonate females had no significant color preference but males preferred white

(the ratio B:W was 20:21 for females and 14:26 for males).

When climbing onto the Y, fifth- and sixth-instar larvae consistently oriented
klinotactically to the black-white interface of the column; larvae paused at the interface
between colors and would repeatedly swing the anterior portion of their body (head,
thorax and the first two abdominal segments) back and forth between the black and white

colors before making a choice. It was also not uncommon to observe larvae climbing

14
Table l. Attraction of gypsy moth larvae on a horizontal surface to black versus white

 

 

 

columns.
Mean number of larvae recovered“

Instar. Black White Other“ MSE‘“
Neonate 12.3a 0.5b 0.2b 0.6
I 13.0a 0.0b 0.0b 0.0
II 11.7a 1.2b 0.2b 2.7
111 12.0a 0.5b 0.5b 1.9
IV 12.2a 0.8b 0.0b 0.6
V+ 11.0a 1.7b 0.3b 1.9
WW 11.8a 0.8 b 0.3b 1.3
I-SEPT 11.0a 1.0b 0.7b 1.7
I-PAP 12.0a 0.5b 0.5b 0.7
III-OAK 13.0a 0.0b 0.0b 0.0

 

" Neonate to V” = larval instars reared on artificial diet in the laboratory; Vmeld = field-
collected fifth- and sixth-instar larvae tested in the field; I-SEPT = first-instar larvae
from eggs randomly selected and held until 15 Sept. ; I-PAP = first-instar larvae tested on
the floor, with new kraft paper between each replication; III-OAK = third-instar larvae
reared on leaves of red oak from eclosion.

” Total possible was 13. Within a row, values followed by the same letter are not
significantly different (Waller-Duncan k-ratio test, k=100).

H" Other represents the number of larvae that either did not climb on a column within 1 h
or that wandered outside the arena.

”’MSE = mean squared error.

15

Table 2. Numbers of gypsy moth larvae of various instars that were attracted to adjacent
versus non-adjacent black or white columns in the laboratory.

 

 

Instar. Adj acent“ Non-adjacent”
Neonate 6 0
I 6 0
II 5 1
III 5 1
IV 5 1
Vt 4 2

 

‘ Larval instars are designated as in Table 1.
“ Adjacent and non-adjacent values total to 6 because only 1 larva was observed for each
replication.

16

Table 3. Influence of instar on larval gypsy moth choice of the black versus the white
arm of a Y after climbing 25 em up a 5-cm-diameter column.

 

 

Instar‘ n % larvae selecting black“
Neonate 100 ‘40
I 94 51
II 46 38
III 76 54
IV 78 53
V 80 “70
VI 34 ”71
V+Pleld 50 "go

 

‘ Larval instars are designated as in Table 1. All larvae were reared on red oak (Quergus
11112123) in the laboratory except instar Vtmd; these were field-collected larvae from the
base of red oak trees from Kaladar, Ontario.

“ Significance determined by a two-tailed binomial test with p = 0.5, or‘ = 0.1, or” =
0.05.

17

along the interface at the top of the Y after initially selecting white and walking for 2 to 3
cm on this arm, pausing, apparently looking back at the black arm and then turning

around and proceeding to the top on the black surface.

Diameter. The number of larvae attracted to a column was strongly influenced by
column diameter (Figure 2). The ranked number of larvae that did not orient to a column
within the allotted time, or that wandered out of the arena, was always significantly less
than the number of larvae attracted to the larger columns and was small compared with
the total number of larvae tested for each population. Larval tracks (Figure 3) for
neonate, third-instar and field-collected fifth- and sixth-instar larvae revealed that older
instars oriented to a column more quickly and turned less frequently (Table 4) than did

early instars.

Regression analysis of the loge transformation of the number of larvae attracted to
a column against 0 yielded similar slopes and elevations for each population, each of
which was significantly different from zero (0t=0.05); the equation for the combined
regressions was: Loge Y = 0.13x + 0.14 (n=225). The proportions of the variation
explained by the regression of “y/r for the combined instars (r1=0.5 8) were, for neonate, I,
111, v+ and Wt“, 0.52, 0.51, 0.61, 0.55 and 0.71, respectively. There was very little
difference among instars in the total number of larvae attracted to each column as a
function of 0 (Figure 4a). I tested the possibility that more larvae might have randomly
encountered larger columns as they left the center of the circle simply because the
probability of encountering a larger column was greater than that of encountering a
smaller one. This probability was defined by the proportions zei and 2'25, where i
represents the angle of the ith column and j represents the jth angle between columns for
6 and for 3. (Figure 5). Summed values for the angles 0 and 3. were 9=44.42° and

l=315.58°. Therefore, by definition, the expected distribution of larvae attracted to

18

Figure 2. The ranked position, by instar, of the number of gypsy moth larvae attracted to
columns of different diameters. Columns A to 1, respectively, represent the smallest to
the largest-diameter columns; X and Z, respectively, represent larvae that exceeded the
30-min test period and larvae that wandered out of the arena. Numbers for columns
above the same horizontal line are not significantly different (Waller-Duncan k-ratio test,
k=100). Numbers below each column are the mean ranked numbers of the first 20 larvae
attracted. Numbers within brackets are the total number of larvae attracted to each
column. (TLT = total larvae tested for each instar until 20 larvae oriented to a column;
instars tested = neonate, I, III, V+ and thidd).

19

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V+1319ch
33 ® 0@ CD. 2 G . . O
3.4 3.4 3.4 3.4 3.4 5.1 7.1 7.5 9.1 10.1 10.1
[0 0 o 0 0 3 3 11 13 30 34]

TLT.-.106

v4-
3.306%) G 2 O
3.2 3.2 3.2 5.2 5.3 5.9 3.0 3.3 7.9 9.7 10.1
[o o o 4 3 3 7 5 12 33 23]

TLT=105

m 0
6) _Z_. @O A G). G .

2.7 3.4 3.5 4.3 4.3 5.3 7.0 7.0 3.0 9.4 10.3
[0 1 2 3 4 5 1o 12 14 13 31]
rtr=1os
I
@ .@A 2 G) .G .
3.4 3.4 3.4 5.1 5.4 3.1 3.5 3.7 3.3 9.5 9.3
[o o 0 o 2 3 4 11 7 33 34]
TLT3105

Neonate Q
2 ®@ at. <9 G .
3.3 3.3 3.3 4.7 4.3 4.3 3.4 7.3 3.5 3.7 10.3
[o 0 o 2 2 3 3 1o 13 22 35]

TLT=102

Figure 3. Tracing of gypsy moth larval tracks from the diameter experiment for (a)
neonate, (b) third-instar and (c) field-collected fifth- and sixth-instar larvae. Ticks
perpendicular to a trace show the posterior position of the larva every 15 sec.

 

 

 

21

 

22

Table 4. Orientation time and accuracy in response to columns, as influenced by instar.”

 

 

Instar‘ Time“ No. turns 2 90° #
Neonate 12.4a 12.7a
I 1 1.2ab 9.1b
III 7.8b 5.7c
V+ 3.7c 2.1d
v+Field 0.4c 0.3a

 

" Larval instars are designated as in Table 1.

“ Data based on the tracks of 20 gypsy moth larvae randomly selected from each instar
tested in the diameter experiment.

# Within columns, means followed by the same letter are not significantly different
(Waller-Duncan k-ratio test, k=100; mean squared errors for time and turns = 34.3 and
25.4, respectively).

Figure 4. Numbers of each instar attracted to columns of (a) different diameters and the
same height and (b) different heights and the same diameter, as a function of the angles 0
and B, respectively, as subtended by a column from the center of the arena where larvae

were released.

 

 

DIAMETER

NUMBER or LARVAE
ATTRACTED TO e.
COLUMN

36

 

24

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12,

 

 

 

 

 

 

0.1
Neonate

 

 

 

I
H EIGHT
sworn or uavxe
ATTRACTED to _
coruuu
39
1 q -

 

26

 

 

 

 

 

 

 

 

 

 

13.

 

 

o- ------
Neonate

 

  
 
 

INSTAR

 

b

 

 

25

Figure 5. An illustration of the angles 0 and A used to calculate the expected proportion
of larvae attracted to the different columns or open spaces between columns.

26

 

e

 

 

27

columns versus those that wandered out of the arena, if columns had no influence, was,
respectively, 13 and 87 out of every 100 larvae tested. The chi-square test comparing
these values left little doubt that the attraction of larvae to a column was not random;
observed and expected values for the Chi-square were 486/65 and 14/435, respectively.
The chi-square value (x2=13.89; 0t=0.05) comparing the expected number of larvae
attracted to each column with the expected distribution of larvae based on the
proportional distribution of 0 for each column was not significantly different from a x2
with 8 df and or=0.05. The observed and expected values were, respectively: 1 versus 2, 5
versus 4, 6 versus 8, 17 versus 16, 46 versus 32, 47 versus 62, 79 versus 91, 130 versus

118 and 169 versus 167.

Height. The height of a column also influenced the orientation behavior of L. dispar
larvae. The ranked number of larvae attracted to columns with different heights
compared with larvae that wandered out of the arena and those that did not respond
within the specified time period (Figure 6) closely paralleled the observations in the
diameter experiment (Figure 2). Regression analysis of the height relationships also
yielded similar slopes and elevations for each population, each of which was significantly
different from zero (0t=0.05); the equation for the combined regressions was LogeY =
0.25x + 0.16 (n=225). The values of r2 for the height regressions, for the same instars as
those tested in the diameter experiment, were, respectively, 0.73, 0.38, 0.73, 0.57 and
0.33; the r2 value for the combined populations was 0.55. Columns that presented the
largest angle (the tallest columns) consistently attracted more larvae (Figure 4b). The
numbers of larvae attracted to a column, expressed as the expected larval distribution
based on the proportional distribution of B for the height of each column, were
significantly different from a x2 with 8 df and oc=0.05. Observed and expected values,

respectively, were 6 versus 2, 12 versus 5, 8 versus 10, 19 versus 19, 26 versus 37, 63

28

Figure 6. The ranked position, by instar, of the number of gypsy moth larvae attracted to
columns with different heights. Columns A to 1, respectively, represent the shortest to
tallest columns; X and Z, respectively, represent larvae that exceeded the 30-min test
period and larvae that wandered out of the arena. Numbers for columns above the same
horizontal line are not significantly different (Waller-Duncan k-ratio test, k=100).
Numbers below each column are the mean ranked numbers of the first 20 larvae
attracted. Numbers within brackets are the total number of larvae attracted to each
column. (TLT = total larvae tested for each instar until 20 larvae oriented to a column;
instars tested = neonate, I, III, V“ and thidd).

 

29

oo— u H: No— “ bah.

Tammaavn-~.~_ H¢~¢.-~.sc

59— mo Na 09 0v 5' 1.. 0' ..v 5.6 an

 

 

 

 

n . . L
hwou '0 cm of. mm an ac ad pi p.

«op u 5......

”8882.3... eeee:
.23 no 2 It. 3 on 2.. on an 2..

 

 

 

man

:.OL~oaw<ox .0:
D _<.

 

 

 

 

3' u 5.45.

Hafittom a en . 3
e22. 2. on mm mm 3 :3 «a mm

“-

 

 

mNodx
D _<

 

 

31... .>

«.nxou
_<U

.zcuw OD

x 0(N
4 _D

 

 

8— n #1:.

7.4.9328“... ..ooeu_
3:: an new. on on: 2” «a

.Gutow

 

ocuuu
_UQ

 

 

30

versus 68, 109 versus 108, 124 versus 120 and 133 versus 132, which produced a chi-

square value of 22.35.

Species of Tree. The numbers of larvae attracted to bolts of authentic tree trunks were
noticeably affected by the species of tree (Table 5). Consistently more larvae (0L=0.05)
were attracted to bolts of red oak than to bolts of either trembling aspen or paper birch,
for each instar tested. The numbers of larvae that wandered out of the arena or that
exceeded the allowed time period were significantly less than the number attracted to red
oak for each population (Ot=0.05). As might be expected from previous diameter and
height results, the youngest larvae (neonates) exceeded the 30-min time period most
frequently. However, the numbers that exceeded the time period or that wandered out of
the arena were also small compared with the total numbers of larvae tested. The results
were almost identical for laboratory-reared instars. Red oak was also preferred by field-
collected fifth- and sixth-instar larvae (a =0.05); however, trembling aspen and paper
birch were selected second and third in order of preference and the preference for each

was significantly different from that for red oak and each other.

Discussion

Larval Preference for Black versus White Artificial Trees. The results in Table 1
clearly demonstrate the strong attraction of all gypsy moth larval instars crawling on a
horizontal surface for a black vertical column. In my experiments, this behavior was not
affected by the storage of egg masses for several months or the use of artificial diet.
First-instar larvae that emerged from eggs held until 15 September (I-SEPT) and larvae
reared to the third instar on‘ red oak (III-OAK) showed the same preference for black
columns as did other populations. The choice of larvae on kraft paper on the floor (I-

PAP) for black columns suggests that larvae do not depend on silk trails, pheromones r

31

Table 5. Larval gypsy moth response, by instar, to vertically positioned bolts 13.8 i 3.1

cm wide by 1 m long, from three naturally occurring host tree species.

 

Mean number of larvae recovered‘

 

 

Treatment Neonate I III V+ V”:ield
red oak 15.2a 16.6a 15.2a 18.0a 7.6a
trembling aspen 1.0c 0.6b 2.4b 1.0b 5.4b
paper birch 0.2c 0.8b 1.6bc 0.8b 2.6c
exceeded time limit“ 3.6b 1.2b 1.2bc 0.0b 0.0d
exited arena 0.2c 0.8b 0.4c 0.0b 5.0b

 

‘ Instars are identified as in Table 1. Within a column, means followed by the same letter
are not significantly different (Waller-Duncan k-ratio test, k=100; Mean squared error,
respectively, for each larval population: Neonate = 1.09, I = 1.68, 111 = 2.69, V+ = 0.83

and V+Fleld = 1.25)
"30 min.

32

chemical substances associated with silk for orientation; these are important cues that
influence the larval behavior of other lepidoptera (Fitzgerald and Peterson 1988).
However, my experiments do not rule out the possibility that silk trails, pheromones and
possible kairomones of the host do not complement host finding and larval orientation.
Liebhold e_t a1. (1986) found that when they removed silk trails from trees by scrubbing
them with a brush, they were unable to eliminate the possibility that larvae used
pheromones or possible chemical substances associated with the trails to find burlap

bands in which to rest.

While climbing on the vertical surface of the Y, significant numbers of
laboratory-reared and field-collected fifth- and sixth-instar larvae selected the black side
of the column (0t=0.05), whereas first, second, third and fourth instars showed no
preference for either color (or=0.1) and neonate larvae preferred white (0t=0.1). The
differences between larval behavior on a horizontal plane (Table 1) and that on a vertical
surface (Table 3) could be explained by several hypotheses. For example, the selection of
black columns by all larval instars on a horizontal surface suggest that larvae orient to
dark objects contrasted against light backgrounds. This scenario occurred both in the
laboratory and the field, where black columns were contrasted against either light beige
walls or the background of sky light. In the field, tree trunks would also be contrasted
against sky light, but it is also possible that larvae discriminate between objects that
present contrasting images in infrared or ultraviolet light. In the laboratory, Hundertmark
(1937a) found that first-instar larvae of the nun moth, Ma mm (L.), never
approached posts whose color was the same as the background. Irrespective of the
mechanism employed, the ability to identify dark columns (tree trunks) by dispersing
first-instar gypsy moth that land on the forest floor (Leonard 1981), late-instar larvae

moving between trees (Barbosa 197 8; Liebhold 91 al. 1986), or larvae accidently

33

dislodged from a tree, is highly advantageous for survival because it allows displaced

larvae to find a prospective host quickly and effectively.

