“)4; , J "‘ ‘ ‘ , . W .
V v i L. V ‘-‘ I ‘l' I ‘
' ‘ . . V , _ _ ~ '” f ’1‘

u
i‘as‘ V dflE—r ‘ :‘ x. nut—J:

“.LL -_ 4.: ,{fi'ig'x‘ . tag-:‘ng‘cfi 1:3.
1%"? ‘ . 4

“w:

"m ':
w“ 'ijfi',"

;
I

." ~<
~Q ‘V‘
v. f'"

 

w“.
111:3." «k .5‘

v6 1"

M.‘ _ 5n}. .4

u
M

'nztz‘flr :1. 5m."

C. “,

 

A 1. «1.4.01:

lvl‘

v v‘ x ‘
l | -. . ‘ ' ' ' ‘ . , . ~ I ~ Q, ‘ - .,..'..r
' ‘. ~ _

I

I." '0 f‘

am“
r31';‘ym;w

.. 1" I
3. 'r';‘r ’3' t-
"" "DJ-é;-

‘ 45'

..,>,-wq,.- ..‘- ‘-.

”a? "51.355 ~
#15:“; :fl— ‘ '7

”w

.u..‘\

 

\
UNIVERSITY UBRARIES f" A ‘ I _ I

 

 

 

 

I

 

 

 

L

 

L‘L‘L

 

ILL

 

 

L

 

 

LLLLL‘" LLLLLLLLLLLLLL LL

3 1293 00910

This is to certify that the

dissertation entitled

AN EVALUATION OF FORMICA EXSECTOI DES FOREL AS A
POTENTIAL BIOLOGICAL CONTROL AGENT OF INSECT PESTS
OF PINES

presented by

Nancy J. Campbell

has been accepted towards fulfillment

 

 

 

 

 

of the requirements for
PI) . D de in En tomol ogg
Major professor
Date 5/ 3] /9 0 (

MS U is an Affirmative Action/Equal Opportunity Institution 0-12'771

 

 

LBBRARY
Michigan State
; University

 

 

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 _
:jL___:_
fl

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
cmimmpma-pd

 

AN EVALUATION OF FORMICA EXSECTOIDES FOREL AS A POTENTIAL
BIOLOGICAL CONTROL AGENT OF INSECT PESTS OF PINES

By

Nancy J. Campbell

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Entomology

1990

ABSTRACT

AN EVALUATION OF E. EXSECTOIDES FOREL AS A POTENTIAL

 

BIOLOGICAL CONTROL AGENT 0F INSECT PESTS 0F PINES

BY

Nancy J. Campbell

This study was designed to explore the potential of utilizing

the predaceous ant, Formica exsectoides Forel as a biological

 

control agent of insect pests of young pine plantations.
The effects of ants on the redheaded pine sawfly, jack pine
budworl, gypsy noth and white pine weevil were examined. 2.

exsectoides readily preyed upon redheaded pine sawflies that were

 

placed on small pine trees around ant lounds. The ants responded
differentially to three instars of sawfly larvae.

g. exsectoides removed both second- and fourth-instar jack
pine budworn fron experimental trees. The budworn's silk chamber
decreased ant predation. The ants also reduced the population of
budworl larvae that were naturally occurring in nature Jack pine
stands. The amount of defoliation of jack pines in both the
overstory and understory decreased as the number of ant mounds per

hectare increased.

E. exsectoides preyed upon gypsy moth larvae, pupae, and

 

adults but did not prey on eggs. Increasing the distance of pupae
from the ant mound negatively affected ant predation.

The number of adult white pine weevils was reduced by ants
foraging on small jack pine trees. The ants indirectly influenced
adult weevils by disrupting their feeding behavior, causing the
weevils to move to new feeding locations on the trees. The
presence of aphids increased ant predation for all of the pests
examined.

A pine-trefoil-ant simulation model was developed that
predicted defoliation of pine trees in the presence and absence of
ants. Model simulations showed that as the number of ant colonies
per hectare increased, the amount of defoliation decreased. The
model also showed that as the soil type improved, the maximum
defoliation decreased independent of the number of ant colonies
per hectare.

Portable ant nests were designed and field tested in a young
stand of red pines located approximately 150 miles from the parent
colonies. Several of the colonies survived in the artificial

nests for over six weeks.

ACKNOWLEDGEMENTS

I am particularly indebted to Gary Simmons for his confidence. generosity and
support. I am also grateful to Cathy Bristow and George Ayers for the
encouragement and assistance they provided during my doctoral program.

I am also grateful to Louis Wilson and Leah Bauer for serving on my guidance
committee.

In addition I wish to acknowledge Stuart Gage for his assistance in developing
the simulation model. The simulation was part of the requirement for the
Systems Science course, 843. In particular I wish to acknowledge several
students from my class team who made the development of the model possible:
Jim Seyler, Ruth Shultz, Kate Thompson, Angela Vandemehr and Martha
Whitaker.

ii

TABLE OF CONTENTS

GENERAL THESIS INTRODUCTION 1
LITERATURE REVIEW 5
Relations between ants and aphids and their effects on trees 10
Stage l-Identifying the pest complex 15
Life cycle, impact and control of the gypsy moth 16
Life cycle, impact and control of the jack pine budworm 17
Life cycle, impact and control of the redheaded pine sawfly 19
Life cycle, impact and control of the white pine weevil 20
Stage 2-Identifying the desirable ant species 22
The life history of F. exsectoides 24
Growth of mounds of F. exsectoides 25
Relationship with Homoptera and other insects 26
Literature Cited 28
MANUSCRIPT l. PREDATION 0F FORMICA EXSECTOIDES 0N FOUR PINE 36
PESTS IN MICHGAN
Introduction 37
Materials and Methods 40
Experiment 1. Ant predation on the redheaded pine sawfly 44
Experiment 2. Ant predation on the jack pine budworm 45
Experiment 3. Ant predation on the gypsy moth 49
Experiment 4. Ant predation on white pine weevil 51
Results 5 3
Experiment 1. Ant predation on the redheaded pine sawfly 5 3
Experiment 2. Ant predation on the jack pine budworm 62
Experiment 3. Ant predation on the gypsy moth 75
Experimental 4. Ant predation on the white pine weevil 76
Discussion 86
Ant predation on the larvae of RHPS, IPBW and GM 86
Ant predation on JPBW and GM pupae 92
Ant predation of adult GM and WPW 95
The influence of aphids on ant predation 99
Literature Cited 104
MANUSCRIPT 2. A SCOTS PINE GROWTH MODEL; AFOCUS ON ECOLOGICAL 107
MANAGEMENT STRATEGIES
Introduction 108

System description 108

iii

Temperature subroutine 109

Tree subroutine 111
Trefoil subroutine 118
Nitrogen subroutine 121
Redheaded pine sawfly subroutin 123
European pine sawfly subroutine 125
Aphid subroutine 127
Ant subroutine 127
Analysis of Output 131
Further model development 147
Additional research 148
Conclusions 149
Literature Cited 152
MANUSCRIPT 3. DESIGNING AND FIELD TESTING PORTABLE 154
COLONIES OF THE PREDACEOUS ANT. F ORMICA
ELSECI‘OIDESFOREL
Introduction 155
Materials and Methods 156
Results and Discussion 158

iv

10

LIST OF TABLES

Mean number of fourth-instar RI-IPS remaining on
experimental jack pine trees 24 h after placing them
on the trees. Initially, each tree contained 60 larvae.

Ant predation on fourth-instar RHPS in the presence
and absence of aphids in Crawford County, Michigan
during 1987. Mean number of larvae remaining on
experimental trees after 24 h. Initially, each tree
contained 60 larvae.

Ant predation on fourth-instar JPBW in Crawford
County, Michigan during 1987. Mean number of
larvae remaining on experimental jack pines trees
24 h after placing them on the trees.

Ant predation on second instar JPBW in the presence
and absence of aphids in Crawford County, Michigan.
Mean number of larvae remaining on jack pine trees
after 72 h.

Ant predation on JPBW pupae in the presence and absence
of aphids in Crawford County, Michigan. Mean number
of larvae remaining on trees after 72 h.

The influence of the JPBW silk chamber on ant predation
in Crawford County, Michigan. Mean number of larvae
remaining on experimental jack pine trees 72 h after
placing them on the trees.

Ant predation on adult WPW in the presence and absence
of aphids. Mean number of WPW remaining on experi-
mental trees after 72 h.

Summary of 24, thirty-minute observation periods
recording the interactions between F. exsectoides

and white pine weevils on jack pines.

The insects and stages of development that F. exsectoides
preyed upon.

Seasonal distribution of insect prey and F. exsectoides.

V

56

58

64

65

66

68

80

82

87

98

LIST OF FIGURES

Experimental arenas were used for all experiments
where ant predation was examined on planted popu-
lations of insect pests. Cages were placed over the
trees to prevent unwanted vertebrate predation.

Workers of F. exsectoides readily preyed on fourth-
instar redheaded pine sawlies.

Mean number of sawflies on jack pine with RHPS
only, RI-IPS and ants, and RI-IPS. ants and aphids.
Differences between treatment means are significant
at 24 h, n=12 (p < 0.05).

Regression of sawflies and ants on jack pines after
24 h. The regression was significant, p < 0.0001,

(y = 1.088 x + 55.180, r2 = 0.63).

Regression of sawflies and ants in the presence of
aphids on jack pine trees after 24 h. The regression
was not significant, p > 0.05. (y = 0.35 x + 14.65,

r2 = 0.025).

Effects of ant predation on three instars of sawflies
over time. There were significant differences between
means at 36 h, n = 4 (p < 0.05). Multiple comparison
showed no differences between instar 5 and instar 1,
and instar 5 and instar 3. respectively.

Relationship between mound index and the number of
JPBW pupae on trees in the overstory, (y = -2.54 x + 20.39.

r2 = 0.42). The regression is highly significant, p < 0.001;
n = 79.

Relationship between the number of mounds and the
number of JPBW pupae on trees in the overstory,

(y = -5.20 x + 19.50. r2 = 0.36). The regression is
highly significant. p < 0.001; n = 79.

vi

43

55

59

60

61

63

70

71

10

ll

12

13

14

15

l6

17
18
19

20

21
22
23
24
25
26

Relationship between mound index and the number of
jack pine branch tips defoliated per 400 tips sampled

(y = -5.75 x 4- 198.14. r2 = 0.08). The regression is
significant, p < 0.001; n = 79.

Relationship between mound index and the number of
tips defoliated per 50 tips sampled in the understory

(y = -0.78 x + 17.14. 1‘2 = 0.07). The regression is
significant, p < 0.05; n = 64.

Regression results of the effect of the distance from the

mound on the percent of adult female gypsy moths re-
moved by ants from jack pine trees. The regression was

significant, P < 0.0001, (y = -1.60 x + 83.32, :2 = 0.441).

Workers of F. exsectoides entering a gypsy moth pupal
case and removing fragments of a pupa.

Persistent and non-persistent interactions between
F. exsectoides and white pine weevils on trees that
contained aphids.

Persistent and non-persistent interactions between
F. exsectoides and white pine weevils on trees that
did not contain aphids.

Workers of F. exsectoides searching for a dislodged
jack pine budworm larva.

Workers of F. exsectoides removing a jack pine
budworm pupa from its silk chamber.

System flow chart.

Michigan Weather.

Weather subroutine.

Tree subroutine.

Trefoil subroutine.

Nitrogen subroutine.

Redheaded pine sawfly subroutine.
European pine sawfly subroutine.
Aphid subroutine.

Ant subroutine.

vii

72

74

77

79

83

84

91

94

110
112
113
115
120
124
126
128
129
132

27
28
29
30
31
32
33
34
35
36
37
38
39

Effect
Effect
Effect
Effect
Effect
Effect
Effect
Effect
Effect

Effect

of
of
of
of
of
of
of
of
of
of

trefoil on tree growth. 1

trefoil on tree growth. 2

trefoil on tree growth. 3

trefoil on tree growth. 4

trefoil density on defoliation. 1
trefoil density on defoliation. 2
trefoil on site classification. 1
trefoil on site classification. 2
soil type and ants on defoliation. 1

soil type and ants on defoliations. 2

Effect of soil type and ants on defoliation. 3

Comparing conventional and "improved” system.

Portable ant colony of F. exsectoides.

viii

134
135
136
137
139
140
142
143
144
145
146
151

160

General Thesis Introduction

This study represents the first steps in the process of exploring the
potential of utilizing the predaceous ant, Formica exsectoides Forel, as a
biological control agent of insect pests in young pine plantations.

Predaceous red wood ants, Formica rufa group, have been exploited in
intensive forestry culture in Europe for over 150 years (Adlung, 1966). Five
species of ants have been studied extensively, F. polyctena Foerst, F. rufa (L),
F. lugubris (Zett.), F. aquilonia (Yar.), and F. nigricans (Em.). Of the five, F.
polyctena is considered to be the most valuable as a biological control agent.

Studies that have been conducted in North America have disclosed that
carpenter ants in the genus Camponorus and mound-building ants in the
genus Formica both feed on insects of economic importance in forestry
(Finnegan et al.. 1978; Campbell and Torgersen, 1982; Torgersen and Mason.
1987). Recently, ants in the genus Solenopsis, have proved to be valuable in
the reduction of insect pests in agricultural systems (Tedders et al.. 1990,
Eskafr and Kolbe, 1990).

Predaceous ants are capable of removing a large number of insects from
trees and shrubs in the forest ecosystem. Wellenstein (1952) counted eight
million insects taken as prey by one medium sized nest of red wood ants
during one season. Seventy-six percent of the scientists that were polled by
Wellenstein (1954) regarded the merits of red wood ants as biological control
agents to be positive. However, the remaining 24 percent regarded red wood
ants to have a negative effect on the forest ecosystem as a whole. Three of
the primary reasons of why the value of red wood ants as biological control
agents continues to be disputed are: 1) for the ants to be effective they must
be present in very large numbers which could result in damage to the

l

ecosystem (Schrotter, 1959), 2) ant attendance of certain aphid species is
known to increase the population of aphids to a level which is detrimental to
the health of certain types of trees (Debach et al., 1951; Wellenstein, 1961),
and 3) very little research has been done that encompasses more than one or
two components of the ant-aphid-insect pest-tree ecosystem.

Mutualism between ants and homoptera is a common phenomenon (Way,
1963). The mutualism may have a positive or negative effect on the forest
ecosystem as a whole depending upon the density of the aphids. the type of
tree that is infested, and the effects of the ant attendance on the natural
enemies of the aphids. Kloft (1959) and Bradley and Hinks (1968) found that
aphid colonies were not randomly distributed but were more numerous near
ant mounds. Whether the increase in the number of ants on trees resulted in
higher ant predation on herbivores is not clear from these studies.

