:-
Mr.-

$.».. .-

..c..rvrul

I: Is: .

.025 . 55:5»: ..

I... l5; 0" ‘.h|..7‘9v -
!..:i . it. .v

. 5

II.
.- .I 9

cittvr, . a

.I 1.... 1
I~.l.;)\.r-.

.rc Olvfillntll’uéirO
r3: [‘3

.2. I. .
‘vllr

vv'yt .! v‘t altl
'Dpll-.,lr Quit lv'lp! .
'of. co. . 7.1:...

9tE-b’nt:
1.9-3; 1‘.

KP.

v0...li

 

 

 

 

 

Cl.

C“-

MlCHl lGAN STATE UNI IEVR

N l M,” I! Hi U!llf’lIHlHHIHIWl

3 12 00914 1023

m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

PHENOLOGICAL VARIATION IN THE SUSCEPTIBILITY
0F WHITE SPRUCE TO THE SPRUCE BUDWORM

presented by

ROBERT KENT LAWRENCE

has been accepted towards fulfillment
of the requirements for

PH . D . . ENTOMOLOGY

degree in

 

 

    

Major professor

l
DateZJfiML

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

 

 

 

F a
MBA!!!
"Wilson State
University J

‘

 

 

It

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

DATE DUE DATE DUE DATE DUE
ll

 

 

 

 

 

 

 

 

 

 

 

fi—fi
=_ Jl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

——I—'—l

MSU Is An Affirmative ActIon/Equal Opportunity Institution
czmmpma-DJ

 

 

 

.,'

PHENOLOGICAL VARIATION IN THE SUSCEPTIBILITY
OF WHITE SPRUCE TO THE SPRUCE BUDWORM

By

Robert Kent Lawrence

A DISSERTATION

Submitted to
Michi an State Universi
in partial ent of the reqmrements
for the degree of

DOCTOR OF PHILOSOPHY

Department ofd Entomology
an
Program in Ecology and Evolutionary Biology

1990

i4

ABSTRACT

PHENOLOGICAL VARIATION IN THE SUSCEPTIBILITY
OF WHITE SPRUCE TO THE SPRUCE BUDWORM

By

Robert Kent Lawrence

Seasonal and maturational changes in plant traits create temporal windows
of susceptibility during which the plant is a suitable host for various herbivores.
Synchrony of insect phenology and host tree phenology has often been suggested
as a factor in the susceptibility of white spruce, Picea glauca (Moench) Voss, and
other host species to the spruce budworm, Choristoneura fimtiferana (Clemens).
Although the concept of phenological windows of susceptibility has long existed,
this study is the first to test the hypothesis comprehensively and to substantiate it.
Phenological variation in white spruce susceptibility was evaluated by measuring
performance of several cohorts of spruce budworms caged on trees at different
phenological stages of the hosts and comparing budworm performance with
phenological variation of 13 foliar traits.

The window of susceptlbility in white spruce was well defined in this study.
Performance is high for budworms that emerge and begin feeding during the 3-4
weeks prior to budbreak and complete larval development prior to the end of
shoot growth. Larval survival averaged greater than 60% for budworms that
began development 1-3 weeks prior to budbreak, but decreased to less than 10%

for budworms that began development two or more weeks after budbreak. More

than 85% of total larval mortality occurred by the time most larvae were third
instars in both early and late budworm cohorts.

Variation in budworm larval survival and length of development (degree
days) among three white spruce seed sources was clearly linked to phenological
differences among seed sources. Variation in budworm weight, growth rate and
length of development (days) among seed sources was positively correlated with
tree growth rate, thereby implying that competition for resources within individual
trees may affect their susceptibility to insects.

Levels of nitrogen, phosphorous, potassium, total sugars and water in
current-year foliage declined, while leaf toughness increased, during spring and
early summer. Calcium, magnesium, copper and zinc declined and later
increased. High water content and low leaf toughness are strongly correlated with
high budworm performance and are the foliar traits most likely to be key factors

defining the window of susceptibility in white spruce.

To Martha, who lovingly made many sacrifices and
contributed to my success in countless ways during a time
when she was laboring on her own doctoral program,
and
To my parents, who first introduced me
to the fascination of the biological sciences

and who have always given their love and support.

iv

ACKNOWLEDGMENTS

I would like to express my sincere thanks to Dr. William J. Mattson, my
major professor, for his guidance and his friendship. I learned a great deal from
him about how to think as well as to do. His unselfish support and encourage-
ment throughout my lengthy research program has been greatly appreciated. He
and his wife, Sheri, have been true friends. Choosing to work with Bill was one of
my better decisions.

I would also like to thank the members of my graduate committee for their
many contributions. Dr. James W. Hanover was exceptionally generous in
allowing me to conduct experiments in his forest genetics plantations and in
providing laboratory facilities for a special research project. I enjoyed taking his
classes and have appreciated his advice and friendly cooperation. Dr. John A.
Witter provided valuable guidance and played a key role as motivator. I
appreciated his occasional friendly suggestions concerning my research velocity. I
am very grateful to Dr. Gary A. Simmons for his advice, encouragement and
willingness to listen to concerns both personal and professional. I enjoyed the
opportunity to assist Gary in teaching, and hope that I can be as enthusiastic and
effective as he in my own future teaching endeavors. My thanks go to Dr. James
R. Miller for providing constructive criticism from the vantage point of one who is
not so close to the trees.

Special thanks are extended to the North Central F orcst Experiment
Station of the U. S. D. A. Forest Service for sponsoring my research program.

The continued financial support throughout my program was deeply appreciated.

V

I am indebted to my co-workers for their assistance, their patience and
their friendship. Dr. Robert A. Haack gave valuable assistance in many areas.

He made important contributions to this research with his observations,
discussions and suggestions during the early phases of my work. He shared many
long days in the field and lab assisting in my experiments as well as maintaining
his own. My sincere thanks goes to Bruce A. Birr for his persistence and patience
in doing all of the foliar nitrogen and mineral analyses. He made my work much '
easier by providing valuable support in many ways in both the field and lab. I
also thank fellow graduate student, Daniel A. Herms, for many stimulating
conversations and helpful suggestions. Thanks also go to Richard Blank for doing
foliar sugar analyses and to Kwei-Fong Hao, Brian Mitchell and Brian
McLaughlin for successfully accomplishing many tedious tasks of sample
preparation and data collection.

I appreciate the advice given by Dr. Carl W. Ramm on techniques of
multivariate statistical analysis. I also thank Tom Stadt and John Vigneron of the
W. K. Kellogg Experimental Forest for providing valuable support to my field
research and Bill Dunn of the Huron-Manistee National Forest for providing

temperature data.

TABLE OF CONTENTS

LIST OF TABLES ............................................. x
IIST OF FIGURES ........................................... xii
WINDOWS OF SUSCEPTIBILITY ................................ 1
Phenological Variation in Plant Traits ......................... 1
SPRUCE BUDWORM AND HOST PHENOLOGY ................... 5
WHITE SPRUCE BIOLOGY .................................... 8
OBJECTIVES ................................................. 9
MATERIALS AND METHODS ................................. 10
Phenological Window of Susceptibility ........................ 10
Spruce Budworm Cohorts . .' ............... ~ ........... 1 1

Tree Phenology .................................... 14

Spruce Budworm Survivorship Curve ......................... 15
Seasonal Variation in Host Nutritional Traits ................... 16
Statistical Methods ....................................... 19

Insect Performance Variables .......................... 19

Analysis of Repeated Measurements .................... 19

Relationship of Host Susceptibility and Host Phenology ...... 21

RESULTS .................................................. 24
Phenological Window of Susceptibility ........................ 24

Relative Timing of Tree Phenology and Budworm Phenology . . 24

Spruce Budworm Performance ......................... 27

vii

Budworm Cohorts and Tree Seed Sources ........... 27

Seed Source Variation in Phenology ............... 33

Tree Growth Rate ............................. 41

Other Factors ................................ 44

Spruce Budworm Survivorship Curve ......................... 51
Seasonal Variation in Host Nutritional Traits ................... 57
Current-Year Foliage ................................ 58
Sugars ...................................... 58

Nitrogen and Minerals ......................... 58

Water content and Leaf Toughness ................ 64
One-Year-Old Foliage ............................... 70
Relationship of Host Susceptibility to Host Phenology ............ 70
Canonical Correlation Analyses ........................ 72
Ordination of Canonical Scores ........................ 76
Regression Analyses ................................ 80
DISCUSSION ................................................ 93
Defining the Window of Susceptibility ........................ 93
Seed Source Effects ...................................... 96
Effects of Tree Growth Rate ............................... 97
Relationship of Host Susceptibility to Host Phenology ............ 98
Key Foliar Traits ................................... 99

Total Sugars ................................. 99

Nitrogen ................................... 100

Minerals ................................... 101

Water ..................................... 102

Toughness .................................. 102

Secondary Compounds ........................ 103

‘viii

Foliar Traits of the Window of Susceptibility ............. 104

Effects of Insect-Tree Phenological Asynchrony ........... 105

Implications for Forest Pest Management ..................... 109
Recommendations for Future Research ...................... 111
CONCLUSIONS ............................................. 113
APPENDIX ................................................ 116
LITERATURE CITED ........................................ 118

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

LIST OF TABLES

Dates and accumulated degree days when each spruce budworm

cohort was placed on white spruce trees in 1985 ..............

Bud phenology rating and percent of buds burst for three white
spruce seed sources during April-May 1985 (N = 32 buds per

seed source) .........................................

Multivariate repeated measures analysis of spruce budworm

performance in 1985 with budworm cohorts as repeated measures . .

Seed source effect on spruce budworm erformance in 1985
represented as mean sums of cohort v ues for budworm

performance variables ..................................

Multivariate repeated measures analysis of spruce budworm
performance in 1985 usin cohorts from phenologically similar
osts (cohorts 1-3 of the ASK seed source and cohorts 2-4 of

the ONT and BC seed sources) as repeated measures ..........

Seed source effect on spruce budworm performance in 1985
using cohorts from phenologically similar hosts (cohorts 1-3

of SASK seed source and cohorts 2-4 of the ONT and BC
seed sources). Seed source effect is represented as mean sums

of cohort values for budworm performance variables ...........

Coefficients of determination (r2) for the relationship of
spruce budworm performance to tree growth rate (cm/yr) for
budworm cohorts started at two weeks prior to budbreak, one

week prior to budbreak and at budbreak in 1985 ..............

Comparison of spruce budworm survival rates among 1985
budworm batches for insects reared on artificial diet in

laboratory studies and for insects in the white spruce field study . . .

Com arison of spruce budworm length of development, adult
wei t and growth rate among 1985 budworm batches for insects
reared on artificial diet in laboratory studies and for insects in

the white spruce field study ..............................

Table 10. Daily average temperatures, daily maximum temperatures and

degree days er day during the period that eight spruce

.13

.25

35

.40

.42

.43

.45

49

.50

budworm co orts were caged on white spruce in 1985 ......... 52

Table 11.

Table 12.

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Appendix 1.

Spruce budworm performance for the two budworm cohorts

in the 1987 survivorship experiment ......................

Average larval stage and distribution of spruce budworm
larvae on ca ed branches in relation to timing of budbreak

during the 1 87 survivorship experiment ...................

Multivariate repeated measures analysis of total sugar, total
nitrogen and mineral content of current-year foliage in
June-September 1985 and water content and toughness of
current-year foliage in May-August 1986 with sample dates as

repeated measures ................................... .

Seed source effects on foliar characteristics represented as
mean sums of repeated measurements for current-year

foliage variables .....................................

Mean values (15E) of 12 foliar characteristics of one-year-old
foliage on three white spruce seed sources sampled less than
one week prior to budbreak on the earliest-flushing seed source

Canonical correlation analysis of insect performance and host
quality using host foliar values during early larval development

and late larval development ............................

Pearson correlation coefficients among 10 major foliar
variables for foliar levels during early larval development

(cohorts 3-8, N = 136) ........................ . ........

Pearson correlation coefficients among 10 major foliar
variables for foliar levels during late larval development

(cohorts 1-8, N = 192) ................................

Range of values for foliar traits during early larval
develo ment and late larval development within the

55

56

59

61

.71

73

9o

91

pheno ogical window of susceptibility .................... 106

Record of Deposition of Voucher Specimens ............ 116

Appendix 1.1 Voucher Specimen Data ........................... 117

Figure 1.
Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

LIST OF FIGURES

Shoot elongation of three white spruce seed sources in 1985 . . . .

Relative timing of eight spruce budworm cohorts and white

spruce budbreak and shoot elongation ...................

Mean percent survival (:SE) of (A) larvae and (B) pupae in

eight spruce budworm cohorts .........................

Mean length of development period (:SE) from second instar
to adult measured in (A) days and (B) degree days for eight

spruce budworm cohorts ..............................

(A) Mean adult dry weight QSE) and (B) mean growth rate
+

SE) for eight spruce budworm cohorts ..................

Mean percent defoliation (:SE) of current-year sheets for
eight spruce budworm cohorts on three seed sources of white

spruce ...........................................

Mean percent larval survival (18E) for (A) spruce budworm

cohorts 1-4 on three white spruce seed sources and for (B)

s ruce budworm cohorts from phenologically similar host trees.
ohorts in (B) from weeks -3 and +1 were not used in

repeated measures analyses but are presented for illustrative

purposes only ......................................

Mean length of development in degree days (:SE) for (A)
spruce budworm cohorts 1—4 on three white spruce seed
sources and for (B) spruce budworm cohorts from
phenologically simrlar host trees. Cohorts in (B) from weeks -3
and +1 were not used in repeated measures analyses but are

26

.28

.29

.31

.32

.34

.36

presented for illustrative purposes only .................... 37

Mean adult dry weight (tSE) for (A) spruce budworm cohorts
1-4 on three white spruce seed sources and for (B) spruce
budworm cohorts from phenologically similar host trees.
Cohorts in (B) from weeks -3 and + 1 were not used in
repeated measures analyses but are presented for illustrative

purposes only ....................................... 38

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.
Figure 21.

Figure 22.

Mean growth rates (:SE) for (A) spruce budworm cohorts

1-4 on three white spruce seed sources and for (B) spruce

budworm cohorts from phenologically similar host trees.

Cohorts in (B) from weeks -3 and +1 were not used in

repeated measures analyses but are presented for illustrative
purposes only ....................................... 39

Relationship of length of development (days) of spruce

budworms and host tree growth rate for budworm cohorts

started two weeks £31m (circles within solid line), one week

prior (triangles wi ' dashed line) and at budbreak (asterisks) . . 46

Relationship of s ruce budworm adult dry weight and host

tree growth rate or budworm cohorts started two weeks prior

(circles within solid line), one week prior (triangles within

dashed line) and at budbreak (asterrsks) ................... 47

 

Relationship of spruce budworm growth rate and host tree
owth rate for budworm cohorts started two weeks prior
circles within solid line), one week prior (triangles within
dashed line) and at budbreak (asterrsks) ................... 48

Spruce budworm survivorship curves (mean percent alive 1 SE)
for an earl cohort (April-June) and a late cohort (June-July)
on the 0 seed source during 1987 ...................... 53

Seasonal variation of total sugar in current-year foliage of
three seed sources in 1985 .............................. 62

Seasonal variation of (A) total nitrogen and (B) phosphorous
in current-year foliage of three seed sources in 1 85 .......... 63

Seasonal variation of (A) potassium and (B) calcium in
current-year foliage of three seed sources in 1985 ............ 65

Seasonal variation of (A) magnesium and (B) manganese in
current-year foliage of three seed sources in 1985 ............ 66

Seasonal variation of (A) copper and (B) iron in current-year
foliage of three seed sources rn 1985 ...................... 67

Seasonal variation of (A) sodium and (B) zinc in current-year
foliage of three seed sources in 1985 ...................... 68

Seasonal variation of (A) water and (B) toughness in
current-year foliage of three seed sources in 1986 ............ 69

Ordination of canonical scores for insect performance and

host quality during late larval development on the SASK seed

source. Numbers within graph indicate cohorts on individual

trees .............................................. 77

Figure 23.
Figure 24.

Figure 25.
Figure 26.
Figure 27.
Figure 28.
Figure 29.
Figure 30.
Figure 31.

Figure 32.

Figure 33.

Ordination of canonical scores for insect performance and

host quality during late larval development on the ONT seed

source. Numbers within graph indicate cohorts on individual 78
trees ..............................................

Ordination of canonical scores for insect performance and

host quality during late larval development on the BC seed

source. Numbers within graph indicate cohorts on individual 79
trees ..............................................

Relationship of percent larval survival and total sugar in
current-year foliage during (A) early larval development and
(B) late larval development ............................. 82

Relationship of percent larval survival and total nitrogen in
current-year foliage during (A) early larval development and
(B) late larval development ..... ' ........................ 83

Relationship of percent larval survival and phos horous in
current-year foliage during (A) early larval deve opment and
(B) late larval development ............................. 84

Relationship of percent larval survival and potassium in
current-year foliage during (A) early larval development and
(B) late larval development ............................. 85

Relationship of percent larval survival and calcium in
current-year foliage during (A) early larval development and
(B) late larval development ............................. 86

Relationship of percent larval survival and water content in
current-year foliage during (A) early' larval development and
(B) late larval development ............................. 87

Relationship of percent larval survival and toughness of
current-year foliage during (A) early larval development and
(B) late larval development ............................. 88

Relationship of percent larval survival and foliar levels of
total nitrogen and water during early larval development ...... 107

Relationship of percent larval survival and foliar levels of
toughness and water during early larval development ......... 108

xiv

WINDOWS 0F SUSCEPTIBILITY

Plant resistance against herbivory involves a combination of many host-
plant traits including nutrients, allelochemics, physical attributes and bio-
associations. However, plant traits are not invariant. They typically vary spatially
and temporally, as does the suitability of a host to herbivores (Mattson et a1.
1982). Spatial variation occurs within and among individual plants and among
populations (Whitham 1981, 1983, Krischik and Denno 1983). Temporal variation
can exist as diurnal, seasonal or ontogenetic changes in plant traits and as changes
induced by biotic or abiotic agents (Mattson et al. 1982, Raupp and Denno 1983).

For each herbivore, there is a range of values for each host-plant trait that
is suitable for growth and survival of the herbivore. Plants may pass through a
temporal suitability/ susceptibility "window" when plant trait values optimally
match with herbivores’ tolerance ranges for plant traits (Feeny 1976, Lawton and
McNeill 1979, Mattson et al. 1982, Mooney et a1. 1983). Such windows of
susceptibility are associated with both phenological and ontogenetic changes in
host quality. The success of an herbivore on a given host plant therefore depends

on the degree of its synchrony with host phenology.

Phenological Variation in Plant Traits

Many plant traits that affect the suitability of a host plant to herbivores

vary seasonally or, in other words, with leaf age (Raupp and Denno 1983, Cates

2
1987, Mattson and Scriber 1987). Among these traits are the foliar levels of water
(Hinckley et al. 1978, Scriber and Slansky 1981), nitrogen (McNeill and
Southwood 1978, Mattson 1980), mineral elements (Mattson and Scriber 1987,
Clancy et al. 1988a), fatty acids (Clark et al. 1975), carbohydrates (Little 1970,
Chapin et al. 1986), phenolics (Feeny 1970, Rossiter et al. 1988), terpenoids (von
Rudloff 1972, Cates and Redak 1988), alkaloids (McKey 1974), epicuticular waxes
(Hanover and Reicosky 1971) and physical attributes such as leaf toughness and
fiber content (F eeny 1970, Coley 1983, Mattson and Scriber 1987).

The levels of water and nitrogen in leaves are positively correlated and
both decline strongly with leaf age (Scriber and Slansky 1981, Scriber 1984b,
Mattson and Scriber 1987). These patterns consistently occur in most plant
growth forms (e.g. grasses, forbs, trees). However, levels of both water and
nitrogen are usually lower in trees than in herbaceous plants, and lower in
evergreen trees than in deciduous trees. The levels of water and nitrogen in very
young leaves of trees approach those found in herbaceous plants (Mattson and
Scriber 1987). Insect herbivore performance (e.g. growth rates and food
utilization efficiencies) is usually positively correlated with foliar levels of both
water and nitrogen (Scriber and Slansky 1981, Scriber 1984b, Mattson and Scriber
1987). Although water and nitrogen are not the only factors affecting insect
performance, they effectively reflect the milieu of plant traits encountered by the
herbivore (Scriber 1984b, Mattson and Scriber 1987), and together these two plant
traits provide an index of plant quality, which can be used to predict maximum
growth performance of insect herbivores (Scriber 1984b).

Plant nitrogen varies qualitatively, as well as quantitatively,
with advancing plant phenology (McNeill and Southwood 1978, Mattson 1980).

Nitrogen content of young plant tissues consists primarily of amino acids, soluble

3
proteins and nitrogen-based secondary compounds, while the nitrogen content of
older tissues consists primarily of structural or insoluble protein (Mattson 1980).

Seasonal variation of foliar mineral concentrations is relatively consistent
for several mineral elements among many woody plant species. Fer example,
phosphorous and potassium levels decrease rapidly during the growing season,
whereas calcium, manganese and aluminum may increase almost linearly (Grigal
et al. 1976, Mattson and Scriber 1987, Clancy et al. 1988a, Hockman et al. 1989).
The seasonal trend for magnesium is less consistent, with foliar concentrations
increasing, decreasing, or sometimes first decreasing and then increasing later in
the season (Grigal et al. 1976, Mattson and Scriber 1987, Clancy et al. 1988a).
Micronutrients such as boron, copper, iron, zinc and sodium may also rapidly
decrease early in the growing season and later increase (Mattson and Scriber
1987).

Seasonal variations in concentrations of carbohydrates in temperate woody
plants reflect the dynamics of photosynthesis, carbon metabolism and
translocation of assimilates (Leach and Little 1973, Bernard-Dagan 1988). Energy
reserves (e.g. starch, sugars, lipids) are metabolized before the start of growth and
translocated from storage sites in roots, stems and shoots to new leaves and other
growing points (Kramer and Kozlowski 1979, Bernard-Dagan 1988). As
photosynthesis increases and new leaves become net producers of carbohydrates,
assimilates are exported from leaves to other parts of the plant. Seasonal patterns
of non-structural carbohydrates differ among plant growth forms (Kramer and
Kozlowski 1979, Chapin et al. 1986). Annual fluctuations in concentrations of
non-structural carbohydrates are much less in evergreen plants than in deciduous
plants. In evergreen plants a large portion of reserve carbohydrates are stored
during the winter in old leaves, and then translocated to new leaves during initial

growth stages in the spring (Ericsson 1978).

4

Foliar levels of allelochemics can vary considerably with plant phenology.
Levels of phenolic compounds typically increase throughout the growing season
(Feeny 1970, Haukioja et al. 1978, Mattson and Palmer 1988). On the other
hand, they may peak at, or within a few weeks after, budbreak and then decline
with leaf age (Coley 1983, Baldwin et al. 1987, Mattson et al. 1988, Rossiter et al.
1988). The total quantity of terpenes in current-year conifer foliage increases
during the growing season (von Rudloff 1972, Cates and Redak 1988). However,
the greatest changes in relative amounts of terpenes in conifer foliage often occur
in buds prior to budbreak or in new leaves just after maturing during the shoot
growth period (von Rudloff 1972, 1975a, 1975b, von Rudloff and Granat 1982).
Levels of alkaloids, cyanogenic glycosides and many other toxins are higher in
younger leaves than older leaves (McKey 1974, Mattson 1980).

Leaf toughness, a largely unappreciated defense, is directly correlated with
leaf maturity (Feeny 1970, Coley 1983, Lowman. and Box 1983). Levels of
structural components (e.g. cellulose, hemicellulose, lignin) that contribute to leaf
toughness increase in maturing leaves (Chung and Barnes 1980, Scriber and

Slansky 1981, Mattson and Scriber 1987).