The examination of black versus white preferences of larval gypsy moth was
initiated as a preliminary investigation of field observations of fifth- and sixth-instar
larvae on paper birch and trembling aspen. Observations indicated that late-larval instars
selected dark areas on the bole of these trees as resting locations during the day. This
behavior is indiscernible on trees with dark bark (red oak) or trees with uniformly
colored bark on which larvae resting locations do not contrast with the adjacent bark. A
possible hypothesis that explains the selection of a dark color by climbing fifth- and
sixth-instar larvae (Table 3) is a behavioral response designed to camouflage late-instar
larvae from predators that hunt visually on lightly colored trees or trees without
protective bark flaps. In one study (Campbell et al. 1975a,b), 590% of the late-larval
instars died, but only 7.5% of the mortality could be attributed to parasites and disease. It
was suggested that predators that hunt visually were the principal mortality factor; larval
survival was attributed to the selection of protected resting locations (Campbell and

Sloan 1976).

The selection of dark areas by late instars may also be influenced by temperature;
caterpillars, especially dark-colored ones such as those of gypsy moth, are susceptible to
heat prostration and desiccation (Doane and Leonard 1975; Dethier 1989). However,
Knapp and Casey (1986) presented a strong argument that gypsy moth caterpillars do not
regulate their body temperature. Further studies would be useful to evaluate how

temperature influences the choice of resting locations by gypsy moth.

The attraction of a significant number of neonate larvae (0t=0.1) to a white

surface while climbing was not exhibited by first-instar larvae that had been fed, or by

34

any of the other instars tested. Neonate larvae are positively phototropic and negatively
geotropic (Leonard 1981; Weseloh 1989); after leaving an egg mass, they climb to the
top of a tree to disperse before feeding (McManus 1973). It is possible that the attraction
of neonate larvae to white is a response to higher levels of reflected ultraviolet light from
a white surface. In a forest stand, the intensity of ultraviolet light above the forest canopy
would be greater than within the stand and would attract larvae to the canopy for
dispersal. However, larval attraction to the forest canopy may be suppressed by feeding,
as when satiated first instars become too heavy to disperse effectively and it is more
important for larvae to prepare a mat of silk to ameliorate melting (Leonard 1967). The
indifference to either color by other larval instars (second, third and fourth) while
climbing is also possibly a result of feeding. In a study that investigated larval orientation
reactions in response to stimulation by light, Wellington (1948) found that third-instar
spruce budworm, Mama Marja (Clem), larvae did not respond to light when
starved. Satiation may induce a similar behavioral response in gypsy moth. However, this
does not explain the indifference of fourth-instar larvae to either color and the difference
in preference for black among fourth, fifth and sixth instars. These instars migrate down
trees in the morning to select resting locations for the day and move back into the crown
at night to resume feeding (Leonard 1981; Roden et al. 1990). The fourth instar of gypsy
moth larvae may represent a transitional period before genes that control the selection of
dark areas on a tree by fifth and sixth instars are activated. It is also possible that the
preference of fifth- and sixth-instar larvae for black is influenced by diet. Lance 91 al.
(1986a) suggested that changes in the feeding-rhythm behavior of late-instar gypsy moth

larvae may be influenced by defoliation-induced changes in leaf quality.

The preferences of larvae on vertical and horizontal surfaces were unaffected by
their sex except for female neonate larvae. It is possible that this discrepancy can be

attributed to the subsequent random mortality of neonate females that chose white and

35

the reduced sample size for comparison that resulted from dividing the initial population

by two when identifying sex.

Diameter, Height and Species of Tree. There was no observed difference among larval
instars for response to either the horizontal or vertical angles presented by a column. A
nonsignificant chi-square value for the proportion of larvae attracted to a given diameter
as a flmction of 0 confirms that attraction to a column (and, presumably, to tree trunks) is
a direct function of the angle. That is, when all other stimuli are equal, larvae will be
attracted to the host with a diameter that presents the largest angle. Of course, the angle
will vary depending on the distance between the caterpillar and the host and where there
are two hosts with equal diameters, the nearest host will present the largest angle and,
hence, be more attractive. Laboratory observations of larval attraction to the nearest

column support this (Table 2).

The chi-square value for height was significantly different (0t=0.05) for the
expected larval distribution as a function of B (Figure 4b). Differences between the
observed and expected values for the two shortest columns contributed 17.8 units to the
chi-square value of 22.35 and indicate that proportionally more larvae were attracted to
the two smallest columns than would have been expected. There are several possible
explanations why the attraction of larvae to the different column heights is not distributed
as a function of B. For example, field-collected larvae accounted for 6 out of the 12
larvae attracted to the 40-mm column. It is possible that this disagreement can simply be
attributed to a type II statistical error. Hungry, field-collected fifth- and sixth-instar
larvae not accustomed to being restrained may have been agitated by being confined
overnight in a Petrie plate. Th‘ese larvae were the most active and spent the least time in
the arena before choosing a column (Table 4). Alternatively, the efficiency of finding a

host may have been adversely affected because the upper portion of larger columns

36

disappeared against the background of the forest canopy and, hence, they did not appear
much taller than shorter columns. It is also possible that the sensitivity of larvae to
vertical angles (height) is less than it is for horizontal angles (diameter) and that height,
as a stimulus, may be less important than diameter for host colonization. Nevertheless,
observations about diameter and height confirm that the horizontal and vertical visual
angles subtended by an object are both important linear dimensions that strongly
influence orientation of larva of L. 511593;. Similar findings of crude form perception
have been reported. Saxena and Khattar (1977) reported that larvae of fiapi_li_g dmgleus
L. recognized differences in the vertical and horizontal angles subtended by vertical
sheets of Citrus leaves. Hundertmark (1937b) also observed that larvae of L. monacha L.
show a greater response to larger objects than to smaller ones offered together at the
same distance and, when larvae were presented with objects of different heights set at
different distances so that the visual angles were the same, almost all larvae showed

equal preferences for the objects.

A significant preference by all larval instars for red oak bolts compared with
trembling aspen and paper birch bolts (Table 5) suggests that gypsy moth larvae
discriminate among different tree species based on color because other studies indicate
that it is unlikely that the olfactory range of caterpillars exceeds a distance of 1 cm
(Saxena and Khattar 1977; Cain 91 al. 1983). Hundertmark (1937a) found that larvae of
the nun moth distinguished between colors. From preliminary behavioral experiments I
have conducted (unpublished data), it appears that larval gypsy moth can also detect

different wavelengths of light.

The ability of gypsy moth caterpillars to distinguish between vertical and
horizontal linear dimensions and the possibility that contrasting images or the wavelength

of light presented by an image, or both, influence larval behavior has important

37

implications for the colonization of many species of trees and may influence the
population dynamics of gypsy moth. For example, species of trees with light bark that is
not strongly contrasted against skylight, or trees with bark that reflects different
wavelengths of light may be less attractive than preferred species (i.e., oak) with darker
trunks, which may be correlated with high-quality food resources. Since significantly
more larvae in my experiment were attracted to red oak than to either paper birch or
trembling aspen (Table 5), it is likely that proportionally more red oak would be
colonized in northern hardwood forests in which red oak, paper birch and trembling
aspen are commonly found together. Where low winter temperatures do not inhibit
overwinter survival (Sullivan and Wallace 1972), gypsy moth larvae at non-outbreak
population levels may colonize red oak in mixed stands, not only because these trees
provide superior nutritional suitability and palatability (Hough and Pimentel 197 8;
Barbosa and Greenblatt 1979) and leaves with a lower pH, which enhances survival
(Keating and Yendol 1987), but because red oaks are also frequently the tallest, darkest
and largest-diameter trees found in a forest. Further, since gypsy moth larvae are
attracted to the largest-diameter and tallest objects, it is likely that the identification of
these stimuli may also explain why "wolf trees" sustain high population levels of gypsy
moth for several seasons. Mason and McManus (1981) speculated that high larval
population levels were maintained on "wolf trees" (a large-diameter branching white oak,
Om 31123 L.,) because the fractured silk threads of dispersing first-instar larvae
became entangled with the many branches on this type of tree and first-instar larvae did
not become dislodged during dispersal. In addition, I suggest that this phenomenon also
occurs because these trees are the tallest and largest-diameter trees in a forest and, hence,
attract more larvae than other trees and, because of their size and microclimates (i.e.,
under bark flaps), which enhance survival (Mason and McManus 1981; Lance and

Barbosa 1982), support large populations.

38

How increasing the distance from a host will affect larval behavior and the
relative efficiency of finding a host at different distances is unlmown. All my
measurements of the stimuli that affect gypsy moth visual foraging behavior were made
at distances of 0.5 m. Dethier (1989) suggested that the effective distance for crude form
perception in caterpillars is not likely to exceed 30 cm, and cited work by Ichikawa and
Tateda (1982) to support this contention. The distance reference cited by Ichikawa and
Tateda was quoted from Hundertmark (1937b) and is the distance at which Hundertmark
found that caterpillars could no longer distinguish between differences in height of 1 cm
for columns with widths between 13 and 19 cm. A distance of 30 cm, however, should
not be considered the maximum distance at which caterpillars can perceive different
crude shapes. Fifth- and sixth-instar larvae in my experiments directly approached
columns from distances of 0.5 m (Figure 3c) and Doane and Leonard (1975) found that
larval gypsy moth are attracted equally to vertical objects at distances of 1 and 2 m; in
their experiment, approximately one-quarter of the larvae tested responded at 3 m, but no
response was obtained from larvae at 4 m. Differences of 1 cm in the height of an object,
or in the angle presented by this difference, are probably not significant in the field. It is
more likely that an optimal height-diameter ratio exists for a perceived object and that
the combination of height and diameter has a synergistic effect, but this ratio and the
distance for effective perception of differences in horizontal and vertical angles by gypsy
moth larvae must be quantified by further research. Such synergistic effects have been

noted for adult Diptera (Moericke e_t al. 1975; Miller and Harris 1985).

My study suggests that host selection by L. dispar larvae involves information
from different sensory modalities. The suggestion that selection of a host by gypsy moth
can be best explained by an inverse relationship between foliage quality and the
frequency of dispersal (Lance 1983) is too naive. The experiments in this chapter clearly

demonstrate that visual cues are important stimuli for this herbivore. Visual and chemir

39

cues may complement each other and influence host selection either simultaneously or
sequentially, or both; however, the mechanism for this is unknown. For example, L.
djspar may respond to olfactory stimuli (kairomones) emanating from the host in '
conjunction with visual stimuli and, on reaching the host, initiate feeding in the presence
of chemotactile stimuli that influence larval acceptance or rejection. Alternatively, only
visual stimuli from the host may be responsible for larval orientation and, on reaching the
host, chemotactile stimuli determine if the host is rejected or accepted. The identification
of visual stimuli that influence host colonization by larval gypsy moth has important
implications for the use of burlap bands in monitoring gypsy moth larval and pupal
populations. Because larvae are influenced by the species, diameter and height of the
host, these variables may be important considerations for standardizing burlap banding in
operational monitoring systems. The color of the burlap material used should also be

considered.

The foraging repertoire and the ecological responsiveness of the gypsy moth is
richer than previously envisioned. In addition to chemical cues, visual cues are important

stimuli that influence host colonization by this herbivore.

CHAPTER TWO

Influence of Current Defoliation and Different Tree Species on Gypsy Moth,
Lymanttla dim (L.), Growth and Development

Introduction

Insect injuries can often elicit physiological changes in plants that render them
less suitable for future herbivore generations (Haukioja and Niemela 1977; Niemela e_t al.
1979; McClure 1977, 1979, 1980). It may also increase within- (Way and Cammell
1970; Whitham 1983) and between-plant variation (Schultz 1983) in nutritional and
defensive chemistry, which can affect herbivore feeding and host acceptance (Feeny
1970; Raupp and Denno 1983; Meyer and Montgomery 1987). Such changes have been
described as rapid induced resistance (RIR) and delayed induced resistance (DIR).
However, their adverse efi'ects on insect populations are far from clear (Wratten et a1.
1990). It has been suggested that RIR may stabilize insects densities (Haukioja 1990),
whereas DIR contributes to population cycles and the decline of outbreaks (Haukioja e_t
a1. 1988). In contrast, some studies have reported negligible effects (Fowler and Lawton
1985), or even improved growth and survival on trees which have been previously

damaged (Niemela e_t fl. 1984; Haukioja et al. 1990).

Several studies have investigated the relationship between the gypsy moth,

Lxmarmia djsnar (L), defoliation and host plant quality (Wallner and Walton 1979;

40

4l

Schultz and Baldwin 1982; Valentine :4 a1. 1983; Rossiter e_t 31. 1988; Barbosa 91 al.
1990a, 1990b). Without exception, they have established that tree species common to the
forests of New England, such as red oak, Quergrs rum L., produce defensive plant
chemicals that can adversely affect larval development. However, little is known about
trembling aspen, PM Emulsides Michx., and paper birch, Bituia papm’fera Marsh,
two very abundent trees in the Great Lakes basin. It has been reported (Witter e_t a1. 1990;
Roden and Surgeoner 1991) that trembling aspen produces more fecund pupae faster than
red oak. It has also been suggested that outbreaks of gypsy moth have occurred in stands
of trembling aspen (Leonard 1981; Witter e1 31. 1990). However, the outbreak potential
of less preferred hosts, such as trembling aspen and paper birch (Houston 1979), has
never been investigated. It is not only important to identify those species of trees in the
forests of the Great Lakes which provide the greatest nutritional potential as hosts for the
development of gypsy moth, but also to determine if damage-induced changes in plant
chemistry occur, and if so, how these host changes affect larval development and growth.
In this study I compared survival, development and fecundity of gypsy moth larvae
reared on trembling aspen and paper birch to those of larvae reared on red oak, at two

levels of defoliation designed to simulate hosts with and without RIR.

Materials and Methods

In 1986 I selected a 2-ha site situated 10 km northeast of Arden, Ontario
(44°43'N, 76°56'W) which was typical of sites where gypsy moth would be expected to
do well (Houston and Valentine 1977 ; Houston 1979). The site consisted of saplings that
had reestablished on abandoned farmland and were inspected in 1986 and 1987 to insure
that they had not been defoliated or infested with other herbivores. In 1987, numerous
forest tent caterpillar, Malamsgma dissm'a an., eggs were found on several trees. These

were removed by hand from experimental trees in the spring of 1988 and then the basr

42

each tree was coated with Sticlcum® (Seablight Enterprises, 4026 Harlan St., Emeryville,
CA) to prevent inter-tree herbivore movement. Study trees were dispersed throughout the
stand; the mean diameter and height were 5.51 i 0.18 cm and 5.9 :l: 0.08 m, respectively.
Each tree was selected so that it had full exposure to the sun because shading may

decrease concentrations of carbon-based secondary metabolites such as phenolics

(Larsson :1 a1. 1986; Mole et a1. 1988).