The jack pine budworm. Charistoneura pinus (Freeman), redheaded pine
sawfly, Neodiprion lecontei (Fitch) and white pine weevil, Pissodes strobi
(Peck) can be serious pests of young pine plantations (USDA, 1983). The
gypsy moth, Lymanrria dispar (L.), prefers to feed on deciduous trees such as
oak and aspen, but will feed on conifers when growing with hardwoods.
Ants have been recorded feeding on the larvae of jack pine budworm and
western spruce budworm, Choristoneura occidentalis Free. (Allen et al. 1970;
Jennings, 1971; Campbell and Torgersen, 1982). Goesswald (1940a) recorded
ants preying on colonies of the redheaded pine sawfly, and found that trees
closer to mounds of F. polyctena had less defoliation. Ants also prey on eggs

and gypsy moth larvae (Weseloh. 1988; Brown and Cameron, 1982).

F. exsectoides is a native North American mound building ant occurring
primarily in the eastern United States. It belongs to the Formica exsecta
group which is a close relative to the Formica rufa group. Workers of F.
exsectoides have been observed feeding on a variety of coniferous forest
pests including certain species of beetles, flies, and Lepidoptera larvae.

F. exsectoides foragers tend aphid colonies for honeydew. but will take dead
or injured aphids back to the mounds as food (Andrews, 1929). Mounds of

F. exsectoides are widely distributed in jack pine stands in the central part of
the lower peninsula of Michigan. These jack pine stands support fluctuating
papulations of jack pine budworm and sawflies. Gypsy moths can be found in
jack pine stands that have an oak and poplar component.

This study examined the interactions between the predaceous .ant, F.
exsectoides. aphids found on jack pine, Cinara sp., and four insect pests of
pines: jack pine budworm, redheaded pine sawfly, white pine weevil, and the
gypsy moth. The effects of ant protection of jack pines resulting from
predation on the jack pine budworm was also measured.

In order to further explore the complex interactions between the ants,
aphids. insect pests. and trees, a simulation model was developed. The model
was used to identify several areas where additional research was needed,
such as ant predation switching values. Very little information is available
on the mechanisms underlying prey switching in ants between insect pests
and honeydew produced by aphids and other homoptera. This prompted a
study where the influence of internal colony conditions on prey selection
was examined in both the laboratory and field during the following year.

These studies are on going, and therefore are not included in this thesis.

The final portion of this study dealt with managing colonies of
F. exsectoides in portable ant nests as a biological form of insect control.
Fruit growers in China have been manipulating colonies of predaceous ants
for many years by overwintering the nests and strategically placing them in
the field in the spring (Huang and Yang, 1987). In North America, Formica
spp. have been transplanted 'into forests to provide protection (Bradley, 1972;

Finnegan, 1975; Kloft et al.. 1979).

Literature Review

Among the insects, many ants are well known for their predatory habits.
Ants are carnivorous and polyphagous, apparently inheriting their
predatory nature from solitary wasp ancestors (Wilson, 1971). The predatory
habits of ants were recognized thousands of years ago as ants are now
thought to have been the first animals exploited for the control of insect
pests (Huang and Yang, 1937). In 304 A. 13.. the benefits derived from the
yellow citrus ant, Oecophylla smaragdiua Fabr., for the protection of oranges
was noted (Huang and Yang. 1987). The citrus ant. has been used to protect
oranges, tangerines, lemons, and pomelos in parts of China. For a brief
period of time during the rise of use in organic insecticides this practice was
abandoned. Pesticide resistance and the rising cost of chemicals prompted
orange growers in a Huangtien village of China to reestablish and once
again expand colonies of the predaceous ant in citrus orchards. Today, ants
are still utilized in China and other parts of the world in biological control
and integrated pest management programs for a variety of insect pests in
agriculture and forestry (Majer, 1986). Among the attributes ants possess as
biological control agents are: 1) diversity and abundance, 2) responsiveness
to spatial variations in prey density. 3) persistence in spite of fluctuations in
their food supply, 4) effectiveness not limited by predator satiation, 5) in
addition to predation, ants can indirectly reduce insect pepulations by
disrupting feeding and oviposition, and 6) potential for manipulation to
maximize contact with pests (Risch and Carroll, 1982). A recent surge in
basic ecological research exploring mutualisms between ants and trees, and

ants and their tended aphids has fueled a renewed interest in pursuing

knowledge that could lead to further exploitation of ant species in plant
protection programs. particularly in North America (Gotwald, 1986).

Ants have been recognized as performing a beneficial function in
European forests for over 150 years (Adlung. 1966). Ants of the Formica rufa
group, particularly Formica polyctena Foerst, have been propagated and
distributed in artificial colonies in intensively managed European forest
plantations (Goesswald. 1951). Large nests of Formica spp. have been
reported to feed on from 65,000 to 100.000 caterpillars per day (Gotwald, 1986).
Their feeding habits. aphid mutualisms and effects of aphid feeding on trees
has been the focus of several European studies (Muller, 1956a, 1956b, 1958;
Kloft, 1953. 1959).

Spurred by the successful exploitation of ants in European forests, several
studies of native and introduced ant species have been conducted in North
American forests. Perhaps the first direct reference to ants and forest
insects is a note by Green and Sullivan (1950) that indicated predation by
ants on the forest tent caterpillar. M at acosoma disstria Huebner. Later, a
study by Sanders (1964) in New Brunswick forests explored the biology of
three species of carpenter ants, Camponotus hcrculeanus (L.),

C. pcunsylvanicus (DeGeer), and C. novaboracensis (Fitch). Although much
of the study was devoted to biology and gallery construction, mention was
made of workers bringing caterpillars back to the nests. Bradley and Hinks
(1968) studied ants and their attended aphids in Manitoba. They found three
species of Myrmicinae, one species of Dolichoderinae, and 17 species of
Formicinae which tended up to four species of aphids on jack pine, Pi nus

banksiana (Lambert). Two ant species were particularly common. F.

obscuripes Forel and Dolochodcrus taschcnbcrgi (Mayr). They tended
several aphid species in the genus Cinara. The researchers noted how the
trees benefitted from aphid tending through removal of defoliating insects
by the ants. Bradley and Hinks (1968) observed that aphids were able to
withdraw large amounts of sap. from the jack pines without seriously
injuring the trees, however, no direct measurements of the effects of aphids
feeding on the trees was done. Subsequent studies showed these species of
ants were present in areas of natural reproduction but absent in plantations.
In order to better exploit the predaceous activities of these ant species,
Bradley (1972) explored the transplantation of nests and proximity of
placement. He observed good establishment of nests and found little
relocation of nests placed 20 m apart, while those placed within 5 m
consistently moved apparently to accommodate the larger territory size of
the colonies. The feeding and foraging behavior of F. obscuripcs was
further explored by Hill (1969) and Headings (1971) in jack pine plantations
in Michigan. Hill (1969) found the ants preying on European pine sawfly,
Neodiprion scrtifer (Geoffroy) and the jack pine budworm, Choristoncura
pinus Freeman, along with a broad group of insect families as they became
available throughout the summer season. He further noted that humidity
and distance from the nest or mound were influential in determining the
number of ants foraging on a given tree. Headings (1971) showed that
species of Aphididae and Miridae were more abundant on jack pine trees
where F. obscuripcs foraged; species of Cercopidae, Coccinellidae, Tortricidae
(particularly jack pine budworm), Plutellidae, Psocoptera and Neuroptera

were significantly fewer on trees where the ants did not forage. The general

activity of F. obscuripcs workers decreased with increased temperature and
rainfall.

Several additional studies have dealt with the role ant predation plays in
the population dynamics of budworm species: jack pine budworm, C. pinus
Freeman, spruce budworm. C. fumiferana (Clem.), and western spruce
budworm, C. occidentalis Freeman. Allen et al. (1970) and Jennings (1971)
noted predation by Formica and Camponotus species on the jack pine
budworm in Michigan and Wisconsin. respectively. McNeil (1978) observed
reduction in spruce budworm populations by F. iugubris Zetterstedt, a species
introduced from Italy, and that spruce budworm was the most common prey
in Quebec. Campbell and Torgersen (1982) investigated 179 potential natural
enemies of western spruce budworm in Oregon. Using ”planted” pupae, they
observed that 85% of the pupae were destroyed by foraging ant species from
two genera: Formica and Camponotus. Youngs and Campbell (1984)
investigated ant species feeding on western spruce budworm pupae in
Oregon and Montana. A number of species of Formica and Camponotus were
found to feed on budworm pupae. Related studies further disclosed the
importance of ants in the population dynamics of budworm (Campbell et al.
1983. 1985). Campbell also studied ants in connection with spruce budworm
in the east.

Recent studies have shown ants to be important in controlling leaf miner
populations. Sato and Higashi (1987) showed that mine density on oaks was
always lower on ant-rich trees than on trees where the ants were not as
common. Faeth (1980) showed that ants contributed significantly to the

mortality of the leaf-miner, Ericraniella sp.. These studies as well as several

others showed consistently that smaller ants were the primarily responsible
for the mortality of leaf miners. Wong et al. (1984) and Wong and Wong
(1988) found that fire ants, Solenopsis gcminara (F.). played an important
role in reducing populations of Mediterranean Fruit Fly, Ceratitis capitata
(Wiedemann). and the Oriental Fruit Fly, Dacus dorsaiis Hendel. The fire ants
preyed on fruit fly larvae in the laboratory and on pupae which were buried
underneath the soil at the base of guava trees. Soicnopsis inivicta Buren, has
also been reported to be an important natural control of larval boll weevil
(Sterling, 1979). and pecan weevil larvae, Curculio caryac (Horn), in Georgia
(Dutcher and Sheppard, 1981).

The introduction of ant species into new habitats for the protection of
forest trees has been tried in North America, but the success of such
introductions has required some protection of the introduced colonies from
predation by other ants and birds. Finnegan (1975, 1977) brought two
species of Formica into Quebec one from Italy, F. lugubris, and one from
Manitoba, F. obscuripcs. In 1977. Finnegan concluded that the ant was well
established and expanding in Quebec (Finnegan. 1977). In 1978, McNeil
studied the seasonal predatory activity of several of the mounds that were
transplanted into Quebec. Prey brought back to the nest by the ants included
insects in the orders, Lepidoptera. Hymenoptera, Diptera, Coleoptera, and
Hemiptera. Lepidopteran insects represented 58 percent of the total prey
brought back with spruce budworm larvae being the most abundant insect
in the sample. The estimated total number of larvae brought back to the nest
was 2,175/day. A series of studies was also done with F. integra Nylander, a

native ”red wood ant" of oak-hickory forests of Georgia. This species was

10

found to be polygenous (have multiple queens) and capable of feeding on a
number of forest insect pests (Wilkinson et al. 1978; Kloft et al. 1979). Nest
transplantation trials resulted in colonies persisting for up to 14 weeks with
aphid tending and predation occurring on pest species. A native carpenter
ant eventually killed out the colonies as was predicted in earlier laboratory
experiments (Wilkinson et al. 1980). Room (1973) cautioned that attempts to
exploit non-native species might not be as good as using native species

primarily because of difficulties in establishment.

Honeydew constitutes the most abundant type of food collected by many
species of ants (Cami and Janzen, 1973; Skinner, 1980). However most ants
also require protein usually in the form of insect prey to complete their diet.
Homoptera benefit by ant attendance in a number of ways. Ants offer
protection to homoptera from natural enemies and remove waste. Certain
species of ants transport Homoptera to feeding sites (Way, 1963).

The effects of the interactions between ants and aphids in relation to
biological control can be either positive or negative depending on a number
of factors. Ant attendance of certain species of Homoptera interferes with
the control exerted by the natural enemies (Flanders, 1943; Samways et al.,
1982). Haney et al. (1987) monitored the activities of the Argentine ant,
Iridomyrmex humilis (Mayr). in three citrus orchards for two years and
found that the presence of the ant inhibited effective predation on the citrus
red mite, Panonychus citri (McGregor). Fritz (1982) found significantly

fewer predators of membracids on branches of black locust with ants than

11

on branches where ants were absent. Protection of aphids by ants can
negatively affect plant health and reproduction also when the potential prey
are plant pollinators or parasitoids of insect herbivores. Bronstein (1988)
found that ants readily attacked four species of wasps associated with fig
trees. One of the species of wasps. is an obligate pollinator of the fig and
predation impacted successful pollination of the plant.

Other studies have shown that ant attendance of certain HomOptera does
not interfere with the natural control of the Homoptera. Flanders (1943)
found that the activities of the parasite, Coccophaguss capcnsis Comp., on
black scale were not inhibited by ants. He later observed heavy infestations
of the black scale, Saisseria olcae (Oliver). that were attended by ants that
were completely parasitized (Flanders, 1951). Stary (1966) concluded from a
study measuring aphid parasitism of several species of aphids, that aphid
parasitism was not negatively effected by the presence of ants. He further
suggested that in certain cases, the ant x aphid x parasite association
effectively limited the population density of aphids.

Green and Sullivan (1950) suggested that ant predation is often a result of
defence of aphid colonies. They observed ants removing a t0tal of 1,800
forest tent caterpillar larvae, M. dissrria. When young larvae passed through
an aphid colony this immediately invoked an attack by the ants. They
observed that when only one to two ants were tending aphids they would
often kill the larvae and leave them or push them off the branch. When
more than two ants were present they would kill the larvae and take them

back to the nest. Goldenrod stems, Solidago spp., that contained membracids

12

Publilia concava, tended by ants, Formica spp., escaped defoliation by two
species of beetles, Trirhabda virgata Leconte and

T. borcaiis LeConte, and showed greater height increment and seed
production than those stems that did not contain membracids (Messina, 1981).
He suggested that plant protection may also be important for membracids
because the ants insure host quality by driving away other defoliators.
Colonies of Homoptera may also have positive effects on the plants by

serving as a nucleus of food to attract natural enemies (Flanders, 1951).

Ant attendance of certain species of Homoptera is known to increase the
population of Homoptera to economical injurious levels. Flanders (1951)
suggested that when an ant-attended Homopteran population is large, it can
be very destructive to the plant. Severe infestations of the black scale have
been reported to be correlated with the presence of the Argentine ant
(Compere. 1940). Increases in the brown soft scale, Coccus hcsperidum L.,
due to ant attendance can cause the population of the scale to reach
economically important outbreak numbers (Debach et al., 1951). Fruit
growers will often band their trees with a sticky barrier substance such as
grease to reduce infestations of aphids (Bremner, 1931). However, Samways
et al. (1982) found that only 2 out of 123 species of ants recorded in South
African citrus orchards were important in precipitating outbreaks of
Homoptera. The other species of ants were relatively harmless and only
caused isolated hompoteran outbreaks. They suggested that the
recommendation to control ants in citrus orchards is probably unfounded in
most situations because most ants are beneficial. An extensive survey of ant-

aphid associations by Samways et al. (1982) showed that even when ant

l3

attendance elevated homopteran levels to economics importance. the
predaceous activity of ants outweighed the occasional damage to shoots of
citrus trees.