SPRUCE BUDWORM AND HOST PHENOLOGY

Interspecific and intraspecific variation in host tree phenology have often
been cited as major factors in the relative susceptibility of host trees to the spruce
budworm (Choristoneura firmiferana (C1emens)) and western spruce budworm (C.
occidentalis Freeman) in North America (Blais 1957, Greenbank 1963, Eidt and
Cameron 1971, Redak and Cates 1984, Volney 1985, Osawa et al. 1986, Blum
1988) and to the European fir budworm (C. murinana Hiibner) in Europe
(Schonherr 1980). Annual variations in host tree susceptibility have been, in
some cases, attributed to climatic effects that alter the synchrony of budworm
phenology and host tree phenology (Schdnherr 1980, Beckwith and Bumell 1982,
Thomson et al. 1984, Shepard 1985).

The spruce budworm is a micro-lepidopteran folivore of boreal North
America. Its primary hosts are balsam fir (Abies balsamea (I...) Mill.) and white
spruce (Picea glauca (Moench) Voss), but the budworm can cause significant
damage to red spruce (P. rubens Sarg.), black spruce (P. mariana (Mill) B.S.P.)
and subalpine fir (A. lasiocarpa (Hook) Nutt.) (Harvey 1985). Periodic
outbreaks of this insect have resulted in many millions of hectares of defoliated
and ldlled trees in northeastern North America (MacLean 1985, Hardy et al.
1986).

Emergence and development of spruce budworm larvae in the spring
generally coincides with budbreak and shoot elongation of its host trees
(McGugan 1954, Greenbank 1963). The larvae overwinter on trees as 2nd instars

and emerge from their hibernacula shortly before budbreak. Larvae initially mine

5

6
either old needles, staminate flowers if present, or unopened vegetative buds.
Most larvae mine one-year-old needles prior to vegetative budbreak (McGugan
1954, Blais 1979). When vegetative buds begin to swell and flush, budworm
larvae move to this new foliage where they generally remain until they complete
development. However, at high population densities some "backfeeding" on old
foliage occurs.

Observations from several field and laboratory studies indicate that
asynchrony of budworm phenology and host tree phenology results in reduced
performance (poorer survival, growth and fecundity) of the insect (Greenbank
1956, Eidt and Little 1970, Eidt and Cameron 1971, Mattson et al. 1983, Thomas
1983, Blake and Wagner 1986, Thomas 1987). Negative effects of asynchrony on
budworm performance have been observed both when insect phenology is
advanced and when it is retarded in relation to tree phenology.

When budworm phenology is advanced in relation to tree phenology (larval
emergence precedes budbreak by several days), the young larvae are not able to
penetrate the tight vegetative buds and must then feed on mature foliage for an
extended period. Larval mortality increases under these conditions (Blais 1957,
Eidt and Cameron 1971, Shepard 1985) and is presumably caused by either
deleterious effects of a less suitable diet or by losses due to increased dispersal of
larvae as they search for acceptable feeding sites.

When budworm phenology is retarded in relation to tree phenology (larval
emergence follows budbreak by several days), reduced performance by the
budworm is reflected by several measures (lower survival, lower body weight,
longer development time, lower relative growth rate, lower efficiency of
conversion of ingested food and lower fecundity) (Greenbank 1956, Mattson et al.
1983, Blake and Wagner 1986, Thomas 1987). Older foliage appears to be less

7
suitable food for budworms, but the precise mechanisms of these effects are not
known.

These patterns of insect performance indicate that a phenological window
of susceptibility to the spruce budworm exists in its host trees. The timing or
phenological limits of this window have not been determined, nor have the causes
of reduced budworm performance been clearly identified. A variety of test
conditions were used in previous studies of spruce budworm phenological
synchrony. In some cases only late-instar larvae and/ or only one or two levels of
asynchrony ‘were used. Excised foliage or potted seedlings were used in
laboratory studies. In some cases, analyses of the phenological variation in plant
traits were minimal or non-existent. No study has examined the question
comprehensively using complete larval development, several levels of asynchrony,
non-juvenile trees under field conditions, and measurements of seasonal variation

in several plant traits.

WHITE SPRUCE BIOLOGY

White spruce, P. glauca, has a transcontinental distribution which extends
from Alaska to Newfoundland and northward to the treeline in northern Canada.
Its range extends southward into the United States from Minnesota to Maine with
isolated populations in Montana, South Dakota and Wyoming (Fowells 1965).
Western portions of the range of white spruce overlap with the ranges of
Engelmann spruce (P. engebnannr’i Parry) and Sitka spruce (P. sitchensis ( Bong.)
Carr.). White spruce hybridizes with these two species in broad zones of
introgression (Roche 1969, Nienstaedt and Teich 1972).

' Genetic variation among white spruce populations has been studied for many
morphological and physiological traits (Wilkinson et al. 1971, Nienstaedt and
Teich 1972, Canavera and DiGennaro 1979, Nienstaedt 1985, Khalil 1985, 1986).
The results of several studies suggest that white spruce consists of two or three
east-west clines (Wilkinson et al. 1971, Nienstaedt and Teich 1972, Khalil 1985).
Seed sources in the Ottawa Valley along the Ontario-Quebec border have
consistently produced the fastest growing white spruce (Nienstaedt 1969,
Wilkinson 1977b, Wright et al. 1977, Genys and Nienstaedt 1979, Khalil 1985).

Vegetative growth phenology of white spruce, specifically the timing of bud
flushing, is under strong genetic control (N ienstaedt and King 1970, Wilkinson
1977a, Nienstaedt 1985). The variation in timing of bud flushing among
individuals in white spruce provenance and progeny tests has ranged from a span

of 12 days to a span of 35 days (Wilkinson 1977a, Blum 1988).

OBJECTIVES

The objectives of this study were to: 1) test the hypothesis that a
phenological window of susceptibility to spruce budworm exists in white spruce, 2)
determine the exact timing and duration of such a window of susceptibility, 3)
describe phenological changes in key host nutritional factors that are related to
host susceptibility, 4) determine how these changes in plant traits are related to
the performance of the spruce budworm and 5) evaluate the effect of early and
late-flushing seed sources on host susceptibility. In addition, there were other
more specific questions addressed in relation to the above objectives. I
hypothesized that (a) early and late-flushing seed sources will differ only in their
shift in timing of phenological events, (b) insect-host asynchrony affects the
budworm survivorship curve primarily by causing a rise in early larval mortality
and (c) phenological changes in plant nutritional factors have different effects on

younger larvae and older larvae.

MATERIALS AND METHODS

A major experiment was conducted in 1985 to test the phenological
window of host susceptibility hypothesis for the spruce budworm and to determine
the effects of insect-host asynchrony on several measures of budworm
performance. A similar experiment in 1987 examined the effects of insect-host
asynchrony on the budworm survivorship curve. Seasonal variation in foliar traits
were measured in 1985 and 1986, which allowed comparison with insect
performance data from 1985 to evaluate the relationship between host
susceptibility and host phenology.

Experiments were conducted in a white spruce provenance plantation located
in Wexford Co., Michigan in the Huron-Manistee National Forest. The trees
were 27 years old in 1985, and were part of the White Spruce Seed Source
Variation Study G-113, which was established by the USDA Lake States Forest
Experiment Station (North Central Forest Experiment Station) in 1962.

Phenological Window of Susceptibility

Phenological variation in host susceptibility was evaluated by measuring
spruce budworm performance at varying levels of synchrony between budworm
and host tree phenology. Synchrony was manipulated by using host trees from
three seed sources that differed in their timing of bud flushing and by introducing
eight cohorts of spruce budworm to the trees at different phenological stages of

the hosts.

10

11

A factorial experiment was used where the main effects were white spruce
seed sources and spruce budworm cohorts. Information on the relative timing of
budbreak among seed sources was obtained from the Dept. of Forestry, Michigan
State University (unpublished data). Seed sources 1665 from Stony Rapids,
Saskatchewan (SASK) and 1677 from Fort McLeod, British Columbia (BC) were
selected as the early and late-flushing seed sources, respectively. Seed source
1663 from Beachburg, Ontario (ONT) is an intermediate to late-flushing seed
source that also was selected because of its unusually good growth characteristics.
It has consistently been one of the fastest growing seed sources among several
plantations in the northeastern US. and Canada (Nienstaedt 1969, Wilkinson
1977b, Wright et al. 1977, Genys and Nienstaedt 1979). Eight trees were
randomly selected as study trees from among the more open-grown trees in each

seed source.

Spruce Budworm Cohorts

Spruce budworm larvae used in this experiment were obtained from a colony
maintained by the Insect Rearing Service of the Forest Pest Management
Institute, Forestry Canada in Sault Ste. Marie, Ontario. Larvae were received
from the stock colony as overwintering 2nd instars enclosed in their hibemacula
on gauze strips. Larvae were stored at ca. 5 °C until used in the field. Eight
cohorts of 2nd-stage larvae were caged on the trees during the spring and summer
of 1985. Each cohort began at a different time in the host tree’s phenology. A
gauze patch containing approximately 30 larvae was enclosed in a small-mesh
nylon sleeve cage (70 cm length) on one mid-crown branch on each tree for each
cohort. Therefore, 24 branches (3 seed sources x 8 trees/ seed source) were caged
for each cohort. Cages were located on all sides of each tree to prevent excessive

destruction of foliage on one side when branches were later removed from the

12
tree. Cage positions were uniformly distributed around the tree circumference
within each seed source and budworm cohort.

I started the first cohort of budworm larvae at approximately the same
time as spring emergence of natural, field populations of budworm 2nd-instars.
After reviewing the literature (Henson 1948, Bean 1961, Cameron et al. 1968,
Miller et al. 1971, Shaw and Little 1973, Thomas 1976), I estimated peak
emergence in the field to occur at 50 to 100 degree days based on a threshold of
5.6 °C and a start date of 1 March. The first budworm cohort was placed on the
trees on 23 April at ca. 173 accumulated degree days (about four calendar days
after the estimated peak of emergence of natural budworm populations). The
next five cohorts were started at weekly intervals through 28 May (Table 1). The
seventh and eighth cohorts were started four and eight weeks later on 25 June
and 23 July, respectively.

Each cohort remained en the trees for six to seven weeks until most of the
budworms were in the pupal stage. At that time all caged branches of that cohort
were cut from the trees and examined in the laboratory. The first cohort was
removed from the trees on 5 June, and the last cohort was removed on 10
September. All budworms were collected from the caged branches and held in
individual plastic vials at ca. 23 °C until adult emergence. Adults were killed by
freezing within 24 hours after emergence, and then dried in an oven at 60 °C for
48 hours prior to being sexed and weighed. Budworms that were still in the final
larval stage when collected from the cut branches were provided with current-year
foliage from the seed source on which they had been caged, and allowed to
complete development in petri dishes. Larvae that did not pupate within four

days after the cohort was removed from the trees, were not included in analyses

of body weight.

13

Table 1. Dates and accumulated degree days when each spruce budworm cohort
was placed on white spruce trees m 1985.

 

 

Cohort IDES: Gre pgan 323221113331
1 113 23 April 173
2 120 30 April 234
3 127 7 May 296
4 134 14 May 406
5 141 21 May 471
6 148 28 May 556
7 176 25 June 900
‘8 204 23 July 1373

 

1 Base = 5.6 °C; Start date = 1 March

14

To provide fresh, vigorous insects for all cohorts, larvae were received in
three batches from the Canadian stock colony and stored (5 °C) no more than five
weeks prior to use in the field. Cohorts 1 to 3 were taken from the first batch
(April), cohorts 4 to 6 from the second batch (May), and cohorts 7 and 8 from the
third batch (June).

The variability of budworm vigor among batches was evaluated by
examining the performance of budworms reared on artificial diet (McMorran
1965) as part of other feeding studies in 1985. These insects came from batches
received from the stock colony in January, March, June and November. They
were used in control treatments in laboratory feeding studies and maintained at

22 °C and a photoperiod of 16:8 hr (lightzdark).

Tree Phenology

The phenological stage of buds and percent shoot elongation on study trees
were monitored at weekly intervals until growth ceased. Bud phenological stage
was rated and shoot elongation was measured on two randomly selected terminal
sheets on both the north side and south side at mid-crown level of each tree. The
phenological stage of the terminal bud on each shoot was rated using a 6-point
system following Nienstaedt and King (1970):

W Rating
Bud in winter condition

Bud beginning to swell

tutor—s

Bud globose; swelling prominently

Bud green; bud scales expanding and
thinmng but still intact 4

Bud scales separating and exposing
green needles 5

Shoot elongating 6

15
The first measurement of shoot elongation was made in April prior to noticeable
swelling of most buds and consisted of total bud length. Elongation
measurements included bud expansion prior to budbreak and elongation of the
shoot after budbreak.
Ambient temperature in the plantation was measured using a recording
hygrothermograph. Cumulative degree days (threshold = 5.6 °C) from 1 March

were computed using a sine-wave method (Allen 1976).

Spruce Budworm Survivorship Curve

The effect of host phenology on the survivorship curve of the spruce
budworm was examined in 1987 in the same white spruce plantation. An ”early"
cohort and a "late" cohort of budworms were placed on 10 trees of the Beachburg,
Ontario seed source (#1663) using techniques similar to those used in the 1985
study. However, in 1987, budworm survival was evaluated at several points during
the developmental period of each cohort, rather than only at pupation.

Budworms were caged on five mid-crown branches of each tree for each cohort.
One caged branch was removed from each tree at four 7 to 8-day intervals during
budworm development. The final caged branch on each tree was removed when
the majority of budworms were in the pupal stage. Branches were examined in
the laboratory under a lighted magnifying lens, and budworm survival was
determined for each sample date. The average developmental stage of budworm
larvae was determined for each sample date by preserving all collected larvae in
70% ethanol and measuring head capsule widths using an ocular micrometer in a
dissecting microscope (McGugan 1954, Bean and Batzer 1957, Crummey and
Otvos 1980).

16
The first cohort in the 1987 experiment was placed on the trees on 10
April at ca. 133 accumulated degree days, slightly earlier than the first budworm
cohort in the 1985 experiment. The second cohort was placed on the trees on 1
June at ca. 760 accumulated degree days, slightly later than the sixth cohort in
1985. Both budworm cohorts in 1987 were taken from the same stock colony
batch.

Seasonal Variation in Host Nutritional Traits

Several periodic measurements were made of the phenological condition of
the host trees during 1985 and 1986. Foliage samples were collected in 1985 at 1-
3 week intervals for analyses of sugar, nitrogen and mineral content. The first
collection was made on 1 May, immediately prior to budbreak on the earliest-
flushing trees (SASK seed source). Samples were collected on that date from
both the previous year’s growth ("old" foliage) and from expanding vegetative
buds. All subsequent samples were collected from current-year foliage only.
Each sample of shoots consisted of ca. 5 g (fresh wt.) of bulked foliage from one
tree. In most cases each sample of buds consisted of less than 5 g. Samples were
sealed in plastic bags, transported to the laboratory in ice-chilled containers, and
stored at -10 °C. Foliage samples were next oven-dried at 60 °C for 48 hours.
Foliage (exclusive of twigs and bud scales) was then ground to a fine powder in a
Wiley mill, and stored in dry conditions at ca. 22 °C until further processing.

Samples were analyzed for soluble sugar content with a high pressure

liquid chromatography (HPLC) technique that extracted 5-, 6- and 12-carbon
sugars (Haack et al. 1991). Each sample was prepared by placing 200 mg of
powdered foliage and 10 ml of 83:17 acetonitrile/water in a centrifuge tube and
shaking the tube for 24 hours in darkness. The sample was then centrifuged

17
briefly and the supernatant passed through a Sep-pak C18 cartridge (Waters
Assoc.) to remove particulate matter. Samples were stored in glass vials.
Analysis was accomplished by injecting 75 ul of sample onto a Waters HPLC
system with a Waters carbohydrate analysis column and differential refractometer
and using a mobile phase of 83:17 acetonitrile/water and a flow rate of 2 ml/min.
A Hewlett Packard integrator performed data integration. Peaks were identified
by comparison with retention times of known standards and quantified by
comparison with a calibration curve.

Total nitrogen and total phosphorous were determined with standard
micro-Kjeldahl techniques. A 0.15 g sample of powdered foliage was digested
with 8 ml concentrated sulfuric acid and a catalyst mixture (1.5 g K2804 and 175
mg HgO) for one hour at 385 °C. Sample volume was then increased to 75 ml
with DI (distilled/ deionized) water. Sub-samples of 3 ml were stored in sealed
plastic vials and then analyzed colorimetrically with a Technicon Auto-analyzer II.

Totals for elemental minerals (K, Ca, Mg, Mn, Cu, Fe, Na, Zn) were
obtained with plasma emission spectroscopy. A 0.40 g sample of powdered foliage
was digested with 5 ml of concentrated HNO3 for one hour at 155 °C. Then 2 ml
of 70% HClO4 was added and digestion continued for two hours at 220 °C. One
ml of 3% LiCl was added to the sample for ionization suppression, sample volume
was increased to 15 ml with DI water, and the sample was stored in polypropylene
vials. Samples were analyzed on a Beckman Spectrometrics Spectraspan Model
III-A Spectrometer.

Accuracy of total nitrogen and phosphorous analyses and elemental
minerals analyses was checked using National Bureau of Standards standard
reference materials; agreement was within 5% of certified values. Precision was
checked by conducting two separate digestions for ca. 2% of samples; average

relative deviations were less than 5%.

18

Measurements of the phenological variation in leaf water content and leaf
toughness were made in 1986 on the same trees studied in 1985. Foliage samples
were collected at one to two-week intervals during April-May and at three to
four-week intervals during June-August. Tree phenological stage (bud stage and
shoot elongation) and ambient temperatures in the plantation were measured in
1986 in the same manner as in 1985.

Foliage samples for analysis of water content were collected on each
sample date at about the same time of day (1100-1500 hrs). Samples were taken
from the previous year’s growth prior to budbreak and from current-year foliage
after budbreak. Each sample consisted of one sheet collected from the mid-
crown level on the south side of each tree. The needles were immediately cut
from each sheet, placed in tightly sealed, pre-weighed plastic vials and transported
in cool, dark conditions to the laboratory. No samples were collected at times
when moisture (dew or precipitation) was visible on the foliage. Fresh weight of
the foliage was determined within 12 hours after collection, and dry weight was
determined after 48 hours at 60 °C.

Samples for analysis of leaf toughness were collected at the same time of day
and at the same general location on the tree as the water content samples. One
shoot was collected per tree, placed in a tightly capped vial, and transported in
cool, dark conditions to the laboratory. Samples were stored at ca. 5 °C until
further processing. Ten needles were selected from each sheet for measurement
of toughness. Toughness was determined by measuring the grams of force
required to break a needle using the type of device developed by Feeny (1970).
This device uses a cylindrical aluminum shaft (2 mm diam.) to puncture the leaf
as force is added to the opposite end of the shaft. A Chatillon fruit puncture

gauge (rather than a beaker of sand as in F eeny’s technique) was used to apply

19
pressure to the end of the shaft and to measure the force at which the needle

broke.

Statistical Methods

Insect Performance Variables

Spruce budworm performance was measured by computing percent larval
survival (2nd instar to pupation) and percent pupal survival (pupation to adult
emergence) for individual caged branches, and length of development period
(days and degree days), adult dry weight (mg) and growth rate (mg/degree day)
for individual insects. To meet the assumption of normality, arcsine
transformation was performed on larval survival data prior to data analysis.

Neuter values (sex effect standardized) were computed for budworm
weight, length of development and growth rate by converting values for individual
males to female-equivalent values. For example, the value of an individual male
weight was converted to its neuter value by multiplying the male’s weight by the
ratio of mean female weight to mean male weight. Neuter values were used in all
analyses of weight, length of development and growth rate. This procedure
eliminated much redundancy in the data analyses and provided more adequate
sample sizes in those treatments with low budworm survival. The patterns of
budworm performance were similar for males and females across seed sources
and cohorts, although females generally had greater body weight and longer

development times.

Analysis of Repeated Measurements
The multivariate analysis technique of repeated measures analysis was used

to evaluate the effects of cohort and seed source on budworm performance data

20
and the effects of sample date and seed source on host foliage data. When
repeated measurements (performance of insect cohorts and analysis of foliar
constituents) are made on the same experimental unit (a tree) over time, the
measurements tend to be correlated with each other. Use of an analysis
technique that ignores this non-independence can result in a large Type I error
(LaTour and Miniard 1983). Repeated measures analysis can be accomplished
with either a univariate or a multivariate approach (Danford et al. 1960, Cole and
Grizzle 1966, Winer 1971, Davidson 1972, Morrison 1976, LaTour and Miniard
1983).

Repeated measures analysis of data in this study was done using the
multivariate approach provided by the GLM procedure of the SAS statistical
computer system (SAS Institute, Inc. 1988). Individual tree values were used in
all repeated measures analyses. For those variables having multiple observations
per tree at each repeated measure (budworm weight, development time and
growth rate) a mean value was computed for each tree. Seed source was the
between-subjects effect. Budworm cohort or foliage sample date was the within-
subject effect. The SAS procedure for repeated measures analysis provides
multivariate test statistics such as Wilks’ lambda criterion (Morrison 1976) for the
repeated measure effect (cohort or sample date) and for the repeated measure x
seed source interaction. F-values are computed from Wilks’ lambda using Rao’s
(1973) procedure. The test for the between-subjects (seed source) effect is a
univariate analysis of variance of tree sums (values for each tree are summed
across all repeated measures).

Data from all eight budworm cohorts were included in repeated measures
analysis of larval survival, but only cohorts 1 to 4 were included in the analyses of

the other budworm performance variables that require data from surviving adults.

21
Because of greatly reduced larval survival in later cohorts, most trees did not have
budworm adults in one or more of the last four cohorts.

Data from all 1986 samples of current-year foliage were used in repeated
measures analysis of water content and leaf toughness. Only foliage samples
collected from June to September in 1985 were used in repeated measures
analysis of sugars, nitrogen and minerals. During budbreak and early shoot
elongation in 1985 when small quantities of new foliage were available, sampling
intensity was reduced to avoid removing large amounts of potential foliage from
the study trees. Therefore, sugar, nitrogen and mineral data are not available for
some trees, and in some cases one or two whole seed sources, for sample dates

prior to 4 June.

Relationship of Host Susceptibility and Host Phenology

The relationship of host susceptibility to host phenology was examined by
comparing the patterns in budworm performance data (1985) with the seasonal
variation in foliar nutritional traits (1985 and 1986). The value of each foliar trait
on each tree was determined for two points in the larval development of each
budworm cohort: at 25% and 75% of the time required for post-overwintering
larval development based on elapsed degree days during each cohort. Foliar
values were determined at each of these two points for each cohort by
interpolation of the foliar data from the two nearest foliar sample dates. The
25% and 75 % points are roughly equivalent to the start of the 3rd larval stage
and early 6th larval stage, respectively, based on the temperature and budworm
development data presented by Regniere (1987). The 25 % point therefore
represents the time in development when most larvae in natural populations are
moving from one-year-old foliage to expanding buds or new shoots. It also

represents a point in post—overwintering development near which most mortality

22
occurs. The 75% point in development occurs near the start of the period of the
greatest food consumption and weight gain.

Foliar values for early development (25 % point) and late development
(75% point) were combined with budworm performance data from each cohort to
form an early development data set and a late development data set. Each data
set consisted of 192 observations (8 cohorts X 24 trees). Relationships between
budworm performance and foliar nutritional traits during early and late
development were evaluated using canonical correlation analysis and regression
techniques.

Canonical correlation analysis is a multivariate technique useful in
analyzing the relationships between two sets of variables (insect performance
variables and host quality variables) in a single procedure. Canonical correlation
produces linear combinations (canonical variates) of the original variables in each
data set that maximize the correlation between the two data sets. Interpretation
of the results is done by examining the "loadings", i.e. the correlations of the
original variables with the new canonical variates (Smith 1981, Sullivan and
Moncreiff 1988). The relationships of budworm performance and host foliar
quality were evaluated using the CAN CORR procedure of the SAS system (SAS
Institute, Inc. 1988) in separate analyses for early and late budworm development.
Seed sources and budworm cohorts were pooled in each analysis. Observations
with missing values for any variable were not included in the analysis procedure.
Missing values occurred in some cases for budworm performance variables due to
low larval survival in cohorts 5 to 8 or for foliar variables due to small foliage
samples during budbreak and early shoot elongation. Therefore, 96 observations
from cohorts 3 to 8 were used in the early development analysis, and 152
observations from all eight cohorts were used in the late development analysis.