I used larvae from gypsy moth egg masses that I collected randomly on 10 April
1988 at the edge of a new infestation at Kaladar, Ontario (44°39'N, 77°07'W). These
were surface-sterilized (Shapiro 1977) and then held a +5°C until incubation so that
larval emergence could be synchronized with gypsy moth emergence in the field. A
randomized complete block design, with a two by three factorial arrangement of
treatments, was used to compare gypsy moth survival, pupal weight and the period of
time required by larvae to pupate. The two experimental factors, level of defoliation (0
and 50%) and tree species (trembling aspen, paper birch and red oak), were each
replicated five times. Gypsy moth larvae, reared on artificial diet (Bell 91 a1. 1981) in the
laboratory until the second instar, were held at 10°C until there were sufficient larvae for
each replication. Five larvae were then selected at random and enclosed with foliage
from a terminal branch in four 50- by 30-cm nylon mesh bags on each tree. During the
larval feeding period, which lasted from 26 of May to 17 July, nylon bags secured around
the opening with fabric ties, were randomly dispersed throughout the crown of each tree
and were moved to new terminal shoots weekly. Foliage in a bag was always sufficient to
insure that larvae never consumed more than 50% of the available food. Shoots that were
used as feeding sites were identified with flagging tape and were not used again. Light
transmission through the nylon bag, measured with a light meter, was reduced by .=_ 15%.
The placement of larvae on the trees was synchronized with second-instar larval

development in the field; replications 1 to 5 were placed on trees on 26, 27, 28, and 30

43

May and on 3 June, respectively. To reduce experimental variation, defoliated and
undefoliated trees within each replication were grouped together by the proximity of '
their location within the stand, their diameter and height, the day that second-instar
larvae were placed on them and the date leaves were torn in an attempt to simulate
defoliation. The mean initial second-instar larval weight (measured on the day larvae
were placed on the tree and after they had been removed from storage for 12 h) was used
as a covariate. Pupae were collected from the nylon bags and weighed within 24 h of

pupation.

I simulated gypsy moth defoliation by tearing leaves by hand (parallel to the leaf
mid-rid) throughout the canopy of each tree on 3 occasions (6-10 June, 15 June and 22
June) at defoliation levels that closely corresponded to the amount of foliage removed by
gypsy moth on trees in a nearby infestation. For the first two tearings, every 10th leaf
was torn in half (5% defoliation); this coincided to the damage that occurred when the
majority of larvae reached third- and fourth-instar development in the field. On the final
tearing date, every remaining leaf was tom (50% cumulative defoliation) coincident with
fifih- and sixth-instar larval development. The tearing of foliage for the first defoliation
was completed at the rate of one replication per day; for the last two tearings, all
replications were completed within a day. Tukey's test (or = 0.05) was used for separation
of the means for gypsy moth pupal weight, development time and survival. Data from the

fifth replication were excluded from the analysis because of high larval mortality.

Results

Pupal weight. Both defoliation and tree species affected female pupal weight, but their

interaction was not significant (Table 6). Female gypsy moth larvae that fed on

defoliated trees produced pupae that weighed ca. 12% less than females that fed on

44

Table 6. Analysis of variance of gypsy moth pupal weights, as influenced by defoliation

and tree species.

 

 

 

 

Female Male
Source df MSE F df MSE F
Total Corrected 35 63
Replication 3 0.072 3.58‘ 3 0.044 7.45”
Defoliation 1 0.186 9.91” 1 0.004 0.78
Tree species 2 0.826 41.16‘“ 2 0.059 10.04“
Defoliation X Tree species 2 0.040 2.01 2 0.002 0.43
Covariate 1 0.148 7 .37‘ 1 0.038 6.46‘
(mean L2 weight/cage)
Experimental error 9 0.020 0.77 14 0.005 2. 18‘
(variance between trees)
Sampling error 17 0.026 40 0.002

(variance between cages)

 

" Significant at or=0.05
“ Significant at 0t=0.01
”‘ Significant at 0t=0.001

45

undefoliated trees (Table 7). The smallest female pupae (0.849 g) came from red oak
(Table 8); these were significantly smaller than pupae from females feeding on paper
birch (1.13 g), which were intermediate and significantly smaller than pupae from
females feeding on trembling aspen (1.43 g). Tree species was the only factor that
affected male pupal weight: oak < paper birch < trembling aspen (Table 7 and 8).
Interestingly, the covariate (mean initial second-instar larval weight) was significant for
both female and male pupal weights (Table 6). This, implies that small differences in
initial size among early instars are still significant and expressed after weeks of feeding

on their host plant.

Development Time and Survival. The number of days required by both male and
female larvae for development from second-instar to pupation was not affected by
defoliation, only by host species (Table 9). Female larvae that fed on trembling aspen
required 4 fewer days to complete development than did females reared on red oak.
Female larval development time on paper birch was intermediate, but not significantly
different from times on either trembling aspen or red oak (Table 10). Similarly, male
larvae that fed on trembling aspen developed about 4 days faster than larvae that fed on

either paper birch or red oak (Table 10).

Gypsy moth mortality, observed bi-weekly, was apparently unaffected by either
defoliation or tree species (Table 11). However, it may have been masked by severe
larval mortality that resulted from Mars plagidus Uhl. (Pentatomidae), which attacked
larvae through the nylon screening as they crawled on the interior surface of the bag. The
number of larvae that escaped from cages in the experiment or that died as a result of

nuclear polyhedrosis virus (NPV) infections was negligible (Table 12).

45

Table 7. Mean pupal weight for gypsy moth larvae reared on undefoliated and defoliated
trees, averaged over host species. Mean standard errors for females and males = 0.02 and
0.005, respectively.

 

 

 

Mean pupal weight (g)‘
Treatment Female Male
Undefoliated 1.144a 0.433a
Defoliated 0.999b 0.423a

 

" values in a column followed by different letters are significantly different (P<0.05,
Tukey's honestly significant difference test).

Table 8. Mean pupal weight for gypsy moth larvae reared on trees of three host species,
averaged over defoliated and undefoliated treatments. Mean standard errors for females
and males = 8.86 and 7.13, respectively.

 

 

 

Mean pupal weight (g)‘
Treatment Female‘ Male‘
Trembling aspen 1.429 a 0.494a
Paper birch 1.130b 0.404b
Red oak 0.840c 0.386b

 

‘ values in a column followed by different letters are significantly different (P<0.05.
Tukey's honestly significant difference test).

47

Table 9. Analysis of variance for gypsy moth development time, as influenced by
defoliation and tree species.

 

 

 

 

Female Male
Source df MSE F df MSE F
Total Corrected 35 63
Replication 3 126.11 14.24” 3 84.27 11.81”
Defoliation l 20.30 2.29 1 4.68 0.66
Tree species 2 40.10 4.53‘ 2 47.08 6.60“
Defoliation X Tree species 2 8.83 1.00 2 0.30 0.04
Covariate l 2.30 0.26 l 20.33 2.85
(mean L2 weight/cage)
Experimental error 9 8.86 2.07 14 7.13 1.84
(variance between trees)
Sampling error 17 4.27 40 3.87

(variance between cages)

 

‘ Significant at or=0.05
” Significant at Ot=0.01

43

Table 10. Mean development time (days) for female and male gypsy moth larvae, as
influenced by host species averaged over defoliated and undefoliated treatments. Mean
standard errors = 8.86 and 7.13, respectively.

 

Mean development time (days)

 

 

Treatment Female‘ Male‘
Trembling aspen 46.1a 40.8a
Paper birch 48.4ab 43.6b
Red oak 50.4bc 44.0b

 

O

values in a column followed by different letters are significantly different (P<0.05,
Tukey's honestly significant difference test).

49

Table 11. Analysis of variance for gypsy moth survival, as influenced by defoliation and
tree species. No results were significant at or=0.05.

 

 

 

 

Female Male
Source df MSE F df MSE F
Total Corrected 35 63
Replication 3 0.747 0.77 3 0.684 0.43
Defoliation 1 0.012 0.01 1 0.459 0.29
Tree species 2 0.970 1.00 2 0.585 0.37
Defoliation X Tree species 2 0.761 0.79 2 1.319 0.83
Covariate 1 1.724 1.78 1 0.001 0.00
(mean L2 weight/cage)
Experimental error 9 0.969 1.45 14 1.584 1.37
(variance between trees)
Sampling error 17 0.669 40 1.153

(variance between cages)

 

Table 12. Source of mortality and number of larvae surviving from second-instar to adult
eclosion for female and male gypsy moth larvae reared on defoliated and undefoliated
host species.

 

 

 

 

Deaths from Survivors
Treatment 11 NPV‘ P.. 213214115 Female Male
Undefoliated trees 240 0 141 27 62
Defoliated trees 240 2 134 29 70

 

‘nuclear polyhedrosis virus

50

Discussion

The primary focus of this study was to determine if two of the most common
deciduous trees found in the forests of the Great Lakes region respond to defoliation with
an immediate or rapid induced response (RIR) that impairs the development of gypsy
moth larvae. At the same time, I also wanted to compare the nutritional suitability of
these species in the field with that of red oak. In the laboratory, Roden and Surgeoner

(1991) reported that trembling aspen and paper birch hosts were superior to red oak for

gypsy moth.

Other studies besides have shown that the nutritional quality of food for gypsy
moth declines after severe defoliation (Wallner and Walton 1979; Schultz and Baldwin
1982; Rossiter it al. 1988). There is evidence that leaf damage occurring in the spring
has a much more detrimental effect on herbivore growth and development than damage
later in the season (Wratten 91 a]. 1984). Moreover, insect-caused damage may elict a
stronger defensive reaction than artificial damage (Hartley and Lawton 1990; Wratten e_t
11. 1990). Finally, all three tree species used in the present study are hosts that are
characterized by leaf chemistry comprised mostly of phenolics (Barbosa and Krischik
1987), which have been shown to vary inversely with gypsy moth pupal weight and
fecundity (Rossiter g 31. 1988). Thus, differences that occurred in the present study can
only be attributed to differences in foliage quality that resulted because of defoliation or

that were characteristic of the host species.

Mean pupal weights for female and male gypsy moth larvae reared on
undefoliated red oak in this experiment fell within the range reported by Wallner and
Walton (1979) for female and male L. dim reared on undefoliated black oak, Quergus
291113114 Lam., 0.90 and 0.41 g versus a range of E 0.75 to 1.05 and E 0.34 to 0.42 g,

51

respectively. Male pupal weights from larvae fed on defoliated oaks in both studies were
also comparable (0.36 g in this study versus an range of E 0.32 to 0.36 g in their study).
However, the mean female pupal weight (0.68 g) of larvae fed on defoliated red oak in
this study was less than the lowest female pupal weight (range 0.69 to 0.81 g) from

larvae fed on comparable black oak and noticeably less than the _=. mean (0.76 g) for the

different sites reported by Wallner and Walton.

Male pupal weights (a range of E 0.35 to 0.42 g) on undefoliated grey birch,
Leyla Ma Marsh, in Wallner and Walton (1979) were comparable to those on
undefoliated paper birch in my study (0.41 g ). Mean male pupal weight on defoliated
paper birch trees was ca 20% higher than that on defoliated grey birch (0.40 g in my
study versus a range of E 0.32 to 0.34 g). Pupal weights of female larvae that fed on
undefoliated and defoliated paper birch in my study were both consistently higher than
those of larvae that fed on gray birch (1.17 and 1.07 g versus a range of E 0.69 to 1.05
and 0.63 to 0.87 g, respectively). Thus, although there are small differences between
female pupal weights among different studies on oak and birch, the male pupal were very

similar on all trees. There are several possible explanations for this.

Because male larvae have only five instars compared to the six of females, it is
possible that the shorter development time, and hence the lower consumption of
defoliated foliage, does not affect male development to the same degree as female
development in the first year of defoliation. A second year of subsequent defoliation
might have exposed male larvae in my study to increased levels of plant phenolics and
other qualitative changes in foliage that may trigger developmental abnormalities and
reduced pupal weight. In the second year of the Wallner and Walton (197 9) experiment,
male pupal weight was significantly affected by defoliation whereas female pupal weight

was not. Another possible explanation is that female development may be affected more-

52

by a nutrient or water imbalance than male development. It has been demonstrated that
slower growth and less efficient utilization of plant material and nitrogen may be
attributed to reduced leaf water content (Scriber 1977 ; Mattson and Scriber 1987).
However, it is also conceivable that differences between my study and that of Wallner
and Walton could be attributable to disparities between the way trees were defoliated.
The defoliation treatment in my experiment was applied similarily to all species while
the one used by Wallner and Walton was different for oak and birch. This discrepancy
may also explain the absence of a defoliation X tree species interaction in my
experiment, which implies that RIR occurred in a parallel way in the tree species
examined; whereas, the presences of a defoliation X tree species interaction in Wallner

and Walton suggests RIR does not occur similarly in all trees species.

The time required for larval development in my study was not affected by
defoliation; however, host species strongly influenced the time required for development
(Table 9). Larvae always developed slower on red oak than on trembling aspen and paper
birch (Table 10). These results differ somewhat from those of Wallner and Walton
(1979), who found that defoliation of oak prolonged larval development; the
development of larvae on grey birch was not affected by defoliation, and this agrees with
my observations on paper birch. My observed development time on red oak for both
sexes was s 10% less than that reported by Wallner and Walton for gypsy moth larval
development on black oak. The times required for female and male larval development
on paper birch were 23 and 20% less than larvae developing on grey birch. Shorter
developmental periods for northern strains of gypsy moth have been reported by Leonard
(1966), who found that gypsy moth from Quebec developed faster than those from
Connecticut. He attributed these differences to the adaptation of larvae to a northern

climate.

53

Pupal weights and development times on undefoliated hosts in this study are E 25
and 32% less, respectively, than the pupal weights and development times reported by A
Roden and Surgeoner (1991) for larvae reared on excised foliage of the same hosts in the
laboratory at constant temperature and humidity. Male and female pupal weights of
larvae reared on undefoliated trembling aspen and red oak in my study were 5 43 and
44% and 30 and 16% heavier, respectively, than pupal weights of gypsy moth larvae
reared on the same tree species in lower Michigan in 1988. Witter e_t a1. (1990) found that
pupal weight varied significantly by year, site, tree species and trembling aspen clone.
Observation of these sources of variation would explain differences in pupal weight
between their study and ours for larvae reared on the same tree species. They would also
explain differences in pupal weight and development time between my study and those
reported by others (Wallner and Walton 1979; Roden and Surgeoner 1991). Disparities
may have occurred because of differences in rainfall, temperature and geographic
location that affected the insects, the trees or both; although, possible differences
between field and laboratory studies that result from laboratory rearings at constant

temperature and humidity should not be discounted.