Determining if an ant-aphid association is positive or negative may also
depend upon the type of tree affected. Dixon (1971a) suggested that with the
exception of the spruce aphid, Elatobium abictinum (Walk.), aphids generally
do not cause extensive defoliation of their host trees except when they are
present in very large numbers. Bradley and Hinks (1968) observed that
aphids are able to withdraw large amounts of sap from jack pines without
seriously affecting the trees. They suggested that this relationship allowed
the aphids to maintain themselves on the same trees or or in the ant's
foraging territory for several years without depleting their food source.
Ants had a positive influence on the survival of mountain birches, Bctuia
pubescens spp. tortuosa (von Ledebour) Nyman), during an outbreak of the
Lepidoptera, Oporinia autumnata (Borkhausen), despite the ant's association
with the aphid, Symydobius oblongus (von der Heyden). Laine and Niemela
(1980) suggested that the cost of the ants eating honeydew may limit the
growth of individual birch trees for several years, but may perhaps insure
its existence over defoliator outbreaks.

Banks and Macaulay (1967), found that bean plants without aphids yielded
56 seeds per plant, those with aphids but no ants yielded 17, and those with
aphids and ants yielded only 8. Dixon (1971a) found that the aphid,
Drcpanosiphum platanoidcs (Schr.), reduced leaf size in sycamore by as
much as 40 percent, and the production of stem wood by up to 62 percent.

Feeding by aphids affected the growth of lime saplings, Tilia vulgaris Hayne,

14

very little above ground but resulted in very poor root growth (Dixon,
1971b). The growth of the aphid infested lime saplings was limited by the
reduced availability of photsynthate.

After reviewing the literature, conducting studies and performing
experiments with ants, both' Youngs (1983) and Majer (1986) encouraged
further research on utilizing predaceous ants in forest ecosystems. Majer
provides specific guidelines for studying potential ant species as biological
control agents. Youngs suggests pursuing studies of both Formica species
and Camponotus species to gain a better appreciation of their potential for
exploitation in biological control under forest conditions.

APPROACH

Majer (1986) provided a formalized procedure for selecting and
developing ants as suppression agents for pest insects. This is the approach
we will follow. Stage 1 requires definition of the pest or the pest complex. To
do this adequately. the biology of the pest insect(s) and the interaction of the
pest insect(s) with the crop needs to be determined. Stage 2 is the process of
identifying ant species that have potential as agents of pest insect
suppression. Stage 3 is the investigation of whether or not an ant species
identified in Stage 2 actually has some suppressing effect on the pest or pest
complex identified in Stage 1. Stage 4 is entered if and when a beneficial ant
species has been identified and involves the manipulation of the ant species
to favor its utilization. Stage 5 is the point at which field trials are conducted
to evaluate the practicality and efficacy of an ant species and the promising
manipulations identified in Step 4. At this point it is possible for the ant-

based biological suppression technique to be combined or integrated into a

15

wider array of pest management methods. Lastly, grower recommendations
are made with instruction about how to apply the ant-based suppression
technique. This stage is implemented with the caution that following its
incorporation in management, continued vigilance is needed to identify new
pest problems that may require modification of the scheme (just as in any
other approach to pest insect suppression). Further development of the
“Approach” section will utilize the first 3 ”stages” in Majer's (1986)

procedure:

STAGE 1 - Identifying the Pest Complex.

Our focus in this program is on insect pest problems associated with
young pines in the states and provinces of the Great Lakes Region. Red pine,
P. rcsinosa Aiton.. and jack pine are utilized for pulp and lumber and are
planted extensively throughout the region. A third species, Scots pine, P .
silvestris L., is the dominant conifer of an extensive Christmas tree industry.
Scots pine Christmas trees are typically grown over a 7 to 10 year rotation,
while jack pine and red pine are managed on rotations exceeding 50 years
for pulp and longer for timber. Despite the differences in rotation lengths.
the primary insect pests of these species affect the trees most when they are
small, very commonly when trees are less than 10 to 15 years of age. The
four insect pests used in this study were: 1) gypsy moth, Lymanrria dispar
(L), 2) jack pine budworm. Choristoneura pinus (Freeman), 3) redheaded
pine sawfly, Neodiprion lccontci (Fitch), and 4) white pine weevil, Pissodcs

strobi (Peck).

16

W

The host complex of the gypsy moth includes over 300 kinds of trees and
shrubs. The favored hosts of the gypsy moth are oak, aspen. apple and
willow. The gypsy moth generally does not feed on pines when other
favored host trees are readily available. However the gypsy moth can be a
problem in stand of pines that are interplanted with hardwoods or near a
stand of susceptible trees.

Hatch and larval activity are both strongly influenced by temperature
(Leonard. 1981). As larvae emerge from egg masses they spin a silk thread.
If the temperature is below 7° C, the larvae will remain on the egg masses.
When the temperature increases, they begin to move to the tops of the trees
and disperse on wind currents. During the first three instars, the larvae
remain on the leaves at all times and feed both during the day and night.
During later larval development, they follow a daily migration pattern
where they move down the trees during the day, and returns to feed only at
night. The gypsy moth remains in the pupal stage for approximately 2 weeks
(Leonard, 1981). The male moths emerge several days before the females.
The males are strong fliers while the females are relatively flightless. Egg
laying begins shortly after the females emerge.

The gypsy moth is considered not only to be a serious pest of forests. but
can also be a serious nuisance to people. Control of this pest in Michigan is
usually limited to areas that contain human dwellings or high-value timber.
Tree mortality can occur after two to three years of repeated, heavy

defoliation. Certain species of oaks. such as white oak. often incur the

17

highest percentage of mortality; up to 50 percent of dominant and
codominant trees (Leonard, 1981). People find it difficult to cope with the
gypsy moth because of the large size of the mature caterpillars, constant
droppings, and often high populations. This pest has the greatest impact on
humans and their outdoor activities during the months of June and July.
The gypsy moth can be adequately controlled with properly timed
applications of the bacterial insecticide, Bt., Bacillus tharingicnsis. Other
insecticides are also used to control this pest. Mechanical techniques such as
removing the egg masses, are also recommended. Both native and introduced
parasites can significantly impact gypsy moth populations. Predators such
as ground beetles (Weseloh, 1988). birds (Brown and Cameron, 1982; Furuta
and Koizumi, 1975) and small mammals are thought be effective at low
population densities of gypsy moth. Campbell et al. (1975) found that the
white-footed mouse, Pcromyscus icucopus, ‘ was the major source of late-
larval and pupal mortality. Parasitism from, Parasctigena silvcstris, was a
major mortality factor during the larval stage at low-density populations
(Elkinton and Liebhold. 1990). At high population densities. the virus, NPV,
Nucleopolyhedrosis virus, can cause the population of gypsy moths to

collapse (Campbell, 1967; Doane, 1970).

11:11 19 lflllli'nfll

The hosts of the jack pine budworm include, Scots, red, jack and Austrian
pines. The jack pine budworm overwinters as first-instar larvae. Depending
upon environmental conditions. the small larvae break diapause at the same

time the male staminate flowers of jack pines begin opening. The larvae

l8

disperse before settling down to begin feeding. The larvae feed on the
flowers and then move to new shoots to continue feeding (Dixon, 1961).
Besides feeding on the foliage, the larvae also use it to build shelters. Larvae
pupate in these shelters from early 'to mid summer. After approximately one
week, the adults emerge. Mating and oviposition occur shortly afterward. In
Michigan, adult jack pine budworms are present from late June to early
August (Graham. 1935). The females lay egg masses, containing
approximately 50 eggs on old growth needles throughout the crown.

The jack pine budworm can cause tree mortality, reduced growth and
pollen production, and top kill (Elliot. 1985). High populations of budworm
may result in up to 90 percent mortality of intermediate and suppressed
trees, and 33 percent mortality of merchantable trees (Batzer and Millers,
1970). Tree mortality is more likely to occur when trees are stressed by
other environmental or biological factors. such as drought or other
defoliators (Batzer and Millers, 1970).

When larvae and pupae are concealed in their feeding shelters, they are
difficult to control (USDA. 1983). Spraying with Bt. is recommended when
the larvae disperse in the spring. Silvicultural techniques are also effective
in controlling the jack pine budworm. Generally, young well stocked stands
are resistant to heavy jack pine budworm attacks (Batzer and Jennings,
1980).

Natural enemies are thought to be important at low budworm densities.
The parasites, Apantcles fumifcranac Viereck, and A. morrisi (Mason),
commonly parasitize jack pine budworm larvae in Michigan (Elliot et al.,

1986). Mortality of small larvae can also result from fall dispersal. Spring

l9

dispersal can significantly impact larger, third-instar budworm. The
population of larger larvae and pupae can be reduced by parasites (Batzer

and Jennings, 1980).

W

The hosts of the redheaded pine sawfly are red, jack and other pines. and
they occasionally attack spruce (USDA, 1983). Redheaded pine sawfly usually
have one generation per year in Michigan. The sawfly overwinters as
prepupae in cocoons in the duff or topsoil beneath its host. Some pupae may
remain in diapause over several seasons. Pupation usually occurs after a
period of warm weather and the adults emerge in approximately two weeks
(MacAloney and Wilson, 1964). In Canada and the Lake states. adults emerge
in June. The females lay approximately 120 eggs in clusters on the current
or previous year’s foliage. In three to five weeks, the larvae hatch and
begin feeding in colonies of just a few to over 100 insects. The larvae feed
for five to six weeks, during the months of July and early August in
Michigan. Early instars feed on the outer tissue of the needles while mature
larvae consume the entire needle (Averill et al., 1982). Mature larvae drop
from the host tree and spin cocoons in the duff or topsoil.

The most susceptible hosts are trees less than 15 feet tall (MacAloney and
Wilson, 1964). Jack and red pine are preferred hosts in Canada and the Lake
states. When foliage is scarce, the larvae will consume the bark. Moderate to
heavy defoliation usually results in reduced height. Complete defoliation

will usually kill young jack and red pines especially when they are grown

20

on poor sites (Averill et al., 1982). Many young trees are deformed or killed
during an outbreak of this pest.

In the Lake states, the redheaded pine sawfly reaches outbreak levels
approximately every five to eight years (Averill et al., 1982). Natural
control factors that impact high population levels include, rodents, extreme
temperatures, and disease. Hanski and Parviainen (1985) found that small
mammals destroyed 70 percent of the population of European pine sawfly,
Neodiprion sertifcr (Geoffroy). cocoons at poor sites. A correlation between
predation rate and the number of predators was not found at fertile sites.
They suggested that this may have occurred because of the abundance of
alternative prey.

In Southern Ontario, most outbreaks of sawflies are initiated by the
discovery of a previously uninfested pine stand that contains few natural
enemies (Lyons, 1977). Averill ct al. (1982) suggested that the most effective
method of controlling the sawfly is through prevention. The first step in
this process is to avoid planting pines in high-risk areas, such as adjacent to
hardwoods on poor soils.

When the population is scattered, mechanical removal of the entire
colony is recommended (USDA, 1983). When the population is high, chemical

control is often necessary.

111:1] ID If! llll' I? ll! '1
The hosts of the white pine weevil include all pines, most species of
spruce and occasionally fir (USDA, 1983). The adults overwinter in the duff

layer beneath host trees. Adults emerge and begin to feed on the terminal

21

growth, below the current year's bud usually in March and April (USDA,
1983). The females lay their eggs beneath the bark shortly after mating.
Larvae emerge and feed on the cambium. As the number of larvae increase,
they aggregate and form a ring around the terminal, causing girdling. The
leader usually dies from a weevil attack and will curve over to one side. This
”shepherds crook” is a good indicator of weevil damage in plantations or
natural stands of pines. Larvae pupate in woody, chip cocoons in the xylem.
The adults remain in these pupation chambers for several weeks before
emerging. Adults feed on the new growth until environmental conditions
force them into overwintering.

The white pine weevil can be an important insect pest in both plantations
and natural stands of pines and occasionally spruce (USDA. 1983). In New
York and New Hampshire, approximately 70 to 90 percent of all eastern white
pines had been damaged by the weevil before reaching the age of 15
(Graham. 1926). Weevil damage often results in poor quality lumber that is
not acceptable for sawlogs (Dirks, 1964). Weevils were found in every county
in the Lower Peninsula of Michigan in 1979 (Michigan Forest Pest Report.
1979). In 1987. weevil damage in the lower peninsula continued to increase
with an average of 43 percent of the leaders damaged (Michigan Forest Pest
Report, 1987). Damage occurs as either a reduction in volume or quality of
lumber (Plummer and Pillsbury, 1929). Weevils rarely kill mature trees.

Two of the more commonly recommended control methods for the white
pine weevil are tip pruning and chemical control. Predation and parasitism
are limited to insects that can mine in the cambium of shoots to locate

weevils in the larval or pupal stage. Eggs of the parasite. Lonchaca corticis

22 _

Taylor, are laid in the feeding and oviposition punctures made by P. strobi
and occasionally directly on the eggs (Harman and Wallace, 1971). Alfaro
and Borden (1980) found that each L. corticis larva consumes almost 3 weevil
pupae.

Silvicultural techniques such as the manipulation of .the overstory or
understory are very effective ways to control the weevils (Graham. 1926).
Planting resistant stock is also recommended to control the white pine weevil

(USDA, 1983 ).

1). STAGE 2 - Identifying Desirable Ant Species.

Finnegan (1971), while searching for potential ant species that might be
exploited in forests, deveIOped a list of desirable criteria: 1) Size. Large size
of individual ants is helpful but can be offset by large colonies where
additional workers can be recruited to help otherwise small workers; 2) Food
requirements and nest populations. Those species that have large nests and
developing broods with insect food needs are thought to place stronger
demands per unit area of land on insect-based food items: 3) Colonial nests.
Those species that have multidomous colonies or produce satellite colonies
over adjacent areas are thought to be helpful from the point of view that
they would increase in numbers and separate sub-colonies would be
compatible; 4) Queens. Longer-lived colonies are thought to be those that
have multiple queens or a queen replacement system; 5) Ant-Homoptera
relationships. Those ant species that tend and encourage undesirable

Homoptera such as aphids are considered less desirable. In addition to

23

Finnegan's set of desired criteria. Majer (1986) added: 6) Species habitat
range. Those species with a broad habitat range might be more amenable to
promotion in areas where they did not already exist. In addition to habitat
range per sc. broad tolerance of wide ranging physical factors such as
temperature, moisture and solar radiation were also considered desirable, and
7) Dominance hierarchy. Useful species are those which are not likely to be
outcompeted by other ants found in association with the crop.