Canonical scores generated in the canonical correlation analysis were used in an

23
ordination procedure (Smith 1981) to map the insect performance-host quality

relationships.

RESULTS
Phenological Window of Susceptibility

Relative Timing of Tree Phenology and Budworm Phenology

The relative timing of phenological events observed among the three white
spruce seed sources in 1985 was consistent with the patterns observed in historical
records for the white spruce provenance test. Bud expansion and budbreak
occurred first in the SASK seed source and followed about one to two weeks later
in the ONT and BC seed sources (Table 2). Budbreak had begun on some trees
in the SASK seed source by Julian date (JD) 120 (equivalent to 234 elapsed
degree days) and was nearly complete by JD 128 (309 degree days). Budbreak in
the ONT seed source occurred during the following week (JD 128-135 and 309-
421 degree days). Timing of budbreak in the BC seed source was similar to that
of the ONT seed source, although 1/4 of the buds measured on the BC seed
source did not break until after JD 135. These results are comparable to the
timing of budbreak in early (228 degree days) and late-flushing clones (322 degree
days) examined by Nienstaedt and King (1970).

Percent shoot elongation in the SASK seed source was one to two weeks
ahead of that in the ONT and BC seed sources (Figure 1). The timing of shoot
growth in the ONT and BC trees was similar, except that the ONT trees appeared
to have a slightly longer growth period. By JD 135, shoot elongation was greater
than 50% on SASK trees, but only 25% and 20% on ONT and BC trees,

24

25

 

 

 

 

Table 2. Bud phenom?! rating and 1piercent of buds burst for three white spruce
seed sources uring April- ay 1985 (N = 32 buds per seed source).
Mean Bud Phenology Percent of
Rating (1 SE) Buds Burst
Julian Accumulated
Date Degree Days SASK ONT BC SASK ONT BC
114 182 2.2 1.0 1.0 0.0 0.0 0.0
(10.1) (10.0) (10.0)
120 234 4.3 3.0 1.9 34.4 0.0 0.0
(10.1) (10.0) (10.1)
128 309 5.8 3.8 3.1 96.9 6.3 0.0
(10.1) (10.1) (10.1)
135 421 6.0 5.9 5.3 100.0 96.9 75.0
(10.0) (10.1) (10.2)
142 481 6.0 5.9 6.0 100.0 96.9 100.0
(10.0) (10.1) (10.0)
149 570 100.0 100.0 100.0

 

26

 

 

 

 

8 ° /
0
d 60 - /
5 o
‘63
m 40 -
|.r.l
D.
./ 0-0....
3 20 _ o A—AONT
2 / / El—EIBC
0 Mr 1 r r r
110 120 13 140 150 160 170 180
JULIAN DATE
APRIL 1 MAY : JUNE

 

 

 

Figure 1. Shoot elongation of three white spruce seed sources in 1985.

27
respectively. Shoot elongation was completed on SASK, ONT and BC trees by
JD 162, JD 176 and JD 169, respectively.

The timing of the placement of the eight spruce budworm cohorts on trees
ranged from three weeks prior to budbreak to about one month after the
completion of shoot growth (Figure 2). Cohort 1 was placed on the trees one to
two weeks prior to budbreak in the early-flushing trees. Cohorts 2 and 3 started
during the budbreak period. Cohort 4 started when new shoots were 50%
elongated on the early-flushing trees, and buds were flushing on the late-flushing
trees. Cohorts 5 and 6 started later in the shoot growth period, and cohorts 7 and
8 followed still later after the cessation of shoot growth. Only cohorts 1, 2 and 3

completed larval development prior to the end of the shoot growth period.

Spruce Budworm Performance

WW3; - Performance of the eight
spruce budworm cohorts was measured by larval survival, pupal survival, adult
weight, length of development period and growth rate. Larval survival averaged
greater than 60% through the first three cohorts, but then substantially decreased
to less than 35% in cohort 4 (Figure 3A). Larval survival further decreased to
less than 10% in each of the last three cohorts. On some trees there were no
survivors in the later cohorts, although this was true for only two of the 24 trees in
the last three cohorts. At least some spruce budworms in each cohort were able
to survive to pupation on each of the seed sources.

Pupal survival was generally high (> 75%) for all cohorts, but was highest
(> 85%) for the later cohorts (Figure 3B). The increase in pupal survival later in
the season may be partly due to the reduced sample size which resulted from low
larval survival. However, it may reflect a reduced effect by natural enemies later

in the season. Remarkably, parasitic ichneumonid wasps were able to sting and

28

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

APRIL MAY JUNE JULY AUGUST SEPT
L . ‘5 I TIMING OF
1 2 I BUDWORM COHORTS
. [ 3 I.
1 4 . 1
1 s; 1
I 6 I

:B l 7 1

5 ud; Shoot :

gBreakg Elongation I 3

 

 

 

 

1 3 1 1 3 1 1 1 1
100 120 140 160 180 200 220 240 260
JULIAN DATE

Figure 2. Relative timing of eight spruce budworm cohorts and white spruce
budbreak and shoot elongation.

 

 

 

 

 

 

§§a
§7
§\\\\\\\\\\\\\\\\\\\\\\&6
.\\\\\\\\\\\\\\\\\\\\\\\\\.5
\\\\\\\\\\\\\\\\\\\\\\\\k3
\\\\\\\\\\\\\\\\\\\\\\\r

R
P

\\\\\\\\\\\\\\\\\\\\\2..
N\\\\\\\\\\\\\\\\\\\.

BDWORM COHORTS

so ._<>_>m:m .52

m.

survival (15E) of (A) larvae and (B) pupae in ei

spruce budworm cohorts.

Figure 3. Mean percent

30
parasitize budworms by inserting their ovipositors through the cloth surface of the
branch cages. Parasitoids accounted for an average of 36%, 64% and 31% of
pupal mortality in cohorts 1, 2 and 3, respectively. In cohorts 4 through 8,
parasitoids emerged from only two budworm pupae.

The length of the development period from 2nd-stage larva to adult was
measured both in number of calendar days and in degree days (Figure 4).
Because development is temperature-dependent, development time measured in
days is confounded by the effects of increasing temperatures as the growing
season progressed. Degree days provide a unit of developmental time
standardized across changing temperature regimes, and therefore allows more
direct comparison of the effects of cohort timing on development. The number of
days required for development was shallowly sigmoidal initially decreasing from
49.8 days (cohort 1) to 47.5 days (cohort 3) and then increasing to 55.2 days in
cohort 6 (Figure 4). The number of degree days required for budworm
development also was slightly sigmoidal, but generally the number of degree days
increased continuously from 614 degree days in cohort 2 to 972 degree days in
cohort 7 (Figure 4).

The pattern of budworm adult weights among cohorts was similar to the
pattern of larval survival. Weights steadily decreased from an average of 26.0 mg
in cohort 1 to 12.4 mg in cohort 8 (Figure 5A) with the largest decrease in adult
weights occurring between cohorts 3 and 4, dropping from 23.7 mg to 16.7 mg,
respectively.

Budworm growth rates were calculated as mg/ degree day. The more
common unit of measure, mg/day, can not be used, because of the confounding
effect of increasing temperatures during the spring and summer. Budworm
growth rates continually decreased from 0.042 mg/ degree day in cohort 1 to 0.013
mg/degree day in cohort 8 (Figure 5B).

31

 

(A) 60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

55 - \~
9
‘o‘

\\ _,
2 3 4 5 6 7 8
(B) 1000

900 -
9
g 800 -
LLI
33
‘13 700 -
o .

600 -

F
500 -
1 2 3

 

 

 

4 5 6 7 a
SPRUCE BUDWORM con-Ioars

Figure 4. Mean length of development period (1SE) from second instar to adult

mplasured in (A) days and (B) degree days for eight spruce budworm
co orts.

 

 

 

 

 

4 5 6 7 8

 

 

 

 

fist

BUDWORM COHORTS

 

.\

.\\\\\\\\\\\\5
\\\\\\\\\\\\\\\..J
\\\\\\\\\\\\\\\\\\\\\.J
\\\\\\\\\\\\\\\\\\\\\\\\.a
§1

5 4 3 2 4| 0
0. O. 0. 0. 0. O.
o O o o O o

3.5 eunuch—\oEv NEE Igomo

S
6

 

 

E

SPRUC

Figure 5. (A) Mean adult dry weight (1SE) and (B) mean growth rate (1SE)
or eight spruce budworm cohorts.

33

A less direct measure of budworm performance in this study is the
variation in defoliation levels. As would be expected with the declining larval
survival, percent defoliation of current-year needles dropped sharply after the
third cohort (Figure 6). Defoliation was generally highest on SASK trees and
lowest on ONT trees. The effect of seed source on defoliation may be partially
due to variation in timing of budbreak, but variation in shoot length and the
quantity of foliage available are also important factors. The average shoot lengths
(based on four mid-crown shoots/tree in 1985 and four mid-crown shoots/ tree in
1986) of the SASK, ONT and BC trees were 11.9 cm, 16.9 cm and 13.4 cm,
respectively. A significant (P < 0.001) negative relationship (Y = 118.11 - 0.46 X,
r2 = 0.34) exists between percent defoliation (Y) and average shoot length per
tree (X) for those early budworm cohorts (1, 2 and 3) that had high larval
survival.

The effects of cohort timing and seed source on budworm performance
were analyzed using repeated measures analysis. Cohort timing was a significant
factor (P < 0.001) affecting all performance variables (Table 3). Seed source was
a significant (P 5 0.05) factor for all performance variables except pupal survival.
The interaction between cohort and seed source was not significant (P > 0.05) for
the survival variables, but was significant for length of development, adult weight
and growth rate. Budworm performance within each cohort was generally lower
(lower larval survival, adult weight and growth rate and longer development
period) on the SASK seed source than on the other seed sources (Figures 7A, 8A,
9A and 10A). The same trends are evident when mean tree sums (sums of cohort
values for each tree) are compared among seed sources (Table 4).

WWW - One possible explanation for the
reduced budworm performance on the SASK seed source is the more advanced

phenology of these trees. As a test of this hypothesis, repeated measures analyses

34

90

 

-SASK
-ONT
CZIBC

DEFOLIATION (x)

 

 

 

 

 

 

 

 

 

.1 - W

2 3 4 5 6 7 8
SPRUCE BUDWORM COHORTS

 

Figure 6. Mean percent defoliation (18E) of current-year shoots for eight
spruce budworm cohorts on three seed sources of white spruce.

35

 

 

 

Table 3. Multivariate repeated measures analysis of spruce budworm
performance in 1985 with budworm cohorts as repeated measures.
Source of Wilks’ F
Dependent Variable1 Variation lambda2 Value Prob.
.Cnhchfi:
Larval Survival (%) Cohort 0.0064 334.76 <0.001
Cohort x Seed Source 0.2876 1.85 0.077
Seed Source 5.64 0.011
mm
Larval Survival (%) Cohort 0.1458 37.10 <0.001
Cohort x Seed Source 0.7258 1.10 0.380
Seed Source 5.39 0.013
Pupal Survival (%) Cohort 0.3441 12.07 <0.001
Cohort x Seed Source 0.8056 0.73 0.634
Seed Source 1.27 0.300
Length of Development Cohort 0.1377 39.65 <0.001
( ays) Cohort x Seed Source 0.2808 5.62 <0.001
Seed Source 7.27 0.004
Length of Development Cohort 0.3053 14.41 <0.001
(degree days) Cohort x Seed Source 0.3543 4.31 0.002
Seed Source 6.72 0.006
Adult Dry Weight (mg) Cohort 0.0659 89.84 <0.001
Cohort x Seed Source 0.4679 2.93 0.019
Seed Source 17.25 <0.001
Growth Rate Cohort 0.0629 94.36 <0.001
(mg/ degree day) Cohort x Seed Source 0.3467 4.42 0.002
Seed Source 19.23 <0.001

 

1 Arcsine transformation performed on all survival data prior to analysis.

2 Wilks’ lambda criterion (Morrison 1976) used to com ute F-values for cohort

effect and cohort x seed source interaction using Rao 5 (1973) procedure.

(A)

LARVAL SURVIVAL (x)

(3)

Figure 7.

LARVAL SURVIVAL (x)

36

 

 

 

 

 

 

 

 

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

eo
-SASK
”E”
a
l I J.
E J-
1 _ 2 - 3 _ 4
SPRUCE BUDWORM COHORTS
eo
-SASK
3° ' -ONT
7o . \. I 1 C380
i l
60 - s
\
50 - §
R
40 - s I
\
30 - S
20 - 3
§
10 - §
0 ‘ - - -
-3 -2 -1 0 +1

TIMING 0F START OF SPRUCE BUDWORM COHORTS
(Number of Week: From Budbreak)

Mean percent larval survival (15E) for (A) spruce budworm cohorts
1-4 on three white spruce seed sources and for (B) spruce budworm
cohorts from phenologically similar host trees. Cohorts in (B) from
weeks -3 and +1 were not used in re eated measures analyses but are
presented for illustrative purposes 0 y.

(A)

DEGREE DAYS

(3)

DEGREE DAYS

Figure 8.

37

 

 

 

 

 

 

 

 

 

 

 

 

 

725

-SASK .
700 . -ONT

[ZIBC
675 -
650 - L

J.
525 . I .r.
600 ~
575 -
550 — — _
1 2 3 4

SPRUCE BUDWORM COHORTS

775 *
-SASK I

75° ' -ONT

725 - [:30

 

700 -
675 -
650 - F
625 - I z
600 -

575 -

5o .. _ 1. _ L.—
5 -3 -2 -1 0 +1

TIMING OF START OF SPRUCE BUDWORM COHORTS
(Number of Weeks From Budbreak)

 

 

 

 

 

 

 

 

 

 

 

 

Mean length of development in degree days (18E) for (A) spruce
budworm cohorts 1-4 on three white spruce seed sources an for (B)
s ruce budworm cohorts from phenologically similar host trees.
ohorts in (B) from weeks -3 and +1 were not used in repeated
measures analyses but are presented for illustrative purposes only.

(A)

ADULT DRY WEIGHT (mg)

(3)

ADULT DRY WEIGHT (mg)

Figure 9.

 

35
- SASK
30 - . ONT
:3 BC

 

 

 

 

 

 

 

 

 

 

 

 

 

-‘ /////////////////////////I

_ ii

2 3
SPRUCE BUDWORM COHORTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

35
-SASK
30 - -ONT
Dec
25 " I 1: F:
20 -
,a
15 - E
10 1-
5 ..
, _ .1 L.
-3 -2 -1 0 +1

TIMING OF START OF SPRUCE BUDWORM COHORTS
(Number of Weeks From Budbreak)

Mean adult dry weight (18E) for (A) spruce budworm cohorts 1-4 on
three white spruce seed sources and for (B) spruce budworm cohorts
from phenologically similar host trees. Cohorts in (B) from weeks -3
and +1 were not used in repeated measures analyses but are
presented for illustrative purposes only.

 

 

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(A)
A -SASK
g 0.05 - § 0NT
s . :160
§ \ = N
0.04- N .2. N :r.
g \ \
\ R N
m 0.03 § S
3 S N \‘ f
N s s
E 0.02 g 3 §
S i i
0.01 S \ S
N i N
m N \ s
o 0'00 1 2 _ 3 4
SPRUCE BUDWORM COHORTS
(B)
" _ -SASK
a?" 0'05 . s -ONT
_ 1"
g. 0.04 z r
' \
\ \
g 0.03 - S
V § I
' \
IE 0.02 - S 1‘
E . t
s
S 0.01 - s
\
5 000' S
' -3 -2' -1' 0 +1

TIMING OF START OF SPRUCE BUDWORM COHORTS
(Number of Week: From Budbreak)

Figure 10. Mean growth rates (1SE) for (A) spruce budworm cohorts 1-4 on
three white spruce seed sources and for (B) spruce budworm cohorts
from phenologically similar host trees. Cohorts in (B) from weeks -3
and + 1 were not used in repeated measures analyses but are
presented for illustrative purposes only.

40

Table 4. Seed source effect on spruce budworm erformance in 1985
represented as mean sums of cohort v ues for budworm performance

 

 

 

 

variables.
Mean Sums By Seed Source2

Dependent Variablel SASK ONT BC
W

Larval Survival (%) 223.3 b 294.5 a 272.6 a
W

Larval Survival (%) 200.0 b 254.1 a 230.0 ab

Pupal Survival (%) 345.7 a 331.0 a 350.5 a

Length of Development

(days) 197.5 a 192.2 b 196.4 a

Length of Development

(degree days) 2599.8 a 2483.5 b 2549.5 ab

Adult Dry Weight (mg) 76.34 c 104.06 a 89.26 b

Growth Rate

(mg/degree days) 0.120 c 0.171 a 0.142 b

 

1 Arcsine transformation performed on all survival data prior to analysis.

2 Cohort values are summed for each tree. Mean sum = mean of eight tree
sums per seed source. Means within the same row followed b the same letter
are not significantly different (P > 0.05, Tukey’s Studentized ange Test).

41
were done using only equivalent cohorts from phenologically similar hosts.
Budworm cohorts 1, 2 and 3 on SASK trees were reclassified as equivalent to
cohorts 2, 3 and 4 on the ONT and BC trees (Figures 7B, 8B, 9B and 108). This
approach was used because the phenology of the SASK seed source was estimated
to be about one week or more ahead of that of the ONT and BC seed sources,
and cohorts 1 to 4 were started at one-week intervals. The combination of
cohorts and seed sources used in this analysis represents three levels of insect-host
synchrony - i.e., budworm cohorts that were started at two weeks prior to
budbreak, at one week prior to budbreak and at the same time as budbreak.

The seed source effect was no longer significant (P > 0.05) for larval
survival and length of development (degree days) when phenologically similar
hosts were used in the analyses (Tables 5 and 6). However, the seed source effect
was still significant for adult weight and growth rate.

1'19: Growth Rage -- Budworm performance was compared with tree
growth rate to indicate how insect performance might be related to the relative
vigor or growth characteristics of the host trees. Tree height varied among seed
sources. The average heights at the end of the 1985 growing season were 3.0, 6.1
and 4.2 m for SASK, ONT and BC trees, respectively. Tree height growth rates
averaged 11.0, 22.7 and 15.7 cm/yr, respectively. These averages are consistent
with height data from the 14 plantations in the white spruce provenance study.
Heights for SASK, ONT and BC trees have been approximately 70%, 130% and
80%, respectively, of plantation mean tree heights (Nienstaedt 1969, Wilkinson
1977b, Wright et al. 1977, Genys and Nienstaedt 1979).

Relationships of budworm performance to tree growth rate were analyzed
with regression techniques. Seed sources were pooled and separate analyses were
done on budworm cohorts from phenologically similar host trees, i.e., cohorts

started two weeks prior to budbreak, cohorts started one week prior to budbreak

42

 

 

Table 5. Multivariate repeated measures analysis of spruce budworm
performance in 1985 using cohorts from phenologically similar hosts
(cohorts 1-3 of the SASK seed source and cohorts 2-4 of the ONT
and BC seed sources) as repeated measures.
Wilks’ F
Dependent Variablel Source of Variation lambda2 Value Prob.
Larval Survival (%) Cohort 0.3194 21.31 <0.001
Cohort x Seed Source 0.6643 2.27 0.079
Seed Source 0.97 0.397
Pupal Survival (%) Cohort 0.2789 25.86 <0.001
Cohort x Seed Source 0.9485 0.27 0.897
Seed Source 0.75 0.487
Length of Development Cohort 0.3048 22.80 <0.001
(days) Cohort x Seed Source 0.5867 3.06 0.028
Seed Source 3.47 0.050
Length of Development Cohort 0.2565 28.99 < 0.001
(degree days) Cohort x Seed Source 0.8124 1.09 0.372
Seed Source 1.16 0.332
Adult Dry Weight (mg) Cohort 0.1465 58.26 <0.001
Cohort x Seed Source 0.3885 6.04 < 0.001
Seed Source 5.56 0.012
Growth Rate Cohort 0.0808 113.82 < 0.001
(mg/degree day) Cohort x Seed Source 0.3593 6.68 <0.001
Seed Source 6.03 0.008

 

1 Arcsine transformation performed on all survival data prior to analysis.

2 Wilks’ lambda criterion (Morrison 1976) used to co
effect and cohort x seed source interaction using Rao

mpute F-values for cohort
s (1973) procedure.

43

Table 6. Seed source effect on spruce budworm performance in 1985 using
cohorts from phenologically similar hosts (cohorts 1-3 of SASK seed
source and cohorts 2-4 of the ONT and BC seed sources). Seed source
effect is represented as mean sums of cohort values for budworm
performance variables.

 

Mean Sums by Seed Source2

 

 

Dependent Variable1 SASK ONT BC
Larval Survival (%) 176.0 a 183.2 a 164.2 a
Pupal Survival (%) 249.9 a 256.2 a 266.3 a
Length of Development (days) 146.4 a 143.1 b 145.0 ab
Length of Development (degree days) 1897.4 a 1869.0 a 1909.7 a
Adult Dry Weight (mg) 62.20 b 74.08 a 64.96 ab
Growth Rate (mg/degree days) 0.100 b 0.122 a 0.104 b

 

1 Arcsine transformation performed on all survival data prior to analysis.

2 Values of three cohorts are summed for each tree. Mean sum = mean of eight
tree sums per seed source. Means within the same row followed by the same
letter are not significantly different (P > 0.05, Tukey’s Studentized Range Test).

44
and cohorts started during budbreak. Significant relationships (P 5 0.01) between
budworm performance and tree growth rate exist (Table 7) for those performance
variables that were significantly affected by seed source even when only
phenologically similar cohorts were used in analyses (Table 5). These
relationships were present only in cohorts that started one or two weeks prior to
budbreak. No significant relationships between budworm performance and tree
growth rate were found for budworm cohorts that started at the time of budbreak.
The length of budworm development in days for cohorts starting one or two
weeks prior to budbreak was negatively related to tree growth rate (Figure 11).
Adult weights and growth rates of these same insects were positively related to
the growth rates of the trees (Figures 12 and 13). A weak relationship with tree
growth rate also existed for larval survival of cohorts that started one week prior
to budbreak.

W .. Further analyses were conducted to determine if factors
other than phenological asynchrony and differences in tree phenology and growth
rates may be confounding the observed cohort and seed source effects. Variation
in insect vigor among batches of budworms received from the stock colony and
seasonal variation in temperatures were examined as potential confounding factors
of cohort effects.

Variation in performance among spruce budworm batches was minimal for
insects reared in controlled conditions as part of laboratory feeding studies.
Coefficients of variation (CV) were less than 10.0 for six of the seven
performance variables measured (Tables 8 and 9). Spruce budworm performance
in the field had a much higher level of variation (CV. as high as 103.0) among
budworm batches. This is the expected result for budworm batches placed on
host trees at different points in the host phenology. The differences in

performance values among batches in the field were in most cases 2 to 14 times

45

Table 7. Coefficients of determination (r2) for the relationshi of spruce
budworm performance to tree growth rate (cm/yr) or budworm
cohorts started at two weeks prror to budbreak, one week prior to
budbreak and at budbreak in 1985.

 

Timing of Start of Spruce Budworm Cohort

 

 

Budworm Performance 2 Weeks 1 Week

Prior to Prior to
(Dependent Variable) Budbreak Budbreak At Budbreak
Larval Survival (%) 0.15 n.s. 0.19 " 0.06 n.s.
Pupal Survival (%) <0.01 n.s. 0.03 n.s. <0.01 n.s.
Length of Development
(days) 0.33 " 0.31 " <0.01 n.s.
Length of Development
(degree days) 0.10 n.s. 0.03 n.s. 0.02 n.s.
Adult Dry Weight (mg) 0.42 "" 0.56 "" <0.01 n.s.
Insect Growth Rate
(mg/degree day) 0.49 "‘ 0.49 “" <0.01 n.s.