Close correspondence of my pupal weights with those of Wallner and Walton
(1979), comparable pupal weights with those reported by Maksimovic (195 8) and
observation about gypsy moth development on trembling aspen that are similar to those
reported by others (Witter et a1. 1990; Roden and Surgeoner 1991) or studies that have
reported that gypsy moth development on oak does not always yield the most fecund
pupae (Barbosa 1978; Lance and Barbosa 1982; ME. Montgomery and J .C. Schultz,
personal communication), impart confidence to my findings and reinforce my belief that
the results I obtained would parallel unfretted larval development in the field. It is
noteworthy, however, that all studies of gypsy moth development on trembling aspen

share one particular finding: pupal weights of larvae fed on trembling aspen, a host not

54

traditionally associated with outbreaks of gypsy moth, are heavier than pupal weights of
larvae that fed on red oak, a traditional host of L. M. Since gypsy moth development
on trembling aspen produced potentially more fecund pupae in a shorter development
time than larvae reared on red oak, an immediate question is whether pure stands of
trembling aspen, paper birch or both will (or can) support outbreaks of gypsy moth in the

forests of the upper Great Lakes.

Survivorship, development and reproductive success of an insect depends in part
on its ability to select appropriate sources of nutrition. Consequently, an insect's host
preference is normally closely correlated with the most nutritionally suitable host in its
environment (Tabashnik 1986). Nevertheless, mismatches between host preference and
nutritional suitability may occur. For example, imprecise correspondence between host
preference and nutritional suitability has been attributed to ecological factors such as
parasitism and predation (Smiley 1978; Price 91 al. 1980) which can make the most
nutritious host inappropriate, or host finding and acceptance behavior that is not directly
related to host suitability because suitable habitats are avoided (Singer 1971; Chew
1981). I suggest that the imprecise correspondence between gypsy moth and hosts that
have been identified as potentially more nutritious exists because gypsy moth
development and potential population increases are restricted on these hosts due to the

effectiveness of NPV and to physical features of the host that affect larval behavior.

The primary natural regulator of North American gypsy moth populations is NPV
(Podgwaite 1981). Gypsy moth survival in my study was not affected by tree species or
defoliation (Table 11); however, these results should be viewed conservatively. Egg
masses used in my study were sterilized to reduce larval mortality as a result of NPV;
only 2 larvae (< 0.05%) showed symptoms that could be attributed to NPV (Table 12). In

naturally infested stands, a much higher percentage of the gypsy moth population woul

55

be infected. Reports of gypsy moth NPV mortality that average 36% are not uncommon
and some estimates are as high as 56% (Podgwaite 1981). Gypsy moth susceptibility to
NPV is influenced, among other things, by the pH of the leaf tissue consumed (Keating
and Yendol 1987). The lower foliage pH of traditional gypsy moth hosts, such as red
oak, is thought to buffer the high midgut pH of gypsy moth (Schultz and Lechowicz
1986) and make larvae less susceptible to the polhyhedral inclusion bodies (PIB) of
NPV, which require alkaline conditions for solubility and infection (Keating and Yendol
1987). Hydrogen or covalent bonding, or both between phenolic compounds and viral
proteins are thought to be responsible for inhibition of NPV (Keating gt a1. 1988; Keating
e_t a1. 1990). Schultz 91 al. (1990) found that the high LDso value for gypsy moth from
ingested PIBs in stands of oak with high hydrolyzable tannin contents was primarily
associated with reduced NPV effectiveness compared to condensed tannins, but small
compared with the LD50 for ingested PIBs on tree species like trembling aspen. For this
reason, I believe that the success of gypsy moth in the Great Lakes basin will be limited
in stands dominated by trembling aspen and paper birch, not only because of the
demonstrated effectiveness of NPV on such hosts, but also because larvae are highly
attracted to oak because of its darker color and its larger diameter and height (Roden e_t
al. 1992). Gypsy moth may invade trembling aspen strands, and population growth may
occur, but I expect that the incidence of NPV will increase rapidly in these populations

and that they will collapse quickly.

The effectiveness of NPV on these populations may also be enhanced by
tremulacin, a phenolic glycoside found in trembling aspen foliage. Lindroth and
Hemming (1990) indicated that increased concentrations of tremulacin in association
with increased levels of defoliation may overload the esterase detoxification capacity of
gypsy moth larvae and lead to impaired survival, growth or both. Although Leonard

(1981) reported that gypsy moth has established and maintained populations on trembl‘

56

aspen in the boreal forest where oaks are less common, specific occurrences were not
cited. Witter e_t a1. (1990) also reported outbreaks of gypsy moth in stands of trembling
aspen, but these were either in stands of aspen mixed with oak or were transient
infestations and support my contention. The lack of reports of gypsy moth outbreaks in
stands of trembling aspen in Maine, where infestations should probably have occurred by
now because of the state's proximity to the North American introduction point for gypsy
moth, or in Ontario, further support my contention. In mixed stands, where species other
than oak are common, larvae will be attracted alter first-instar dispersal to species with
darker trunks (Roden 91 a1. 1992), such as maple (Acg); several studies have shown that
species of maple are not optimal hosts for gypsy moth (Hough and Pimentel 1978 ; Lance
and Barbosa 1982; Roden and Surgeoner 1991), therefore, larval development will be
impaired. In mixed stands that contain trembling aspen and oak, it is still too early to
forecast about the magnitude of the potential threat from gypsy moth At low population
levels, research suggests that most larvae will be attracted to species of oak (Roden e_t _a_1.

1992); however, further research is required to clarify this issue for several reasons.

First, Meyer and Montgomery (1987) reported that cottonwood, 1101mm
dejtgjdes Marsh, an indeterminate species, concentrates its defenses in young leaves,
whereas species of oak, a determinate species, retain one set of heavily defended leaves
over the entire growing season. However, the production of defensive chemicals and
their association with age-related characteristics are poorly understood (Mooney e_t 31.
1983; Meyer and Montgomery 1987). A tree that can replace leaf area lost to herbivory
throughout the season may be less affected by damage than one that cannot (Coley e_t 31.
1985). Second, Schultz 91 a1. (1990) speculated that NPV is likely to play a major
regulatory role for L. dim in stands dominated by weakly inhibitory trees and a minor
role in stands comprised mainly of tree species that are inhibitory. They suggested that

the ratio between condensed and hydrolyzable tannins was responsible for the influenc

57

of gypsy moth NPV in such stands. Third, the ratio between condensed and hydrolyzable
tannins may influence or may be influenced by the type of defence response employed by
the tree (i.e., RIR versus DIR). All species in the present study evoked plant defences
that retarded gypsy moth development, and all species are characterized by a leaf
chemistry that is comprised mostly of phenolics (Barbosa and Krischik 1987). To better
understand the potential for gypsy moth development in forest stands that contain these
species, the ratio between condensed and hydrolyzable tannins should be established,
including how this relationship varies within and between tree species and how it is
influenced by defoliation and resistance. In addition, the manner in which visual stimuli
interact and influence inter-tree larval behavior should be assessed. I lost 5 60% of the
larvae in my experiment to attacks by P. plagidus (Table 12); therefore, the potential of
this predator should also be investigated in terms of its influence on gypsy moth

populations in the Great Lakes basin.

CHAPTER THREE

Laddering: A Climbing Behavior of the
Gypsy Moth, Laramie 11131131: (L)

Introduction

While investigating the potential use of ABS plastic pipe as artificial trees in a
study of the movement of gypsy moth, magma dim (L.), larvae, Roden et a1. (1990,
1992), I noticed that many larvae had difficulty climbing vertically on a smooth surface.
Some larvae, however, experienced no difficulty, although their rate of climb was slow
because they would pause frequently and move their head from side to side. Close
examination revealed that the larvae were spinning 1- to 2-cm-long sequential strands of
silk and using these like the rungs of a ladder to ascend the "tree". Although healthy
gypsy moth larvae are known to trail a single silk thread (Leonard 1967, McManus and
Smith 1972), descriptions of the use of silk as a mechanism for gaining a foothold
(Balfour-Browne 1925; Dethier 1980) are not well documented quantitatively or
experimentally. This chapter reports on laboratory experiments to describe the climbing

behavior, here termed "laddering".

Materials and Methods

In my study I used several populations of gypsy moth between January and

September from 1987 to 1989 to investigate the possibility that gypsy moth larvae were

58

59

using silk to construct a "ladder" to assist climbing on a smooth surface. Gypsy moth
larvae were obtained from egg masses collected annually in October near Kaladar,
Ontario (44°39'N, 77°07'W), and held at +5°C until diapause was completed, then were
surface sterilized (Shapiro 1977) and incubated at 22°C. Larvae were reared individually
on artificial diet (ICN Biochemicals, Cleveland, OH 44128) in 150- by 25-mm screened
plastic Lab-Tek Petrie plates at 70% RH with a 16L:8D photoperiod. To investigate the
frequency of laddering and how it was affected by the surface being climbed, a
randomized complete block design, blocked by time with five replications of 10 larvae
each, was used to observe larval climbing behavior on two 50-cm pieces of 5-cm-
diameter ABS plastic pipe. A mixture of 250 mL of black paint (Color Your World,
Toronto, Ontario M8W 3R5; paint # 5900) and 150 mL of sand sifted through a 20-mesh
screen was applied to one piece of pipe to provide a rough surface; the other piece of
pipe was covered with the same paint but without sand. The populations tested were:
neonate larvae 24 to 48 h old; fed first-instar larvae older than 48 h; second-, third-, and
fourth-instar larvae; and pooled fifth- and sixth-instar larvae. All larvae tested were held
overnight without food to increase their activity (Leonard 1967). For testing, larvae were
gently coaxed from a Petrie plate with a camel-hair brush onto a piece of wooden
doweling angled at ca. 300 against the vertical column being tested. As larvae began to

climb on the column, the doweling was gradually removed from below.

After determining the frequency of laddering for the different instars, a second
experiment, with a two X two factorial arrangement of treatments in a randomized
complete block design was conducted with third-instar larvae (because this instar had the
highest frequency of laddering) to test the hypothesis that silk was actually used as
substrate while climbing a smooth surface. The two factors examined were a larva's
spinneret (sealed with glue versus no glue) and the surface (smooth versus rough),

providing four treatment combinations: larvae with the spinneret sealed with Crazy Glue®

60

(to make it inoperative) and larvae without the spinneret glued, climbing on the previously
described smooth or rough columns. The number of larvae that were able to climb onto a
column and then climb vertically for at least 10 cm was considered an indication that silk
production was not necessary for climbing vertically; these data were analyzed with a
contingency table. The mean times for larvae completing the 25-cm climb were compared
by ANOVA. Each larva was allowed 30 min to complete the climb. For each of the 25
replications, four larvae were randomly selected from the appropriate instar and then
anesthetized by placing them in a refrigerator at 4°C for 30 min before randomly selecting
two of the four larvae for glueing. The remaining two larvae were handled by placing
them under the microscope and touching the spinneret in the same manner used to apply

glue to the other two larvae.

In 1988 I attempted to confirm the experiments conducted in 1987 with third-
instar larvae, but I was unable to reproduce the observed laddering behavior. After
examining larvae from eg masses collected at several different locations, I hypothesized
that the likely explanations for the change in behavior between the two years were: (i) I
had discovered an uncommon behavior in a particular population of gypsy moth that
might be difficult to find again, or (ii) the artificial diet (Bell e_t a1. 1981) substituted for
the ICN diet in 1988 had influenced the behavior. (I had replaced the ICN diet with Bell's
diet because larvae reared on the ICN diet exhibited developmental problems.) To
investigate these possibilities, I compared the frequency of laddering for third-instar
larvae randomly selected from several egg masses that were collected from areas near
Kaladar with high and low population levels of gypsy moth Larvae from each source
were subdivided into two groups and reared on either Bell's or the ICN diet. The
frequency of laddering and the time required to climb 25 cm on the smooth column were
analyzed separately for a randomized complete block design with a two X two factorial

arrangement of treatments, blocked by time and replicated 25 times with population level

61

(high versus low) and diet (Bell's versus ICN) as the two factors. After finding that the
ICN diet was influencing laddering, I used ICN diet to rear gypsy moth larvae for a
randomized complete block design experiment, blocked by time and diameter, with four
replications of 10 larvae on each treatment to compare the frequency of laddering and the
time required to climb 25 cm on 30-cm bolts of trembling aspen (11993th tremulgides
Michx.), paper birch (Ma WMarsh.) and red oak (925L913: 3111a L.). Because
I could not see the strands of silk used for laddering on the wooden bolts, except with
great difficulty under microscopic examination, I used the characteristic side-to-side head
movement of the larvae that I had identified in the previous experiments as an indicator
of laddering and silk production. The wood bolts, sealed at each end with paraffin to
retard desiccation, were positioned 1 111 off the floor by inserting a 2-cm-diameter piece
of galvanized pipe into a slightly smaller hole drilled into the center of the base of each
bolt. The galvanized pipe was then supported vertically by screwing the other end of the
pipe into a floor flange that was fastened with screws to an 18- by 18-cm base of
plywood. Bolts were randomly placed 1 m apart in a line between banks of overhead
fluorescent lighting in the room. Each replication was repeated with bolts collected from
different trees; bolts were randomly arranged within the line to remove the possibility
that lighting and/or drafts from the ventilation system influenced the results. The bolts
were always selected from what appeared to be the smoothest portion of the main stem
below the foliage. Larvae for each replication were released individually from one of 10
L-shaped cardboard platforms previously fastened around the base of each bolt with
Elmer's Glue-All® at 36° intervals; the order in which a platform and a treatment were
selected for larval release was chosen at random. Between trials, each belt was wiped
with a cloth to remove any silk deposited during a previous test. A Student-Newman-

Keuls test (01:0.05) was used for separation of the means.

62

Results and Discussion

In the initial experiment in 1987, first-, second-, third- and fourth-instar larvae
(Table 13) laddered. However, the frequency of this behavior was only significantly
different between the smooth and the rough surfaces for third- and fourth-instar larvae.
Neonate larvae were never observed laddering; from my observations, it does not appear
that they experienced difficulty climbing on either column. Although no laddering was
observed for fifth- and sixth-instar larvae during the experiments described in this
chapter, laddering was later observed (Figure 7) and recorded on VHS videotape for a
single sixth-instar larva climbing on a smooth column of ABS plastic pipe. Copies of the
event, which included a visible ladder, are available from the author through the Forestry

Canada, Ontario Region, library.

The experiment with the glued spinneret clearly demonstrated that larval gypsy
moth deposit silk as a substrate to assist climbing. The chi-square values for the main
effects (spinneret and surface), 7.40 and 19.58, respectively, for the number of larvae
climbing 10 cm or more were both statistically significant (p<0.05); the chi-square value
for their interaction (1.90) was nonsignificant (p>0.1). All of the 25 larvae on the smooth
column with the spinneret glued were unable to climb and repeatedly fell from the
column when attempting to do so, whereas 64% of the larvae without the spinneret glued
completed the 25-cm climb successfully. Most larvae climbing on the rough column did
not experience difficulty climbing; 92% of the larvae with the spinneret glued and 100%
of the larvae without the spinneret glued completed the 25-cm climb. Further, because
both factors were independent, this indicates a difference in climbing ability between
larvae with and without the spinneret sealed. This difference was noticeable on both the

smooth and the rough columns; this suggests that larvae also use silk somehow to assist

63

Table 13. Mean number of larvae per replicate, by instar, that demonstrated laddering

behavior on smooth and rough columns of ABS plastic pipe. All larvae were reared on

ICN diet.