Finnegan (1971) surveyed the existing literature and records of about 80
species of ants indigenous to Quebec and concluded that none of the species
present possessed enough of the characteristics he considered desirable for
exploitation against forest pest insects. In his discussion, however, be
indicated species in the genera Camponotus and Formica possessed many of
the desired characteristics except polygyny or the quality of having multiple
queens. Finnegan stated ”That the sole queen must always be included and
kept in good condition, or the nest will perish." This prompted the
importation of the non-native species. F. lugubris from Italy and F.
obscuripcs from Manitoba that possessed multiple queens.

Of the two species brought to Quebec, F. obscuripcs seems to have
established well. Later studies by Bradley (1972) indicated colonies of F.
obscuripes could be transplanted, and after a period of reorganization lasting
2-4 days, colonies would re-establish without difficulty. Work by Hill (1969).
Headings (1971) and Bradley (1972) provided additional information on
general biology. colony size estimation. propagation techniques, aphid
relationships and predator-prey relationships. Other studies done in other

ecosystems have also shown F. obscuripcs to prey on forest pests of interest

24

(Youngs 1983; Youngs and Campbell, 1984). Thus, a species from the genus
Formica, would seem a logical choice to complete Stage 2 in Majer's selection
procedure. The ant species used in this study is the Allegheny mound ant,
Formica cxscctoidcs.

Pierson (1922) and Haviland (1947) described the life cycle and seasonal
history of F. cxscctoidcs. Depending upon seasonal cues and geographical
location, F. exsectoides hibemates in burrows deep below the mound from
approximately October to April (Haviland, 1947). Temperature appears to be
the most important factor in determining the length of hibernation. Queens
begin to appear in late April. A large number of small, white elliptical eggs
can be found in the mound in April and egg laying continues throughout the
warm season in the field (Pierson. 1922). The length of the egg stage in the
field appears to be approximately 2 weeks. Larvae of all sizes were found in
Maryland in the mound in May (Haviland, 1947). Large larvae deveIOp into
reproductives and smaller larvae become workers. Throughout the season,
most of the colonies' activities occur within the top 6 inches of the mound
proper. The adult workers move brood throughout the mound depending
upon both the temperature and humidity. Pupae are present in the mound
from approximately June through September. The mature larvae spin, white,
papery cocoons and remain in this stage for about 2 weeks. The last
reproductives are usually seen in mid-July in Maryland. In general, a large
number of queens and kings swarm at the same time and the kings die soon

after flight. Some of the queens are seized by workers who immediately start

25

to build a new nest and. begin colony formation (Pierson, 1922). There are
several generations per season and queens of some species of ants have been
known to survive for up to 15 years (Pierson, 1922). By August, colony
activity has begins to decrease and by September some of the workers are
down below the mound entering into hibernation. Only delated queens and
workers are found in the mound during hibernation. The ants do not store
food of any type, including live aphids or aphid eggs in the mound during
this period.

Haviland (1947) reared F. cxsectoidcs in the laboratory from the egg to the
adult stage. In 30 artificial nests, eggs hatched anywhere from 10 to 27 days,
with an average of 19.4 days at 70° F. In nine of the nests, it took 21 to 53
days, with an average of 37.1 days for complete larval development at 70° F.
The pupal period lasted an average of 26.25 days at 70° F. The total time
required from the egg to the callow stage ranged between 54 to 107 days. with
an average of 75 days.

Wetsuits

Early work detailed the distribution and growth of mounds of F.
cxsectoides in pine plantations (Pierson, 1922; Andrews, 1925, 1926; Haviland
1938). The mounds occur in groups fairly close together and their size is a
fairly accurate index to their age (Pierson, 1922). Examinations showed that
galleries extended 20 feet or more from nests and penetrated to a depth of 6
feet. The mounds are built largely of coarse sand supported somewhat by
pine needles and twigs and a one to two inch thatch covers the mounds and
provides protection from the elements. The mounds in a group are usually

connected to other mounds nearby by a network of tunnels. One of these

26

mounds is usually the mother mound and the others are termed subsidiary
mounds. Haviland (1938) showed that a mound 19 inches high contained
237,103 workers and 1,407 queens. They estimated that 12 million ants were
present on 10 acres and averaged seven mounds per acre.

Andrews (1925) followed the growth of several mounds of F. cxscctoidcs
over 19 years. The growth of the mounds were not constant but fluctuated
over this time. In the early years. growth is very slow and increases in rate
in later years. However, Andrews concluded that the rate of growth of a
mound of F. cxscctoides is strongly influenced by the individual character of
each community under its own complex environment.

In general. the survival and growth of F. cxscctoides is dependent upon
certain stages of forestation (Andrews, 1926). Shading by old growth trees
appears to be an important factor in the survival of mounds of F. exsectoidcs.
In the natural succession of forests, F. cxscctoides thrives in areas in which
the forest is temporarily interrupted or destroyed by fire or wind. Therefore
the length of time mounds of F. cxsectoidcs can be found in a forest appears
to correlate well with the time required for a tree to reach maturity. 30 years
or so. Haviland (1938) could not find a correlation between the dispersal of
ant mounds and several factor she studied (distance from the trees, height of
trees. presence of undergrowth. and the amount of sunlight). However, two
out of the three locations where new mounds were found. were on the edge
of a clearing where the light intensity was the greatest.

Headley (1943) noted workers of F. cxsectoidcs tending green aphids on

trees. Haviland (1947) observed the ants tending treehoppers and aphids

27

during two consecutive summers. No evidence was found that the workers
feed on the aphids. In 1929, Andrews examined the relationship between
membracids and ants. Membracids supply honeydew to the ants chiefly
when they are young, but the ants eat injured or dead membracids. The ants
tend membracids during the day and all through the night. F. exscctoidcs
was found to make succursals (small mounds of sand) at the bases of small
trees which were used by the membracids. The succursal provides both
shelter and protection for the membracids and the ants benefit chiefly by
limiting the freedom of the membracids. Andrews (1929) found that sugar
flies associated with the membracids where effectively driven off by F.
cxscctoides in an attempt to monopolize the honeydew produced by the
membracids. In this account of membracid-ant associations, Andrews
mentions that this ant feeds on living and dead beetles. flies, bees, bugs and
especially lepidoptera larvae. Workers of F. cxscctoides were also noted
preying on beetles, sowbugs, spiders, crickets, grasshoppers, termites, and
unidentified Lepidoptera and Diptera from April through September
(Haviland, 1947). Allen et al. (1970) observed the ants preying on budworm
larvae that were on the ground or at other times when they were outside
their protective silk chambers.

In Michigan jack pine stands where this ant species is frequently
found, intensive management methods would not be cost-effective. However,
manipulating this ant as a biological control agent in managed high value

pine plantations may be worthwhile.

28

Literature Cited

Adlung, K. G. 1966. A Critical Evaluation of the European research on use of
red wood ants (Formica rufa Group) for the protection of forests against
harmful insects. Z. ang. But 57: 167-201.

Alfaro, R. I. and J. H. Borden. 1980. Predation by Lonchaca corticis Diptera
Lonchaeidae on the white pine weevil Pissodes strobi Coleoptera
Curculionidae. Can. Entomol. 112: 1259-1270.

Allen, D. C., F. B. Knight and J. L. Foltz. 1970. Invertebrate predators of the
jack-pine budworm, Choristoncura pinus, in Michigan. Ann. Entomol.
Soc. Amer. 63: 59-64.

Andrews, E. A. 1925. Growth of ant mounds. Psyche. 32: 75-87.

Andrews. E. A. 1926. Sequential distribution of Formica cxsectoidcs Forel.
Psyche. 33: 126-150.

Andrews, E. A. 1929. The mound building ant, Formica exscctoides F.,
associated with tree hoppers. Ann. Entomol. Soc. Amer. 22: 370-391.

Averill, R. D., L. F. Wilson and R. F. Fowler. 1982. Impact of the redheaded
pine sawfly (Hymenoptera: Diprionidae) on young red pine plantations.
Great Lakes Entomol. 15: 65-91.

Banks. C. J. and E. D. M. Macaulay. 1967. Effects of Aphis fabae Scop. and of its
attendant ants and insect predators on yields of field beans (Vicia faba
L). Ann. appl. Biol. 60: 445-453.

Batzer, H. O. and D. T. Jennings. 1980. Numerical analysis of a jack pine
budworm outbreak in various densities of jack pine. Environ. Entomol.
9: 514-524.

Batzer, H. O. and I. Millers. 1970. Jack pine USDA forest Service pest leaflet. 7.

Bradley, G. A., and J. D. Hinks. 1968. Ants, aphids and jack pine in Manitoba.
Can. Entomol. 100: 40-50.

Bradley, G. A. 1972. Transplanting Formica obscuripes and D. lichodcrus
tashcnbcrgi (Hymenoptera: Formicidae) colonies in jack pine stands of
southeastern Manitoba. Can. Entomol. 104: 245-249.

Bremner, O. H. 1931. Relation of ants to aphis control. California Cult. 77:
488-489.

Bronstein, J. L. 1988. Predators of fig wasps. Biotropica. 20: 215-219.

29

Brown, M. W. and E. A. Cameron. 1982. Natural enemies of Lymantria dispar
(Lep. : Lymantriidac) eggs in central Pennsylvania, U. S. A., and a
review of the world literature on natural enemies of L. dispar eggs.
Entomophaga 27: 311-322.

Campbell, R. W. 1967. The analysis of numerical change in gypsy moth
populations. For. Sci. Monogr. 15: 1-33.

Campbell, R. W., and T. R. Torgersen. 1982. Some effects of predaceous ants
on western spruce budworm pupae in north central Washington.
Environ. Entomol. 11: 111-114.

Campbell. R. W., T. R. Torgersen, and N. Srivastava. 1983. A suggested role for
predaceous birds and ants in the population dynamics of the western
apmce budworm. For. Sci. 29: 779-790.

Campbell, R. W., N. Srivastava. and T. R. Torgersen. 1985. Predation by birds
and ants affects aggregation of the western spruce budworm. Proc.
IUFRO Conf. in: MW
W3. Can. For. Serv.. Pacific For. Res.Centre. pgs. 164-168.

Carroll, C. R. and D. H. Janzen. 1973. Ecology of foraging by ants. Annual
Review of Ecology and Systematics 4: 231-257.

Compere, H. 1940. Parasites of the black scale, Saissctia oleac. in Africa.
I-Iilgardia 13: 387-425.

DeBach, P.. C. A. Fleschner, and E. J. Dietrick. 1951. A biological check
method for evaluating the effectiveness of entomOphagous insects.
J. Econ. Entomol. 44: 763-766.

Dirks. C. O. 1964. The white pine weevil in Maine: its biology and dispersal
and the effect of prompt clipping of infested leader on trunk form. Me.
Agr. Exp. Sta. Bull. 625. 23 pp.

Dixon, A. F. G. 1971a. The role of aphids in wood formation. J. appl. Ecol. 8:
165-179.

Dixon, A. F. G. 1971b. The role of aphids in wood formation II. J. appl. Ecol. 8:
393-399.

Dixon, J. C. 1961. The biology and ecology of the jack pine budworm in
Wisconsin with special reference to insect parasites. Unpublished Ph. D.
Thesis. Univ. of Wisconsin, Madison.

Doane, C. C. 1970. Primary pathogens and their role in the development of
an epizootic in the gypsy moth. J. Invertr. Pathol. 15: 21-33.

30

Dutcher, I. D. and D. C. Sheppard. 1981. Predation of pecan weevil larvae by
red imported fire ants. J. Ga. Entomol. Soc. 16: 210-213.

Elkinston, J. S. and A. M. Liebhold. 1990. P0pulation dynamics of gypsy moth
in North America. Annu. Rev. Entomol. 35: 571-596.

Elliot. N. 1985. Studies of the parasites of the jack pine budworm: Life
history studies; hyperparasitism; and the application of time series
analysis to the study of adult activity patterns. Unpublished Ph. D.
thesis. Michigan State University.

Elliot, N. C., G. A. Simmons and D. L. Haynes. 1986. Mortality of pupae of jack
pine budworm (Lepidoptera: Tortricidae) parasites and density
dependence of hyperparasitism. Environ. Entomol. 15: 662-668.

Eskafi, F. M. and M. M. Kolbe. Predation on larval and pupal Ccrtitis capitata
(Diptera: Tephritidae) by the ant Solenopsis geminata (Hymenoptera:
Formicidae) and other predators in Guatemala. Environ. Entomol. 19:
148-153.

Faeth, S. H. 1980. Invertebrate predation of leaf-miners at low densities.
Ecol. But. 5: 111-114.

Finnegan, R. J. 1971. An appraisal of indigenous ants as limiting agents of
forest pests in Quebec. Can. Entomol. 103: 1489-1493.

Finnegan, R. J. 1975. Introduction of a predaceous red wood ant Formica
lugubris (Hymenoptera: Formicidae) from Italy to eastern Canada. Can.
Entomol. 107: 1271-1274.

Finnegan, R. J. 1977. Establishment of a predaceous red wood ant Formica
obscuripcs (Hymenoptera: Formicidae) from Manitoba to eastern
Canada. Can. Entomol. 109: 1145-1148.

Finnegan, R. J. 1978. Predation by Formica lugubris (Hymenoptera:
Formicidae) on Choristoncura fumifcrana (Lepidoptera: Tortricidae)
Can. For. Serv.. Bi-Mon. Res. Note 34: 3-4.

Flanders, S. E. 1943. The Argentine ant versus the parasites of the black
scale. Calif. Citrograph 28: 117- 128.

Flanders, S. E. 1951. The role of the ant in the biological control of
hymenopterous insects. Can. But. 83: 93-98.

Fritz, R. S. 1982. An ant-treehopper mutualism: effects of Formica subscricea
on the survival of Vanduzzca arquata. Ecol. Entomol. 7: 267-276.

Furuta. K. and C. Koizumi. 1975. The mortality factors of experimental
populations of Lymantria dispar in a larch plantation and natural
forest. J. Jpn. For. Soc. 57: 432-435.

31

Goesswald, K. 1940a. Observations on the protection of pine plantations
against Diprion pini. A. f Forst-u. Jagdwesen 72: 370-371.