 

n.s., not si ' cant; ‘, significant at P
"’, si ' cant at P5 0.001; for Y =

(cm/71r .

5 0.05; ”, significant at P 5 0.01;
a +

b with X = tree growth rate

46

35.23 56525
3 new Ace: @263 565 no SE: SE 3.003 28 A25 38 553 SEE SE 8.003 93 nor—Sm $.28

59.6.5 E8 38 526 005 .8: one Eek—DB 88% we $.33 80an—3% we Smoo— ue Ema—Beauty—

:383 w._.<m I._.>>Om6 mmmh

 

 

mN 4N ON or N_. m V

_ . _ . . 3.
\.< ......................................... ul-
4 ........... m
. 00 4 o 4 4 , - 3 m
.x.‘ w 0 G H
, 4 4 % Q 4 % mu.
0 3.4mm x o 4 o 4 - 3. m
0 ............................ O %V O A
no ..... x o o x a
* 0 g ....... 4 m
x O 4 . on w
a. a; m a
x x , .......... m.
.w.z 3.8565 2 x .. mm m
5. u N. x to I mde u > ._o_..n_ x025 _. 4 m
mm." N L x _..OI 0.00 n > ._o_._n_ 9325 N O I

 

 

 

#m

.2 25E

47

v8.8; $.28 .5963 he 3a.. :58» out “no: and Emma? bu :23 .5363 82.. m «o Eamcousom .NH 2399

mN

€363 m._.<m 1.55010 mmm...

3.523
x3525 “a cum 3:: 3:93 555 mpg—want: BE x33 28 A05— 38 £55 86:0 SE. $.33 95

 

 

 

cm on o. a. m ...
q d 4 _ 4
% Wm .V . 1 Q
4 ..
... ... ...-.o. m *8 ...
.......... Q 0 AV 4 1
.. ....................... . w. o . .
u.... 4 ‘ O Q 4 6 «0.... l
.. G 4 ..... 02W ...... O
.. Aw ....... Q
o 4 4 .. HQ
.. O Q fl 1
Q. © .
o 0
av 1
4 .. .. .m.z €9.85 2 ...
8.0 ... a. x ed +2: u > BE v.8; P 4 -
«to ... a. x m.o+ at u > .25 9.82, u o

 

 

NF.

0..

ON

VN

mN

N0

m0

0v

(5“!) iHSIBM ABC! unav

48

Eggs €05.25 3 ea 3:: BE 55? 835:: .23 use; 28 .3: 28 555 8.23 SE

$.33 93 @833. 38:8 88:63 Sm 2a.. 538» no: so: 23 BS 336% .5363 33% mo NEEBEBM .3 22E..—

wN

€353 m._.<m I._.>>Om@ mum...

 

 

«a on 9 3 m
_ a _ . ..- _
.. .W. .4.
x. .1. ... ... .4. 4 my
l. o mum M 04
4 AV Q 4. O Q
.. 4 O 4 6 4....
4 4 ........ O...® O
.O 4 ..O. .......... O
.. 4 O O
.. 4 ..Q
o @O 4....
.o
a; O \\
.m.z xmosnzm 2 Wm
med u m . x 606+ «mod n > .25 .325 P 4
3.0 u N . x 505+ Rod .... > .25 9.32, N o

 

 

No.0

00.9

40.0

00.0

00.0

50.0

(‘90 691590/5‘“) SLVH HlMOHS WHOMCIHG

49

Table 8. Comparison of spruce budworm survival rates among 1985 budworm
batches for insects reared on artificial diet in laboratory studies and
for insects in the white spruce field study.

 

Batches Used in Laboratory Studies1

 

January March June2 November Mean 1 SD C.V.

 

No. Insects 400 400 200 400

W
Larvae 70.5 71.3 - 63.8 68.5 .i 4.2 6.1
Pupae 82.5 78.8 -- 94.9 85.4 _+_ 8.4 9.8
Total 58.2 56.2 69.0 60.5 61.0 _4; 5.6 9.2

 

Batches Used in Field Study3

 

 

April May June
Cohorts Cohorts Cohorts
1-3 4-6 7-8 Mean i SD C.V.
No. Insects 2180 1445 1120
W
Larvae 64.9 18.7 6.3 30.0 i 30.9 103.0
Pupae 82.7 933 91.8 89.2 _+_- 5.7 6.4
Total 53.3 17.6 5.8 25.6 i 24.7 96.5

 

1 Values for laboratory studies are based on one observation per batch.

Datia for larval and pupal survival not available for June batch in laboratory
stu y.

3 Values for field study are averages based on 72, 71 and 47 caged branches,
respectively, for the three batches.

50

Table 9. Comparison of spruce budworm length of development, adult weight
and growth rate among 1985 budworm batches for insects reared on
artificial diet in laboratory studies and for insects in the white spruce

 

 

 

 

 

 

field study. ‘
Batches Used in
Laboratory Studies

Jan. March June Nov. Mean _t SD C.V.
No. Insects 233 218 138 242
Length of Development1
(days) 40.5 35.8 41.5 38.6 39.1 i 2.5 6.5
Length of Development
(degree days) 666.6 588.1 682.6 634.2 642.9 1 41.7 6.5
Adult Dry Weight (mg) 23.20 27.04 23.29 21.62 23.79 i 2.30 9.7
Growth Rate
(mg/degree day) 0.035 0.047 0.035 0.035 0.038 i 0.006 15.6

Batches Used in Field Study

April May June
Cohorts Cohorts Cohorts

1-3 4-6 7-8 Mean 1, SD C.V.
No. Insects 1111 230 64
Length of Development1
(days) 48.5 50.7 53.7 51.0 _4; 2.6 5.1
Length of Development
(degree days) 622.3 698.5 944.2 755.0 1 168.2 22.3
Adult Dry Weight (mg) 24.85 16.40 12.90 18.05 _+_- 6.14 34.0
Growth Rate
(mg/degree day) 0.040 0.024 0.014 0.026 _j; 0.013 50.0

 

1 Batch values for each insect performance variable are averages of individual

insect values.

51

as great as the standard deviation among batches in the laboratory studies. Thus,
differences among budworm batches in field tests were assumed to be due to
causes other than variation among stock colony batches. Pupal survival was the
only performance variable that had very small differences among batches in the
field (differences < 1.5 times the standard deviation in the laboratory). An
interesting point in these comparisons of laboratory and field-reared insects is that
the performance values for the first budworm batch in the field (cohorts 1 to 3 -
i.e., those insects that were most closely synchronized with tree phenology) were
quite similar to the mean performance values for the laboratory-reared insects.

Repeated measures analysis of field temperature data revealed that the
daily average temperature, daily maximum temperature and number of degree
days per day did not differ significantly (P > 0.05) among cohorts 1 to 4 (Table
10). However, when cohorts 5 through 8 were added incrementally to the
analysis, there was in each case a significant difference among cohorts for each of
the three temperature variables. Daily maximum temperatures greater than 30 °C
occurred only once prior to 1 July (JD 182). During the period in which cohort 7
was on the trees (JD 176-226), 17 of 50 days had maximum temperatures greater
than 30 °C. However, no temperatures greater than 33.5 °C were recorded during
the course of this study. Higher temperatures later in the season are not
considered to be a serious confounding factor, because they were present several

weeks after the major changes in budworm performance occurred.
Spruce Budworm Survivorship Curve
The pattern of the spruce budworm survivorship curves is the same for

both early (April-June) and late (June-July) cohorts in the 1987 survivorship

experiment (Figure 14). More than 85% of the total larval mortality occurred by

52

Table 10. Daily average temperatures, dail maximum temperatures and degree
days per day during the period at eight spruce budworm cohorts

were caged on white spruce in 1985.

 

 

 

 

 

 

 

 

 

Daily . Daily
Average Maximum
Tem . Tem . De ee Days
(°C (°C er Day
# Days
Cohort in Field Mean SE. Mean SE. Mean 8.1-3.
1 45 13.3 i 0.58 22.8 i 0.61 11.3 _t 0.55
2 43 13.7 _+_ 0.58 23.3 .i: 0.61 11.6 _+_ 0.54
3 42 14.0 i 0.59 23.4 i 0.64 12.0 _i 0.57
4 43 13.9 i 0.52 23.2 i 0.62 11.8 i 0.51
5 49 15.4 i 0.55 24.8 i 0.59 13.3 i 0.52
6 51 16.4 _4; 0.57 25.6 i 0.60 14.2 _+_- 0.55
7 51 19.3 i 0.40 28.1 _+_; 0.43 17.1 _+_; 038
8 50 19.0 i 0.42 25.5 i 0.53 17.0 i 0.41
F F F
Cohort Effect1 Value Prob. Value Prob. Value Prob.
Cohorts 1-4 0.49 0.688 0.21 0.886 0.51 0.677
Cohorts 1-5 3.81 0.011 2.71 0.045 4.32 0.006
Cohorts 1-8 28.91 < 0.001 13.93 < 0.001 31.88 < 0.001

 

1 Repeated measures analysis using cohorts as repeated measures.

53

 

 

 

 

 

 

‘00 ' 9 o—o APRIL-JUNE '
A—A JUNE-JULY

so - -
5 6° " d
'2 g, .
LIJ
o 1\ 5
E 40 ' <15 1\6 d
D. .L

20 - ~

_ 9\£ -

 

L2 L3 L4- L5 L6 PUPA ADULT
AVERAGE STAGE OF BUDWORM DEVELOPMENT

Figure 14. Spruce budworm survivorship curves (mean percent alive 1: SE for
an early cohort (April-June) and a late cohort (June-July) on e
ONT seed source during 1987.

54
the time most budworms had become 3rd instars, thus demonstrating a "type A"
survivorship curve (sensu Price 1984). These results are comparable to the
survivorship curve developed by Morris and Miller (1954) for field populations of
spruce budworms.

The trends in budworm performance in this 1987 experiment agree with
the results of the 1985 experiment. Larval survival and adult weight were much
higher and the degree days required for development were much fewer in the
early cohort than in the late cohort (Table 11). Larval survival in the early cohort
in 1987 (43%) was, however, about one-third less than larval survival in the first
three cohorts in 1985 (Figure 3A). This may be partly due to the earlier start of
the first cohort in 1987, which caused the young budworm larvae to be on the
trees for one to two weeks longer prior to budbreak than were larvae of the first
cohort in 1985. The early cohort in 1987 was placed on the trees at 40 fewer
accumulated degree days (133 vs 173) and at one to two weeks earlier in terms of
host phenology (four to five weeks vs three weeks prior to budbreak on ONT
trees) than cohort 1 in 1985. However, peak pupation occurred at ca. 85% shoot
elongation for both the early cohort in 1987 and cohort 1 in 1985.

The timing of the late cohort in 1987 was slightly later than that of the
sixth cohort in 1985. The late cohort in 1987 was placed on the trees at ca. 90%
shoot elongation. Cohort 6 in 1985 was started at ca. 60% shoot elongation.
Peak pupation occurred at ca. three weeks after the end of the shoot growth
period for both the late cohort in 1987 and cohort 6 in 1985.

The within-branch locations of the budworm larvae were recorded during
branch examination in 1987 as an indication of larval feeding sites (Table 12).
Prior to budbreak, most 2nd and 3rd—stage larvae of the April-June cohort were

either mining the previous year’s needles or were situated in webbing among the

55

Table 11. S ruce budworm performance for the two budworm cohorts in the

1 87 survivorship experiment.

 

 

 

 

 

 

April-June J une-July
Cohort Cohort

Budworm Performance N Mean (_tSE) N Mean LtSE)
Larval Survival (%) 10 43.0 (15.8) 10 6.2 (i 1.8)
Pupal Survival (%) 10 84.2 (_+_-5.6) 7 81.0 (114.3)
Adult Dry Weight (mg) 77 30.9 ($0.7) 10 13.8 (i 1.2)

N Medianl N Median

Length of Development (days) 77 61.0 10 47.9
Length of Development
(degree days) 77 780.5 10 905.2

 

1 Median values are given for length of develo ment because more than one-
third of surviving insects in the April-June co ort had completed development

prior to the time of final branch examination.

56

Table 12. Avera e larval stage and distribution of spruce budworm larvae on
caged ranches in relation to timing of budbreak during the 1987
survivorship experiment.

 

 

 

 

 

 

 

A ril-June J une-July
ohort Cohort
Number of Number of
Weeks Prior Weeks After
to Budbreak Budbreak
-1 0 +4 +5 +6 +7
Average Larval Stagel 2.9 4.6 2.8 33 4.6 6.0
Total Number Larvae 102 93 50 24 13 16
Lemme
1W
Old Needles - Mines 14.7 1.1 2.0 0.0 0.0 0.0
Old Needles -- External 39.2 10.7 0.0 0.0 0.0 0.0
Vegetative Budsz 40.2 64.5
New Needles - Mines --- -- 22.0 16.7 0.0 0.0
New Needles - External m -- 54.0 50.0 76.9 81.2
Moving or Not On Branch 5.9 23.7 22.0 33.3 23.1 18.8

 

1 Average instar number.

2 Vegetative buds were present only during the April-June cohort. Expanded
shoots with new needles were present only during the June-July cohort.

57
bases of those needles. As the vegetative buds became swollen, more larvae were
located next to or within the buds. After budbreak the older larvae fed primarily
on foliage of the newly expanding shoots.

Larvae of the June-July cohort were rarely found on or in the previous
year’s foliage (Table 12). Almost all of the needle mining done by young larvae
in this cohort occurred in the new needles of the expanding shoots. As the larvae
matured they continued feeding almost exclusively on new foliage.

Larval counts from the first two sample dates of the early cohort were
lower than all later larval counts in that cohort. Apparently many larvae were not
detected during those early branch examinations because of the extreme difficulty
of finding the smallest 2nd instars as they mined old needles and vegetative buds.
Data from these two dates have not been included in Figure 14 and Table 12.
This problem did not exist in the late cohort, because buds were not present at
that time, and due to warmer temperatures most larvae that were alive at the first

sample date were more easily found because they were larger 3rd instars.

Seasonal Variation in Host Nutritional Traits

The timing of budbreak and shoot growth phenology were very similar for
1985 and 1986. Budbreak in the early-flushing SASK seed source occurred
between 28 April and 8 May, or at 200-300 accumulated degree days in both
years. Fifty percent shoot elongation on SASK trees occurred at ca. 13-14 May
and 400-410 degree days in both years. The same between-year similarities
occurred for ONT and BC trees, however the timing of phenological events in
these seed sources was 7-10 days and 100-125 degree days later than that of the

SASK seed source.

58
Current-Year Foliage

All measured nutritional traits in current-year foliage varied significantly (P
5. 0.001) among sample dates (Table 13). For most of these traits the greatest
quantitative changes occurred during the period of active shoot growth. A seed
source effect and a sample date x seed source interaction were also significant (P
_g 0.05) for several traits (Table 13). Mean sums of repeated measurements were
used to compare individual seed sources (Table 14).

Spas -- Total foliar sugar (fructose + glucose + sucrose) in new foliage
was at its highest levels near the middle of the shoot elongation period and then
generally decreased during the summer months (Figure 15). The earliest sample
date for which sufficient foliage was available for sugar analyses was 17 May (JD
137). The highest total sugar levels observed on SASK trees (15.0%) occurred on
that date. SASK trees at that time had reached 60% shoot elongation. ONT
trees had reached 30% shoot elongation at that time and had a total sugar level
of 14.4%. Twelve days later at about 60% shoot elongation the total sugar level
in ONT trees peaked at 15.5%. This difference in the timing of the peak levels of
total sugar may be the result of the difference in timing of phenology between the
SASK and ONT seed sources. However, the pattern does not appear to hold for
the late-flushing BC seed source which has a trend similar to that of the early-
flushing SASK seed source.

WW - Total nitrogen and phosphorous levels in current-
year foliage were highest in the swelling buds, declined rapidly at the time of
budbreak, and then remained low through the summer (Figure 16). Prior to
budbreak, ONT trees had the highest levels of both nitrogen and phosphorous.
However, the slightly lower levels in SASK trees may be due to the advanced
phenology of that seed source. Data for the first sample date (1 May) are based

on a sub-sample of trees (three trees each for SASK and ONT sources and one

59

Table 13. Multivariate repeated measures analysis of total sugar, total nitrogen
and mineral content of current-year foliage in June-September 1985
and water content and toucghness of current-year foliage in May-
August 1986 with sample ates as repeated measures.

 

 

Dependent Source of Wilks’ F

Variable Variation lambdal Value Prob.
Total Sugar Date 0.1818 15.30 < 0.001
Date x seed source 0.4121 1.90 0.080

Seed source 1.97 0.165

Nitrogen Date 0.2524 13 1.30 < 0.001
Date x seed source 0.1898 4.40 <0.001

Seed source 2.67 0.093

Phosphorous Date 0.0266 124.58 < 0.001
Date x seed source 0.2833 2.99 0.008

Seed source 11.79 < 0.001

Potassium Date 0.0278 118.94 < 0.001
Date x seed source 0.3555 2.30 0.034

Seed source 3.53 0.048

Calcium Date . 0.0549 58.53 < 0.001
Date x seed source 0.4068 1.93 0.075

Seed source 1.05 0.368

Magnesium Date 0.0777 40.34 <0.001
Date x seed source 0.2997 2.81 0.012

Seed source 1.80 0.190

Manganese Date 0.1141 26.40 < 0.001
Date x seed source 0.1530 5.29 <0.001

Seed source 8.80 0.002

Copper Date 0.1097 27.58 <0.001
Date x seed source 0.6150 0.94 0.514

Seed source 3.47 0.050

Iron Date 0.2999 7.94 < 0.001
Date x seed source 0.3751 2.15 0.047

Seed source 3.59 0.046

Sodium Date 0.1859 14.89 < 0.001
Date x seed source 0.5124 1.35 0.245

Seed source 1.95 0.167

Table 13 (cont’d.).

60

 

 

Dependent Source of Wilks’ F
Variable Variation lambdal Value Prob.
Zinc Date 0.1660 17.09 < 0.001
Date x seed source 0.3397 2.43 0.026
Seed source 1.53 0.239
Water Content Date 0.0044 1009.45 <0.001
Date x seed source 0.0402 17.94 <0.001
Seed source 6.58 0.006
Toughness Date 0.0072 623.22 <0.001
Date x seed source 0.0727 12.19 <0.001
Seed source 1.93 0.170

 

1 Wilks’ lambda criterion (Morrison 1976) was used to compute F—values for date

effect and date x seed source interaction using Rao’s (1973) procedure.

61

Table 14. Seed source effects on foliar characteristics represented as mean sums

of repeated measurements for current-year fohage variables.

 

Mean Sums By Seed Source]

 

 

Foliar Trait Units SASK ONT BC

Total Sugar % dw 63.2 a 69.5 a 68.1 a
Nitrogen % dw 7.38 a 7.44 a 7.99 a
Phosphorous mg/ g dw 10.46 a 10.67 a 12.82 b
Potassium mg/ g dw 54.30 a 60.23 ab 66.47 b
Calcium mg/g dw 20.93 a 20.17 a 18.17 a
Magnesium mg/g dw 6.61 a 7.11 a 7.12 a
Manganese mg/ g dw 2.92 a 3.83 ab 4.76 b
Copper ug/g dw 27.8 a 31.9 a 32.6 a
Iron ug/g dw 239.6 a 195.2 b 203.8 ab
Sodium leg/g dw 171.0 a 177.6 a 195.6 a
Zinc rig/g dw 2603 a 278.3 a 240.6 a
Water Content % fw 332.1 a 340.2 b 336.6 ab
Toughness g 2878.8 a 3095.4 a 2817.2 a

 

1 Sum of repeated measurements = sum of six sample dates (total sugar, total
nitrogen and minerals in June-September 1985) or five sample dates (water

content and toughness in May-August 1986) per tree.

N = 8 trees for each mean. Means within the same row followed b the same

letter are not significantly different (P > 0.05, Tukey’s Studentized

ange Test).

62

 

 

 

 

17 I I I
, . . O—OSASK
A 16 - : : A : A—AONT
i I A I
25 . . .
g 13 - : : g\ \A\
12 ' : i D\
D I
m 11 . : . @3856 g
_l I I I
< l I I
10 - g0
'5} laud: Shoot : °
9 - :Break: Elongation :
8 I L I 1 41 1 l 1
110 130 150 170 190 210 230 I 250
JULIAN DATE

Figure 15. Seasonal variation of total sugar in current-year foliage of three seed

sources in 1985.

(A)
3
E
as.
E
8
E
Z
i
.'§

(B)

PHOSPHOROUS (mg/g Dry Wt.)

Figure 16.

63

 

 

 

 

 

'8 I Bud I Shoot I o—o SASK
7 _ :Break: Elongation : A—AONT

l : : U—U BC
6 ' I I

' I I
5 I I '

_ l I
4 - I I

_ I l
3 I I

' I

. I

I

2- 08a. - _
o ’ : 1 :I 1 1 1

1 10 150 150 170 190 210 230 250

 

   

 

 

_ IBud I Shoot i O—OSASK
IBreakI Elongation I A—AONT
12 IA : : u—cch
1o . . I I
I I I
' Io I I
8 ' I I I
u I | I
s- ' : :
u I I l
4- I I
. I ~ I
2 - I I W
_ l I I
o | I I II I I I

 

1 10 130 150 170 190 210 230 250
JULIAN DATE

Seasonal variation of (A) total nitrogen and (B) ghosphorous in
current-year foliage of three seed sources in 198 .

64
tree for BC source), however these higher levels of nitrogen and phosphorous in
the swelling vegetative buds agree with seasonal patterns generally observed in
plant foliage (Mattson et a1. 1983, Mattson and Scriber 1987).

The levels of potassium in ONT and BC trees decreased from averages of
17-18 mg/g (dw) in mid-May to less than 10 mg/g by mid-July (Figure 17A). In
contrast, potassium levels in SASK trees increased during shoot elongation from
12.6 to 15.5 mg/g, and then decreased to levels during the rest of the summer that
were significantly lower (P _<, 0.05) than those of BC trees (Table 14).

Levels of calcium, magnesium, copper and zinc in current-year foliage
decreased and then increased over the course of the spring and summer (Figures
17B, 18A, 19A, 20B). The decreases occurred during shoot elongation. The
increase in calcium began in early June at the end of the shoot elongation period,
and by September reached levels 60% to 90% higher than those of mid-May.
Increases in magnesium, copper and zinc occurred later in the summer and
yielded end-of-season levels not substantially different from those of mid-May.

Trends in the levels of manganese, iron and sodium are less clear, although
the levels of manganese and iron generally increased during the summer (Figures
188, 19B, 20A). SASK trees had significantly (P _<_ 0.05) lower levels of
manganese than were found in BC trees and significantly higher levels of iron
than were found in ONT trees (Table 14).

WW -- Water content of the current-year
foliage in 1986 peaked at ca. 80%, soon after budbreak in mid-May and then
steadily decreased during the shoot growth period to less than 60% by mid-July
(Figure 21A). The patterns of variation in water content were very similar among
seed sources, however, the peak and decline of water content in the early-flushing
SASK seed source were about one to two weeks advanced over that of the other

two seed sources.

65

 

 

 

 

 

 

 

(A) f: . I I I O—OSASK
A I- I I I A—AONT
5 16'- E EA<\2 E Cl—DBC
g 14 :- I I ,4 \tl
S 12 - I I00 I \

5 10 - I I to\A
: a“. I I I \ \gfi/flfl
2 l I I I 0\ /C-f
"’ 6~ : : : °
3 4 L I I I
8 P I Bud I Shoot I
2 . :Broak: Elongation :
’ I I I
o k 1 1 1 -1 L 1 1
110 1.30 150 170 190 210 230
(a) 5 " ' t'_—_Budkt Elsmatfi . as
IBM I ong on I .
S 4 I I I In
I I I a
b ' I I I /D
o | I / / E
Q 3 ' I I
a. I A\
33 ’ I P .
2 .. I I /I
2 I I I
2 I I I
g I I I O—OSASK
1 ' ' ' ' A—AONT
I I I I D—DBC
L l E L {I l L j
110 130 150 170 190 210 230
JUUAN DATE

Figure 17. Seasonal variation of (A) potassium and (B) calcium in current-year
foliage of three seed sources in 1985.