 

Number of larvae‘

 

Smooth column

Rough column

 

Instar (n=10) (n=10) MSE
Neonate 0.0a 0.0a 0.0

I 0.4a 0.0a 0.40
11 0.8a 0.0a 0.85
111 6.6a 0.4b 1.35
IV 4.2a 0.0b 1.35
V + VI 0.0a 0.03 0.0

 

‘ values in a row followed by different letters are significantly different (P<0.05, Student-

Newman-Keuls test).

Figure 7. Silk ladder used by a sixth-instar gypsy moth larva to climb on a smooth
column of ABS plastic pipe.

 

 

65

 

66

climbing on rough surfaces even though these surfaces provide structures for their
crochets to grasp. Individual spinneret treatment responses, averaged across surface
types, showed that 46% of larvae with the spinneret sealed completed the climb versus
84% of larvae without the spinneret sealed; 96% of larvae on the rough column versus
32% of larvae on the smooth column completed the climb when responses to the two
surfaces were averaged across the spinneret treatments. The main effects and their
interaction'were each significant with respect to the time required for larvae to complete
the 25-cm climb (Table 14). Larvae without the spinneret glued completed the climb
substantially faster on the rough (2.4 min) than on the smooth column (4.86 min). Since
larvae with the spinneret glued were unable to climb on the smooth column, no times
were recorded. Interestingly, larvae with the spinneret glued required substantially longer
to climb the rough column (12.61 min) than larvae without the spinneret glued on the
smooth column (4.86 min), which supports previous observations that larvae also use silk

to assist climbing on rough surfaces.

In the experiment comparing laddering by third-instar larvae from low- and high-
density populations of gypsy moth reared on ICN or Bell's diet, laddering was affected
only by diet. The chi-square value (25.99) for the main effect, diet, was significant
(p<0.001); chi-square values for the population effect and the diet-population interaction,
0.76 and 0.01, respectively, were both nonsignificant (p>0.1). Averaged over the main
effect of population density, significantly more of the larvae reared on the ICN diet used
laddering to assist climbing on the smooth column compared with larvae reared on Bell's
diet (66 and 10%, respectively). The time for the larvae to climb 25 cm was also} affected
only by the main effect of diet. Larvae reared on Bell's diet from both the high- and the
low-density populations completed the climb significantly faster than larvae reared on the

ICN diet (Table 15). The effect of diet on climbing behavior may be related to larval

67

Table 14. Mean time for gypsy moth larvae with and without the spinneret sealed with
Crazy Glue® to climb 25 cm on a smooth or rough 50-cm column of 5-cm-diameter ABS
plastic pipe. Standard error of the mean = 0.55.

 

Mean time (min)

 

 

Column surface Spinneret glued Spinneret without glue
Smooth - 4.86

(0) ( 17)
Rough 12.61 2.40

(23) (25)

 

t

numbers within brackets = number of larvae (total possible = 25) that completed the
25-cm climb within 30 min.

68

Table 15. Percent of larvae laddering and mean time for gypsy moth larvae reared on
ICN and Bell's artificial diets (averaged over low and high population levels) climbing 25
cm in < 30 min on a smooth column of ABS pipe. Standard error of the mean time =
0.10.

 

 

Diet % laddering Time (min)‘ No. of larvae“
Bell's 10 1.42 43
ICN 66 5. 19 47

 

differences were significant at P<0.05 (Student-Newman-Keuls test).
number of larvae that completed the 25-cm climb, out of a total of 50.

.0

Table 16. Measures of fitness for 15 larval gypsy moth reared on Bell's and ICN diets.

 

Population statistic ICN diet‘ Bell's diet‘

 

Mean head-capsule size (mm 3: SE)

Neonate 0.56 :1: 0.01a 0.55 i 0.01a
111 1.52 d: 0.03a 1.62 :i: 0.03b
VI 2.47 :i: 0.063 5.40 :1: 0.10b

Mean pupal weight (g :1: SE)

Males 0.35 :h 0.03a 0.62 :i: 0.02b
Females 0.94 i 0.09a 1.94 :1: 0.14b
Mean larval development time (days i SE)

Males 53.6 i 0.59a 31.7 d: 0.42b
Females 69.6 i 1.433 35.4 i 0.5%

 

 

Values in a row followed by different letters are significantly different (P<0.05, t-test).

69

fitness; larvae reared on Bell's diet are larger and develop faster than larvae reared on

ICN diet (Table 16).

Frequency of laddering and the speed at which larvae climbed 25 cm on the three
species of trees were inversely related. The results for red oak were also significantly
different from those for the other two species (Table 17). On red oak larvae completed
the 25-cm climb in the shortest time (2.35 min) and were not observed laddering. Larvae
climbing on paper birch appeared to have the greatest difficulty climbing; they required
the longest time (8.18 min) to complete the 25-cm climb and laddered most frequently
(8.3). The climbing time (5.13 min) and frequency of laddering on trembling aspen (7.0)

was intermediate between the other two treatments.

My discovery of the laddering behavior was accidental. I have not had an
opportunity to search for it thoroughly in the field. Although laddering appears to be
induced by an inadequate diet in the laboratory and occurs mostly when climbing on a
smooth surface, I feel that because larvae reared on Bell's diet also exhibited the same
behavior, there is reason to believe that laddering may occur in the field. Strong :1 a1.
(1984) indicated that many caterpillars spin silk thread to aid their attachment as they
move about on smooth plant surfaces. Unfortunately, specific references were not cited
by the authors; however, other references about the use of silk to aid attachment have
usually considered this to be a mechanism that insects have evolved to enable them to
skirt plant defences and exploit a defended resource. For example, larvae of the butterfly
Meghanjg} isthmaj Bates (Ithomiidae) successfully avoid trichomes on their spiny host
(39m spp.) by spinning a network of silk scaffolding across the underside of the leaf,
hanging below the spines on silk threads, and feeding safely on the unprotected edges
(Rathcke and Poole 1975). Laddering by gypsy moth larvae appears to be a specific

larval adaptation that is "triggered" by surfaces that are difficult to climb. The behavior

70

Table 17. Mean number of gypsy moth larvae per replicate that exhibited laddering
behavior on 30-cm bolts of red oak, paper birch and trembling aspen (MSE = 0.86) and
the mean time per replicate for larvae to climb 25 cm (MSE = 0.38).

 

 

Treatment Number laddering‘ Mean time‘ 11“
Red oak 0.0a 2.353 40
Trembling aspen 7.0b 5.13b 34
Paper birch 8.3b 8.180 33

 

values in a column followed by different letters are significantly different (P<0.05,
Student-Newman-Keuls test).

“ number of larvae that completed the 25-cm climb within 30 min. Total number
possible = 40.

71

has not been documented before; however, it too would be advantageous for survival of a
polyphagous herbivore such as gypsy moth because it would enable larvae to better
exploit hosts that are difficult to climb. The experiment that examined laddering on red

oak, trembling aspen and paper birch supports this possibility.

The bark surface of red oak is more irregular than the bark surfaces of trembling
aspen and paper birch. It is possible that the smoother bark surface of trembling aspen
and paper birch may have been too difficult for the less-fit larvae reared on the ICN diet
to grasp firmly with their crochets; as a result, larvae used silk to construct a ladder to
assist climbing. If this hypothesis is correct, then laddering in the field may be more
common in wild populations with decreased fitness that result, for example, from
reduced food quality of the host due to wound-induced plant defences (Schultz and
Baldwin 1982; Mattson and Scriber 1987). A clear understanding of why laddering
occurs, however, is confounded because of the similarities between the two diets and the

developmental differences that occurred in larvae reared on the ICN diet.

The data sheet provided with the ICN diet lists ingredients and proportions
identical to those that are used in Bell's diet. The reduced growth and longer development
times that were observed with the ICN diet suggest an inadequate source of essential fatty
acids and sterols (Chapman 1982a; McFarlane 1985). It is possible that the ICN diet may
have been damaged during storage or shipping, although I observed similar symptoms of
reduced larval fitness with it when it was purchased on two separate occasions more than
a year apart; alternatively, for some other unrecognized reason, it is possible that the
dietary requirements of the specific Ontario population of larvae used in the study were
not met by the ICN diet. Further studies would be useful in clarifying how laddering is
influenced by diet and the possible role that silk production may play in host-plant

interactions of the gypsy moth.

Chapter Four

A Laboratory Technique to Study a Change in Feeding

Behavior Between Small and Large Larvae of the

Gypsy Moth. Lymantria diaper (L)

Introduction

The gypsy moth, Lmamfia £11593; (L.), is an important defoliator of deciduous

forests that was introduced into eastern North America from Europe (Forbush and

F ernald 1896). After eclosion and air-home dispersal in the spring (McManus 1973;
Mason and McManus 1981), first-instar larvae become established in the crown of a tree,
where they remain until the fourth instar (Leonard 1981). Feeding during this period of
larval development occurs primarily in the early morning, after temperatures have
increased, and secondarily in the late afternoon (Leonard 1981). However, fourth-instar
larvae exhibit a profound change in their feeding behavior, marked by the migration of
larvae towards the base of a tree each day at dawn to seek shelter and then ascent into the
crown after dusk to resume feeding. (In this chapter, "feeding behavior" implies a change
in microhabitat selection and should not be confused with "feeding rhythm", which refers
to the time of day when larvae feed.) At high larval densities such a change in the
feeding behavior may not occur (Leonard 1970, 1974; Campbell e_t a1, 1975a). The
Change in behavior of late-instar larvae also involves movement from one host tree to

another (Barbosa 1978; Liebhold et aL 1986). To investigate these matters further, I

72

73

developed a lighting system for the laboratory that would elicit field-like changes in

larval behavior. Construction of the lighting system and artificial tree stems is described

in this chapter.

Materials and Methods

Lighting System. A change in the intensity of light at dawn and dusk is thought to be the
cue that large gypsy moth larvae use to begin their diurnal movement on a tree (Leonard
1981; Weseloh 1989). In Ontario, the movement of larvae down a tree begins at
approximately 0300 h EST, about 15 min before dawn; larval migration peaks ca. 1 h

later, and is complete, except for stragglers, by 0700 hours (Figure 8).

To model the spectral properties and the change in intensity of light that occur in
the field, I used a combination of incandescent and fluorescent light in a 4- by 5-m
rearing room with a 17L:7D photoperiod. Incandescent lights were used to approximate
the changes in intensity and spectral properties that occur at dawn and dusk because
peaks in the spectroradiometric curve at this time shift from the blue wavelengths at
midday to the red wavelengths predominant in incandescent light (Moon 1961).
Incandescent lights also provide an inexpensive way to regulate the intensity of light.
The color temperature of the bulb was also selected to approximate the color temperature
of the sun 1 h after sunrise because the color of light produced affects the way in which
the color of an object is perceived (Brill 1980). The color temperature of the sun at
sunrise is ca. 1800'K; this increases to 3500°K within the first h and approaches 5000°K
by noon (Brill 1980). Changes in spectral properties that occur as the sun rises were
simulated with fluorescent lamps with spectral-distribution curves similar to those of
measurements of midmoming and midday solar radiation (Riordan et a1, 1989). This was

accomplished with four banks of two in-line and two adjoining 122-cm two-lamp

Figure 8. Mean number of larvae (i SE) descending below 1.5 m on three white oak,
Om alha L., trees (mean dbh = 38.1 :1: 4.59 cm; mean height = 17.2 i 0.61 m) during
subsequent 15-min intervals between 0200 and 0800 hours EST on 30 June 1989
(Kaladar, Ont).

 

MEAN NUMBER OF LARVAE

 

L A A A A A A A

75

 

 

 

 

rTTTWYUUY'

I UUUUUUU TY—l'YI'Ilfivi[UTrTrtr'UITYTTVVVTTITYUj’TfYTTr

0300 0400 0500 0000 0700 0800
TIME EST

76

fluorescent fixtures. These were arranged symmetrically on either side of a central line
of three 100-watt incandescent bulbs. The set of fluorescent lamps on either side of the
incandescent lights (General Electric F40CWX) simulated the spectral properties and
color temperature (4175'K) of light during the early morning and late afternoon; the
second set (General Electric F40/C75) simulated the spectral properties and color
temperature (7500°K) at midday, when sunlight is more intense. Photoperiods of 13L:9D
and 4L:20D, respectively, for each set of fluorescent lamps were regulated with
mechanical timers (Paragon Electric Company, Two Rivers, WI 54241; Model Number
4001-00).

The increase and the decrease in the intensity of light that occur at dawn and
dusk, respectively, were simulated with three Phillips 130-volt incandescent lOO-watt
bulbs, which were lined up in a row fastened to the ceiling in plastic incandescent light
fixtures. The rated color temperature of each bulb at an applied voltage of 110 to 115
volts was 2800°K. The intensity of light and the rate at which the intensity changed were
regulated by a variable rheostat (Superior Electric Company, Bristol, CN) driven by a
reversible geared motor (Hurst Instrument Motors, Hurst Manufacturing Division,
Emerson Electric Company, Princeton, IN 47670; Model Number A-SP 2881) that
completed 0.5 revolutions per hour (Figure 9). The chosen gear ratio of the motor was
based on a compromise between the period during which I considered larvae to be active
in the field at dawn (0300-0700 EST) and a desire to match the color temperature and
period of elapsed time after first visible light to the color temperature of the bulb

(2800’K), a period of approximately 50 min.

Incandescent lights were not turned off completely at night (< 30% = off). I
found, through trial and error, that a change in larval behavior did not occur when the

room was completely dark. However, a level of light in the room at night such that the

Figure 9. Schematic of the brass union that coupled the drive motor to the handle of the
rheostat: a, drive motor; b, rheostat handle; c, brass union; (1, 5-mm pin; e, brass rod; 1',
rheostat; g, set screws; h, drive-motor mounting bracket.

 

78

.............
..........

      

,

        

1
.......
.......
........
........
........
o e .
.

 

 

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

79

silhouette of an object could just be differentiated corrected the problem; this was
accomplished with a rheostat setting of 36% power. In my judgment, this approximates

the level of light in a forest on an overcast night.

Power to the red and black wires on the motor that reversed the rheostat was
controlled by two PET 71-120 digital timers (Paragon Electric Company). The precision
of a digital timer was employed to insure that the distance traveled by the motor was
exactly the same when power was increased or decreased; otherwise, an error of more
than 1 or 2 min in the power applied to the drive motor would have forced the handle of
the rheostat past the stop at full power (100%). As an additional margin for error, I
further reduced the maximum rheostat setting to 99%. This also allowed for the error
that occurred when the two timers were being synchronized; I was never able to reduce
this to below 10 sec. Components for the incandescent light circuit (rheostat, drive
motor, mounting-bracket for the drive motor, two digital timers and a terminal strip with
10 terminals for connecting wire) were all mounted inside a transformer box (Bell

Products Inc., Montreal, Que. Model Number MC-302010) that was fastened to the wall.