Goesswald. K. 1951. WWW.
Forstwirtschaftliche, Bedeutung, Nutzung, Lebensweise, Zucht.
Vermehrung und Schutz. M. Kinau Verlag, Lunegerg. 160 pp.

Gotwald, W. H., Jr. 1986. The beneficial economic role of ants, pgss. 290-313.
In: W Vinson.S B (Ed)

Graham. S. A. 1926. Biology and control of the white pine weevil, Piss-odes
strobi (Peck). Cornell Univ. Agr. Exp. Sta. Ithaca, N. Y. Bull. 449. 32 pp.

Graham, S. A. 1935. The spruce budworm of Michigan pine. Univ. Mich.
School For. and Cons. Bull. 6.

Green, G. W., and C. R. Sullivan. 1950. Ants attacking larvae of the forest tent

caterpillar, Malacosoma disstria an. (Lepidoptera: Lasiocampidae). Can.
Entomol. 82: 194-195.

Haney, P. B., R. F. Luck and D. S. Moreno. 1987. Increases in densities of the
citrus red mite Panonychus citri (Acarina: Tetranychidae), in
association with the argentine ant, ridomyrmcx humilis (hymenoptera:
Formicidae) in southern California citrus. Entomophaga 32: 49-57.

Hanski, I. and P. Parviainen. 1985. Cocoon predation by small mammals, and
pine sawfly population dynamics. Oikos 45: 125-136.

Harman, D. M. and J. B. Wallace. 1971. Description of the immature stages of
Lonchaca corticis, with notes on its role as a predator of the white pine
weevil, Pissodcs strobi. Ann. Entomol. Soc. 64: 1221-1226.

Haviland, E. E. 1938. Mound changes after ten years in colonies of
Formica cxsectoidcs Forel. Ann. Ent. Soc. Amer. 31: 50-57.

Haviland, E. E. 1947. Biology and control of the Allegheny mound ant.
J. Econ. But. 40: 413-419.

Headings, M. E. 1971. The tree-foraging activity of Formica obscuripes
Forel. and its influence on arthropods of jack pine. Unpublished M. S.
Thesis. Dept. Entomol., Michigan State Univ., E. Lansing. 117 pp.

Headley, A. E. 1943. The ants of Ashtabula County, Ohio (Hymenoptera:
Formicidae). Ohio J. Science 43: 22-31.

Hill, H. R. 1969. The effect of the ant. Formica obscuripes Forel. on jack pine
arthropods. Unpublished M. S. Thesis, Dept. Entomol., Michigan State
Univ., E. Lansing. 60 pp.

32

Huang, H. T. and P. Yang. 1987. The ancient cultured citrus ant.
BioScience 37: 665-671.

Jennings. D. T. 1971. Ants preying on dislodged jack-pine budworm
larvae. Ann. Entomol. Soc. Amer. 64: 384-385.

Kloft. W. J. 1953. Wood ants and plant aphids. Allg. Forstz. 8: 529.

Kloft. W. J. 1959. Attempt to analyze the trophic relationship of ants to
aphids. Biol. Zbl. 78: 863-870.

Kloft. W. 1., R. C. Wilkinson, W. H. Whitcomb, and E. S. Kloft. 1979.
Formica integra (Hymenoptera: Formicidae) 1. Habitat, nest
construction, polygyny, and biometry. Fla. Entomol. 56: 67-76.

Laine, K. J. and P Niemela. 1980. The influence of ants on the survival of
mountain bitches during an Oporinia autumnata (Lep., Geometridae)
outbreak. Oecologia 47: 39-42.

Leonard, D. E. 1981. In:
W Forest Service Technical Bulletin 1584. pp 9- 28.

Lyons, L. A. 1977. On the population dynamics of Neodiprion sawflies. Minn.
Agri. Exp. Stn. Bull. 310: 48-55.

MacAloney, H. J. and L. F. Wilson. 1964. The red-headed pine sawfly. US.
Department of Agriculture Forest Pest Leaflet 14. 5 pp.

Majer, J. D. 1986. Utilizing economically beneficial ants, In: W
W (Ed) 8 B Vinson pp 314-331

McNeil, J. N., J. Delisie and R. J. Finnegan. 1978. Seasonal predatory
activity of the introduced red wood and, Formica lugubris
(Hymenoptera: Formicidae), at Valcartier, Quebec, in 1976.

Can. Entomol. 110: 85-90.

Messina, F. J. 1981. Plant protection as a consequence of ant ant-
membracid mutualism: interactions on goldenrod (Solidaga sp.)
Ecology 62: 1433-1440.

Montgomery, B. A. 1988. Michigan Forest Pest Report. Mich. Cooperative For.
Pest Management Program. 81 pp.

Mosher, D. G. 1980. Michigan Forest Pest Report. Mich. Cooperative For. Pest
Management Program. 32 pp.

Muller, H. 1956a. Can honeydew producing aphids cause damage to their host
trees? 2. and. But. 39: 168-177.

33

Muller, H. 1956b. Mass increase of some honeydew producing aphids in
1954. Insectes Sociaux. 3: 75-90.

Muller, H. 1958. Contribution to the knowledge of damages caused by
lachnids to their host trees. 2. ang. But. 42: 284-291.

Peirson, H. B. 1922. Mound building ants in forest plantations.
J. of Forestry. 20: 325-336.

Plummet, C. C. and A. E. Pillsbury. 1929. The white pine weevil in New
Hampshire. NH. Exp. Sta. Bull. 247, pp.32.

Risch, S. J. and C. R. Carroll. 1982a. The ecological role of ants in two
Mexican agroecosystems. Oecologia 55: 114-119.

Room P. M. 1973. Control by ants of pest situations in tropical trees: A
strategy for research and development. Paupua New Guin. Agri. J. 24:
98-103.

Sanders, C. J. 1964. The biology of carpenter ants in New Brunswick. Can.
Entomol. 96: 894-909.

Sato, H. and S. Hegashi. 1987. Bionomics of Phyllonorycter
(Lepidoptera, Gracillariidae) on Qucrcus. II. effects of ants.
Ecol. Res. 2: 53-60.

Samways, M. J., N. Magda and A. J. Prins. 1982. Ants (Hymenoptera:
Formicidae) foraging in citrus trees and attending honeydew-
producing Homoptera. Phytophylaciica 14: 155-157..

Schotter, H. 1959. The effects of F. polyctcna. Forst u. Jagd. 9: 40-41.

Skinner, G. J. 1980. The feeding habits of the wood-ant Formica rufa
(Hymenoptera: Formicidae) in limestone woodland in nort-east England.
J. of Animal Ecology. 49: 417-433.

Stary, P. 1966. Aphid parasites (HYM., Aphidiidae) and their relationship to
aphid attending ants, with respect to biological control. Insectes
Sociaux 13: 185-202.

Sterling, W. L., D. Jones and D. A. Dean. 1979. Failure of the red
imported fire ant to reduce emtomophagous insect and spider
abundance in a cotton agroecosystem. Environ. Entomol. 8:
976-981.

Tedders. W. L., C. C. Reilly, B. W. Wood, R. K. Morrison and C. S. Lorgren. 1990.
Behavior of Solenopsis invicta (Hymenoptera: Formicidae) in pecan
orchards. Environ. Entomol. 19: 44-53.

34

Torgersen, T. R. and R. R. Mason. 1987. Predation on egg masses of the
douglas-fir tussock moth (Lepidoptera: Lymantriidac). Environ.
Entomol. 16: 90-93.

US. Department of Agriculture. Christmas Tree Pest Manual. North-central
Forest Exp. Sta., St. Paul, MN. 108 pp.

Way, M. J. 1963. Mutualism between ants and honeydew-producing
Homoptera. Ann. Rev. But. 8: 307-344.

Wellenstein, G. 1952. On the nutrition and biology of red wood ants. Z.
Pflanzenkr. 59: 430-451.

Wellenstein, G. 1954. What can we expect from the red wood ant for the
forest protection? Beitr. But. 4: 117-138.

Wellenstein, G. 1961. More results on the beneficial use of red wood ants for
bee farming. Atti IV Congresso U.I.E.I.S. Pavia 12: 60-73.

Weseloh, R. M. 1988. Effects of microhabitat, time of day. and weather on
predation of gypsy moth larvae. Oecologia 77: 250-254.

Wilkinson, R. C., A. P. Bhatkar, W. J. Kloft. W. H. Whitcomb, and E. S. Kloft.
1978. Formica integra 2. Feeding. tropholaxis, and interspecific
confrontation behavior. Fla. Entomol. 61: 179-187.

Wilkinson, R. C., A. P. Bhatkar, W. H. Whitcomb, and W. J. Kloft. 1980.
Formica integra (Hymenoptera: Formicidae) 3. Trial introduction into
Florida. Fla. Entomol. 63: 143-146.

Wilson, E. O. 1971. MW. Belknap Press of Harvard Univ.
Press, Cambridge, MA. 548 pp. '

Wong, T.T. Y., D. O. Mcinnis, J. I. Nishimoto, A. K. Ota, and V. C. 8. Chang. 1984.
Predation of the Mediterranean fruit fly (Diptera:Tephritidae) by the
Argentine ant (HymenOptera: Formicidae) in Hawaii. J. Econ. Entomol.
77: 1454-1458.

Wong. M. A. and T. Y. Wong. 1988. Predation of the Mediterranean fruit fly
and the Oriental fruit fly (Diptera: Tephritidae) by the fire ant
(HymenopterazFormicidae) in Hawaii. Proc. Hawaiian Entomol. Soc. 31:
169-177.

Youngs, L. C. 1983. Ant predators of the western spruce budworm
Choristoncura occidentalis Freeman): Species composition and
patterns of occurrence in eastern Oregon and western Montana.
Unpublished Ph. D. thesis, Dept. Entomology. Oregon State Univ.,
Corvallis.

35

Youngs, L. C., and R. W. Campbell. 1984. Ants (Hymenoptera: Formicidae)
preying on pupae of the western spruce budworm (Lepidoptera:
Tortricidae) in eastern Oregon and western Montana. Can. Entomol.
116: 1665-1669.

MANUSCRIPT l.

Predation of Formica exscctoidcs on four pine pests in Michigan.

36

37

Introduction

Predaceous red wood ants, Formica rufa group, have been exploited in
intensive forestry culture for over 150 years (Adlung, 1966). European efforts
have been aimed at encouraging the growth, mass rearing. and dispersal as
well as protection of native ant species in forest ecosystems. Researchers in
North America have tried introducing species from Europe to provide
protection of forests from insect pests, such as the spruce budworm (Finnegan,
1975; McNeil et al., 1978).

Studies that have been conducted in North America have disclosed that
native carpenter ants in the genus Camponotus and mound-building ants in
the genus Formica both feed on insects of economic importance in forestry
(Finnegan et a1. 1978; Campbell and Torgersen, 1982; Torgersen and Mason,
1987). More recently, ants in the genus Solenopsis, have proved to be valuable
in reducing insect pests in pecan and coffee and orange orchards,
respectively (Tedders et al., 1990; Eskafi and Kolbe, 1990).

Many researchers have investigated utilizing predaceous ants because
they are capable of removing a large number of insect prey. Wellenstein
(1952) estimated eight million insects taken as prey by one medium-size mound
of red wood ants during one season. In addition to preying on insects, many
Species of ants obtain honeydew by tending aphids (Way, 1963). The effects of
the interactions between the ants and homoptera in relation to biological
control can be either positive or negative, depending upon several factors.
Ant attendance of certain species of aphids is known to increase the
population of homoptera to economically injurious levels (DeBach et al., 1951;
Flanders, 1951). In addition to damaging the host plant, ant attendance can

negatively effect the natural enemy complex of aphids (Samways et al., 1982;

38

Haney et al., 1987). Other studies have shown that ant attendance of homoptera
does not interfere with natural control but instead, increases predation (Stary,
1966). Flanders (1943) observed that heavy infestations of black scale that
were attended by ants were completely parasitized.

Several North American studies have focused on the predatory behavior
of the thatching mound ant, Formica obscuripcs Forel, while the mound ant.
Formica cxsectoides Forel, has received little attention. Early studies examined
the growth and distribution of mounds of F. cxsectoidcs (Pierson, 1922;
Andrews, 1925, 1926; Haviland, 1938). Haviland (1947) examined the life history
and control of F. exscctoides. In the context of this study, she noted workers of
F. cxsectoidcs preying upon beetles, spiders and other insects.

F.I exscctoides occurs primarily in the eastern United States but is also
scattered throughout other parts of the country (Andrews. 1926). It can
construct as many as 100 to 400 mounds per acre (Allen et al., 1970). It is a
large, aggressive ant that will feed on both live and dead arthropods and tend
aphids for honeydew (Andrews, 1929). Allen et al. (1970) observed F.
exsectoidcs preying upon jack pine budworm larvae that were outside their
feeding chambers. F. cxsectoides possesses several other characteristics that
Finnegan (1971) deemed important if a predaceous ant species were to be
successful as a biological control agent in a forest ecosystem. It forages in all
strata of the forest throughout the day and night and is polygynous (multiple
queens), which increases the longevity of the mound.

F. exsectoidcs is frequently found in association with jack pine forests
in parts of the lower peninsula of Michigan. This study examined the
interactions between F. cxscctoidcs, aphids, Cinara sp., and four potential

insect pests of jack and red pines,: jack pine budworm, Choristoncura pinus

39

(Freeman), redheaded pine sawfly, Neodiprion lccontci (Fitch), white pine
weevil. Pissodcs strobi (Peck), and the gypsy moth, Lymantria dispar (L). Ant
predation was examined on several developmental stages for each insect. Ant
predation on the gypsy moth was included in this study because ant mounds
were located near stands of hardwoods that were interspersed with jack and
red pines. The effects of ant protection of jack pines resulting from predation

of the jack pine budworm was also measured.

40

Materials and Methods

This study was conducted in a mature jack pine forest in the Huron-
Manistee National Forest in Crawford County, Michigan. The overstory was
predominantly jack pine, Pinus banksiana Lamb., interdispersed with both
white pine, Pinus strobus L., and oak, Quercus spp. Mounds of F. cxscctoides
were scattered throughout the study area.

Ant predation was assessed by direct measurements of the number of
insects removed from trial trees and indirectly through periods of
observation. Ant predation was examined on several developmental stages for
each insect. In 1987 and 1988, when these studies were conducted, the
following pest conditions were reported for the Huron-Manistee forest:

1) Jack pine budworm populations were at low levels region wide with a trap
catch of 23 male moths at a site in Crawford County; 2) In 1987, the median
percent weeviling rate for the Lower Peninsula was 43 percent;

3) Redheaded pine sawflies were not detected at the study sites. Small sawfly
larvae were collected at a mixed jack and red pine stand in Alpena County,
Michigan and utilized in this study; and 4) The gypsy moth was present in
Crawford County during 1987 but at levels below what is considered to cause
significant defoliation (Michigan Forest Pest Report, 1987). However, the
population of gypsy moth has increased lOO-fold since 1987, and will cause
some noticeable defoliation in Crawford County in 1990.