 

250

 

250

66

 

(A) 1.6

 

‘ ‘ i i—XSIST"
I I —
S; 1'5 ' : :A : n—nBc
z~ 1.4 - ' ' '
o : : :
Q 1.3 " : : ll : I
g I I I I
V 1.2 " I I A ' I l o
z I I°~ ' ' ' .
:7 1.1 . ° _—
g I I . A .
z 1.6. : : : '
g IBud . Shoot .
0-9 * :Broak: Elongation :
O a E I | 1

 

 

110 130 150 170 190 210 230 250

 

(a) 1" laud: Shoot
IBroakI Elongation

1.2

1

. 1.0

0.8

V r V

0.6

7W

 

0.4

0—0 SASK
A—A ONT
El—EI BC

MANGANESE (mg/g Dry Wt)

0.2

' 1

 

 

D----------

D---
1-----

o.o 1 1 1 1 1
1 10 130 150 170 190 210 230 250

JULIAN DATE

 

Figure 18. Seasonal variation of (A) magnesium and (B) manganese in
current-year foliage of three seed sources in 1985.

67

 

(A) 8

  
 

  

—------1

/°\3

COPPER (ug/g Dry Wt)
01

 

-----------‘

 

4 - I
r I
I _
3 . Bud I Shoot : 2—2glANISTK
_ Break: Elongatlan I n—n BC
| I
2 L 1 1 1

 

110 130 150 170 190 210 230 250

 

(B) 55

    
  

     

 

 

 

Iaud I Shoot I o—OSASK
50 _ :Broak: Elongation I A—AONT
| ' ' EI—UBC o
A I I I
i 45- : : :
g I I I A
6 4° . I I I
\ I I
O
3 35 ' I I I
I I I
§ ao- : : :,
- I I A\A/I
25 - I I I
I I I
I I I
20 - 1 - 1 -L L 1 1
110 130 150 170 190 210 230 250
JUIJAN DATE

Figure 19. Seasonal variation of (A) copper and (B) iron in current-year foliage
of three seed sources in 198 .

68

 

      
 

 

 

 

 

 

 

 

 

(A) 60 . . .
I I I O—OSASK
: : = raga
3“ 5°" 3 : i
E I I I
40 .
Q I I I
3 I I I
2 30 b : ' ' \
2 I I
8 I I
In 20 - I I I
I BUd I Shoot I
:Broak: Elongation I
10 1 1 a 1 -1 1 1 1
110 130 150 170 190 210 230 250
B . . .
( ) 7O tBud I Shoot I o—o SASK
65 _ IBroaltI Elongation I A—AONT
: : : EI—DBC
3; 6°" 5 E E
6 55~ : : :
g 50 - I I I
3 I I0\ I
0 45 ’ I I
6 .0. = =
I I
35 . I I
30 E 1i 1 -L 1 1 1
110 130 150 170 190 210 230 250
JULIAN DATE

Figure 20. Seasonal variation of (A) sodium and (B) zinc in current-year foliage

of three seed sources in 1985.

69

 

 

    

 

 

 

 

   

 

 

(A) 85 . . .
I I I O—O 33K
A I I I A—A
3 5° - : ' n—cl ac
.c O '
§ 75 - I I
L I I
5 I : :
E 70 I I
I l
I; 65 ~ :
o I
O 60 ' I I
a: | I
I5 55 _ I Bud I Shoot I A
§ :Broak: Elongation :
50 I 1 I I I 1 1 1 1
110 130 150 170 190 210 230 250
(a) I Bud I Shoot I
1000 * :Broak: Elongation I
I
A 800 - I
3 I
I
g 600 - I
z I
I I
8 400 ~ I
2 :
I O—O SASK
20° ' : A—A ONT
I L'l—EI BC
0 I 1 1 1 1
1 10 130 150 170 190 210 2.30 250
JULIAN DATE
Figure 21. Seasonal variation of (A) water and (B) toughness in current-year

foliage of three seed sources in 1986.

70

Toughness of the newleaves, on the other hand, increased rapidly during
the period of shoot elongation (Figure 213). The more advanced phenology of
the SASK seed source was again indicated by an earlier increase in toughness.
However, the pattern of increasing toughness in SASK trees differed from that of
the other two seed sources. Toughness increased rapidly in SASK leaves up to
582 g in late May (JD 146), when new shoots were 90% elongated. Leaf
toughness continued to increase but at a slower rate throughout the summer until
reaching a peak of 843 g in mid-August (JD 224). In contrast, leaves on ONT
and BC trees increased in toughness only during the period of shoot elongation
(early June). However, leaf toughness at the end of this period on these seed
sources peaked at higher values (965 g and 899 g, respectively) than occurred a

fill] two months later on SASK trees.

One-Year-Old Foliage

One-year-old foliage was sampled from all trees at less than one week
prior to budbreak on the earliest-flushing trees (equivalent to one to two weeks
prior to budbreak on late-flushing trees). These samples were analyzed as was
current-year foliage for nitrogen, nine minerals, water and toughness (Table 15).
The levels of many of these foliar traits were similar to the levels observed in
current-year foliage at the end of the summer (Figures 1621). However, water
was one of the major exceptions to this trend. Water content of old foliage was

ca. 46%, about one-fifth less than the mid-August level of 57% in new foliage.

Relationship of Host Susceptibility to Host Phenology

The relationships between budworm performance and host quality were

examined in two ways using host foliar values occurring during early larval

71'

Table 15. Mean values (18E) of 12 foliar characteristics of one-year-old foliage
on three white spruce seed sources sampled less than one week prior
to budbreak on the earliest-flushing seed source.

 

 

 

 

< 1 Week 1-2 Weeks
Prior to Prior to
Budbreak Budbreak
N per
Foliar Trait1 Units Mean SASK ONT BC
Nitrogen % dw 3 1.12 1.21 1.22
(1 0.04) (1 0.08) (1 0.02)
Phosphorous mg/ g dw 3 1.46 1.45 1.49
(1 0.07) (1 0.12) (1 0.09)
Potassium mg/ g dw 3 6.97 6.80 6.70
(1 0.79) (1 0.61) (1 0.69)
Calcium mg/ g dw 3 5.24 5.77 4.61
(1 0.96) (1 0.41) (1 0.73)
Magnesium mg/g dw 3 1.09 1.19 1.15
(1 0.06) (1 0.03) (1 0.01)
Manganese mg/ g dw 3 1.01 1.36 1.44
(1 0.29) (1 0.22) (1 0.21)
Copper reg/g dw 3 3.51 4.53 4.11
(1 0.16) (1 0.44) (1 0.14)
Iron ug/g dw 3 55.2 47.2 39.0
(1 11.2) (1 11.6) (1 0.73)
Sodium ug/ g dw 3 22.4 22.0 28.6
(1 0.7) (1 2.4) (1 3.2)
Zinc ug/g dw 3 60.5 68.0 54.6
(1 10.0) (1 7.6) (1 5.2)
Water Content % fw 8 46.0 46.5 45.7
(1 0.005) (1 0.002) (1 0.003)
Toughness g 80 859.4 965.5 890.5
(1 12.2) (1 6.2) (1 9.6)

 

1 Nitrogen and minerals sampled on 1 May 1985. Water content and toughness

samp

ed on 28 April 1986.

72
development and those during late larval development. All data used in these
analyses except foliar levels of water and toughness were collected in 1985.
Values of water and toughness were determined for each observation in the data
set by interpolation of water and toughness data from 1986 using the appropriate

elapsed degree day values for early and late larval development.

Canonical Correlation Analyses

Budworm performance and host quality data were first analyzed using
canonical correlation techniques. The relationships revealed by these analyses are
similar for foliar traits during both early and late larval development, although the
relative importance of some variables differed (Table 16).

Four pairs of canonical variates (linear combinations of original variables)
were produced for each of the two analyses (early foliar traits and late foliar
traits). Each pair of canonical variates consists of an insect performance variate
and a host quality variate. However, biologically meaningful relationships were
demonstrated in only the first pair of canonical variates in each analysis. Lack of
meaningful relationships in the second, third and fourth pairs of canonical variates
was indicated by low loadings (absolute values < 0.50) for almost all variables
and by low levels of redundancy. Redundancy is a measure of the proportion of
variance in one set of original variables (here, insect performance) that is
explained by the canonical variate of the other data set (host quality) (Smith
1981). The redundancy estimates for the first pair of canonical variates were
69.4% in the early foliar trait analysis and 76.1% in the late foliar trait analysis.
Redundancy was 3% or less for each of the other three pairs of canonical variates
in each of the two analyses. The canonical correlation between the insect
performance variate and the host quality variate was 0.95 for the first canonical

variate in both the early and the late foliar trait analyses.

73

Table 16. Canonical correlation analysis of insect performance and host quality
using host foliar values during early larval development and late larval

 

 

 

development.
Correlation of Original Variable

With First Canonical Variate

Early Larval Late larval

Development Development

Variables (Cohorts 3-8)’ (Cohorts 1-8)

Wage:
Larval Survival 0.877 0.909
Length of Development1 -0.905 -0.920
Adult Dry Weight 0.779 0.882
Growth Rate 0.932 0.959
W:

Total Sugar 0.622 0.588
Nitrogen 0.964 0.692
Phosphorous 0.919 0.737
Potassium 0.679 0.788
Calcium -0.309 -0.736
Magnesium 0.541 0.138
Manganese -0.195 -0.558
Copper 0.478 0.241
Iron 0.092 -0.347
Sodium 0.255 -0.136
Zinc 0.628 0033
Water Content2 0.858 0.984
Toughnessz -0.900 -0.854

74
Table 16 (cont’d.)

 

 

Early Larval Late Larval

Development Development

(Cohorts 3-8) (Cohorts 1-8)
Wanna 0.952 0.950
Rednndangy’ 0.694 0.761

 

1 Length of development measured in degree days.

2 Water content and toughness data collected in 1986. All other data collected in

1985.

3 Proportion of variance in original insect performance variables explained by

host quality canonical variate.

75

The patterns of loadings for insect performance variables in the first
canonical variate were consistent between the early foliar trait analysis and the
late foliar trait analysis. Larval survival, adult dry weight and growth rate were
positively correlated, whereas length of development was negatively correlated
with the first canonical variate. All correlations of insect performance with the
first canonical variate had absolute values of greater than 0.77 in each analysis.
Thus, the first canonical variate for these four variables provides a strong linear
representation of insect performance.

In the analysis of foliar traits during early larval development, the host
quality variables that had the highest loadings (absolute values greater than 0.85)
were nitrogen, phosphorous, water content and toughness. Nitrogen, phosphorous
and water content were positively correlated whereas toughness was negatively
correlated with the first canonical variate. A second grouping of host quality
variables includes total sugar, potassium and zinc with positive loadings of 0.62,
0.68 and 0.63, respectively. A weak positive relationship is also indicated for
magnesium with a loading of 0.54.

In the analysis of foliar traits during late larval development, water content
and toughness again had high loadings (0.98 and -0.85) in the first canonical
variate. However, nitrogen and phosphorous with loadings of 0.69 and 0.74 were
less strongly correlated with the first canonical variate than in early larval
development. Potassium had a high positive correlation of 0.79, and calcium had
a high negative correlation of -0.74. Weak relationships were indicated for total
sugar and manganese with loadings of 0.59 and -0.56, respectively.

The results of the canonical correlation analyses indicate that insect
performance was positively correlated with foliar levels of water, nitrogen,
phosphorous, potassium and total sugar and negatively correlated with leaf

toughness during both periods of early and late larval development. There also

76
were positive relationships between insect performance and magnesium and zinc
levels during periods of early larval development, and negative relationships
between insect performance and calcium and manganese levels during late larval
development. Insect performance did not appear to be related to levels of
copper, iron or sodium.

Bivariate plots of data used in canonical correlation analysis were
examined to evaluate the data for non-linear relationships that might affect the
level of correlation. Insect performance had slightly non-linear relationships with
water content and toughness during early larval development and with calcium
during late larval development. However, all of these relationships approximated
linearity, and these host quality variables had high loadings on the first canonical

variate.

Ordination of Canonical Scores

An ordination technique using two-dimensional plots of canonical scores
(Smith 1981) provides a useful way to compare insect performance and host
quality relationships among seed sources and budworm cohorts. Separate plots
were prepared for each seed source using canonical scores from the canonical
correlation analysis of late larval development (Figures 22-24). Each point on a
plot represents one cohort on one tree (number indicates cohort). Three groups
of cohorts are indicated on each plot: cohorts 1 to 3 were started at weekly
intervals from 23 April to 7 May; cohorts 4 to 6 were started at weekly intervals
from 14 May to 28 May; and cohorts 7 and 8 at monthly intervals on 25 June and
23 July.

The ordination technique clearly illustrates the positive relationship
between host nutritional quality and insect performance. Host quality and insect

performance both decrease with progressively later budworm cohorts. These

77

 

Late April to
Mid-June

 

 
   

 

Mid-May to
Mid-July

INSECT PERFORMANCE

Late June to
_2 _ Early September

 

 

L l

 

 

-2 -1 0 1
HOST QUALITY

Figure 22. Ordination of canonical scores for insect performance and host
uality durintfifilate larval development on the SASK seed source.
umbers wi ' graph indicate cohorts on individual trees.

78

 

 
   

 

   

2 ...
Late April to
Mid-June
w 1 7'
0
Z
<
E
Mid-May to 4
O Mid-Jul/ 4 4
II. V 4
m 0 4 44
LIJ
a.
I'-
0
III
0)
E
-1 _—
8 Late June to
2 Early Septem'ber
l l

 

 

 

 

-2 -1 o 1
HOST QUALITY

Figure 23. Ordination of canonical scores for insect performance and host
gluality duririfilfilte larval development on the ONT seed source.
umbers wi ' graph indicate cohorts on individual trees.

78

 

 
   

 

   

2
Late April to
Mid-June
l..l.l 1 '—
0
Z
<
E
Mid-May to
E MId‘JUIY/4 4414
CE 0 4
LIJ
a.
'—
0
IL!
0)
.2.
-1 _—
8 Late June to
2 Early September
I l

 

 

 

 

-2 -1 0 1
HOST QUALITY

Figure 23. Ordination of canonical scores for insect performance and host
quality durin late larval development on the ONT seed source.
Numbers wi ' graph indicate cohorts on individual trees.

79

 

 

 

    
     
 

2
Late April to
Mid-June
LIJ 1 '
0
Z
<
E
O /44
LI- 44‘4
“-1 Mid-May to 4 4
°- Mid-July 4
I-
O
LIJ
(D
Z
-1 -
Late June to
-2 ' Early September
I l

 

 

 

 

-2 -1 O 1
HOST QUALITY

Figure 24. Ordination of canonical scores for insect performance and host
guality durintfillflte larval development on the BC seed source.
umbers wi ' graph indicate cohorts on individual trees.

80
changes occur quickly during the spring (cohorts 1 to 6), but only slight changes
occur during the summer (cohorts 6 to 8). Seed source differences can also be
seen among the plots. Cohorts 1 to 3 had the best host quality and performed the
best on the ONT seed source (Figure 23). Comparison of the locations of cohorts
1 to 4 on the three plots reveals the effects of the earlier phenology of the SASK
seed source. These points on the SASK plot are shifted more to the lower left
than on the ONT and BC plots. Among ONT and BC trees, insect performance
in cohorts 1 to 3 changed only slightly while host quality was decreasing.

Regression Analyses

Regression analysis of budworm performance and host quality was
performed using larval survival and individual foliar traits. Bivariate plots of
budworm performance data with foliar trait data revealed that the relationships of
larval survival, adult weight and budworm growth rate with individual foliar
variables were nearly identical. The relationships of length of development with
foliar variables were, as expected, opposite to that of the other three budworm
performance variables. Therefore, larval survival is presented here as an overall
indicator of budworm performance in regression analysis.

The results of regression analysis support the conclusions of the canonical
correlation analyses. The strongest relationships between survival and foliar
variables were with levels of nitrogen, phosphorous, water and toughness during
early larval development and levels of phosphorous, potassium, calcium, water,
and toughness during late larval development. These results were obtained using
the same data set used in the canonical correlation analyses. However, some
additional data, that were not included in canonical correlation analyses because
of the requirement of a balanced data set, were available for regression analysis.

These data included observations from branches that had 0% larval survival in

81
some later cohorts and data for nitrogen, phosphorous, water and toughness
during early development in cohorts 1 and 2. The following discussion of
individual foliar variables is based on regression analyses of data sets including
these additional observations.

Larval survival was positively associated with foliar levels of total sugar,
nitrogen, phosphorous, potassium and water and negatively associated with foliar
levels of calcium and toughness (Figures 25-31). The general trends of these
relationships were the same for foliar levels during early larval development and
during late larval development. However, in several cases the strength of
association or the pattern (linear vs. non-linear) of the relationships did vary
between early and late development. All regression equations illustrated in
Figures 25-31 are significant (P < 0.01). As indicated by canonical analysis, the
positive relationship between larval survival and total sugar (Figure 25) was
relatively weak (r2 < 0.40). This relationship was similar for foliar levels of sugar
during both early larval development and during late larval development.

The relationships of larval survival with levels of nitrogen and phosphorous
were very similar (Figures 26 and 27). Foliar levels of nitrogen and phosphorous
varied much more during early larval development than during late larval
development. Nitrogen had a range of 1% to 6% and phosphorous had a range
of 1 to 10 mg/ g during early larval development. Larval survival was positively
associated with nitrogen, up to nitrogen levels of about 2.5% to 3%. Thereafter,
survival appeared to level off or possibly decline, although there were few
observations in this category. A similar non-linear relationship existed between
larval survival and the level of phosphorous during early larval development.
Larval survival was greatest at phosphorous levels of at least 4 to 5 mg/ g. The
levels of nitrogen and phosphorous during late larval development varied over

much smaller ranges. Nevertheless, regression analysis revealed highly significant

82

(A) 100

 

. v - -49.63 + 5.66 x
80 - r2 = 0.36 ° °

LARVAL SURVNAL (It)

 

 

 

 

 

_ Y - -63.65 + 8.23 x 8 6
- O
r2 0.32 o 00 80

8
o

LARVAL SURVIVAL (at)
3 8

N
O

 

 

 

 

6 8 1o ' H 12 ' - 14 16 18
TOTAL SUGAR (95 Dry Wt)

Figure 25. Relationship of percent larval survival and total su ar in current-year

foliage during (A) early larval development and (B late larval
deve opment.

83

 

    
  

  

 

 

 

(A) 100 o
O
O
80- 0 o o
8 . o
g 60- :8 o
. 0 .
a O: o
40- O
i _ 8.
E 20 0 Y--72.64+75.60X—9.47 x2
o r2-0.76
o t " l l I 1
0 1 2 3 4 5 6

 

(a) 100

LARVAL SURVIVAL (x)
8

 

 

 

 

 

4O ..
2 0 o = -66.29 + 72.51 x
r2 - 0.39
O 0 . 1 n
0 3 4 5 6

TOTAL NITROGEN (a: Dry Wt)

Figure 26. Relationship of percent larval survival and total nitrogen in current-

year foliage durmg (A) early larval development and (B) late larval
development.

84

 

  
 

 

 

(A) 100 o
O
O
a 80" O O O
V 0 00
g 60- 0
D b
U) 40*-
i
E 20 Y= -49.21+ 34.77x - 2.45 x2
r2=0.71
o : ' :3. n . 1 . I a L
0 2 4 6 8 10

 

 

 

 

 

(B) 100
S? 80
g 60
3
m 40
i
E 20 Y=-4-4.31+36.4-4X
r2=0.50
o 1 . l . 1
6 8 1O

PHOSPHOROUS (mg/g Dry Wt)

Figure 27. Relationship of percent larval survival and phosphorous in current-

year foliage dunng (A) early larval development and (B) late larval
development.

 

85

 

 

 

 

 

 

 

  
  
   

 

 

(A) 100
Y = -16.05 + 2.64 x
S? 50 r2-0.25 o
E
D
(I)
g
22
(B) 100 o
Y- -36.66+ 6.16x 0°C 0° 0

a O
9 80 P r2 0.57 0 %0 Q38
V o 0Q)
. «96°
5 80 ,, a: 0%
6 ° ‘5 o
:2 ‘° oo°
g o

20
O
o '.: . :: 2' z: :3 : - l ‘ ‘ ‘ ' ‘
4 6 6 10 12 14 16 16 20 22

POTASSIUM (mg/g Dry Wt)

Figure 28. Relationship of percent larval survival and potassium in current-year

ftzli

e during (A) early larval development and (B) late larval
veo.pment

86

 

 

 

 

 

 

 

(A) 100
, Y= 33.63 - 5.36 x
51? 60- 0 o 12-0.06
V o
i b 8 O O O
60- 0 o
o
E o 08
a O o 0 00
i 40' 0 0 0° 0 00 o
g cog: °
0 oo
20' o 800% o
. cg o
‘90 @OQDQ’O 3932 0a 02
O
1 2 3 4 5
(B) 100 o 2
_ °6’ 69° Y-155.02-63.42X+6.63X
o r2-0.53
9 80? 0 $00
V . ° 00 00 O
. . o
g ... 0% go. .
. ,‘o o
a o 009ao
d 40 ’ o o 3 .~‘oo o
+ 0 q, .(o 0
E o ‘ ° 0 o
20" o 0% O. %
o o 0
0° 0 (12900 ,0 . m
0 ° uno' o
1 2 3 4 5

CALCIUM (mg/g Dry Wt)

Figure 29. Relationship of percent larval survival and calcium in current- ear

foli 13 during (A) early larval development and (B) late larv
deve opment.

 

87

(A) 100

 

_ Y - 625.62 - 20.036 x + 0.161 x2 0°63

2- o
80_ r 0.59 08°9

LARVAL SURVIVAL (x)
8

 

 

 

 

 

4,0 ..
20 -
0‘
50
(B) 100 000
Y - —177.32 + 3.16 x 0 <9 0
8 0 8
60 r2 -0.76 00 >

 

3

LARVAL SURVIVAL (x)
8
‘b

N
O

 

 

 

50 55 65 70 75 60 65
WATER CONTENT (95 Fresh Wt)

Figure 30. Relationship of percent larval survival and water content in current-

year foliage dunng (A) early larval development and (B) late larval
development.

88

 

 

 

 

 

(A) 100
7.8 Y - 67.31 - 0.16 x + 0.0001 x2
9 80 f‘; e . r2 - 0.79
v "t: o
i 322.0
E 50 ' :3 o
D 5’. .. . 8
U) 3Q} ‘ o 0
g 40 _ Q) 0 o 0 ° 0
°: 0 o
g 0 ° 0 o o o 0°
20 F o .' . o. o O 0
80° .-‘°°. %
o O . ' ‘0‘ :. .\‘4.: 0 .g; 3.
0 200 400 600 600 1000
B 100
< ) 0 ° 0 o Y = 65.31 - 0.06 x
' ° 8 ° r2-059
60 ° 0 0° '

LARVAL SURVIVAL (x)
3 8

N
O

 

 

 

 

0 200 400 ' 600 P 600 ”000
TOUGHNESS (g)

Figure 31. Relationship of percent larval survival and toughness of current-year

foli e during (A) early larval development and (B) late larval
deve opment.

89
relationships (P < 0.0001) between larval survival and the foliar levels of these
nutrients during late larval development.

Larval survival was also positively associated with the levels of potassium
but negatively associated with the levels of calcium during late larval development
(Figures 28 and 29). A weak relationship also existed between survival and
potassium levels during early larval development.

The relationships of larval survival with foliar levels of water and toughness
were non-linear for foliar levels during early larval development and linear for
levels during late larval development (Figures 30 and 31). larval survival was low
(< 40%), when water content was < 75% and toughness was > 300 g during
early larval development.

Among the other foliar nutrients measured, regression analyses revealed
minor relationships (r2 < 0.25) between larval survival and foliar levels of
magnesium, manganese and zinc. Survival was positively associated with
magnesium and zinc levels during early larval development and negatively
associated with manganese levels during late larval development.