The cycle used to light the room with incandescent and fluorescent lamps
overlapped so that the intensity of light in the room either increased or decreased,
respectively, as each set of lights turned on or off. The first digital timer (henceforth
referred to as the "sunrise" timer) regulated the clockwise rotation of the motor and
began to increase the level of incandescent light in the room at 0600 h. When the level
of light reached full power (99%) at 0715 h, the motor stopped; however, the
incandescent lights remained on throughout the day until 2145 h, when the second digital

timer (henceforth referred to as the "sunset" timer) began the counterclockwise rotation

80

of the motor. This decreased power over the same period (1.25 h) to the level of light
used at night (36%). Midmorning fluorescent lights came on at 0800 and went off at A

‘2100; midday fluorescent lights came on at 1230 h and went off at 1630 h.

The procedure used to wire the time clock for each bank of fluorescent lights
should not present a problem for the do-it-yourself entomologist. However, the wiring
for the incandescent light circuit and for the motor that operated the rheostat (Figure 10)
was more intricate. Instructions for wiring the incandescent light circuit can be found in

Appendix A.

Artificial Trees. I fabricated 12 artificial trees from 5-cm ABS plastic pipe to serve as
hosts in my tests to determine whether or not the lighting system induced a change in
larval behavior. Each tree (Figure 11), an inverted Y-shaped column 1.5 min height,
was supported by two 0.75-m vertical arms inserted at the base of a toilet flange that was
fastened with wood screws to a 50-cm square of 18.5-mm plywood. The upper portion
of the tree, the inverted Y, was constructed by gluing two 15-cm pieces of pipe at a 45°
angle with ABS cement and then welding these to a single 0.6-m upper vertical stem with
a hot-air welder (Leister-Kombi, Type: Triac, purchased from Johnson's Industrial
Plastics, Toronto, Ont.). The 15-cm arms were then inserted into two 45° elbows at the
top of each lower vertical supporting stem. A 6-cm circular plate, cut from 5-mm
Plexiglass and glued in place with ABS cement, sealed the top of the tree. For a feeding
station I used a 20-mL coffee creamer filled with artificial diet that was suspended by the
rim of the creamer at the top of each tree in an opening cut in the center of the circular
plate. The creamer was replaced daily. Strips of 2.5- by 15-cm black felt were used to
fabricate artificial bark flaps to provide a refuge for late larval instars. Several strips of
felt, attached with Elmer's Glue-All, were fastened around each of the 45° elbows. The

thicker plastic on the elbow positioned the strips of felt away from the surface of the

81

Figure 10. Wiring diagram for the incandescent light circuit: DM, drive-motor; VR,
variable rheostat; CP, 0.25 MFD capacitor.

 

 

rm! CLOCK
turnouts 1"
I A to D)

'l

 

SUNSET
TIME CLOCK

M

" "”3:
$5 ‘

__E£L

 

 

82

 

 

WHITE

 

 

ILAEK

 

POWER

SUPPLY ’l “ER [MflL

Swrtch

 

 

SUNRISE
TIME CLOCK

 

 

 

 

 

 

 

 

L UN
WNITE (Neutral!
r - .1
l 1 3
"humus/l,
( t to 71 l
L--- ’

-
--

 

 

 

 

 

 

 

   
 

 

 

‘O 60

Tennmai

‘- j/SNID

 

 

BLACK ' L

LIGHTS

 

 

 

83

Figure 11. Artificial tree used to observe a change in feeding behavior.

 

 

 

85

artificial tree and provided a space between the tree and the felt beneath which larvae
could rest beneath during the day. In addition, inverted U-shaped pieces of cardboard '
(bark flaps) were also fastened to the 15-cm arms to provide a larval retreat. The surface
of each artificial tree was also textured with a mixture of 150 mL of sand sifted through a
20-mesh screen and 250 mL of black latex paint. Previous experiments showed that

larvae preferred a dark color with a rough surface (Roden, unpublished data).

Larval gypsy moth used in the study were obtained from egg masses collected on
20 October 1987, from Kaladar, Ont. (44°39'N, 77°07'W). These were held at 5°C until
4 January 1987, then were surface-sterilized (Shapiro 1977) and incubated in the rearing
room described above at 22°C and 70% RH. Electric lights increased the temperature
during the photophase by ca. 6°C; I did not attempt to rectify this increase because the
diurnal variation in temperature firrther simulated conditions in the field. Larvae that
emerged from eggs were reared individually on artificial diet (Bell :1 a1, 1981) in
screened 150- by 25-mm plastic Lab-Tek Petri plates until the second instar. When
sufficient second-instar larvae were obtained, 36 were randomly selected and three were
placed on each of the 12 artificial trees. The trees were centered approximately 1 m apart
with six on either side of the central incandescent bank of lights to balance the effect of
room lighting. Larval positions on each tree were recorded daily during the photophase
at three times: 0900, 1500 and 2100 h. Periodically, positions were also recorded during
scotophase to confirm feeding and observe activity. Whenever a molt occurred, the
instar of that particular larva was recorded. Instances when gypsy moths in late larval
instars descended the artificial trees for a day (0900 until 2100 h) to seek refuge, and
remained motionless under the black strips of felt, the cardboard flaps, or the crevice of

the Y, were considered indicative of a change in feeding behavior.

86

Results and Discussion

The lighting system assembled to change the intensity and the spectrum of light
automatically in the laboratory clearly influenced the feeding behavior of late larval
gypsy moth on artificial tree stems. When larvae were placed on the trees, they moved
about on their new hosts; however, excessive larval movement ceased within 24 h. For
the following 8 days, second- and third-instar larvae were observed feeding only during
daylight hours (0600 to 2300) and remained within 25 cm of the top of each tree; larval
resting locations were on the rim of or inside the creamer of diet, or on the side of the
tree. The first change in larval behavior was recorded on day 10. By day 22, the day
before pupation began, 85% of the population (N = 30) migrated down the artificial tree
stems each day to seek shelter, and remained nearly motionless under the black strips of
felt, the cardboard flaps, or the crevice of the Y; only fourth-, fifth- and sixth-instar
larvae exhibited this behavior. Eventually, however, all larvae changed their feeding
behavior. During sunrise (the period between 36% and 99% power), most larvae had
begun moving for shelter by the time the rheostat had reached 70% power. All migrating
larvae attained shelter before the midmoming fluorescent lights came on. In previous
experiments I was unable to change the feeding behavior of larvae with only the on/off
effect of fluorescent lights; however, all larvae in the current experiment selected resting
locations below the felt for pupation. Obviously, midmoming and midday fluorescent
lights were not responsible for initiating a change in feeding behavior, although an
increase in the intensity and a change in the spectrum of fluorescent light may have
ensured that larvae remained inactive throughout the photophase. Furthermore, I cannot
be certain that the diurnal temperature change in the room did not also contribute to the

success of my lighting system.

87

Larvae usually began moving up a tree when the rheostat decreased to
approximately 40 to 45% power. These larvae were active through scotophase and
frequently changed trees during the night and then clustered on another tree for the next
resting period. Similar inter-tree larval movement was reported by Liebhold e_t 3L (1986)
and may indicate a larval searching behavior for suitable hosts with resting locations that
increase survival (Campbell 1981). Large larvae were not observed on the walls of the
chamber and did not select resting locations that were off the trees during photophase;
larvae were also not observed feeding during photophase. Lance e_t 31, (1986b) reported
similar feeding observations and suggested that the agreement between field and
laboratory data imply that it would be possible to study the influence of environmental
factors that affect feeding rhythm. My results support their observations. The increase in
variance (Table 18) after a change in feeding behavior shows that there is a tendency for
the distribution of larvae to become more random over time; however, the maximum
variance (1.91 on day 18; mean = 2.50) did not depart from that of a normal distribution.
It is possibly that a larger sample size would have shown a negative binomial distribution
of larvae among trees. I did not mark larvae individually; therefore, I was unable to
discern any particular differences or patterns of behavior related to sex or individual
uniqueness. Movement of fourth-, fifth- and sixth-instar larvae between tree stems
occurred only at night. However, movement between hosts in the field often occurs
during the day among high population levels (Leonard 1970). I speculate that larval
movement between trees that occurred only at night in the laboratory is typical of larval

behavior at low population densities.

As a result of the compromise between the period during which larvae were
active at dawn (4.0 h) and the gear ratio of the motor designed to match the color
temperature to natural light, natural color temperature was not accurately simulated at the

time of day when the bulb reached full power. Incandescent lights operated at lower than

88

Table 18. Variance and maximum/minimum number of larval gypsy moth per tree
(n=12) on 5 subsequent days before (days 5-9) and after (days 16-20) a change in
feeding behavior was observed. During photophase on days 16-20, 80-85% of the
population (n=30) changed feeding behavior.

 

 

Day Variance Maximum/minimum no. larvae/tree
5 0.27 3/2
6 0.27 3/2
7 0.27 3/2
8 0.27 3/2
9 0.27 3/2
16 0.46 4/2
17 1.18 4/1
18 1.91 5/1
19 1.36 5/1

20 1.61 5/0

 

89

the rated voltage will reduce the color temperature of a bulb (Brill 1980). Changes that
occurred took place over 75 min instead of 50 min. The sensitivity of gypsy moth larvae
to flickering light is also unlmown. Insect flicker-fusion frequencies (FFF) between 20
cps (Chapman 1982b) and 300 cps (Wigglesworth 1974) have been reported. An insect
with a F FF > 120 cps would be able to detect the on-off effect of my AC-powered
fluorescent lamps and incandescent light and this could possibly disrupt normal behavior.
Nevertheless, differences between my lighting system and natural daylight were not
sufficient to prevent the change in feeding behavior that occurs in the field between small

and large larval instars of gypsy moth.

In conclusion, the lighting system and the artificial tree stems developed for use
in the laboratory will make it possible to study various aspects of the biology and feeding
behavior of the gypsy moth that could not have been investigated before. For example,
by manipulating the physical features of an artificial tree in the laboratory, it may be
possible to learn about the key factors that influence the insect's choice of a host.
Barbosa (1978) speculated that the physical features of a host play an important role in
host selection by penultimate and ultimate larval instars. At this point, very little is
known about how physical traits of the host affect the host preference and feeding
behavior of larval gypsy moth. Clearly, chemical cues are important stimuli (Barbosa
and Capinera 1977; Hough and Pimentel 197 8; Barbosa e_t 3L 1979); however, visual or
structural cues, or both, may be equally important. Studies with adult Diptera (Harris
and Miller 1983; Moericke gt 31, 1975; Prokopy 1977) have demonstrated that host
selection is a very complex process that involves host information from many different
sensory modalities. The ecological significance of the responsiveness of larval gypsy

moth to different stimuli remains to be elucidated.

SUMNIARY AND CONCLUSIONS

The objectives of this research were: (1) to evaluate the host potential that
trembling aspen and paper birch offer for gypsy moth, (2) to compare these hosts to red
oak in the Great Lakes basin, and (3) to investigate the possible influence of a host's
physical features on larval behavior. Over a 6-year period, it was shown that the physical
features of a host strongly influence the larval behavior of all gypsy moth instars. In
laboratory and field experiments, with artificial and authentic tree trunks, larval attraction
to an object was positively correlated with the angle at which the diameter and height
were presented. Research studies also demonstrated that larval attraction to a host was
strongly influenced by tree species. The attraction of larvae to red oak was frequently
fifteen-fold greater than to paper birch or trembling aspen. Since larvae are influenced by
diameter, height and species of the host, these variables should be important
considerations in standardizing burlap banding in operational monitoring systems; they

may also help explain the gypsy moth "wolf tree" phenomenon.

Research in this thesis also demonstrated that gypsy moth development on both
defoliated and undefoliated trembling aspen and paper birch is superior to larval
development on a traditional host such as red oak. Female gypsy moth larvae fed on
defoliated red oak produced significantly smaller pupae (0.84 g) than females feeding on
paper birch (1.13 g) or trembling aspen (1.43 g). Tree species was the only factor that
affected male pupal weight. Male larvae reared on red oak yielded pupae that weighed
less (0.37 g) than male larvae reared on paper birch (0.40 g), and both weighed

significantly less than pupae from male larvae reared on trembling aspen (0.49 g).

90

91

The number of days required by both male and female larvae for development
was only affected by host species. Female larvae that fed on trembling aspen required
significantly fewer days (46) to complete development than females reared on red oak
(50); the number of days for female larval development on paper birch (48) was
intermediate, but not significantly different from the times for either trembling aspen or
red oak. Male larvae fed on trembling aspen also required significantly fewer days (41)

for development than larvae that fed on paper birch (44) and red oak (44).

Gypsy moth mortality in the experiment was not affected by either host species or
the level of defoliation. However, these results should be viewed cautiously. Predatorial
attacks by Mm 313915135 Uhl. (Pentatomidae) from outside the cage through the
screening on the ventral surface of larvae seriously reduced the number of larvae in all
treatments by E 60% and may have masked possible mortality effects that could have
been attributed to the different tree species. The number of larvae that escaped from

cages in the experiment, or that died from injestion of NPV was less than 4%.

A previously undescribed behavior of gypsy moth larvae used for climbing on a
smooth vertical surface was serendipitously discovered. By spinning a "ladder of silk" on
surfaces that do not provide structures that can be grasped by the crochets, larvae are able
to climb vertically. The use of silk for climbing occurred more frequently in populations
that were less fit (as measured by head-capsule size, reduced pupal weight and increased
development time). For example, the mean female pupal weight for fit and unfit
populations were, respectively, 0.94 versus 1.94 g. I suggest that the use of silk for
climbing in the field may be associated with decreased larval fitness that results from

wound-induced plant defences.

92

A lighting system with incandescent and fluorescent light in the laboratory was
also developed that induced a change in feeding behavior similar to the change that
occurs in the field. The change in feeding behavior was observed on artificial "tree
stems" constructed from 5-cm ABS plastic pipe and fitted with felt and cardboard "bark
flaps". On the day before pupation, 85% of the population migrated down the artificial
stems to seek shelter under the bark flaps; only fourth-, fifth-, and sixth-instar larvae
were observed exhibiting this behavior. The development of the lighting system and the
artificial tree stems should make it possible to identify other key factors that influence

gypsy moth's choice of a host under controlled conditions.

RECOMIVIENDATIONS

The identification of visual stimuli that influence the behavior of larval gypsy
moth and the knowledge that the nutritional benefits of trembling aspen exceed those of
red oak, a traditional gypsy moth host, pose as many questions as they answer. Questions

that should be addressed by future research are:

1) All measurements of stimuli that affected the visual foraging behavior of larval
gypsy moth in this research were conducted at distances of 0.5 m. What is the
maximum distance at which these stimuli (such as diameter, height and species)

affect larval behavior?

2) Is there a synergistic affect between diameter and height that makes specific
combinations or ratios of a diameter and height more attractive? Such effects have

been noted for adult Diptera (Moericke 3t 31. 1975; Miller and Harris 1985).
3) What are the other stimuli that affect gypsy moth larval behavior? For example,
from preliminary studies that were initiated, but not completed, it is apparent that

gypsy moth larval behavior is also influenced by the wavelength of light.

4) What affect do diameter, height and species of a tree have on operational monitoring

system that use burlap bands to quantify population levels?

93

5)

94

What effect does the color of the material used for tree bands have on population

' measurements? Based on research in this thesis, it is evident that larvae do not

respond equally to all colors.