Experimental arenas were used for all of the experiments that examined
ant predation on planted pepulations of insects. Large mounds of F. exsectoides
that appeared very active, were selected non-randomly, to be the center of

each experimental arena. The selected mounds were relatively free of plant

41

growth on the surface which usually is a good indicator that a mound is active.
Secondly, mounds that were selected had a minimum of two "major" foraging
trails.

Eight or nine jack pines were planted 2.8 m from the edge of the nest,
encircling the mound. Oak saplings were planted in the same pattern where
ant predation was examined on the gypsy moth. The trees were planted
equidistantly, except when a naturally occurring large tree interfered with
this pattern (See Fig. 1). Trees that were > 2 m tall were removed from the
experimental arenas. Control trees (three to four depending upon the
experiment) were wrapped with a 8 cm sticky tape barrier to prevent ant
foraging; trial trees were three to four trees where the ants foraged freely.
The lowest whorl of branches of the trees were removed. making the trunk the
only way the ants could access the trees.

Before the tests, the experimental trees were shaken vigorously and
then carefully searched to remove foliage-feeding insects, predaceous insects
and spiders. To prevent unwanted vertebrate predation, l-cm hardware cloth
cages were placed over the trees with a 5 cm clearance at the bottom, thus
allowing the ants access to the trees.

A one-way analysis of variance was used to test for differences between
the number of larvae remaining on trees after 24 h. Duncan's multiple range
test and Fisher's PSLD were used to test for specific differences between
treatments. Regression analysis was used to examine the relationship between

the number of ants and insect pests on the trees.

42

Figure 1. Experimental arenas were used for all experiments where ant
predation was examined on planted populations of insect pests. Cages
were placed over the trees to prevent unwanted vertebrate predation.

 

43

 

44

Experiment 1. WWW

Ant predation was examined on larvae and pupae of the RHPS. The influence
of aphids and larval size on ant predation was also examined.

A. Larval Stage:

Two colonies each with 30 fourth-instar RHPS were placed on opposite sides of
the tree crown on each of the four trial trees for each experimental arena
during July of 1987. The treatments were: 1) ants and JPBW; and 2) JPBW only.
The parameters measured on each tree were the number of larvae and the
number of ants found respectively every 12 hours for 3 days. Immediately
after placing the larvae on the trees and then every 15 minutes, observations
were made to determine how quickly the ants would discover the larvae. The
experiment was conducted at three different ant mounds (experimental
arenas) when fourth instar RHPS were naturally occurring in the field
making a total of 12 trial and 12 control trees, each with 60 fourth-instar

larvae.

B. Influence of Aphids:

Everything was done as in the above experiment except two small pine aphid
colonies. Cinara sp.. were transferred to the terminal leader of each of three
trial trees for each experimental arena. This was done by field collecting pine
shoots containing aphids tended by F. cxscctoidcs nearby. The clipped shoots
were attached to the terminal leader of the tree in such a way that as the twig
dried out, the aphids moved to the live shoot. An aphid colony consisted of 30
to 50 individual aphids. The larvae were placed on the trees after the ants
began tending the aphids. The treatments were: 1) ants and RHPS; and 2) ants,

RHPS, and aphids. Control trees contained two colonies of RHPS and were

4S

banded with sticky tape to exclude ants. Each treatment was replicated on
three jack pines and at four different ant mounds; making a total of 24 trial

and 12 control trees, each with 60 fourth-instar larvae.

C. Influence of Larval Size:

Three distinct stages of RHPS larvae were placed on each of three trees for
each experimental arena. The treatments were: 1) first instar RHPS and ants;
2) third instar RHPS and ants; and 3) fifth instar RHPS and ants. Controls were
three trees containing larvae of the three respective sizes that were handed
with sticky tape to prevent foraging by the ants. For the trial trees, the larvae
were allowed to establish feeding sites before the ants were permitted to
forage. Each treatment was replicated on three jack pines and at four
different ant mounds: making a total of 12 trial and 12 control trees, each with

60 larvae.

Experiment 2. WWW

Ant predation was examined on second and fourth instars and pupae of JPBW.
The influence of aphids and the silk chamber of the JPBW on ant predation
was also examined. JPBW larvae and pupae were obtained by collecting
individual JPBW and their associated webbing from mature jack pine trees

near the study sites.

A. Larval Stage:
Thirty, fourth-instar JPBW were placed on each of eight trees during June of
1987 for each experimental arena Four of the eight trees were banded with

sticky tape and served as controls. The larvae were allowed 48 hours to

46

establish before the ants accessed the trial trees. The parameters measured on
each tree were the number of larvae and the number of ants found
respectively every 12 hours for 3 days. This experiments was conducted at

four different ant mounds; making a total of 16 trial and 16 control trees.

B. Influence of Aphids:
Ant predation on second instar and JPBW pupae was examined during the
spring and summer of 1988.

Fifteen, second instar JPBW were placed on each of nine trial trees for
each experimental arena. The treatments were: 1) ants and JPBW, 2) ants.
JPBW, and aphids. The controls were three trees banded with sticky tape each
containing 15 larvae. The parameters measured on each tree were the
number of larvae and the number of ants found respectively every 12 hours
for 3 days. In addition to direct measurements, a total of 26 hours of
observations consisting of 15 minute blocks, were conducted on thourse
interactions between thourse ants and JPBW. The following interactions were
recorded: 1) ant made antennal contact with a larva or a silk chamber, 2) ant
made mandibular contact with a larva or a silk chamber, 3) ant(s) carried a
larva down a tree, 4) the JPBW dropped from the tree in response to ant(s). and
5) ant encountered a JPBW or its silk chamber and appeared to ignore it. Each
treatment was replicated on 3 jack pines and at 3 different ant mounds; making
a total of 18 trial and 9 control trees.

To examine ant predation on the pupal stage of the JPBW, 15 pupae and
their associated webbing were placed on each of nine trees for each
experimental arena. The treatments were the same as above except pupae

were used instead of second instar JPBW. The parameters measured on each

47

tree were the number of pupae and the number of ants found respectively
every 12 h for 3 days. Ten hours of observations consisting of 15-minute
blocks were conducted. The following interactions were recorded: 1) ant made
antennal contact with a pupa or its silk chamber, 2) ant made mandibular
contact with a pupa or its silk chamber, 3) ant(s) carried a pupa down the tree,
and 4) ants removed fragments of a pupa. Each treatment was replicated on 3
jack pines and at 3 different ant mounds; making a total of 18 trial and 9

control trees.

C. The Influence of the Silk Chamber of the JPBW:

For this study the treatments were: 1) exposed: larvae without a silk chamber,
and 2) intact (control); larvae that were protected inside their silk chambers.
Fifteen larvae from the intact treatment were placed on each of four trees and
15 larvae from the exposed treatment were placed on each of four trees for
experimental arena. Intact larvae were placed on the trial trees 2 days prior to
placing the ”exposed” larvae on trees. This was done to insure that the larvae
had enough time to construct silk chambers. The trees were handed with
sticky tape during this period to prevent ant predation. Exposed larvae were
placed on the experimental trees approximately 1 h before the start of the
study. The majority of the new silk produced by the exposed larvae was
removed at 3 h intervals throughout the study. The parameters measured for
each tree were the number of larvae and the number of ants found
respectively at every 12 h for 3 days. Each treatment was replicated on 4 jack
pines and at 3 different ant mounds; making a total of 12 trial and 12 control

ITOOS .

48

D. Protection of Jack Pine Trees by Ants from JPBW Defoliation:

The objectives of this study were to examine the influence of ant predation on
field populations of JPBW and to determine the amount of protection resulting
from the presence of ants on mature jack pine trees.

This study was conducted at three jack pine stands where there was
considerable variation in the number of ant mounds per hectare. The
characteristics of both the ant mounds and the stand were measured at 80
different sites within the jack pine stands. The sites were selected by walking
a minimum distance of 12 m in any cardinal direction from the last ant mound
sampled. The first mound encountered, was used as the base mound. All
additional mounds found within an 8 m radius from the base mound also were
evaluated for mound characteristics for that site. The following stand
characteristics also were measured at each site: 1) stand basal area. 2) heights
of all trees from which branch samples were collected. 3) the number of tips
defoliated by the jack pine JPBW, and 4) the percent defoliation of the current
year’s growth due to JPBW feeding.

One branch sample was cut from the upper mid crown section of each
of four trees at the 80 different sites. One hundred tips per branch were
evaluated according to the Montgomery et al. (1982) method for assessing
defoliation due to feeding by spruce budworm. Choristoneura fumifcrana
(Clemens). The number of pupae found in the branch samples was also
recorded. The height of each tree that branch samples were taken from was
visually estimated. Stand basal area was estimated using a lO-factor prism.

Three variables were measured for all of the trees (<0.1 m DBH) in the
understory: 1) the number of pupae. 2) the number of tips that were defdliated

by JPBW, and 3) the number of aphid colonies.

49

Both the number and the size of all ant mounds located within the 8 m
radius from the base mound, were determined. Each mound was visually
‘ inspected and quantitatively rated for size from 1 to 3, with 3 being the largest.
The values from the size of the mounds and the number of mounds were used to
create a mound index for each site (mound size A + mound size B, etc). The
values for the mound indices ranged from 0 to 13. Regression analysis was

used to examine the relationships between stand and mound characteristics.

Espcriment 3. WW
Ant predation was examined on egg masses, fourth instar larvae, pupae, and
adult female GM. The influence of the setae of GM caterpillars on ant

predation was also tested.

A. Adult Stage:

During the summer of 1987, 108 adult female GM were collected in the field and
placed on oak saplings around ant mounds in the experimental arenas. Three
GM adults were placed on each of nine trees for each experimental arena. The
treatments were: 1) ants and GM: and 2) ants, GM, and aphids. Control trees
were handed with sticky tape and each contained three female adult gypsy
moths. I The ants were allowed to access the trees only after all of the gypsy
moth adults began depositing eggs. The parameters measured on each tree
were the number of GM found every 2 h until there were no GM remaining.
Each treatment was replicated on 3 jack pines and at 4 ant mounds; making a

total of 24 trial and 12 control trees.

50

B. Egg Mass Stage:

After the ants had removed the majority of the adults from the trial trees. the
remaining adults were removed while leaving the egg masses intact. The
percent of individual egg masses that appeared to have been tampered with or

removed by the ants was estimated every 6 h for a 24 h period.

C. Influence of the Setae of Fourth Instars:

The influence of the setae in deterring ant predation was examined by

clipping back the setae of fourth instar GM. The treatments were:

1) caterpillars with clipped setae; and 2) caterpillars with setae intact. The
control trees were handed with sticky tape and each contained 10 larvae. Three
hundred sixty caterpillars were collected in the field and brought into the
laboratory where they were anesthetized with carbon dioxide. The setae of one
third of the caterpillars were clipped back to approximately one fourth of the
original length with a razor blade. Ten larvae were placed on each of the trial
trees in experimental arenas. Each treatment was replicated on 3 jack pines
and at four different ant mounds; making a total of 24 trial and 12 control
trees. The parameters measured for each tree were the number of caterpillars

found respectively every 6 hours for a 24-hour period.

D. Pupal Stage:

During the summer of 1988, ant predation was examined on gypsy moth pupae.
The influence of distance of the prey item from the mound was also examined.
Jack pine trees that were less than 4 m tall were randomly selected at varying

distances from mounds of F. cxscctoides that appeared active. Four pupae were

51

placed on each of 28 trees in hand-crafted pupation chambers constructed out
of yarn and designed to mimic natural gypsy moth pupation sites.
Approximately 60 percent of the trees were between 0 < 5 m, 25 percent were
between 5 < 10 m, and 14 percent were > 10 m from the mound. The parameters
measured for each tree were the number of pupae found respectively every 6
hours for a 48-hour period. Regression analysis was used to infer the

influence of distance from the mound on ant predation of gypsy moth pupae.

531”me 4- WWW
Ant predation was examined on adult WPW in the presence and absence of
aphids. Two hundred forty adult WPW were collected from a nearby young
jack pine plantation. Adult WPW were trapped in nylon bags as they emerged
from pupation. On August 15, 1987, 20 WPW were placed on each of six trial
trees. On two of the trees, aphid colonies were transferred to the terminal
leaders before the WPW were placed on the trees. The treatments were: 1) ants
and WPW; 2) ants, WPW, and aphids. Control trees were handed with sticky tape
and each contained 20 weevils. To prevent unwanted vertebrate predation,
nylon bags were placed over the trial trees. The bags were drawn closed by a
cord underneath the bottom whorl of branches. thus allowing only a small
opening at the trunk for the ants to access the trees. Each treatment was
replicated on two trees and at two ant mounds; making a total of eight trial
trees and four control trees.

The number of WPW remaining on the trees was recorded every 24 h for
3 days. In addition, 24 observation periods, consisting of 30-minute blocks,
were conducted where the following interactions between the ants and WPW

were recorded: 1) the WPW was feeding; _2) the ant made antennal contact with

52

a WPW; 3) the ant made mandibular contact with the WPW; 4) the ant ignored
the WPW (the ant walked on or by the WPW); 5) the WPW stopped feeding
because of harassment from the ant(s); 6) the WPW stopped feeding on its own;
and 7) the ant stopped tending aphids and responded aggressively toward the
WPW. A series of interactions were classified as either persistent or non-
persiStent. A persistent interaction was one where the ant continued to harass
the WPW until the WPW stOpped feeding, moved to a new feeding location or
the ant removed it from the tree. A non-persistent interaction was one in
which the ant ignored the WPW even after walking on the weevil or nearby

the weevil.

53

Results

Experiment 1- WWW

During the experiment, F. cxsectoides was repeatedly seen attacking and
removing fourth instar RHPS from the experimental trees (Fig. 2).

F. exsectoidcs discovered the colonies of fourth-instar RHPS within 3 hours
after they first ascended the trees. After 12 hours. control trees (without
ants). showed an average of 57.67 1- 0.36 larvae per tree while trial trees had an
average of 44.33 i 2.60 larvae per tree. After 24 hours, the ants significantly
reduced (p<0.0005) the number of RHPS larvae on trial trees to an average of
19.33 -_+_ 4.94 while control trees retained an average of 55.42 1 2.11 (Table 1).
The average number of larvae remaining on trial trees after 36 hours was 0.46
i 0.29 versus an average of 54.17 3 0.72 on control trees.