Many models of relationships between budworm performance variables and
two or more foliar variables were developed with multiple regression techniques.
For each budworm performance variable (survival, weight, growth rate and length
of development), there were always several significant (P _<_ 0.05) models. It was
difficult to select only a few models that best described the relationships of
budworm performance with host quality. Three and four-variable models that
explained a high proportion of the variation in budworm performance (r2 > 0.7)
frequently excluded some foliar variables that previous analyses (canonical
correlation and individual regressions) had indicated were important. This
multiplicity of highly significant models is due to a high level of intercorrelation

among many of the foliar variables measured in this study (Tables 17 and 18).

90

Table 17. Pearson correlation coefficients1 among 10 major foliar variables for
foliar levels during early larval development (cohorts 3-8, N = 136).

 

 

Tough-

N P K Ca Mg Mn Zn Water ness
Total
Sugar .54 .53 .42 -.38 .22 -.17 .24 .68 -.62
N -- .95 .62 -.27 .55 "‘ .56 .75 -.84
P - -- .72 -.39 .54 "‘ .49 .81 -.86
K -- -- -- -.52 .49 -.27 .30 .78 -.72
Ca -- -- -- -- .29 .50 .43 -.62 .46
Mg -- -- - -- - .22 .74 .32 -.40
Mn -- - -- - -- -- .19 -.39 .31
Zn -- -- -- -- -- - -- .27 -.37
Water -- -- - -- -- -- -- -- -.91

 

1 All correlations, except those marked with an asterisk, are significant

(P5 0.05).

91

Table 18. Pearson correlation coefficients1 among 10 major foliar variables for
foliar levels during late larval development (cohorts 1-8, N = 192).

 

 

Tough-
N P K Ca Mg Mn Zn Water ness

Total

Sugar .39 .38 .38 -.42 “ ‘ "' .58 -.46
N -- .91 .64 -.40 .36 " .21 .71 -.76
P -- -- .76 -.54 .35 -.21 ‘ .79 -.79
K -- -- -- -.57 .31 -.33 " .80 -.70
Ca -- -- -- -- .36 .51 .59 -.72 .55
Mg -- -- -- -- -- .35 .65 " -.19
Mn -- -- -- -- -- -- .24 -.47 .40
Zn __ __ __ __ .. __ __ . .
Water -- -- -- -- -- - -- - -.87

 

1 All correlations, except those marked with an asterisk, are significant

(P 5 0.05).

92
Thus, several foliar variables may be good "predictors" of budworm performance,
but the determination of which foliar variable(s) are biologically important to
budworm performance will be best done using experimental diets (Harvey 1974,
McLaughlin 1986, Clancy et al. 1988b, Mattson et al. 1990). Perhaps the
canonical analyses provide the best representation of the relationship between
budworm performance and foliar properties. These analyses incorporate,
simultaneously, most of the variables that one would expect to be important in

regulating the growth and development of folivores.

DISCUSSION
Defining the Window of Susceptibility

Although the concept of a "window of susceptibility" in plants to herbivores
has been in the literature for years, and in the everyday parlance of entomologists
and ecologists (Mattson et a1. 1982), this study is the first to comprehensively test
the hypothesis, and to furthermore offer substantive support for it. Owing to the
rather widespread and casual use of the term, window of susceptibility, there
seems to be no precise definition for it except as implied in Mattson et al. (1982),
i.e. growth and survival of the herbivore outside of the putative "window" is
ecologically negligible. In other words, very few insects are capable of surviving
and reproducing if they initiate or terminate feeding outside of the boundaries of
the putative window.

In the case of white spruce and the spruce budworm, the window for
successful development begins shortly before budbreak and terminates near the
end of the period of active shoot elongation. This means that to achieve
ecologically meaningful grth and reproduction, budworms must begin and finish
growing between these temporal limits of the window. For example, budworm
larval survival, adult weights and growth rates were highest for the first three
cohorts that started feeding prior to or during budbreak and ended feeding and
larval development by the time that shoot elongation ended. Later cohorts were
substantially less successful than the first three, owing to greatly diminished

survival and growth.

93

94

The window of susceptibility in white spruce may begin as early as four
weeks prior to budbreak. The earliest cohorts in 1985 and 1987 began three and
four weeks, respectively, before budbreak and had good success as measured by
all performance variables. However, the survival of the cohort starting four weeks
ahead of budbreak was less than expected from those starting three weeks or less
before budbreak. Outbreaks of western spruce budworm epidemics in British
Columbia have been associated with years when larval emergence preceded
budbreak by 17-18 days (Thomson et al. 1984). Synchrony of larval emergence
and host budbreak is also important for many other species of tree-feeding
defoliators (Embree 1965, von Schiinbom 1966, Wickrnan 1976, Witter and
Waisanen 1978, Leather 1986, Turgeon 1986, Du Merle 1988).

The end of the window of susceptibility is associated with the end of active
shoot growth. Budworms placed on the trees at the time of budbreak, in most
' cases, did not complete larval development until shortly after the end of shoot
elongation. These insects, which include cohort 3 on SASK trees and cohort 4 on
ONT and BC trees, were developing at the later limits of the window of
susceptibility. The largest decreases (38%-53%) in larval survival occurred
between cohort 3 and cohort 4 in each seed source (Figure 7). Only seven days
separated the starting time for development of these two cohorts in the field.
These results indicate a rather abrupt end of the window of susceptibility.
Shepherd (1985) similarly observed high survival when western spruce budworm
larvae emerged prior to swelling of buds, but very low survival when larval
emergence occurred nearer to the time of budbreak. Strong et al. (1984) cited
several examples in which insect-host asynchrony of only one or two weeks caused
greatly reduced performance in insect herbivores. Thus, the phenological window

can be interpreted as a race between the herbivore and the host plant. The insect

95
feeds on a plant that is rapidly deteriorating in food quality owing to nutrient
dilution and buildup of tissue fiber content, and perhaps defenses too.

The beginning of the window of susceptibility in white spruce coincides
well with the normal peak of spring emergence of overwintering larvae (ca. 100
accumulated degree days). However, budworm-tree asynchrony can occur with
budworm phenology either too advanced or too retarded in relation to tree
phenology (Blais 1957, Beckwith and Burnell 1982, Thomson et al. 1984, Raske
1985, Shepherd 1985). The cause of asynchrony may be related to differences in
mechanisms controlling the emergence of larvae and the initiation of vegetative
growth. larval emergence is determined primarily by ambient air temperatures
(Thomas 1976), although photoperiod also plays a role (Harvey 1958). On the
other hand, initiation of vegetative growth is determined by soil temperatures as
well as air temperatures (Cleary and Waring 1969, Lavender et al. 1973, Beckwith
and Burnell 1982, Volney 1985).

Budworm-tree asynchrony was associated with the collapse of two western
spruce budworm epidemics in British Columbia (Thomson et al. 1984).
Asynchrony likewise occurred for three consecutive years (1981-83) during the
collapse of the recent spruce budworm epidemic in Newfoundland (Raske 1985).
In the latter case, larvae became active in the spring of each year at about two
weeks after budbreak on balsam fir. This timing would be approximately
equivalent to one week after budbreak on white spruce. Although parasitoids and
diseases accounted for some larval mortality in these collapsing populations, many

larvae died of unknown causes during the later instars.

96

Seed Source Effects

A seed source effect was present for all budworm performance variables
except pupal survival (Table 3). In most cases budworm performance was lower
on the early-flushing SASK seed source (Table 4). Is seed source effect related to
different timing of phenology among the seed sources? Analyses of
phenologically similar hosts (Tables 5 and 6) indicated that phenological
differences among seed sources accounted for most of the seed source effect on
two budworm performance variables: larval survival (Figure 7) and length of
development in degree days (Figure 8). The effect is particularly evident for
length of development. The shortest budworm development time (in degree days)
occurred in cohort 1 on SASK trees, in cohort 2 on ONT trees and in cohorts 2
and 3 on BC trees (i.e. the same relative order of tree phenology). Thus,
budworm development required the fewest degree days for those insects that
began spring larval development at two weeks prior to budbreak. The seed
source effect on other budworm performance variables (growth rate, adult weight
and length of development in days), however, was apparently related to some
factor other than, or in addition to, phenological differences.

The seed sources used in this study are from widely separated parts of the
white spruce range and have several factors other than phenological differences
that could affect insect performance. As with other plant species, many
morphological and physiological traits vary among widely separated populations of
white spruce (Wilkinson et al. 1971, Nienstaedt and Teich 1972, Nienstaedt 1985,
Khalil 1986). The BC seed source is from a spruce hybrid zone in the western
part of the white spruce range and may be affected by introgression with
Engelmann spruce. Variation in susceptibility to insects among rangewide spruce

seed sources has been recorded for the eastern spruce gall adelgid (Adelges abietis

97
(L.)) on white spruce (Canavera and DiGennaro 1979) and the green spruce

aphid (Elatobium abietinum (Walker)) on Sitka spruce (Day 1984).

Effects of Tree Growth Rate

Despite the many inherent differences possible among the seed sources, a
single variable, tree growth rate, appeared to account for much of the seed source
effect that was not due to phenological differences. As noted above, a significant
seed source effect still existed for length of development in days, growth rate and
adult weight of budworms even after phenological differences were removed.
These three variables were strongly associated with tree growth rate. Length of
budworm development in days was negatively correlated and budworm growth
rate and adult weight were positively correlated with tree growth rate (Figures 11,
12 and 13). These relationships existed for insects that were placed on host trees
prior to budbreak, but not for those placed on trees during the time of budbreak.

One possible explanation of the effect of tree growth rate on insect
performance is competition within the host tree for allocation of resources to
growth and differentiation processes. Plants that are differentiation-dominated
are inherently slow-growing and adapted to resource-limited environments. They
generally have tougher, less digestible tissues owing to higher levels of cellulose
and lignin and defensive compounds such as polyphenols and terpenoids (Coley et
al. 1985, Bazzaz et al. 1987, Coley 1987, 1988, Loehle 1988). The SASK trees had
the slowest growth rates and are probably adapted to drier sites than are trees of
the other seed sources in this study. The relative growth rates of the seed sources
correspond to the relative amounts of average annual rainfall at their sites of

origin. Annual rainfall was reported by Nienstaedt (1963) to be 28 cm/yr for the

98
SASK seed source (Slowest-growing), 50 cm/yr for the BC seed source, and 85
cm/yr for the ONT seed source (fastest-growing).

Temporal variation in the allocation of resources may explain why tree
growth rate affected insects that began feeding prior to budbreak, but did not
significantly affect those that began feeding during budbreak. The proportions of
plant resources allocated to growth and to differentiation vary seasonally (Mooney
et al. 1983, Lorio 1986, 1988). Allocation of resources to purely growth processes
is highest in the early spring. Later, as new leaves finish growing, differentiation
processes dominate and more resources are allocated for maturation and
defenses. Budworms that started feeding prior to budbreak completed larval
development during the period of active shoot growth. Those that started feeding
during budbreak completed development at or shortly after the end of shoot
elongation. It is possible that this latter group was exposed to less digestible and
better defended tissues for a longer period.

Therefore, insects placed on trees during or after budbreak would
encounter much less suitable food resources during their later larval stages. But,
these are the crucial stages in which the greatest increases in body weight occur
(Retnakaran 1983). Plant maturation processes that reduce the digestibility of
foliage would be expected to negatively affect folivore weight gain and growth rate
and to increase time of development. These are, in fact, the budworm

performance variables most affected by tree growth rate.
Relationship of Host Susceptibility to Host Phenology
Results of canonical correlation and regression analyses both suggest that

budworm performance is positively correlated with foliar levels of total sugar,

nitrogen, phosphorous, potassium and water, and negatively correlated with

99
calcium and leaf toughness. Because of the intercorrelation among many of these
foliar traits, it is difficult in a field study such as this to "tease out" and identify
key factors affecting host susceptibility. However, several strong, consistent
relationships do exist.

Comparison of the timing of changes in budworm performance and
changes in levels of foliar traits can be used to identify potential relationships.
The greatest reductions in larval survival occur during early larval stages. At this
time, nitrogen, phosphorous, water and toughness are the foliar traits that exhibit
the greatest variation and, therefore, could be considered as a suite of candidates
for key factors affecting larval survival. Most of larval weight gain occurs during
the last larval stage (Retnakaran 1983). At this time, phosphorous, potassium,
calcium, water and toughness are the foliar traits that have the greatest variation
and, therefore, could be considered as a suite of candidates for key factors
affecting weight gain. However, the magnitude of effect that changes in foliar
quality have on budworm performance may vary with larval age. For example,
one hypothesis states that the performance of young larvae tends to be much
more sensitive to changes in host quality than is that of older larvae (Scriber and

Slansky 1981, White 1984).

Key Foliar Traits

Imam - The levels of total sugars in new foliage were highest near
the middle of the shoot elongation period. This pattern is consistent with earlier
observations of foliar sugars in white spruce (McLaughlin 1986) and balsam fir
(Little 1970). A weak positive relationship exists between budworm performance
and levels of total sugars. Sugars are potent feeding stimulants for spruce
budworm larvae (Heron 1965, Albert and Jerrett 1981, Albert 1982, Albert et al.

1982, Albert and Parisella 1985). Budworm weight and development rate increase

100
as dietary levels of sugars increase to an optimal level (Harvey 1974; McLaughlin
1986; F. Lo, unpublished data). Harvey (1974) determined that glucose, fructose
and sucrose (the most common sugars of current-year foliage in conifers) have
equivalent nutritional value for the spruce budworm. Thus, from a nutritional
point of view, the total amount of sugars are probably more important than the
proportion of individual sugars.

Foliar levels of sugars in this study were adequate for budworm
development in all cohorts. It is not likely that variation in sugar levels was a
major determinant of variation in budworm performance among cohorts.
However, increased sugar levels during shoot growth may have elicited increased
larval feeding in the early cohorts.

Nitrogen - Quantitative variation in nitrogen followed the pattern typical
of most plants (Mattson 1980, Mattson et al. 1983, Mattson and Scriber 1987).
Nitrogen levels were highest in swelling buds and decreased rapidly during early
shoot growth. The importance of nitrogen to insect performance has been well
documented for many phytophagous insects (McNeill and Southwood 1978,
Mattson 1980, Scriber and Slansky 1981, Scriber 1984a, 1984b, White 1984,
Mattson and Scriber 1987) as well as for spruce budworms specifically (Harvey
1974, Mattson et al. 1983, Montgomery and Czapowskyj 1985, Brewer et al. 1985,
1987, McLaughlin 1986, Cates et al. 1987, Mattson et al. 1990).

Budworm performance was strongly correlated with nitrogen levels,
particularly with levels occurring during early larval development. Larval survival
was positively correlated with nitrogen levels during early larval development up
to ca. 2.5% to 3% (dw) for total nitrogen. Only a few observations were available
at higher nitrogen levels, but generally larval survival did not increase with higher

levels. Similarly, Brewer et al. (1985, 1987) reported that performance of the

101
western spruce budworm peaked at foliar nitrogen levels of 2.4% to 2.5% and
decreased at both higher and lower levels.

Nitrogen in plant tissue varies qualitatively as well as quantitatively during
a growing season. Young tissue has higher levels of amino acids, soluble proteins
and nitrogen-based secondary compounds, while older tissue has higher levels of
structural or insoluble protein (Mattson 1980). Seasonal variation of several
amino acids in white spruce foliage has been documented (Durzan 1968).
L-proline and amino acid/base fractions of foliage from white spruce and other
host species have been shown to be mild feeding stimulants for the spruce
budworm (Heron 1965, Albert 1982, Albert and Parisella 1985). Arginine is the
foliar amino acid that makes up the largest portion of protein in budworm larvae,
although it does not act as a feeding stimulant itself (Heron 1965, Durzan and
Lopushanski 1968).

Minerals - Seasonal variation among mineral elements was consistent with
patterns observed in other conifers (Chapin and Kedrowski 1983, Clancy et al.
1988a, Hockman et al. 1989) and most other plants (Mattson and Scriber 1987).
Among the mineral elements measured, increased budworm performance was
most closely correlated with increased levels of phosphorous, potassium and zinc
and with decreased levels of calcium. The pattern of variation of phosphorous is
an exponential decrease during the growing season that closely follows the change
in levels of nitrogen. Thus, the same phenological relationships existing between
budworm performance and nitrogen also exist for budworm performance and
phosphorous. Potassium decreases at a slower rate during the growing season
than do nitrogen and phosphorous, and therefore budworm performance was
more closely associated with levels of potassium during late larval development

rather than early larval development. Budworm performance was positively

102
associated with zinc levels during early larval development and negatively
associated with calcium levels during late larval development.

Little information is available on the levels of mineral elements in insect
tissues and the dietary requirements of insects for mineral elements. However,
results of most studies indicate that phosphorous, potassium and magnesium are
the dominant mineral elements in insect tissues (Mattson and Scriber 1987).
Phosphorous, copper, zinc, iron and sodium in spruce budworms (Mattson et al.
1983) and copper and zinc in western spruce budworms (McLean et al. 1983)
have been measured at levels that are at least twice that of host tree foliage.

Eater -- Water content of current-year foliage peaked at more than 80%
soon after budbreak and decreased to less than 60% by early July. Chrosciewicz
(1986) observed a nearly identical pattern for white spruce and similar patterns
with slightly lower peaks for balsam fir, black spruce and jack pine (Pinus
banksiana Lamb.). Budworm performance in this study was highly correlated with
water content of foliage during both early and late larval development. The
positive relationship between leaf water content and the performance of
folivorous insects has been well documented for many insect species (Scriber
1977, 1978, Scriber and Slansky 1981, Scriber 1984b).

The general trend of decreasing water content in maturing foliage occurs in
most, if not all, plants (Scriber 1984b). It is closely correlated with declining
levels of nitrogen. Because of the consistency in phenological variation in water
and nitrogen levels, and their crucial roles in governing the course of plant
physiology (Agren 1985, McIntyre 1987), these two variables are valuable as
indicators of the entire chemical and physiological "gestalt" of a host plant
(Mattson and Scriber 1987).

Ionglmess -- Leaf toughness increased dramatically during shoot

elongation and was strongly negatively correlated with decreased budworm

103
performance. Toughness is related to the levels of compounds that together
comprise "fiber" (lignin, silica, pectin, cellulose, hemicellulose, etc.) and structural
components such as sclerenchyma fibers and bundle sheaths (Huang 1975, Huang
and Fuhrer 1979, Vincent 1982, Hagen and Chabot 1986, Mattson and Scriber
1987). Leaf toughness has frequently been associated with reduced levels of
herbivory or reduced insect performance (F eeny 1970, Huang 1975, Huang and
Fuhrer 1979, Coley 1983, Lowman and Box 1983, Raupp 1985, Hagen and Chabot
1986). Bauce and Hardy (1988) reported that increased rawfiber content of
current-year foliage was related to previous defoliation of balsam fir and caused
decreases in pupal weight, larval developmental rate and survival of Spruce
budworms.

Leaf toughness and water content are generally negatively correlated and
together can be considered as a measure of sclerophylly (Lechowicz 1983).
Seasonal variation of these two variables more clearly demonstrated the
phenological differences among seed sources in this study than did any other
foliar traits. Thus, leaf toughness and water content may be highly important
factors that affect larval survival and length of development in degree days, i.e.
those budworm performance variables that are associated with phenological
differences among seed sources.

W - The potential role of secondary compounds was
not evaluated in this study; however, the evidence available to date does not
indicate a strong relationship between secondary compounds and resistance to the
spruce budworm. Phenolic compounds seem to have little or no deleterious effect
on spruce budworm or western spruce budworm performance at the levels found
in host foliage (Cates et al. 1983a, 1983b, Mattson et al. 1983, Wagner and Blake
1983, Cates and Redak 1988). The role of foliar terpenes is less clear. In some

cases budworm performance in field tests has been negatively correlated with

104

measures of various terpene levels during later larval instars, although the trends
are inconsistent (Cates et al. 1983a, 1983b, Mattson et a1. 1983, Cates and Redak
1988). Laboratory tests with artificial diet containing levels of monoterpenes
comparable to those found in host foliage have had results that differ between
insect species. Bornyl acetate concentrations were negatively correlated with
western spruce budworm performance (Cates et al. 1987), but bornyl acetate and
six other monoterpenes had little or no effect on spruce budworm performance

(Mattson et al. 1990).

Foliar Traits of the Window of Susceptibility

The extent of the window of host susceptibility to the spruce budworm can
be described in terms of temporal events (bud and shoot phenology, dates and
degree days), as discussed above, and in terms of foliar traits. Ordination of
canonical scores (Figures 2224) provides one view of the extent of the window of
susceptibility along a composite axis of foliar traits. The zero-point on the insect
performance axis closely (and somewhat coincidentally) defines the window of
white spruce susceptibility to the spruce budworm. Points located in the upper
two quadrants of the ordination plots (Figures 22-24) lie within the window of
susceptibility. These points include cohorts l, 2 and part of 3 for the early-
flushing SASK seed source and cohorts 1, 2, 3 and part of 4 for the ONT and BC
seed sources. The lower limit of the window of susceptibility in terms of foliar
traits corresponds to a value of ca. -0.3 on the host quality axis.

A multi-dimensional estimate of the window of susceptibility can be
constructed by examining the ranges of individual foliar traits for which the
herbivore can successfully survive and grow (Mattson et al. 1982). A summary of
ranges of foliar trait values associated with points in the upper quadrants of the

ordination plots (Figures 22-24) provides a 13-dimensional estimate of the window

105

of white spruce susceptibility to the spruce budworm (Table 19). The ranges of
most foliar trait values within the window of susceptibility are comparable to
optimal dietary levels that have been estimated for the spruce budworm (Harvey
1974, Mattson and Koller 1983) and other folivorous Lepidoptera (Mattson and
Scriber 1987). One exception is the level of calcium which is 4 to 17 times
greater than the estimated minimally optimal level (Mattson and Scriber 1987).

A two-dimensional foliar trait space has often been used to describe insect
herbivore-host plant relationships. Foliar levels of nitrogen and water are good
indicators of relative growth rate and approximate digestibility for many insect
herbivores (Scriber and Slansky 1981, Scriber 1984a, 1984b). Foliar levels of
nitrogen and water that existed during early larval development accurately define
the window of white spruce susceptibility to the spruce budworm (Figure 32).
Larval survival was highest when water content was greater than 75% and the
level of total nitrogen was greater than 2%. A clear relationship is also
demonstrated for leaf water content and toughness (Figure 33), the two foliar
traits likely to be the key factors in determining white spruce susceptibility to the
spruce budworm. Larval survival was greatest when water content and toughness
during early larval development were greater than 75% and less than 200 g,

respectively.

Effects of Insect-Tree Phenological Asynchrony

When spruce budworm phenology is too advanced in relation to host tree
phenology, the 2nd-stage larvae must remain on older foliage for an extended
period prior to budbreak. Older foliage clearly has lower nutritional value (Table
15 and Mattson et al. 1983). Nutrient levels in one-year-old foliage in early spring
are in many cases similar to values existing at the end of summer for current-year

foliage. Moreover, water content is even lower than the end-of-summer levels.