6.) My data suggest that both defoliated and undefoliated trembling aspen and white

7)

8)

birch are more nutritious, and would yield potentially more fecund gypsy moth
females than a traditional host such as red oak. However, I propose that outbreaks of
gypsy moth will not occur in stands that contain primarily these species. Better
gypsy moth performance on trembling aspen and paper birch is attributed to an
imprecise correspondence between host and herbivore that inhibits naturally
occurring outbreaks of gypsy moth on these tree species. I suggest that the gypsy
moth NPV, which influences populations levels, and physical features of the host
that effect gypsy moth larval behavior are responsible for this imprecise
correspondence. Since Schultz e_t 31. (1990) suggest one of the keys to understanding
the susceptibility within a tree species may be the relationship between condensed
and hydrolyzable tannins, this aspect should be investigated for trembling aspen and

paper birch.

What is the predatorial potential of the pentatomid, 13. 913913113? The frequency of
attacks on caged gypsy moth larvae in this study suggest that the impact of this

predator should be investigated.

What is the relationship between first-instar weight and pupal weight? The
significant covariate for male and female larvae observed in this thesis suggests that
the largest first-instar larvae also produce the largest and potentially the most fecund

pupae.

95

9) What effect do plant-induced defences have on the physical ability of larvae to
exploit a host? The influence of diet on the construction of a silk ladder to assist
climbing suggests that the ability of larvae to exploit a host may be physically

impaired by wound-induced plant defenses.

10) What is the potential of laboratory studies for measuring and quantifying gypsy
moth larval behavior? The development of a laboratory lighting system and a
suitable artificial tree stem suggest that it may be possible to rigorously quantify

various aspects of the biology and feeding behavior of the gypsy moth.

APPENDIX

APPENDIX A

Instructions for Wiring The Incandescent Light Circuit

When following the incandescent wiring instructions outlined below, the reader
should refer to Figure 10. A black or red wire indicates a wire carrying a load (power)
and white or blue, a wire that is neutral. I used only the first seven of the ten terminals on
the terminal strip for connecting wires to the transformer box. To clarify the numerical
designation between the 7 terminals on the terminal strip that I used and the 8 terminals
on the time clock, terminals on the time clock were assigned a designation A through H
from left to right. Note, however, that the last three (8, 9 and 10) on the terminal strip

and the last four on each time clock (E to H) were not used.

To bring power to the terminal strip fastened to the transformer box, I used Type
SW l4-gauge 2-wire with a male plug that inserted in a female wall receptacle; I elected
to use the male plug instead of the switch shown in Figure 10 simply because it was more
appropriate for my installation. The black and white wires that entered the transformer
box fastened, respectively, to 1 (load) and 2 (neutral). The third wire, green, (note:
Canadian electrical numerical wiring designations do not include the ground wire)
fastened to ground on the mounting plate. This plate attached to the back of the
transformer box with machine screws and anchored the rheostat, the drive-motor
mounting plate and the time clocks. Grounding for each of these and for the motor that
fastened to the drive-motor mounting plate was accomplished by scraping away paint

between points of contact. Power to each time clock was provided by black wires that ran

96

97

from 1 to A, the power input terminal on each time clock. A common point (D) on each
time clock between an internal set of contacts received power through a black wire from
A. The neutral terminal on each time clock, B, was connected by separate white wires to
2. Terminal 2 was also connected by a white wire to the two blue neutral leads from the
motor. A black wire from 1 provided power to 5; the terminal that supplied power to the
input on the rheostat. Power from the output on the rheostat supplied power to 6. This
terminal was connected to the load (i.e., the black wire that delivered power to the
incandescent lights); the neutral wire (white) to the lights was connected to 7. Terminal 7
was also wired to 2. Terminals 3 and 4 connected with black and red wires, respectively,
to C on each time clock. These terminals also connected, respectively, to the black and
red wires from the motor. When the program for the sunrise time clock completed the
circuit, power was supplied from C through 3 to the black wire from the motor; this
initiated clockwise rotation of the rheostat and increased the voltage, subsequently
increasing the intensity of light. Similarly, power regulated by the sunset time clock
supplied power from C through 4 to the red wire from the motor; this initiated
counterclockwise rotation of the motor and reduced the intensity of light. A capacitor of
0.25 MF D (supplied with the motor) was inserted between terminals 3 and 4. This
reversed the phase sequence of voltage applied to the stator on the motor, thereby

reversing the direction of the magnetomotive force.

LIST OF REFERENCES

LIST OF REFERENCES

Balfour-Browne, F. 1925. The evolution of social life among caterpillars. pp. 334-340 in
K. Jordan and W. Horn (Eds), Proceedings 3rd. Intern. Cong. of Entomol,
Zurich.

Barbosa, P. 1978. Host plant exploitation by the gypsy moth, M3 dispar. mt. m.
Applic. 24: 28-37.

Barbosa, P., and J. L. Capinera. 1977. The influence of food on developmental

characteristics of the gypsy moth, mm 313133; (L.). Q33. ,1. L391. 55:
1424-1429.

Barbosa, P., and J. Greenblatt. 1979. Suitability, digestibility and assimilation of various
host plants of the gypsy moth, LEM 313331; (L.). W3 43: 111-119.

Barbosa, P., and VA. Krischik. 1987. Influence of alkaloids on feeding preference of
eastern deciduous forest trees by the gypsy moth, M3 313331. American
Naturalist 130: 43-69.

Barbosa, P., J.A. Greenblatt, W. Withers, W. Cranshaw, and EA. Harrington. 1979.
Host plant preferences and their induction in larvae of the gypsy moth,

231mm Emmet. Exit. and App]. 23; 180-188.

Barbosa, P., P. Gross, G.J. Provan, D.Y. Pacheco, and FR. Stermitz. 1990a.
Allelochemicals in foliage of unfavored tree hosts of the gypsy moth, M
313323; (L.). 1. Alkaloids and other components of W 311mm L.
(Magnoliaceae), Am 1111mm L. (Aceraceae) and £91333 £19333 L. (Comaceae).
1. Chm. £913. 16: 1719-1730.

98

99

Barbosa, P., P. Gross, G.J. Provan, D.Y. Pacheco, and RR Stermitz. 1990b.
Allelochemicals in foliage of unfavored tree hosts of the gypsy moth, Lymggtria

dispar (L.). 2. Seasonal variation of saponins in Her; m and identification of
saponin aglycones. 1. gm. B391. 16: 1731-1737.

Beck, SD. 1965. Resistance of plants to insects. Ann Rey, Entgmgl. 10: 207-232.

Bell, R.A., C.D. Owens, M. Shapiro, and LR. T ardif. 1981. Mass rearing and virus
production. pp. 608 111 CC. Doane and ML. McManus (Eds), The Gypsy Moth:

Research Towards Integrated Pest Management. USDA For. Serv., Washington,
DC, Tech. Bull. 1584.

Brill, TE. 1980. Light. Plenum Press, New York. 287.

Cain, M.L., J. Eccleston, and RM. Karevia. 1983. The influence of food plant dispersion
on caterpillar searching success. E991. E31. 10: 1-7.

Cameron, E.A., M.L. McManus, and C.J. Mason. 1979. Dispersal and its impact on the
population dynamics of the gypsy moth in the United States of America.
Mitteilungen der schweizerischen entomologischen gesellschaft 52: 169-179.

Campbell, R.W. 1979. Gypsy moth: Forest influence. U.S.D.A. Info. Bull. No. 423.

Campbell, R.W. 1981. Population dynamics: Historical review. pp. 65-86 i_rl C.C. Doane
and ML. McManus (Eds), The Gypsy Moth: Research Towards Integrated Pest
Management. USDA For. Serv., Washington, D.C., Tech. Bull. 1584.

Campbell, R.W., D.L. Hubbard, and R.J. Sloan. 19753. Patterns of gypsy moth
occurrence within a sparse and numerically stable population. Eggnog. Em. 4:
535-542.

Campbell, R.W., D.L. Hubbard, and R.J. Sloan. 1975b. Locations of gypsy moth pupae
survival in sparse, stable populations. Envirgn. Em. 4: 597-600.

Campbell, R.W., and R.J. Sloan. 197 6. Influences of behavioral evolution on gypsy moth
pupal survival in sparse populations. Envirgg. E31. 5: 1211-1217.

100

Campbell, R.W., and R.J. Sloan. 1977. Forest stand responses to defoliation by the gypsy
moth Fir. $31. Mgnogr. 19: 1-34.

Capinera, J .L., and P. Barbosa. 1976. Dispersal of first-instar gypsy moth larvae in
relation to population quality 95991313 26: 53-60

Chapman, RP. 19823. Nutrition. pp. 84-99 111 The insects: Structure and function. Third
Ed. Harvard Press. Cambridge, MA.

Chapman, R.E. 1982b. The eye and vision. pp. 642-673 i_n_ The insects: Structure and
function. Third Ed. Harvard Press. Cambridge, MA.

Chew, SP. 1981. Coevolution of pierid butterflies and their cruciferous foodplants.

16339313311 Mg 118: 655-672.

Coley, P.D., J .P.Byrant, and RS. Chapin. 1985. Resource availability and plant
antiherbivore defence. $911733; 230: 895-899.

Conover, W.J., and R.L. Iman. 1981. Rank transformation as a bridge between
parametric and nonparametric statistics. Amer. _S_t31. 35: 124-133.

 

Day, R.W., and GP. Quinn. 1989. Comparison of treatments after an analysis of
variance in ecology. E99193. Mgnggraphs 59: 433-463

Dethier, V.G. 1971. A surfeit of stimuli: A paucity of receptors. m. 8J1. 59: 706-715.

Dethier, V.G. 1980. The world of the tent-makers. A natural history of the eastern tent
caterpillar. pp. 61 Univ. Mass. Press. Boston, MA.

Dethier, V.G. 1982. Mechanism of host-plant recognition. E_n_t. m. Applig. 31: 49-56.

Dethier, V.G. 1987. The feeding behavior of a polyphagous caterpillar (Di3cg'si3
nggjnig3) in its natural habitat. Q33. ,1. Z931. 66: 1280-1288.

101

Dethier, V.G. 1989. Patterns of locomotion of polyphagous arctiid caterpillars in relation
to foraging. E931. E31. 14: 375-386.

Doane, CC, and DE. Leonard. 1975. Orientation and dispersal of late-stage larvae of
Pgrthgtri3 333331(Lepidoptera: Lymantriidae). £33. ELI. 107: 1333-1338.

Duncan, DB. 1975. t Tests and intervals for comparison suggested by the data.
Bigmggrjgs 31: 330-359.

Feeny, P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring
feeding by winter moth caterpillars. My 51: 565-581.

Fitzgerald, TD, and SC. Peterson. 1988. Cooperative foraging and communication in
caterpillars. B19391. 38: 20-25.

F orbush, E.H., and CH. Femald. 1896. The gypsy moth. Wright and Potter Printing Co.,
State Printers, Boston.

Fowler, S.V., and J. H. Lawton. 1985. Rapidly induced defences and trees: The devils
advocate position. I11; Arum M31131 126: 181-195.

Harris, M.O., and JR. Miller. 1983. Color stimuli and oviposition behavior of the
onion fly. 21113131E319m31. 9: 145-155.

Harris, M.O., and J .R Miller.199l. Quantitative analysis of ovipostional behavior:
Effects of a host-plant chemical on the onion fly (Diptera: Anthomyiidae). l.

Ingggt 391133391 4: 773-792.

Hartley, SB, and J .H. Lawton. 1990. Damaged-induced changes in birch foliage:
Mechanisms and effects on insect herbivores. pp. 147-155 13 AD. Watt, S.R.
Leather, M.D. Hunter, and NA. Kidd (Eds), Population dynamics of forest
insects. Intercept, Hampshire.

Haukioja, E. 1990. Positive and negative feedbacks in insect-plant interactions. pp. 113-
122 in AD. Watt, S.R. Leather, M.D. Hunter, and NA. Kidd (Eds), Population
dynamics of forest insects. Intercept, Hampshire.

102

Haukioja, E., and P. Niemela. 1977 . Retarded growth of a geometrid larva after damage
to leaves of its host tree. A911. Z991. Fenn. 14: 48-52.

Haukioja, E., and S. Neuvonen. 1985. Induced long-term resistance of birch foliage
against defoliators: defensive or incidental? E9019gy 66: 1303-1308.

Haukioja, E., S. Neuvonen., P. Niemela, and S. Hanhimaki. 1988. The autumnal moth
E91913 3991199313 in Fennoscandia pp. 167-178 in AA. Berryman (Ed)
Dynamics of forest insect populations. Plenum Press, New York.

Haukioja, E., K. Ruohomaki, J. Senn, J. Suomela, and M. Walls. 1990. Consequences of

herbivory in the mountain birch (@313 999mm: ssp t9r_t99s3): importance of
the functional organization of the tree. 1299919213 82: 23 8-247.

Hough, J .A. and D. Pimentel. 1978. Influence of host foliage on development, survival
and fecundity of the gypsy moth Envjr99. E9t9m91. 7: 97-102.

Houston, DR. 1979. Classifying forest susceptibility to gypsy moth defoliation. USDA.
Agric. Handbook No. 542.

Houston, DR, and MI. Valentine. 1977. Comparing and predicting forest stand
susceptibility to gypsy moth. fig. ,1. £91. 399. 7: 447-461.

Hundertmark, A. 19373. Helligkietis- und Farbenunterscheidungsvermégen der Eiraupen
der Nonne (Lym399‘13 91191139113 L). L. V9rgl. P4193191. 24: 42-57.

Hundertmark, A. 1937b. Das F ormenunterscheidungsvermégen der Eiraupen der Nonne

(Lmantria menacha L.) L 19121- Ehysial. 24: 563-582.

Ichikawa, T., and H. Tateda. 1982. Receptive field of the stemmata in the swallowtail
butterfly 2911191129199. P11319191. 146: 191-199.

Jones, D. 1984. Use, misuse and role of multiple-comparison procedures in ecological
and agricultural entomology. Eng/1:93. Ent9m91. 13: 635-649.

103

Keating, ST, and W.G. Yendol. 1987. Influence of selected host plants on gypsy moth
(Lepidoptera: Lymantriidae) larval mortality caused by a baculovirus. Environ.
Ent9m9l. 16: 459-462. -

 

Keating, S.T., W.G. Yendol, and J .C. Schultz. 1988. Relationship between susceptibility
of gypsy moth larvae (Lepidoptera: Lymantriidae) to a baculovirus and host plant
foliage constituents. E9vir_9n. Ent9m9l. 17: 952-958.

Keating, S.T., M.D. Hunter, and J.C. Schultz 1990. Leaf phenolic inhibition of the gypsy
moth nuclear polyhedrosis virus: the role of polyhedral inclusion body

aggregation. ,1. firm. E99] 16: 1445-1457.

Knapp, R., and TM. Casey. 1986. Thermal ecology, behavior, and growth of gypsy
moth and eastern tent caterpillars. E9919gy 67: 598-608.

Lance, D. 1983. Host-seeking behavior of the gypsy moth: the influence of polyphagy
and highly apparent host plants. pp. 208 19 S. Ahmad (Ed), Herbivorous insects:
Host-seeking Behavior and Mechanisms. Academic Press, New York.