The ants removed the RHPS larvae more quickly when aphids were
present. F. exscctoidcs discovered the colonies of RHPS on trial trees
containing aphids within 1 hour. while it took as long as 4 hours for the ants
to discover RHPS on trial trees without aphids. Three times as many ants were
observed foraging on trees with aphids present as compared to trees with
RHPS larvae only.

After 12 hours, the ants reduced the number of larvae on trees with
RHPS only to an average of 49.56 :t 1.81 while the average number remaining
on trees with aphids present was 40.78 :t 4.23. The control trees retained an
average of 56.11 :1; 0.74 larvae. After 24 hours. the ants significantly reduced
(p<0.0001) the number of RHPS larvae on trees with RHPS only to an average of

54

Figure 2. Workers of F. exsectoides readily preyed on fourth-instar redheaded
pine sawfly.

55

 

56

Table 1. Mean number of fourth-instar RHPS remaining on experimental jack
pine trees 24 h after placing them on the trees. Initially. each tree
contained 60 larvae.

 

 

t-te st
Mean Std. Error Count P
Treatment 19.33 4.94 12 < 0.0005

Control 55.42 0.61 12

 

57

42.42 i 7.90, and on trees with RHPS and aphids to an average of 1.33 .1: 2.31
(Table 2). There were significant differences between treatment means at 24 h
(p<0.05) (Fig. 3).

Regression of ant numbers on sawfly numbers was significant
(p<0.0001) on trees with RHPS only, while not significant (p<0.50) on trees with
both RHPS and aphids present. On trees with RHPS only, the number of
sawflies decreased as the number of ants increased (Fig. 4). There was not a
direct relationship between the number of ants and the number of sawflies on
trees when aphids were present. The number of ants on the trees remained
relatively high throughout the experiment regardless of the density of
sawflies (Fig. 5). Even after the majority of sawflies were removed from the
trees, the ants remained on the trees tending aphids. The total number of ants
found on trees that contained both aphids and RHPS was higher than on trees
with RHPS only.

The ants preyed more readily on smaller sawfly larvae than on the
larger fifth instars. Initially, the ants removed more fifth-instar RHPS than
either first or third instars. After 6 h, the ants reduced the number of fifth-
instar RHPS remaining on the trees to an average of 42.5 i 6.65. The average
number of third instar RHPS remaining was 46.0 3;. 4.64 and the average
number of first instar RHPS remaining was 49.0 1 4.22. The average number of
first instar RHPS remaining on trees was 38.67 16.00; the average number of
third instar RHPS remaining was 37.67 :l; 8.19; and the average number of fifth
instar larvae remaining was 39.0 i 8.54. However, after 24 h the trend
reversed, there was an average of 32.0 :t 4.93 fifth instar RHPS remaining;
19.25 :3; 0.75 third instar RHPS remaining; and an average of 22.75 ;_1-_ 3.71 first

instar RHPS remaining on the trial trees. After 36 h, an average of 29.25 1;

58

Table 2. Ant predation on fourth-instar RHPS in the presence and absence
of aphids in Crawford County, Michigan during 1987. Mean number
of larvae remaining on experimental trees after 24 h. Initially, each
tree contained 60 larvae.

 

 

ANOVA
Group: Count: Mean: Std. Error: F
With 12 1.33 2.31 394.09
aphids
Without 12 42.42 7.90 p < 0.0001
aphids

Control 12 55.17 2.17

 

59

70

60

 

 

50"
40-

30-

 

20-

   

——o— RHPS

AVE. NO. SAWFLIES

 

 

 

10: —A— RHPS-APHIDS
. —il— Control (no ants)
o . . . . . .
0 1 0 20 3 0
TIME (hours)

Figure 3. Mean number of sawflies on jack pines with RHPS only, RHPS and
ants, and RHPS, ants and aphids. Differences between treatment means

are significant at 24 h, 11: 12 (p < 0.05).

 

60

NUMBER OF SAWFLIES

 

 

 

NUMBER OF ANTS

Figure 4. Regression of sawflies and ants on jack pines after 24 h. The
regression was significant, p < 0.0001, (y = 1.088 x 4» 55.180, r 2: 0.63).

61

0'1
0

1
II

NUMBER OF SAWFLIES
8
a

 

I
0__..._.,_I_..._-_1. . I I , '
0 10 20 30 40

NUMBER OF ANTS

 

Figure 5. Regression of sawflies and ants in the presence of aphids on jack
pine trees after 24 h. The regression was not significant, p > 0.05.

(y = 0.35 x + 14.65, t2 = 0.025).

62

3.68 fifth instar RHPS were remaining on the trees while an average of 9.0 1’.
3.58 third instar RHPS were remaining and 9.0 1; 4.10 first instar RHPS were

remaining (Fig. 6).

Esperimcnt 2. W

F. easectoidcs significantly reduced (p<0.0001) the population of fourth instar
JPBW 72 h after placing them on trees. Trial trees contained an average of
10.19 :t 1.32 fourth instar JPBW while control trees retained on the average
20.63 1; 1.30 fourth instar JPBW (Table 3).

The ants removed second instar JPBW more quickly from trees that
contained aphids than trees with JPBW only (Table 4). The JPBW population
was significantly reduced (p<0.0001) by the ants 72 h after placing the second
instar RHPS on trees. The trees that contained both aphids and JPBW had an
average of 4.67 1; 0.76 larvae remaining as compared to trees that had JPBW
only had an average of 8.44 i 0.93. After 72 h, the control trees retained on the
average 12.22 .21; 0.52. There were significant differences (p<0.05) between
treatment means after 72 h.

The ants significantly reduced (p<0.0001) the population of JPBW pupae
(Table 5). There were significant differences between treatments (p<0.05) 72 h
after stocking the pupae as inferred by Fisher's multiple comparison test
(Statview 512, 1986). The average number of pupae remaining on trees with
JPBW and aphids was 3.11 1 1.24 while trees that did not contain aphids had 7.22
i 1.43 pupae remaining after 72 h. The number of pupae on the control trees
remained relatively high, 13.11 i 0.51. During the study we observed 16 pupae

being carried down the trees by foragers of F.cxscctoidcs, 11 of the 16 were on

63

.EE as...
_o=coo-m 85:.
ES has
_o::oo-m Lama.
.25.; .98.
Sacco; 86:.

805.

 

  
  

\\\\\\\\\\\\\\\\\\\\\\‘

 

 

  
 

 

 

 

em

_ u\\\\\\\\\\\\\\\\\\\\\\\‘

 

 

 

.2038on8 .m 365 can n
355 can J .335 28 m 365 5232. 3258:? :5 m ecu _ 33%:

5038p nouns—3:8 E moon—20:6 o: e032? cemEnEoo 23:32
Amod v 3 v u: .: em “a £52: 5232. moor—20:6 2.35:9.“ 803
.082 .85 353% .8 SEE: 09:: co c2333 :3 no flootm

APSE: WEE.

NF

  

9k\\\\\\\\\\\\\\\\\\\\\\\\\\\‘

 

 

 
     

 

  

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

 

 

 
 

R\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\‘k\\‘i

 

 

om

ov

co

om

.o 83E

SEII'IJAAVS 'ON 'EIAV

64

Table 3. Ant predation on fourth-instar JPBW in Crawford County, Michigan
during 1987. Mean number of larvae remaining on experimental jack

pine trees 24 h after placing them on the trees.

 

 

t-test
Mean: Std. Error: Count: P:
Treatment 10.19 1.31 16 0.0001

Control 20.63 1.30 16

 

65

Table 4. Ant predation on second-instar JPBW in the presence and absence of
aphids in Crawford County, Michigan. Mean number of larvae
remaining on jack pine trees after 72 h.

 

 

ANOV
Group: Count: Mean: Std. Error: F:
with
aphids 9 4.67 0.76 24.91
without 9 8.44 0.93 p < 0.0001
aphids

control 9 12.22 0.52

 

66

Table 5. Ant predation on JPBW pupae in the presence and absence of
aphids in Crawford County, Michigan. Mean number of larvae
remaining on trees after 72 h.

 

 

ANOV
Group: Count: Mean: Std. Error: F:
with
aphids 9 3.11 1.24 19.68
without 9 7.22 1.43 p < 0.0001
aphids

control 9 13.11 0.51

 

67

trees where aphids were present. Besides removing whole pupae, the ants also
returned to the mound with only fragments of pupae.

During the observation periods, a total of 157 ants were seen on trees
with aphids, ants and JPBW compared to a total of only 56 ants seen on trees
with ants and JPBW. The ants preyed readily upon both dislodged larvae and
larvae that were protected inside their silk chambers. Only on one occasion
did we see an ant encounter a budworm inside a silk chamber and ignore it.
Ants were seen carrying larvae down trees on three separate occasions during
the observation periods. Thirty five encounters were recorded between the
ants and JPBW, and for nine of these the larvae dropped from branches on silk
threads in response to the ants.

F. cxsectoidcs preyed upon both dislodged larvae and those that are
protected inside their silk chambers, however, the silk chambers did reduce
ant predation. Significantly more (p<0.0005) larvae were removed by the ants
that did not have silk chambers than those that were protected inside silk
chambers within 72 h after stocking the larvae on the trees (Tables 6). The
mean number of JPBW larvae remaining on trees that were protected by silk
chambers was 7.67;}; 0.88 compared to a mean number of 3.08 3; 0.82 remaining
on trees that did not have silk chambers.

During this study, 32 larvae were recorded dropping from branches in
response to ants for the ”exposed” treatment and 10 larvae were carried down
the tree by the ants. For the ”intact“ treatment, only 15 larvae dropped from
branches in response to ant probing and four larvae were seen being carried
down trees by ants. A total of 369 ants were recorded on trees with "exposed”
larvae during the 53 observation periods, while a total of 511 ants were seen on

trees where the larvae were protected inside their silk chambers.

68

Table 6. The influence of the JPBW silk chamber on ant predation in Crawford
County, Michigan. Mean number of larvae remaining on experimental
jack pine trees 72 h after placing them on the trees.

 

 

t-test
Mean Std. Error Count P
With chamber
(control) 7.67 0.88 12 p < 0.0005

Without
chamber 3.08 0.82 12

 

69

Field Survey of Pupal Densities. The regression of mound index vs.
number of pupae. r2 = 0.41 (Fig. 7) and the regression of the number of
mounds vs. the number of pupae. r2 = 0.35 (Fig. 8). both were highly
significant. but mound index was the better predictor of pupal densities. At
the sites where there were no ant mounds, the number of pupae ranged from 7
to 53 per 400 branch tips. When the mound index was greater than 8, no pupae
were found in the trees except in one instance.

The regression of mound index against average percent defoliation of
jack pine tips was significant (p<0.05), r2 = 0.056. As the value for the mound
index increased, the percent defoliation in the overstory decreased. At sites
where there were no ant mounds present the percent defoliation of jack pines
ranged between 18 and 50% and averaged 33% of the current-year's growth.
When the mound index was greater than 7. the percent defoliation ranged
from 15 to 35% and averaged 27% of the current-year's growth.

More variation was explained (r2 = 0.08) when the total number of tips
defoliated by JPBW was regressed against mound index (Fig. 9). A trend was
found similar to that for mound index vs. the number of pupae, suggesting that
as mounds increased in number and size, predation on larvae increased
resulting in reduced defoliation. Relative to the number of branches in the
overstory that were at least partially defoliated due to JPBW feeding, few pupae
were actually found. The regression was significant at (p<0.0001). When there
were no ant mounds present at a site, the number of tips damaged ranged from
107 to 330, and averaged 209.58 ;|-_ 12.80 per 400 tips sampled. When the mound
index was greater than 7, the number of tips damaged ranged from 57 to 247,

and averaged 154 :1: 12.52 per 400 tips sampled.

7O

.2 u e :85 v e .eeeecawe been
mm commmocwou 2F .Amvd u «a .amdu + xvnd- u .0 (”.8596 05 5 «out no
398 3?: .3 .8982. 2: e5 .82: 2508 c0222. 3:25:23..— .5 03mm..—

 

 

 

xwoz_OZDOs_
...-N...... m. m 4 w o N.
t . . . QT.
1‘ o
O .
O N
O O . .ow .0
O n M
O O O O 5N M
o . a
e .3 m
.ov mu.
. um
.om

 

71

.2. u e :86 v 9 350553 >23: mm commmocmoc
2F .Gmd u «a 609 + mend- n .O 5035.6 05 E «8.: no 082. 39:.
.E confine 05 e5 3588 Co .362. 05 5953 fiancee—om .w 2:3“—

 

 

 

 

 

(AHOlSHEA013Vdnd 'ON

 

72

«l.

.2. u c "Sod v a .685:me m commmewoc 2:.
.806 u «u .Eda— + awed- u b BEES an: 09. Ba cog—£3 a: 555
3E xufi. .8 39:5: 05 e5 Seem 2:88 5253 3:30:2an .o oSmE

N
1-
o
a-
"m
(D
'V’
"N
O
N
r

 

 

 

00v I oaivnoaao Sdll 'ON

 

73

Mound index was correlated with the number of pupae in the
understory (r2 = 0.21), but did not correlate as strongly with the percent
defoliation in the understory due to JPBW feeding (r2 = 0.06). However. there
was considerable variability in defoliation of the understory, with a weak
trend toward a decrease in defoliation as the value for the mound index
increased (Fig. 10). Despite the large number of tips that were actually
defoliated by JPBW. few pupae were actually found.

The basal area of the trees at the sites sampled in the jack pine
plantations ranged from 0 to 110. The basal area of the stand did not appear to
influence ant predation on JPBW (r2 = 0.002). When pupal densities were
regressed on basal area for only those sites where there were no ant mounds,
the results were not significant at the (p< 0.05) even though a stronger linear
trend was evident (:2 = 0.056).

Tree height had no signifith effect (p<0.05) on the density of JPBW
pupae. These results suggest that F. exsectoides forages on trees of varying
heights and height does not interfere with predation on JPBW larvae and
pupae.

In the field experiments where we used planted JPBW, the presence of
aphids increased ant predation. There was not a significant relationship
between the regression of pupal densities vs. aphid densities in the field
survey (r2 = 0.01). However, because aphid colonies did not occur on many of
the trees included in the survey. it is not certain whether the aphids actually

influenced predation on JPBW pupae from this experiment.

74

.3 u a “med v a .anamcwfi fl :23“:on 2F .de u «a
6:.— + and. u .0 b90095 05 E «yo—9:3. an: on BA 38:80..
an: .«o 595:: 2: can .82: 2508 c332. qucoufiom .o— aswi

 

 

 

 

 

”on
new
.3.
.3.