106

Table 19. Ran e of values for foliar traits during early larval development and
late arval development within the phenological window 0

 

 

susceptibility.
F oliar Trait Units Egzlégggt I-fiaettfefirmaelnt
Total Sugar % dw 11.6 - 17.4 9.80 - 16.7
Nitrogen % dw 1.70 - 5.84 1.12 - 2.12
PhOSphorous mg/ g dw 2.47 - 9.94 1.67 - 3.88
Potassium mg/g dw 10.9 - 20.7 10.9 - 19.9
Calcium mg/g dw 1.43 - 5.18 1.27 - 3.47
Magnesium. mg/g dw 1.05 - 1.89 0.98 - 1.48
Manganese mg/ g dw 0.225 - 0.833 0.233 - 0.946
Copper ug/g dw 5.14 - 17.14 4.35 - 10.20
Iron rig/g dw 24.5 - 42.4 20.3 - 43.4
Sodium rig/g dw 24.1 - 80.7 18.0 - 48.7
Zinc rag/g dw 41.6 - 72.1 30.7 - 57.8
Water Content % fw 71.7 - 82.8 65.9 - 81.8
Toughness g 5.0 - 263.0 27.0 - 802.0

 

107

RES 380 adv—.6 533 can comp—z: :83 «o £9,0— Ezou 98 B>E=m RES 8083 «o minnows—om .Nm 2&5

om

$>> amoun— ¢$ ...Zm._.ZOO mw._.<>>

cm

on

00

guano—90c

 

 

$o¢10

 

r_<>_>m3m

 

 

.._<>m<r_

 

 

 

(1M Mo %) NBS-OBLIN 111101

108

4588—05—0
BEE >13 was“. 833 new 32.—Eng me 326— 328 98 3323 BE_ 8083 Lo minnows—om .mm Bani

3>> gawk... a3 ...thZOO EMH<>>

 

 

 

 

om om on 00 on
_ (II-”lama“. _ I I _ _ _ _ O
III-III II
..u. room
g. of , 1 . .
I I I II l
I. .. 1 00.? O
... .. . n
I I II IIIII I “H9
Q OVA I II II I . III
. . .86 w.
. . u . S
I I I III IIII 1 00m “DI”
$9410 . ..n....~....
IIII" II
. . .. , coop
4.4);me ._<>m_<r_
ooww

 

 

109
One cause of larval mortality during this period is increased larval dispersal as
larvae search for suitable feeding sites (Beckwith and Burnell 1982). In this study,
where larvae were caged on branches and- prevented from dispersing, the typical
pattern of high mortality levels among young larvae still existed. Thus, I
hypothesize that extended feeding on poor nutritional quality diets is also a major
cause of young larval mortality.

When spruce budworm phenology is retarded in relation to host tree
phenology, reduced budworm performance is obvious for many performance
variables. Survival, growth rates and weights are dramatically reduced. Increased
development time allows an increased period of exposure to predators.
Decreased body weight is correlated with decreased fecundity (Miller 1957).
Larvae feeding on older current-year foliage generally have a higher relative
consumption rate, a lower relative growth rate and lower efficiency of food

utilization (Blake and Wagner 1986, Thomas 1987, 1989).
Implications for Forest Pest Management

The results of this study impact on at least two areas of forest pest
management: modelling pest population dynamics and tree improvement
strategies. Although numerous models of spruce budworm and western spruce
budworm population dynamics have been developed in recent years (Simmons et
al. 1984, Sheehan et al. 1989), better understanding of the effects of insect-host
asynchrony on budworm survival and fecundity is needed, because it will allow
more precise modelling of budworm population fluctuations. Models could be
developed that describe temperature effects on the levels of insect-host
asynchrony, the phenological limits of the windows of susceptibility for all host

tree species and the effects of varying levels of asynchrony on budworm survival

110
and fecundity. Much of the data required for a spruce budworm-white spruce
phenological synchrony model has been provided by this study.

The effects of seed source differences in phenology and grth rate on
spruce budworm performance have important implications for future tree
breeding strategies. White spruce is a popular species in tree improvement
programs in north central and northeastern United States and throughout Canada
(e.g. Bongarten and Hanover 1982, Carter and Simpson 1985, Rauter 1985,
Fowler 1986, Nienstaedt and Kang 1987, Fowler and Morgenstern 1990).
However, reduced interest in planting white spruce in some areas has resulted
from, among other reasons, the susceptibility of white spruce to damage from the
spruce budworm and damage from late spring frosts (Rauter 1985, Blum 1988).
Breeding for delayed budbreak has been suggested as both a method of inducing
budworm-tree asynchrony, and thereby increasing resistance to the spruce
budworm (Blum 1988), and as a method to reduce damage due to spring frosts
(Nienstaedt and King 1970, Nienstaedt 1985).

Breeding for delayed budbreak may not be a practical solution to these
problems. Nienstaedt (1985) and Blum (1988) concluded that some gain in
lateness of budbreak could be achieved, although it would be costly. It is also
questionable whether enough gain could be achieved to overcome the
phenological plasticity of the spruce budworm. In my study, budworms were able
to perform well even when budbreak occurred three to four weeks after the larvae
were placed on the trees. Other studies indicate that there is wide variability ~
within and among populations of the spruce budworm and the western spruce
budworm for rates of spring emergence in response to temperature (Rose and
Blais 1954, Thomas 1976, Volney et al. 1983, Volney 1985).

Based on the results of my study, it might be suggested that advancing the

host window, rather than delaying it, may be more effective in reducing

111
susceptibility to the budworm. The end of the phenological window is very
abrupt, and reduction in budworm performance is dramatic. However, advancing
the end of the window, would involve selecting for an earlier date of growth
cessation, a trait that is controlled largely by photoperiod (Nienstaedt 1974), and
may result in the undesirable trait of reduced tree height growth.

Loehle and Namkoong (1987) have raised another concern, i.e. that
breeding for increased growth rate may adversely affect defensive capabilities of
trees. Resources that are allocated within the tree to growth are not available for
defense or other differentiation processes (Coley et al. 1985, Bazzaz et al. 1987,
Coley 1988, Loehle 1988). The positive relationship between tree grth rate and
spruce budworm performance in my study supports this hypothesis. In contrast,
apparent negative relationships between tree growth rate and insect performance
exist for some other herbivore guilds, e.g. bark beetles (Hard 1985). However,
negative relationships involving bark beetles are largely based on phenotypic
variation in tree vigor and growth rates in response to variation in forest stand
density (Mitchell et al. 1983, Hard 1985, Matson et al. 1987). Further studies are
needed to determine the roles that genotypic variation in tree growth rates and
seasonal variation in tree growth-differentiation relationships (see Lorio 1988)
play in tree susceptibility to bark beetles. Nevertheless, the results of my study
underscore the need for caution in tree breeding programs when selections are

based primarily on tree grth traits.
Recommendations for Future Research
There are several areas where we lack the knowledge required to fully

develop models of budworm-tree synchrony. Comprehensive studies should be

undertaken to determine the phenological windows in other host tree species

112
(balsam fir, red spruce, black spruce). More detailed information is needed about
the role of phenological windows in the differential susceptibility of host tree
species. Regniere (1987) has suggested several areas in which we lack
fundamental knowledge about budworm and host tree phenology - e.g.
mechanisms controlling budworm diapause termination, variability in budworm
responses to temperature and developmental responses of larvae feeding on old
foliage and various host species. A‘ more detailed examination of the plasticity of
spring emergence rates among and within budworm populations should be done
prior to large investments in breeding programs for delayed budbreak.

A better understanding is needed of the allocation of resources to tree
growth and differentiation processes as it relates to tree susceptibility to the
spruce budworm. What are the mechanisms causing reduced performance of
asynchronous budworms? Sclerophylly (tough leaves with low water content)
appears to be the major factor reducing budworm performance in my study. Are .
there specific defensive compounds produced by host trees that also reduce
budworm performance? If so, patterns of phenological variation of such
compounds and their roles in the competitive allocation of resources need to be

determined.

CONCLUSIONS

White spruce has a sharply defined phenological window of susceptibility to
the spruce budworm, a period of about eight weeks in Michigan.

The phenological window for successful development of the spruce budworm
commences about three weeks prior to budbreak and ends at the cessation of
shoot elongation, Spruce budworm performance was also high for cohorts
that started feeding as early as four weeks prior to budbreak and no later
than budbreak. Total performance was best for those insects that started
feeding two weeks prior to budbreak. larvae that began during this period
were able to complete larval development prior to the end of Shoot

elongation.

At least a few spruce budworms were able to survive and develop to adults
during all test periods between late April and early September, demonstrating

the physiological plasticity within the species.
Variation among white spruce seed sources in their suitability for larval

survival and length of development (degree days) was clearly linked to their

differences in phenology.

113

114
Variation among white spruce seed sources in their suitability for total
budworm growth, growth rate and length of development (days) was clearly
linked to tree growth rate. Competition within host trees for allocation of
resources to growth processes and to maturation and defense processes may

explain the effect of tree growth rate.

Although budworm survival was much higher in April-June, budworms
developing in June-July had the same survivorship curve: more than 85% of
total larval mortality occurred by the time most surviving larvae were 3rd

instars.

When young larvae were placed on trees three to four weeks prior to
budbreak, the majority of them fed on old foliage. On the other hand, when
young larvae were placed on trees at about three weeks after budbreak, they
fed almost exclusively on current-year foliage. The latter group was observed
mining new needles in the manner typical of larvae mining old needles in

early spring.

Increased budworm performance was most strongly correlated with higher
foliar levels of water and lower levels of leaf toughness. Water content and
leaf toughness were the two foliar traits that most closely reflected the
relative differences in timing of phenology among the three white spruce seed
sources. Together, these two foliar traits represent a measure of sclerophylly
and are likely to be the key factors that define the phenological window of

susceptibility.

10.

115
Among the other foliar traits, increased budworm performance was strongly
correlated with higher levels of total nitrogen and phosphorous. Weaker
positive relationships existed between budworm performance and higher

levels of total sugars, potassium and zinc and lower levels of calcium.

In some cases the relationship between budworm performance and foliar
traits differed between foliar levels present during early larval development
and the levels of the same traits present during late larval development.

For example, budworm performance was more closely correlated with early
than late foliar levels of total nitrogen, phosphorous, zinc and toughness. On
the other hand, performance was more closely correlated with late than

early foliar levels of potassium and calcium.

APPENDIX

APPENDIX 1

Record of Deposition of Voucher Specimens*

The specimens listed on the following sheet(s) have been deposited in
the named museum(s) as samples of those species or other taxa which were
used in this research. Voucher recognition labels bearing the Voucher
No. have been attached or included in fluid-preserved specimens.

Voucher No.: 1988'03

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

Phenological Variation in the Susceptibility
of White Spruce to the Spruce Budworm

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

Other museums: None

Investigator's Name (3) (typed)

 

Robert K. Lawrence

 

 

Date 27 July 1990

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

Deposit as follows:

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

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

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

116

APPENDIX 1.1
voucher Specimen Data

1 of 1 Pages

Page

. when

a x N

auumuo>wca manum :mw«:0H: osu cu afimoaov

rrhwszAWWJ avsaamrmmurr

uwwwwno

cam“ an:1 Nu sumo

 

 

00:: awofioeoucm

coconzom .x 99030:

 

now mamswooam voumfia 0>0nm 0:0 mo>«ouuz

 

mormmmfi .02 nosu=0> AvonauV

Amvusmz m.u0umw«umo>:u

Aausumuuoc mu uuoozu ancouuuevm away

 

Roma Hana< .ouaaaao no;

“newsman.

 

 

 

 

 

 

 

 

 

 

3m: m m o« ow cuocmu .0wumpco .ownaz .opm uaaom acououfieau muaocoumfiuozo
.d 0.. 0332—00 was .000: .3 “5800300 .5qu .350 no IoaooA—m
m e r r m m e .m u macawooam you munv Huang
e r 0.0 e .1 .1 a 0. w s
an 2.. m a. .... w m .. ...”
“n w.d.1 .u .A .A n. n" r. w.

 

 

II
“It

0 Moses:

 

 

117

LITERATURE CITED

LITERATURE CITED

Agren, 6.1. 1985. Theory for growth of plants derived from the nitrogen
productivity concept. hysiol. Plant. 64: 17-28.

Albert, PJ. 1982. Host plant preferences in larvae of the eastern spruce budworm,
Chon'stoneura fumiferana. p. 19-24. Proc. 5th Int. Symp. Insect-Plant
Relationships. Pudoc, Wageningen.

Albert, PJ., C. Cearleg, F. Hanson and S. Parisella. 1982. Feedin responses of
eastern spruce udworm larvae to sucrose and other carbo ydrates. J.
Chem. Ecol. 8: 233-239.

Albert, PJ. and RA. Jerrett. 1981. Feeding preferences of spruce budworm
(Chorirtoneura fumiferana Clem.) larvae to some host-plant chemicals. J.
Chem. Ecol. 7: 391-402.

Albert, PJ. and S. Parisella. 1985. Feeding preferences of eastern spruce
budworm larvae in two-choice tests with combinations of host-plant
extracts. Ent. Exp. Appl. 38: 221-225.

Allen, J .C. 1976. A modified sine wave method for calculating degree days.
Environ. But. 5: 388-396.

Baldwin, I.T., J .C. Schultz and D. Ward. 1987. Patterns and sources of leaf tannin
variation in yellow birch (Betula alle henienszlr) and sugar maple (Acer
saccharum). J. Chem. Ecol. 13: 106 -1078.

Bauce, E. and Y. Hardy. 1988. Effects of drainage and severe defoliation on the
rawfiber content of balsam fir needles and growth of the spruce budworm
(Lepidoptera: Tortricidae). Environ. But. 17: 671-674.

Bazzaz, F.A., N.R. Chiariello, P.D. Coley and LF. Pitelka. 1987. Allocating
resources to reproduction and defense. BioSci. 37: 58-67.

Bean, J .L. 1961. Predicting emergence of second-instar spruce budworm larvae
fiom hibernation under field conditions in Minnesota. Ann. Ent. Soc.
Amer. 54: 175-177.

Bean, J .L and H0. Batzer. 1957. Mean head width for s ruce budworm larval
instars in Minnesota and associated data. J .Econ. nt. 50: 499.

Beckwith, RC. and DO. Burnell. 1982. Spring larval dispersal of the western

s ruce budworm (Le idoptera: Tortric1dae) in north-central Washington.
nviron. Bat. 11: 82 -832.

118

119

Bernard-Dagan, C. 1988. Seasonal variations in energy sources and biosynthesis of
terpenes in maritime pine. p. 93-116. In: W.J. Mattson, J. Levieux, and C.
Bernard-Dagan (eds.) Mechanisms of Woody Plant Defenses Against
Insects - Search for Pattern. Springer-Verlag, New York.

Blais, J .R. 1957. Some relationships of the s ruce budworm, Chon'stoneura
firmiferana (Clem.), to black spruce, icea mariana (Moench) Voss. For.
Chron. 33: 364-372.

Blais, J .R. 1979. Rate of defoliation of balsam fir in relation to s ruce budworm
attack and timing of spray application. Can. J. For. Res. : 354-361.

Blake, EA. and MR. Wa er. 1986. Foliage age as a factor in food utilization by
the western spruce udworm, Chon'stoneura occidentalis. Great Basin Nat.
46: 169-172.

Blum, BM. 1988. Variation in the phenology of bud flushing in white and red
spruce. Can. J. For. Res. 18: 315-319.

Bongarten, B.C. and J.W. Hanover. 1982. Hybridization among white, red, blue,
and white x blue spruces. For. Sci. 28: 129-134.

Brewer, J .W., J .L. Capinera, R.E. Deshon, Jr. and M.L. Walmsley. 1985. Influence
of foliar nitrogen levels on survival, development, and re roduction of
western spruce budworm, Choristoneura occidentalis (Lepldoptera:
Tortricidae). Can. Ent. 117: 23-32.

Brewer, J .W., K.M. O’Neill and RE. Deshon, Jr. 1987. Effects of artificially
altered foliar nitrogen levels on development and survival of youn mstars
of western spruce budworm, Choristoneura occidentalis Freeman. . Appl.
Ent. 104: 121-130.

Cameron, D.G., G.A. McDougall and CW. Bennett. 1968. Relation of spruce
ll)3udwt1))rrn8 gl7ev8eslopment and balsam fir shoot grth to heat units. J. Econ.
nt. 1: - 8.

Canavera, BS. and J. DiGennaro. 1979. Characteristics of eastern spruce gall
a hid attack among 24 white sgrruce seed sources in central Maine. Proc.
ortheast. For. Tree Improv. onf. 26: 96-101.

Carter, K.K. and J .D. Simpson. 1985. Status and outlook for tree improvement
programs in the Northeast. North. J. Appl. For. 2: 127-131.

Cates, R.G. 1987. Influence of biological rhythms, tissue development, and
Ehysiological state of plants and insects on their interactlons. p. 175-182.
: V. labeyrie, G. Fabres and D. lachaise (eds.) Insects - Plants. Dr. W.
Junk Publ., Dordrecht, Netherlands. .

Cates, R.G., C.B. Henderson and RA. Redak. 1987. Responses of the western
gagugti‘sbudworm to varying levels of nitrogen and terpenes. Oecologia 73:

120

Cates, R.G. and RA. Redak. 1988. Variation in the terpene chemist?I of Douglas-
fir and its relationship to western s ruce budworm success. p. 17-344.
In: K.C. Spencer (e .) Chemical ediation of Coevolution. Acad. Press,
San Diego,-Calif.

Cates, R.G., R. Redak and CB. Henderson. 1983a. Natural product defensive
chemistry of Douglas-fir, western spruce budworm success, and forest
management practices. Z. Ang. Ent. 96: 173-182.

Cares, R.G., R. Redak and CB. Henderson. 1983b. Patterns in defensive natural
product chemistry: Douglas-fir and western spruce budworm interactions.
. 3-1‘9‘.’ IE: IDAC Hedin (ed.) Plant Resistance to Insects. Amer. Chem.
0., as ., . .

Chapin, PS, 111, and RA Kedrowski. 1983. Seasonal changes in nitrogen and
phos horous fractions and autumn retranslocation in evergreen and
deci uous taiga trees. Ecology 64: 376-391.

Chapin, F .S., III, J .D. McKendrick, and DA. Johnson. 1986. Seasonal changes in
carbon fractions in Alaskan tundra plants of differing growth form:
Implications for herbivory. J. Ecol. 74: 707-731.

Chrosciewicz, Z. 1986. Foliar moisture content variations in four coniferous tree
species of central Alberta. Can. J. For. Res. 16: 157-162.

Chung, H. and R.L. Barnes. 1980. Photosynthate allocation in Pinus taeda. 11.
Seasonal aspects of photosynthate allocation to different biochemical
fractions in shoots. Can. J. For. Res. 10: 338-347.

Clancy, K.M., M.R. Wagner and R.W. Tinus. 1988a. Variation in host foliage
nutrient concentrations in relation to western spruce budworm herbivory.
Can. J. For. Res. 18: 530-539.

Clancy, K.M., M.R. W er and R.W. Tinus. 1988b. Variation in nutrient levels as
a defense: Iden ' ' 5 key nutritional traits of host lants of the western
s ruce budworm. g. 03-213. In: WJ. Mattson, J. vieux, and C. Bernard-
agan (eds. Mec anisms of Woody Plant Defenses Against Insects -
Search or attern. Springer-Verlag, New York.

Clark, E.W., J .D. White and E.L. Bradle . 1975. Seasonal changes in foliar fatty
acids and sterols in Carya glabra Mill.) Sweet and Quencus falcata Michx.,
370 1105383 gfg 41336 elm spanworm, Ennomos subsignarius (Hiibner). Environ.
nt. 2 - .

Cleary, B.D. and RH. Waring. 1969. Temperature: collection of data and its
finalyziis flog7t11133interpretation of plant growth and distribution. Can. J.
at. : - .

Cole, J .W.L. and J .E. Grizzle. 1966. Applications of multivariate analysis of
variance to repeated measurements experiments. Biometrics 22: 810-828.

Coley, RD. 1983. Herbivory and defensive characteristics of tree species in a
lowland tropical forest. Ecol. Monogr. 53: 209-233.

121

Coley, RD. 1987. Between-species differences in leaf defenses of tropical trees.
. 30-35. In: F.D. Provenza, J.T. Flinders and ED. McArthur (eds.) Proc.
gympos. on Plant-Herbivore Interactions. USDA For. Serv. Gen. Tech.
Rept. INT -222.

Coley, RD. 1988. Effects of plant growth rate and leaf lifetime on the amount and
type of anti-herbivore defense. Oecologia 74: 531-536.

Coley, P.D., J .P. Bryant, F.S. Chapin, III. 1985. Resource availability and plant
antiherbivore defense. Science 230: 895-899.

Crummey, HR. and LS. Otvos. 1980. Biolo and habits of the eastern spruce
budworm, Choristoneura fumiferana ido tera: Tortricidae) in
Newfoundland. Environ. Canada, Ne mm and For. Res. Ctr. Info. Rept.
N-X-181, 33p.

Danford, M.B., H.M. Hughes and RC. McNee. 1960. On the analysis of repeated-
measurements experiments. Biometrics 16: 547-565.

Davidson, M.L. 1972. Univariate versus multivariate tests in repeated-measures
experiments. Psychol. Bull. 77: 446-452.

Day, K. 1984. S stematic differences in the population density of green spruce
aphid, E obium abietinum, in a rovenance trial of Sitka spruce, Picea
srtchensis. Ann. Appl. Biol. 105: 05-412.

Du Merle, P. 1988. Phenological resistance of oaks to the een oak leafroller,
Tortrix viridana (Le idoptera: Tortricidae). p. 215- 6. In: W.J. Mattson,
J. Levieux, and C. mard-Dagan (eds.) Mechanisms of Woody Plant
Defenses Against Insects - Search for Pattern. Springer-Verlag, New York.

Durzan, DJ. 1968. Nitrogen metabolism of Picea glauca. 1. Seasonal chan es of
free amino acids in buds, shoot apices, and leaves, and the metabo ' m of
uniforml labelled 14C-L-arginine by buds during the onset of dormancy.
Can. J. 0t. 46: 909-919.

Durzan, DJ. and SM. Lopushanski. 1968. Free and bound amino acids of s ruce
budworm larvae feeding on balsam fir and red and white spruce. J. ect
Physiol. 14: 1485-1497.

Eidt, DC. and MD. Cameron. 1971. Dela ed budbreak and spruce budworm
survival. Can. For. Serv. Bi-Mon. es. Notes 27(4): 28-29.

Eidt, DC and C.H.A. Little. 1970. Insect control through induced host-insect
asynchrony: A progress report. J. Econ. But. 63: 1966-1968.

Embree, DO. 1965. The opulation dynamics of the winter moth in Nova Scotia,
1954-1962. Mem. nt. Soc. Can. 46, 57p.

Ericsson, A. 1978. Seasonal changes in translocation of 14C from different e-
classes of needles on 20-year-old Scots pine trees (Pinus silvestris). P ysiol.
Plant. 43: 351-358.

122

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

F eeny, P. 1976. Plant apparency and chemical defense. Rec. Adv. Phytochem. 10:
1-40.

Fowells, HA. 1965. Silvics of forest trees of the United States. USDA For. Serv.
Agric. Hdbk. 271, 762p.

Fowler, DP. 1986. Strategies for the genetic improvement of important tree
species in the Mantimes. Can. For. Serv. - Maritimes, Info. Rept.
M-X- 156, 30p.

 

Fowler, DP. and E.K. Morgenstern. 1990. Tree im rovement research and
development in Canada. For. Chron. 66: 9 -101.

Genys, J .B. and H. Nienstaedt. 1979. Variation in white spruce from 24 different
rovenances studied in M land’s piedmont plateau. Proc. Northeast. For.
ree Improv. Conf. 26: 48- 7.

Greenbank, DO. 1956. The role of climate and dispersal in the initiation of
outbreaks of the s ruce budworm in New Brunswick. 1. The role of climate.
Can. J. Zool. 34: 53-476.

Greenbank, DO. 1963. Host 5 cies and the spruce budworm. p. 219-223. In:
R.F. Morris (ed.) The amics of Epidemic Spruce Budworm
Populations. Mem. Ent. Soc. Can., No. 31.

Grigal, D.F., L.F. Ohmann and RB. Brander. 1976. Seasonal dynamics of tall
shrubs in northeastern Minnesota: Biomass and nutrient element changes.
Forest Sci. 22: 195-208.

Haack, R.A., R.K. Lawrence and WJ. Mattson. 1991. Impact of root pruning on
white spruce and its herbivore the spruce budworm. In peer review.

Hagen, RH. and J .F . Chabot. 1986. Leaf anatomy of maples (Acer) and host use
by Lepidoptera larvae. Oikos 47: 335-345.

Hanover, J .W. and DA. Reicosky. 1971. Surface wax deposits on foliage of Picea
pungens and other conifers. Amer. J. Bot. 58: 681-687.