Lance, D., and P. Barbosa. 1982. Host tree influences on the dispersal of late instar gypsy

moths, M39193 91593;. 121199 38: 1-7.

Lance, D.R., J .S. Elkinton, and GP Schwalbe. 1986. Feeding rhythms of gypsy moth
larvae: effects of food quality during outbreaks. E9919gy 67 : 1650-1654.

Lance, D.R., J .S. Elkinton, and CF. Schwalbe. 1986a. Two techniques for monitoring
feeding of large larval lepidoptera, with notes on feeding rhythms of late-instar

gypsy moths (Lepidoptera: Lymantriidae). A119. 59199191. 399. m. 79: 390-394.

Larsson, S.A., A. Wiren, L. Lundgren, and T. Ericsson. 1986. Effects of light and
nutrient stress on leaf phenolic chemistry in $3113 9333191399; and susceptibility to
fi31em99113 1199913 (Coleoptera). 91km 47: 205-210.

Leonard, DE. 1966. Differences in development of strains of the gypsy moth, 11931191113
Slim (L.). Conn. Agric. Exp. Stn. Bull. 680: 1-31.

104

Leonard, DE. 1967. Silking: behavior of the gypsy moth, P9rtl1etri3 disp31. Q.
Ent9m91. 99: 1145-1149.

Leonard, DE. 1970. Feeding rhythm in larvae of the gypsy moth. ,1. E90n. Entomol.
63: 1454-1457.

Leonard, DE. 1974. Recent developments in ecology and control of the gypsy moth.
A311. Rex. Ent9m91. 19: 197-229.

Leonard, DE. 1981. Bioecology of the gypsy moth pp. 9-29 mCC. Doane and ML.
McManus (Eds), The Gypsy Moth: Research Towards Integrated Pest
Management. USDA For. Serv., Washington, DC, Tech. Bull. 1584.

Liebhold, A.M., J.S. Elkinton, and WE. Wallner. 1986. Effect of burlap bands on
between-tree movement of late-instar gypsy moth, W 91993;
(Lepidoptera; Lymantriidae). E9vir99. £91. 14: 373-379.

Lindroth, R.L., and J .D.C. Hemming. 1990. Responses of the gypsy moth (Lepidoptera:
Lymantriidae) to tremulacin, an aspen phenolic glycoside. Enyir9n. Ent9m91. 19:
842-847.

Maksimovic, M. 1958. Experimental research on the influence of temperature upon the
development and the population dynamics of the gypsy moth (L1p3m 91993; L).
29391299, 115131119 B191, 1991, NR, $112119 3: 1-115. (Trans. from Serb-Croatian
1963. 91S 61-11203: 1-95).

Mason, C.J., and ML. McManus. 1981. Larval dispersal of the gypsy moth. mCC.
Doane and ML. McManus (Eds), The Gypsy Moth: Research Towards
Integrated Pest Management. USDA For. Serv., Washington, D.C., Tech. Bull.
1584.

Mattson, W.J., and J .M. Scriber. 1987. Nutritional ecology of insect folivores of woody
plants: nitrogen, water, fiber and mineral considerations. pp. 105-146 19 Slansky,
F., and J .G. Rodriguez (Eds), Nutritional ecology of insects, mites and spiders.
John Wiley and Sons, New York.

105

McClure, MS. 1977. Population dynamics of red pine scale, MW 99391939,

(Homoptera: Margarodidae): the influence of resinosis. Envir99. Ent9m91. 6:
789-795.

McClure, MS. 1979. Self-regulatiuon in populations of the elongated hemlock scale,
F iorinia ex_te1113 (Homoptera: Diaspididae). Qecol9g1a 39: 25-64.

McClure, MS. 1980. F oliar nitrogen: a basis for host suitability for elongated hemlock
scale, F19rjni3 9319113 (Homoptera: Diaspididae) E9919gy 61: 72-79.

McFarlane, J .E. 1985. Nutrition and digestive organs. pp. 59-89 i_n Blum, M.S. (Ed),
Fundamentals of insect physiology. John Wiley and Sons, New York.

Mc Manus, M.L., and HR. Smith. 1972. Importance of the silk trails in the diel behavior
of late instars of the gypsy moth. Envir99. Entom9l. 1: 793-795.

McManus, ML. 1973. The role of behavior in the dispersal of newly hatched gypsy
moth larvae. USDA For. Serv., Northeastern For. Exp. Stn., Hamden, CT, Res.
Pap. NE-267.

McManus, M.L., and T. McIntyre. 1981. Introduction. pp. 1-7 111_C.C. Doane and ML.
McManus (Eds), The Gypsy Moth: Research Towards Integrated Pest
Management. USDA For. Serv., Washington, DC, Tech. Bull. 1584.

Meyer, G. A., and M. E. Montgomery. 1987. Relationship between leaf age and the food

quality of cottonwood foliage for the gypsy moth, W 91393;. 1192919213
72: 527- 532.

Miller, J .R., and MO. Harris. 1985. Viewing behavior-modifying chemicals in the
context of behavior: lessons from the onion fly. pp. 166-168 i_n_ T.E. Acree and
D.M. Soderlund (Eds), Semiochemistry: Flavors and Pheromones. Walter de
Gruyter and Co., New York.

Miller, J .R., and KL. Strickler. 1984. Finding and accepting host plants. pp. 127-157 19
W.J. Bell and R.T. Cardé (Eds), Chemical Ecology of Insects. Sinauer Associates
Inc., Sunderland, MA

Mills, NJ. 1983. Gypsy moth (Lmantria 9ispar): Work in Europe in 1983. C.I.B.C.
Report. European Station, Delémont, Switzerland.

106

Moericke, V., R.J. Prokopy, S. Berlocher, and G.L. Bush 1975. Visual stimuli eliciting
attraction of Rh3g919 's 99m9n9113 (Diptera: Tephritidae) flies to trees. Ent9m91.
Exp. A991. 18: 497-507.

Mole, S., and J .A. Ross and P.G. Waterman. 1988. Light-induced variation in phenolic
levels in foliage of rain-forest plants 1. Chemical changes. 1. Chem. EC9l. 14: 1-
21.

Mooney, H.A., S.L. Gulmon, and ND. Johnson. 1983. Physiological constraints on plant
chemical defenses. pp. 21-36 19 RA. Hedin (Ed) Plant resistance to insects. Am.
chem. Soc. Washington DC

Moon, P. 1961. The scientific basis of illuminating engineering. Dover, New York.

Niemela, P., E.M. Arc and E. Haukioja. 1979. Birch leaves as a resource for herbivores:
damaged-induced increases in leaf phenols with trypsin inhibiting effects. R99.

43K v .SllbamticReseatphStatiQn 15: 37-40.

Niemela, P., J. Toumi, R. Mannila, and P. Ojala. 1984. The effect of previous damage on
the quality of Scots pine foliage as food for Diprionid sawflies. Z9it§9hrifi &

3113913111139 Emphasis 95: 33-43.

Perkel, DH, and TH. Bullock. 1968. Neural coding. Neurosciences Res. Program Bull.
6: 221-348.

Perry, J .N. 1986. Multiple-comparison procedures: A dissenting view. 1. E999. E9t9m9l.
79: 1149-1155.

Peterson, R.G. 1985. Separation of Means. pp. 77 19 Design and analysis of experiments.
Dekker Inc. New York.

Pritchett, W.R. 1975. The potential geographic diffusion of the gypsy moth (P99heri3
913931;) into southern forests. W W 14: 47-55. ‘

Podgwaite, J .D. 1981. Natural disease within dense gypsy moth populations. pp. 125-134
i_n C.C. Doane and ML. McManus (Eds), The Gypsy Moth: Research Towards
Integrated Pest Management. USDA For. Serv., Washington, DC, Tech. Bull.
1584.

107

Price, P.W., C.E. Bouton, P. Gross, B.A. Mcpherson, J.N. Thompson, and AB. Weis.
1980. Interactions among three tropic levels: Influence of plants on interactions
between insect herbivores and natural enemies. A99. R99. E901. S451. 11: 41-65.

Prokopy, R.J. 1977. Attraction of Rh3g9letis 99m9nell3 flies (Diptera: Tephritidae) to
red spheres of different sizes. Cg. Ent9m91. 65: 1444-1447.

Rathcke, BL, and R.W. Poole. 1975. Coevolutionary race continues: butterfly larval
adaptation to plant trichomes. 3919199 187: 175-176.

Raupp, M.J., and RF. Denno. 1983. Leaf age as a predictor of herbivore distribution and
abundance. pp. 95 111 RP. Denno and MS. McClure (Eds), Variables Plants and
Herbivores in Natural and Managed Systems. Academic Press, New York.

Riordan, C.J., Myers, DR, and Hulstrom, R.L. 1989. Spectral Solar Data Base
Documentation. Solar Energy Research Institute. Rep. Tr-215-3513B. Golden,
CO.

Roden, DB, and GA. Surgeoner. 1991. Survival, development time and pupal weights
of larvae of gypsy moth reared on foliage of common trees of the upper Great
Lakes region. N91. ,1. A991. F91. 8(3): 126-128.

Roden, D.B., J .R. Miller, and GA. Simmons. 1992. Visual stimuli influencing
orientation by larval gypsy moth, 199939993 913931. Q39. Ent9m91. 124: 287-304.

Roden, D.B., J .C. Kimball, and GA. Simmons. 1990. A lighting system to induce a
change in feeding behavior between small and large larvae of gypsy moth,

Lasagna-1'3 dim (L.). $12.11. Ent9m91. 122: 617-625.

Rossiter,M.C., and J .C. Schultz. 1988. Relationships among defoliation, red oak
phenolics and gypsy moth growth and reproduction. E99195: 69: 267-277.

Rossiter, M.C., J .C. Schultz, and LT. Baldwin. 1988. Relationships among defoliation,
red oak phenolics, and gypsy moth growth and reproduction. E9919gy 69: 267-
277.

108

Saxena, K.N., and P. Khattar. 1977. Orientation of P39ili9 9999993 larvae in relation to

size, distance, and combination pattern of visual stimuli. 1. 1ns99t Phy§i9l. 23:
1421-1428. -

Schultz, J .C. 1983. Habitat selection and foraging tactics of caterpillars in heterogeneous
trees. pp. 62 19 RF. Denno and M.S. McClure (Eds), Variables Plants and
Herbivores in Natural and Managed Systems. Academic Press, New York.

Schultz, J .C., and LT. Baldwin. 1982. Oak leaf quality declines in response to defoliation
by gypsy moth larvae. 3919199 217: 149-151.

Schultz, J .C., and M.J. Lechowicz. 1986. Hostplant, larval age, and feeding behavior

influence midgut pH in the gypsy moth (@3993 919931;). mm 71: 133-
137.

Schultz, J .C., M.A. Foster, and ME. Montgomery. 1990. Host plant-mediated impacts of
a baculovirus on gypsy moth populations. pp. 303-313. 19 AD. Watt, S.R
Leather, M.D. Hunter, and NA. Kidd (Eds), Population dynamics of forest
insects. Intercept, Hampshire.

Scriber, LM. 1977. Limiting effects of low leaf-water content on the nitrogen utilization,

energy budget and larval growth of Hy31991m 9991:9913 (Lepidoptera:
Satumiidae). 999919213 23: 269-287.

Shapiro, M. 1977. Gypsy moth mass rearing. Evolution of methods for surface
sterilization of eggs and/or pupae. Laboratory Report (October 1976 - March
1977). USDA. Anim. Plant Health. Insp. Serv., Otis AFB, Mass.

Shorey, H.H. 1977. Introduction. pp. 1 19 H.H. Shorey and J .J . McKelvey (Eds),
Chemical control of insect behavior. John Wiley and Sons, New York.

Singer, MC. 1971 Evolution of foodplant preference and the butterfly mm
991913. Ev9l9ti9n 25: 383-389.

Smiley, J. 1978. Plant chemistry and host specificity: New evidence from 1191199n4i1s and
9359111913. 3919199 201: 745-747.

109

Sterner, TE, and A.G. Davidson. 1981. (Eds) pp. 15 in Forest insects and disease
conditions in Canada 1980. Dep. Environ, Can. For. Serv., Ottawa, Ont.

Strong, D.R., J.H. Lawton, and TR. Southwood. 1984. Insects on plants. pp. 21 Harvard
Univ. Press., Cambridge, MA.

Sullivan, CR, and DR. Wallace. 1972. The potential northern distribution of the gypsy
moth, P9rth9tri3 919931(Lepidoptera: Lymantriidae). C39. Em. 104. 1349-1355.

Tabashnik, BE. 1986. Evolution of host plant utilization in M39 butterflies. pp 173-
184 19 MD. Huettel (Ed), Evolutionary genetics of behavior: Progress and
prospects. Plenum Press, New York.

Thomsteinson, A]. 1960. Host selection in phytophagus insects. A99. F991. 9999191. 5:
193-318.

Valentine, H.T., W.E. Wallner, and RM. Wargo. 1983. Nutritional changes in host
foliage during and after defoliation, and their relation to the weight of gypsy

pupae. Q99919g13 7: 298-302.

Waller, RA, and DB. Duncan. 1969. A Bayes rule for the symmetric multiple
comparisons problem. ,1. Arm; $13931. A999. 64: 1484-1503.

Wallner, W.E., and GS. Walton. 1979. Host defoliation: a possible determinant of gypsy
moth population quality. 5199. 29191991. 599. A19. 72: 62-67.

Way, M.J., and M. Cammell. 1970. Aggregation behavior in relation to food utilization
by aphids SEE B9. E991. S99. 10: 229-247.

Wellington, W.G. 1948. The light reactions of spruce budworm, (999151999993
9mm Clemens (Lepidoptera: Tortricidae). 939. E91. 130: 56-82.

Weseloh, RM. 1989. Behavioral responses of gypsy moth (Lepidoptera: Lymantriidae)
larvae to abiotic environmental factors. 591/1199. E919mg1. 18: 361-367.

110

Whitham, T.G. 1983. Individual trees as heterogeneous environments: adaptation to
herbivory or epigenetic noise? pp. 9-27 i_n Denno, RF. and H. Dingle (Eds),
Insect life history patterns: Habitat and geographic variation. Springer-Verlag,
Berlin and New York.

Wigglesworth, VB. 1974. The principles of insect physiology. Seventh ed. Wiley, New
York.

Witter, J .A., CA. Chilcote, J. L. Stoyenoff, and ME. Montgomery. 1990. Host
relationships of gypsy moth in aspen and oak stands: Management applications.
pp. 63-74 19 Proc. Annu. Gypsy Moth Rev. Ottawa, Ontario.

Wratten, S.D., P.J. Edwards, and I. Dunn. 1984. Wound-induced changes in the
palatability of 999113 999escens and _B_. @9913. Q99919g13 61: 372-375.

Wratten, S.D., P.J. Edwards, and AM. Barker. 1990 Consequences of rapid feeding-
induced changes in trees for the plant and the insect: Individuals and populations.
pp. 137-145 19 AD. Watt, S.R. Leather, M.D. Hunter, and NA. Kidd (Eds),
Population dynamics of forest insects. Intercept, Hampshire.

Zar, J .H. 1984. Biostatistical Analysis, 2nd ed. Prentice-Hall, N.J.

"illlllrllbillet