Hard, 8501985. Spruce beetles attack slowly growing spruce. For. Sci. 31: 839-

Hardy, Y., M. Mainville and D.M. Schmidt. 1986. An atlas of s ruce budworm
defoliation in eastern North America, 1938-80. USDA or. Serv. Misc.
Pub. 1449, 52p.

Harvey, G.T. 1958. A relationship between photoperiod and cold-storage
treatment in the spruce budworm. Science 128: 1205- 1206.

Harvey, G.T. 1974. Nutritional studies of eastern s86ruce budworm (Lepidoptera:
Tortricidae) I. Soluble sugars. Can. But. 1 : 353-365.

123

Harvey, G.T. 1985. The taxonomy of the conifertgziragous Choristoneura
glepido tera: Tortricidae): A review. p. 1 8. In: C]. Sanders, R.W.
tar .J. Mullins and J. Murph (eds.) Recent Advances in S ruce
Budworms Research, Proc. CANIJSA Spruce Budworms Res. ympos.

Haukioja, E., P. Niemela, L. Iso-Iivari, H. Ojala and E. Aro. 1978. Birch leaves as
a resource for herbivores. 1. Variation in the suitability of leaves. Rep.
Kevo Subarctic Res. Stat. 14: 5-12.

Henson, W.R. 1948. Influence of meteorological factors on behaviour of the
s ruce budworm. Domin. Dept. Agnc.; Sci. Serv.-Div. Ent; Ottawa,
da; For. Insect Invest. Bi-Mon. Prog. Rept. 4(6): 2.

Heron, RJ. 1965. The role of chemotactic stimuli in the feedin behavior of
spruce budworm larvae on white spruce. Can. J. Zool. 3: 247-269.

Hinckley, T.M., J .P. lassoie and SW. Running. 1978. Temporal and spatial
variations in the water status of forest trees. For. Sci. Monogr. 20, 72p.

Hockman, J .N ., J .A. Burger and D.W. Smith. 1989. Spatial and temporal
gastrigggh 3 51f foliar nutrient levels in Fraser fir Christmas trees. For. Sci.

Huang, P. 1975. Relationship between the needle structure of Picea excelsa and
51:9 behaviour of some needle eating insect species. Z. Ang. Ent. 77: 264-

Huang, P. and E. Fuhrer. 1979. As to food quality of spruce needles for forest
dama ' g insects. 12. Variation in epidermis and hypodermis structure. Z.
Ang. nt. 88: 231-245.

Khalil, M.A.K. 1985. Genetic variation in eastern white sgxuce (Picea glauca
(Moench) Voss) populations. Can. J. For. Res. 1 : 444-452.

Khalil, M.A.K. 1986. Variation in seed qualiz and some juvenile characters of
white spruce (Picea glauca (Moench) oss). Silvae Gen. 35: 78-85.

Kramer, PJ. and T.T. Kozlowski. 1979. Physiology of woody plants. Acad. Press,
Orlando, Fla., 811p.

Krischik, VA. and RF. Denno. 1983. Individual, Iptg‘pulation, and Egogra hic
patterns in plant defense. p. 463-512. In: . . Denno and .S. cClure
(eds. Variable Plants and Herbivores in Natural and Managed Systems.
Aca . Press, New York.

IaTour, SA. and PW. Miniard. 1983. The misuse of repeated measures analysis
in marketing research. J. Marketing Res. 20: 45-57.

Lavender, D.P., G.B. Sweet, J .B. Zaerr and R.K. Hermann. 1973. Spring shoot
growth in Douglas-fir ma be initiated by gibberellins exported fiom the
roots. Science 182: 838- 39.

124

Lawton, J.H. and S. McNeil]. 1979. Between the devil and the dee blue sea: On
the problem of being a herbivore. Symp. Brit. Ecol. Soc. 2 : 223-244.

Leather, SR. 1986. The effect of neonatal starvation on the growth, development
and survival of larvae of the pine beauty moth, Panolis flammea (D. & S.).
Oecologia 71: 90-93.

Lechowicz, MJ. 1983. Leaf quality and the host preferences of moth in the
northern deciduous forest. . 67-82. In: R.L. Talerico an . Montgomery
(eds.) Forest Defoliator -- ost Interactions: A Comparison Between
Eggs Moth and Spruce Budworms. USDA For. Serv. Gen. Tech. Rept.

Little, C.H.A. 1970. Seasonal changes in carbohydrate and moisture content in
needles of balsam fir (Abies balsamea). Can. J. Bot. 48: 2021-2028.

Loach, K. and C.H.A. Little. 1973. Production, storage, and use of photosynthate
clitligilng sérsoot elongation in balsam fir (Abies balsamea). Can. J. Bot. 51:
-11 .

Loehle, C. 1988. Tree life history strategies: the role of defenses. Can. J. For.
Res. 18: 209-222.

Loehle, C. and G. Namkoong. 1987. Constraints on tree breeding: Growth
tlrggdleoffs, growth strategies, and defensive investments. For. Sci. 33: 1089-

l.orio, P.L., Jr. 1986. Growth-differentiation balance: A basis for understanding
southern pine beetle - tree interactions. For. Ecol. Mgmt. 14: 259-273.

Lorio, P.L., Jr. 1988. Growth differentiation-balance relationships in pines affect
their resistance to bark beetles (Coleoptera: Scol 'dae). p. 73-92. In: W.J.
Mattson, J. Levieux, and C. Bernard-Dagan eds. Mechanisms of Woody
Plan]: Defenses Against Insects - Search for attern. Springer-Verlag, New

or

Lowman, MD. and J .D. Box. 1983. Variation in leaf toughness and phenolic
«ignztsent among five species of Australian rain forest trees. Aust. J. Ecol. 8:

MacLean, DA. 1985. Efiects of s ruce budworm outbreaks on forest growth and
yield. I! 148-175. In: C.J. anders, R.W. Stark, EJ. Mullins and J. Murphy
éeds.) ecent Advances in Spruce Budworms Research, Proc. CANU SA

pruce Budworms Res. Sympos.

Matson, P.A., F.P. Hain and W. Mawby. 1987. Indices of tree susceptibility to
bark beetles vary with silvicultural treatment in a loblolly pine plantation.
For. Ecol. Mgmt. 22: 107-118.

Mattson, WJ. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev.
Ecol. Syst. 11: 119-161.

125

Mattson, W.J., R.A. Haack, R.K. lawrence and SS. Slocum. 1990. Considering
the nutritional ecology of the spruce budworm in its management. For.
Ecol. Mgmt., In Press.

Mattson, WJ. and C.N. Koller. 1983. Spruce budworm performance in relation to
matching selected chemical traits of its hosts. p. 138-148. In: A.S. Isaev
l(:ed.) The Role of Insect-Plant Relationship in the Population Dynamics of

orest Pests. Proc. Inter. Sympos., Irkutsk, U.S.S.R.

Mattson, W.J., N. Lorimer and RA. Leary. 1982. Role of plant variabilig gtrait
vector dynamics and diversity) in plant/herbivore interactions. p. 9 -303.
In: H.M. Heybroek, B.R. Stephan and K. von Weissenberg (eds.)
Resistance to Diseases and ests in Forest Trees. Pudoc, Wageningen.

- Mattson, WJ. and SR. Palmer. 1988. Changes in levels of foliar minerals and
phenolics in tremblin aslplen, Popular tremuloides, in response to artificial
defoliation. . 157-16 . : W.J. Mattson, J. Levieux, and C. Bernard-
Dagan (eds. Mechanisms of Woody Plant Defenses Against Insects -
Search or Pattern. Springer-Verlag, New York.

Mattson, WJ. and J .M. Scriber. 1987. Nutritional ecolo of insect folivores of
woody plants: Nitro en, water, fiber, and miner considerations. . 105-
146. In: F. Slansky r. and JG. Rodriguez (eds.) Nutritional Eco ogy of
insects, hitches, Spiders, and Related Invertebrates. John Wiley & Sons,

ew or

Mattson, W.J., G.A. Simmons and J .A. Witter. 1988. The spruce budworm in
eastern North America. p. 309-330. In: A.A. Berryman (ed.) Dynamics of
{lorestYIniect Populations: Patterns, Causes, Implications. Plenum Press,

ew or

Mattson, WJ., S.S. Slocum and C.N. Koller. 1983. S ruce budworm
(Chonlrtoneura gerana) performance in re ation to foliar chemistry of its
host lants. p. 5-6 . In: R.L. Talerico and M. Montgomery (eds.) Forest
Defo iator - Host Interactions: A Comparison Between G Moth and
Spruce Budworms. USDA For. Serv. Gen. Tech. Rept. NE- 5.

McGugan, BM. 1954. Needle-mining habits and larval instars of the spruce
budworm. Can. Eat. 86: 439-454.

McIntyre, G.I. 1987. The role of water in the regulation of plant development.
Can. J. Bot. 65: 1287-1298.

McKeyég. 1974. Adaptive patterns in alkaloid physiolog. Amer. Nat. 108: 305-

Mclaufii’lnin, BM. 1986. Performance of the spruce budworm, Chon’stoneura
ifemna, in relation to dietary and foliar levels of sugar and nitrogen.
MS. Thesis, Michigan State University, 86p.

126

McLean, J.A., P. laks and TL. Shore. 1983. Comparisons of elemental profiles of
the western s ruce budworm reared on three host foliages and artificial
medium. p. 3 -40. In: R.L. Talerico and M. Montgomery (eds.) Forest
Defoliator - Host Interactions: A Comparison Between G sy Moth and
Spruce Budworms. USDA For. Serv. Gen. Tech. Rept. NE- 5.

McMorran, AR. 1965. A synthetic diet for the spruce budworm, Chon’stoneura
fumiferana (Clem.) (Lepidoptera: Tortric1dae). Can. Ent. 97: 58-62.

McNeil], S. and T.R.E. Southwood. 1978. The role of nitrogen in the development
of insect/plant relationships. p. 77-98. In: J.B. Harborne (ed.) Biochemical
Aspects 0 Plant and Animal Coevolution. Acad. Press., London.

Miller, CA. 1957. A technique for estimating the fecundity 0f natural populations
of the spruce budworm. Can. J. Zool. 35: 1-13.

Miller, CA, D.C. Eidt and G.A. McDougall. 1971. Predictin spruce budworm
development. Canad. For. Serv. Bi-mo. Res. Notes 2 : 33-34.

Mitchell, R.G., R.H. Waring and GB. Pitrnan. 1983. Thinning lodgegole ine
goat-Eases tree vigor and resistance to mountain pine beetle. or. '. 29:
11.

Montgomery, ME. and M.M. Czapowskyj. 1985. Nitro en as an indicator of
foliage quality to the s ruce budworm. . 494-4 5. In: C.J. Sanders, R.W.
Stark, EJ. Mullins an J. Murph geds. Recent Advances in S ruce
Budworms Research, Proc. CANIJ A pruce Budworms Res. ympos.

Mooney, H.A., S.L. Gulmon and N .D. Johnson. 1983. Physiolo 'cal constraints 0n
lant chemical defenses. p. 21-36. In: RA. Hedin (ed.) lant Resistance to
ects. Amer. Chem. Soc., Wash. DC.

Morris, RF. and CA. Miller. 1954. The development of life tables for the spruce
budworm. Can. J. Zool. 32: 283-301.

Morrison, DR 1976. Multivariate statistical methods. 2nd. ed. McGraw-Hill, New
York, 415p.

Nienstaedt, H. 1963. The establishment of white spruce seed sources from the
entire range of the species. USDA For. Serv., Unpublished Report.

N ienstaedt, H. 1969. White spruce seed source variation and ada tation to 14
lanting sites in northeastern United States and Canada. oc. 11th
eeting Comm. For. Tree Breeding in Canada, p. 183-194.

N ienstaedt, H. 1974. Genetic variations in some phenological characteristics of
forest trees. p. 389-400. In: H. Lieth (ed.) Phenology and Seasonality
Modeling. Springer-Verlag, New York.

Nienstaedt, H. 1985. Inheritance and correlations of frost injury, growth, flowering,
and cone characteristics in white spruce, Picea glauca (Moench) Voss.
Can. J. For. Res. 15: 498-504.

127

Nienstaedt, H. and H. Kang. 1987. Establishin a Picea glauca Moench) Voss
base breeding population for the lake tates regron of e United States.
Silvae Gen. 36: 21-30.

Nienstaedt, H. and J .P. King. 1970. Breeding for delayed budbreak in Picea glauca
Moench Voss - Potential frost avordance and growth gains. Proc.
A0/ 0 2nd. World Consult. For. Tree Breeding, p. 61-80.

Nienstaedt, H. and A. Teich. 1972. The genetics of white spruce. USDA For.
Serv. Res. Pap. WO-15, 24p.

Osawa, A., C.J. Spies and J .B. Dimond. 1986. Patterns of tree mortali during an
uncontrolled s ruce budworm outbreak in Baxter State Park, 1 83. Maine
Agric. Exper. tn. Tech. Bull. 121, 69p.

Price, PW. 1984. Insect ecology. 2nd. ed. John Wiley, New York, 607 p.

Rao, CR. 1973. Linear statistical inference and its applications. 2nd. ed. John
Wiley & Sons, New York, 625p.

Raske, AG. 1985. Colla sing budworm pulations. p. 141-142. In: C.J. Sanders,
R.W. Stark, EJ. ullms and J. urphg (eds.) Recent Advances in Spruce
Budworms Research, Proc. CANU SA pruce Budworms Res. Sympos.

Raupp, MJ. 1985. Effects of leaf toughness on mandibular wear of the leaf beetle,
Plagiodera versicolora. Ecol. Ent. 10: 73-79.

Raupp, MJ. and RF. Denno. 1983. Leaf a e as a predictor of herbivore
distribution and abundance. .91-12 . In: R.F. Denno and MS. McClure
(eds. Variable Plants and erbivores in Natural and Managed Systems.
Aca . Press, New York.

Rauter, RM. 1985. Tree improvement policy guidelines for Ontario with s cial
emphasis on white and black spruce. p. 193-202. In: D. Schmitt (ed),e
S ruce-fir Management and Spruce Budworm. USDA For. Serv.
ortheastern For. Exper. Stn., Gen. Tech. Rept. NE-99.

Redak, RA. and R.G. Cates. 1984. Douglas-fir (Pseudotsuga menziesii)-spruce
budworm (Charistoneura occidentalis) interactions: the effect of nutrition,
chemical defenses, tissue phenology, and tree physical parameters on
budworm success. Oecologia 62: 61-67.

Regniere, J. 1987. Temperature-dépendent development of eggs and larvae of
Chonlstoneura fitmiferana ( em.) (Lepido tera: Tortric1dae) and simulation
of its seasonal history. Can. Ent. 119: 71 -728.

Retnakaran, A. 1983. Spectrophotometric determination of larval ingestion rates
in the spruce budworm (Lepidoptera: Tortricidae). Can. Ent. 115: 31-40.

Roche, L. 1969. A genecolo 'cal study of the genus Picea in British Columbia.
New Phytol. 68: 505- 54.

128

Rose, AH. and JR. Blais. 1954. A relation between April and May temperatures
and spruce budworm larval emergence. Can. Ent. 86: 174-177.

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

SAS Institute Inc. 1988. SAS/ STAT user’s guide, Release 6.03 ed. SAS Institute
Inc.; Cary, NC. 1028p.

Schiinherr, J. 1980. Dispersal of the European spruce budworm, Chon'stoneura
murinana Hb. (Lepidoptera: Tortricidae). p. 128-133. In: A.A. Berryman
and L. Safianyik (eds.) Proc. 2nd. IUFRO Conf. on Dispersal of Forest
Insects: Evaluation, Theory and Management Implications.

Scriber, J .M. 1977. Limiting effects of low leaf-water content on the nitrogen
utilization, energy budget, and larval grth of Hyalophora cecropia
(Lepidoptera: Saturniidae). Oecologia 28: 269-287.

Scriber, J.M. 1978. The effects of larval feeding s ecialization and plant growth
form on the consumption and utilization 0 plant biomass and nitrogen: an
ecological consideration. Ent. Exp. Appl. 2 : 694-710.

Scriber, J .M. 1984a. Nitrogen nutrition of plants and insect invasion. p. 441-460.
In: R.D. Hauck (e .) Nitrogen in Crop Production. Amer. Soc. Agron.,
Madison, Wisconsin.

Scriber, J .M. 1984b. Host-plant suitabili . p. 159-202. In: W.J. Bell and R.T.
Starde (eds.) Chemical Ecology o Insects. Sinauer Assoc.; Sunderland,
ass.

 

Scriber, J .M. and F. Slansky Jr. 1981. The nutritional ecology of immature insects.
Ann. Rev. Ent. 26: 183-211.

Shaw, G.G. and C.H.A. Little. 1973. Dispersal of second instar spruce budworm.
Canad. For. Serv. Bi-mo. Res. Notes 29: 30-31.

Sheehan, K.A., W.P. Kemp, J .J . Colbert and N.L. Crookston. 1989. The western
s ruce budworm model: Structure and content. USDA For. Serv. Pacific
orthwest Res. Stn., Gen. Tech. Rept. PNW-GTR-241, 70p.

Shepard, RF. 1985. A theory on the effects of diverse host-climatic environments
in British Columbia on the dynamics of western s ruce budworm. p. 60-70.
In: C.J. Sanders, R.W. Stark, E.J. Mullms' and J. Ath (eds.) Recent
Advances in Spruce Budworms Research, Proc. C gA Spruce
Budworms Res. Sympos.

Simmons, G.A., W. Cuff, B.A. Monggomery and J .M. Hardman. 1984. Projection
Methods. p. 142-165. In: D. . Schmitt, D.G. Grimble and IL. Searcy
(Tech. Coord. Spruce Budworms Handbook - Managing the S ruce
Budworm in astem North America. USDA For. Serv., Agr. dbk. 620.

129

Smith, KG. 1981. Canonical correlation analysis and its use in wildlife habitat
studies. p. 80-91. In: D.E. Capen (ed.) The Use of Multi-variate Statistics
in Studies of Wildlife Habitat. USDA For. Serv. Gen. Tech. Rept. RM-87.

Strong, D.R., J.H. lawton and R. Southwood. 1984. Insects on plants -
Icgmmgmgy patterns and mechanisms. Harvard Univ. Press, Cambridge,
ass. 1 p.

Sullivan, MJ. and CA. Moncreiff. 1988. A multivariate analysis of diatom
community structure and distribution in a Mississippi salt marsh. Bot.
Marina 31: 93-99.

Thomas, AW. 1976. Effects of temperature on emer ence of second-instar spruce
$§zvgam8larvae Can. For. Serv., Maritimes or. Res. Ctr., Info. Rept.
- - ’ p.

Thomas, AW. 1983. Folia e consumption by 6th-instar spruce budworm larvae,
Chon'stoneura ierana (Clem. , feeding on balsam fir and white spruce.
. 47-48. In: .1. alerico and . Montgomery (eds.) Forest
efoliator - Host Interactions: A Comparison Between Gypsy Moth and
Spruce Budworms. USDA For. Serv. Gen. Tech. Rept. NE-85.

Thomas, AW. 1987. The effect of age of current- ear shoots of Picea glauca on
survival, development time, and feeding .e cien of 6th-instar larvae of

Chon'stoneura fizmiferana. Ent. Exp. Appl. 43: 1-260.
Thomas, AW. 1989. Food consumption and utilization by 6th-instar larvae of

. spruce budworm, Chonlrtoneura fumzfemna: a comparison on three Picea
(spruce) species. Ent. Exp. Appl. 52: 205-214.

Thomson, A.J., R.F. Shepard, J .W.E. Harris and R.H. Silversides. 1984. Relating
weather to outbreaks of western spruce budworm, Chon'stoneura
gggicggntalis (Lepidoptera: Tortricidae), in British Columbia. Can. Ent. 116:

- 1.

Turgeon, J .J . 1986. The henological relationship between the larval development
of the spruce bu 0th, Zeimphera canadensis (Lepidoptera:
Ollgtlgr‘tesugiglge), and white spruce in northern New Brunswick. Can. Ent.

1 : - .

Vincent, J.F.V. 1982. The mechanical design of grass. J. Mater. Sci. 17: 856-860.

Volney, W.J.A. 1985. Comparative population biolo °es of North American spruce
budwormsdg. 71-84. In: C.J. Sanders, R.W. tark, EJ. Mullins and J.
M h (e .) Recent Advances in Spruce Budworms Research, Proc.
C SA Spruce Budworms Res. Sympos.

Volney, W.J.A., W.E. Waters, R.P. Akers and AM. Liebhold. 1983. Variation in
6 ring emergence patterns among western Chon'stoneura spjsa. (Lepidoptera:
ortrrcidae) populations in southern Oregon. Can. Ent. 11 : 199-209.

130

von Rudloff, E. 1972. Seasonal variation in the composition of the volatile oil
(5)f the leaves, buds, and twigs of white spruce (Picea glauca). Can. J. Bot.
0: 1595-1603.

von Rudloff, E. 1975a. Seasonal variation in the terpenes of the foliage of black
spruce. Phytochem. 14: 1695-1699.

von Rudloff, E. 1975b. Seasonal variation of the terpenes of the leaves, buds, and
twigs of blue spruce (Picea pungens). Can. J. Bot. 53: 2978-2982.

von Rudloff, E. and M. Granat. 1982. Seasonal variation of the terpenes of the
128638V2682, létsids, and twigs of balsam fir (Abies balsamea). Can. J. Bot. 60:
- 6 .

von Schiinborn, A. 1966. The breediérbg of insect-resistant forest trees in central
and northwestern Europe. p. -27. In: H.D. Gerhold, R.E. McDermott,
E.J. Schreiner and J .A. Winieski (eds.) Breeding Pest-Resistant Trees.
Pergamon Press, London.

Wagner, MR. and EA. Blake. 1983. Western spruce budworm consum tion --
Effects of host species and foliage chemistry. p. 49-54. In: R.L. alerico
and M. Mont gomery (eds.) Forest Defoliator - Host Interactions: A
Comparison Ietween 6&5), Moth and Spruce Budworms. USDA For.
Serv. Gen. Tech. Rept. -85.

White, T.C.R. 1984. The abundance of invertebrate herbivores in relation to the
availability of nitrogen in stressed food plants. Oecologia 63: 90-105.

 

Whitham, T.G. 1981. Individual trees as heterogeneous environments: Adaptation
to herbivory or ep‘ifinetic noise? p.9-27. In: R.F. Denno and H. Dingle
geds.) Insect Life ' tory Patterns - Habitat and Geographic Variation.

pringer-Verlag, New York.

Whitham, T.G. 1983. Host manipulation of parasites: Within-plant variation as a
defense a ainst rapidly evolving pests. p.15-41. In: R.F. Denno and M.S.
McClure Eeds.) Variable Plants and Herbivores in Natural and Managed
Systems. Acad. Press, New York.

Wickman, BE. 1976. Phenology of white fir and Douglas-fir tussock moth egg
hatch and larval development in California. Environ. Ent. 5: 316-322.

Wilkinson, R.C. 1977a. Inheritance of budbreak and correlation with early height
owth in white s ruce (Picea glauca) from New England. USDA For.
erv. Res. Pap. -391.

Wilkinson, R.C. 1977b. Height growth performance of white spruce (Picea glauca
(Moench) Voss) provenances in central Maine. Proc. Northeast. For. Tree
Improv. Conf. 25: 87-93.

Wilkinson, R.G., J .W. Hanover, J.W. Wright and R.H. Flake. 1971. Genetic
1813ari9a)tion in the monoterpene composition of white spruce. For. Sci. 17:

131

Winer, BJ. 1971. Statistical principles in experimental design. 2nd. ed. McGraw-
Hill, New York, 907p.

Witter, J.A. and LA. Waisanen. 1978. The effect of differential flushing times
among trembling as en clones on tortricid caterpillar populations.
Environ. Ent. 7: 13 -143.

Wright, J.W., H. Nienstaedt, W.A. Lemmien, J.N.Bright, M.W. Da and R.L.
Sajdak. 1977. Better white spruce for Michigan. Michigan tate Univ.
Agric. Exper. Stn. Res. Rept. 316, 8p.