‘c;
N
(AHOLSHBONO) omvnoaao SdLL 'ON

 

75

Echrimcnt 3. W

F. exsectoides readily preyed upon adult female gypsy moths. Within 1 hour,
the ants removed nearly all of the moths from the trees where aphids were
present. The ants significantly (p<0.05) reduced the number of adults on trees
both in the presence of aphids and on trees that contained only adult moths.
The adult female moths remained on the control trees during this time and
continued to deposit egg masses.

The ants forced gypsy moth adult females to drop from trees by
repeatedly biting them. Moths dropping to the ground, were immediately
seized by the ants and torn into pieces. Occasionally. the ants would drag an
intact moth back to the nest.

The ants did not display much interest in the egg masses of the gypsy
moth. Six hours after the ants were allowed to ascend the trees, the egg masses
remained intact despite moderate to high numbers of ants on the trees. Even
after 24 h. the ants continued to ignore the gypsy moth egg masses. During
the observations. only a total of five ants were seen investigating egg masses
and individual ants did not spend more than a few seconds on each egg mass.

The ants significantly (p<0.0002) reduced the population of fourth instar
GM on trial trees after 24 b. However, there were no significant differences
between the number of "intact” caterpillars and the number of caterpillars
that had their setae mechanically reduced remaining on trial trees. The
average number of caterpillars recorded on trees that were ”intact” was 3.17 i
0.91 while the number of caterpillars remaining on trees that had their setae
reduced was 2.42 i 2.43. The control trees where ants were not permitted to

forage contained an average of 7.0 :L 0.52 fourth instar GM.

76

In this experiment, the caterpillars responded to an attack from an ant
by moving to a different feeding place on the tree or dropping to the ground. .
F. exsectoides killed more gypsy moth caterpillars after they fell to the ground
than when they were on the trees. Thirty nine interactions were recorded
between ants and gypsy moth caterpillars. In 19 interactions, the
caterpillars drOpped from the trees in response to the ants. Twelve of these 19
caterpillars were captured by other ants that were foraging on the ground
near the base of the trees.

As the distance from the mound increased. the percentage of pupae
remaining on the trees also increased (r2 = 0.441) (Fig. 11). This can be
explained in part by the high concentration of ants occurring near the ant
mounds. The ants removed 60 out of 112 pupae from the trial trees within 24 h
after placing the pupae on the trees. Very few pupae were removed and few
ants were found on trees that were more than 10 m from the mounds.

Gypsy moth pupae are large relative to F. exsectoides. Therefore, instead
of removing an entire pupa. the ants would pierce a hole in the end of the
pupal case and individual workers would remove pupal fragments (Fig. 12).
After the initial discovery of a pupa. the ants would continue to forage on the
pupa for up to 2.5 h. Only on two occasions did we see the ants remove an

entire pupa from a tree.

Echrimcnt 4. W

The ants significantly reduced the number of adult WPW on trial trees 72 h
after placing the WPW on the trees. The average number of adult WPW
remaining on trees with aphids was 2.00 + 0.91 while the average remaining

on trees with WPW only was 6.25 _-t_-_ 1.93 (Table 7). Control trees retained on the

_, \"Z

77

% GYPSY MOTHS REMOVED
8

 

 

 

 

 

1

0 1 0 20
DISTANCE FROM MOUND (m)

Figure 11. Regression results of the effect of the distance from the mound on
the percent of adult female gypsy moths removed by ants from jack
pine trees. The regression was significant. P<0.0001, (y = -1.60x + 83.32.

r2 = 0.441).

78 -

Figure 12. Workers of F. exsectoides entering a gypsy moth pupal case and
removing fragments of a pupa.

79

 

80

Table 7. Ant predation on adult WPW in the presence and absence of aphids.
Mean number WPW remaining on experimental trees after 72 h.

 

 

ANOVA
Group: Count: Mean: Std. Error: F:
With 4 2.00 0.91 22.16
aphids
Without 4 6.25 1.93 p < 0.003
aphids

Control 4 13.75 0.50

 

81

average 13.75 :1: 0.58 adult WPW. Fisher's PLSD showed significant (p<0.05)
differences between treatments after 72 h.

A total of 74 interactions were observed and recorded between the ants
and WPW on the trial trees. An additional 22 observations were recorded on
the feeding behavior of WPW adults on control trees (Table 8). Forty
interactions were recorded between weevils and ants on trial trees with aphids
present (Fig. 13). Out of the 40, there were 16 interactions in which the weevil
stopped feeding because of persistent harassment from an ant(s). In six
interactions. the weevil stopped feeding because of the ants and moved to a
new location on the tree. In six other instances. we found that the weevil
stopped feeding for a brief period following harassment from the ants but
shortly afterwards relocated. In two out of the 16 persistent interactions. the
ants captured the weevils and removed them from the experimental trees.

Frequently, even after an ant made antennal or mandibular contact
with a weevil. the weevil did not relocate or discontinue feeding. We observed
21 non-persistent interactions in which the ants did not respond aggressively
to weevils even after walking on or nearby the weevils. In five of these 21
non-persistent interactions. the weevils stopped feeding and moved to new
locations on the tree.

On trees that did not contain aphids (weevils and ants only), we
observed 34 interactions between ants and weevils (Fig. 14). Eighteen of 34
interactions were persistent in which the ant harassed the weevil until the
weevil responded by discontinuing feeding or by moving. In five of the
persistent interactions, the weevils stopped feeding and moved to new

locations on the tree following repeated attacks by the ants. In 12 persistent

82

5:3 .8 mega.“ 9 32% 95:8

EB.“ :25» 35 a 3388 28 cotomno 83 H

 

 

3:2. 9 25
magnum 38m 395

 

2 -- --
o 3 on
m 2 2
:30 :0 3%» 88.5.0
mam—88m 39m 3E5 «883 E< 2:56.32

88.80
358:2

A35 25

«a 32.8
S. 82?.
5:5

82?.

E .855

.8825
we a 25882.“.

 

228825 2: mam—:32 meow—on 52:53.8 2:58-5hfi .3 no mean—E3

.mocfi 8.0% cc 2:603 2:: SE? ecu 8383.38 .k 503.2.

.w 032.

 

83

With Aphids Present

4 0
(Persistent) 16 /2 1 (non -persistent)
(weevil stopped (weevil stopped (ant removed (weevil stopped (weevil continues
feeding) feeding and weevil from feedinz) feeding)
relocated) tree)

Figure 13. Persistent and non-persistent interactions between F. cxscctoides
and white pine weevils on trees that contained aphids.

84

Ant and Weevils Only

34
12 5 3 10
(weevil flopped (weevil stopped (weevil stopped (weevil continues
feeding) feeding and feeding) feeding)

relocated)

Figure 14. Persistent and non-persistent interactions between F. exsectoides
and white pine weevils on trees that did not contain aphids.

85

interactions. the weevils did n0t relocate and resumed feeding shortly after
the ant left the area.

Thirteen of the 34 interactions that were observed between the ants and
weevils were non-persistent. These interactions were classified as non
persistent because the ant walked on or near a weevil without responding to it.
However. even though the ant did not respond aggressively, in three of these
interactions. the weevils stopped feeding while the ants were in the immediate
area. Ten of these interactions did not disrupt the weevil sufficiently enough

to cause them to stop feeding or move to a new location on the tree.

86

Discussion
This study shows that F. exsectoides will prey on a wide variety of insect
pests. This species of ant also can reduce the population density of adult WPW
by disrupting their feeding and forcing them to move to new feedings

locations (Summarized in Table 9) .

Ant Predation on the Larvae of RHPS, JPBW and GM

Little is known about predatory activity of F. exsectoides even though it
is considered to be the most common mound building ant of North America
(Wheeler. 1910). However, it is a close relative to the Formica rufa group
where ant predation has been studied more extensively in both Europe and the
United States for species in this group. In a study examining the relationship
between F. exsectoides and membracids in the Eastern United States, Andrews
(1929) observed that the ants preyed on beetles, flies, bugs and especially
LepidOptera larvae. Workers of F. exsectoides were also noted preying on
beetles. sowbugs, spiders, crickets, grasshoppers. termites. and unidentified
Lepidoptera and Diptera from April to September in Maryland by (Haviland,
1947).

In this study, F. exsectoides readily removed fourth-instar RHPS from
trial trees. Despite the two obvious defense mechanisms of RHPS, the colonial
feeding habit and the resin-rich substance that is stored in the sawflies
diventricular pouch and discharged when they are disturbed by predators
(Eisner et al., 1974), F. cxsectoides removed 240 sawflies at one mound within 36

h. This represented only a small fraction of the total number of workers

87

doze—anon Son 885 2: v8.62 Eucaucmcmmm £5 2: :2: 236.: «antenna 2:.

 

 

3% 3m?

e? e? Z a? 20

hr hr 39:.
.9 2mm ,

22::— outs mam :35. 82::

 

con—D uoaem $32888 .k .2: Eon—@2269 we woman was 382: 2.... .o 039—.

88

foraging in the forest during this time. Several European studies have shown
that ants in the Formica rufa group significantly reduced populations of
sawfly larvae (Wellenstein 1957a; Goesswald, 1940; Bruns and Schrader. 1955).

In the experiment comparing predation on the three stages of RHPS
larvae, the ants initially removed fifth-instar sawflies more readily than
either first or third instars. After 24 h. fewer first and third instar RHPS
remained on the trial trees than fifth instars. Traniello (1987) found that
when workers of F. schaufussi were given a choice between 50 and 400 mg
prey, the ants selected the smaller prey. He suggested that smaller prey items
have a' greater net profitability because they do not require group retrieval
and that there is less chance that smaller prey will be lost through
interference from other predators. especially other ant species. A similar
phenomenon may be true for F. exsectoides because individual workers were
seen removing and carrying both first- and third-instar RHPS, while the
removal of fifth instar RHPS usually required. more than one and as many as
seven workers of F. cxsectoldes. F. exsectoides may have discovered the larger
fifth-instar lavae more quickly because of their relative size, but then
removed the smaller instar RHPS more quickly because of the reduced
handling time.

Bradley and Hinks (1968) observed workers of F. obscuripes carrying
JPBW larvae back to their nests. Allen et al. (1970) observed workers of F.
exscctoides attacking larvae of JPBW that fell from trees to the ground.
however. they concluded that predation by F. exsectoides did not occur on
JPBW larvae that were protected in their silk feeding chambers.

This study shows that F. exscctoides preys on JPBW larvae whether

inside or outside their silk chambers. But those on the outside dropped from

89

branches in response to ants more than twice as often. After a larva dropped
from a branch the ants would either wait for the larva to climb back onto the
branch via its silk thread before attacking or continue to pursue the larva
down its silk thread (Fig. 15). The ants also responded aggressively to the silk
chambers of the JPBW and repeated attacks would often result in the larvae
backing out of its silk chamber and dropping on a silk strand. Often, ants were
present at both ends of the silk chamber and the larva would be captured as it
attempted to escape. These results suggest that the first line of defense for
JPBW larvae is to remain in their protective silk chambers and drop from
branches only when the ants continue to pursue them.

A few studies have been conducted where predation by ants, Formica
spp., and ground beetles. Calosoma spp., have been recorded attacking GM
larvae. In a recent study, Weseloh (1988) found that ants readily preyed on
tethered GM larvae that occurred in the soil litter while small mammals
consumed the majority of larvae that were on the holes of trees. Tethered
larvae in the tree canopy were not preyed on as readily by invertebrates. The
results of this study support Weseloh's findings in that more GM caterpillars
were killed by ants that fell to the ground than those that remained on the
trees.

My original hypothesis that the setae of fourth-instar GM is an effective
deterrent against ant predation was not supported by the findings of this
study. The direct measurements of predation and the observations suggest
that it was not the setae that deterred ant predation but the large size of
fourth-instar GM and their associated mobility. Sixty-three percent of the
caterpillars that dropped to the ground were attacked by ants. The other

caterpillars appeared to escape in the dense underbrush of the jack pine

90

Figure 15. Workers of F. cxsectoides searching for a dislodged jack pine
budworm larva.

 

92

forest. In the study area, ants were found covering much of the ground and
foliage and the possibility that the caterpillars were captured later seems quite

high.

Ant Predation on JPBW and GM Pupae

Several studies have suggested that predaceous ants may play an
important role in the population dynamics of the western spruce budworm
(Campbell and Torgerson, 1982; Youngs and Campbell, 1984). These studies
examined ant predation on the pupae of the western spruce budworm.

F. exsecrot’des preyed on JPBW pupae that were placed on trees
surrounding ant mounds in experimental arenas (Fig. 16). The field survey of
JPBW pupal densities and ant mound densities indicated that moderate to high
numbers of mounds of F. exsectoides have a significant negative effect on
JPBW pupal densities.

The field survey also provided a measure of the effects of ant predation
on defoliation due to JPBW feeding. At sites where the ants foraged. defoliation
decreased 3 to 30 times compared to sites where the ants were not present. In a
similar study, Warrington and Whittaker (1985) found that sycamore and ash
trees suffered approximately 5 times as much defoliation in areas where ants
(Formica rufa group) did not forage as compared to areas where ants foraged.

Although the presence of ants in the jack pines stands resulted in a
significant decrease in the number of pupae found and defoliation of jack
pines, there were still more tips defoliated than pupae actually found. One
interpretation of this is that the ants removed budworm pupae more easily
than budworm larvae, possibly due to the halt in silk production during the

pupal stage of development. A second explanation could be that there was a

93

Figure 16. Workers of F. exsectoides removing a jack pine budworm pupa from
its silk chamber.

94

 

95

source of alternative prey which was preferred by the ants available in the
field during the time JPBW larvae were present. Perhaps the relationship
between the number of pupae found and the amount of defoliation is simply a
reflection of ant predation extending over the larval and pupal stages of
development of the JPBW.

The ants readily preyed upon the pupae of the GM. Because the distance
between the trial trees with pupae and the mound increased, the percentage of
pupae remaining on the trees also increased suggesting that the pupae were
outside the normal foraging range of the ants. Few ants were found on trees
that were more than 10 m from the mounds, and thus few pupae were removed.
Weseloh (1985) found that Calosoma larvae destroyed 70 percent of tagged
gypsy moth pupae under burlap bands on trees near the ground. In this study,
F. exsectoidcs destroyed 56 percent of the GM pupae within 24 h after placing
them on trees. Smith and Lautenschlager (1981) also found that ants were
responsible for removing a significant proportion of both healthy and

parasitized gypsy moth pupae.

Ant Predation on Adult GM and WPW

The ants preyed on both adult GM and WPW, however, the GM were more
readily removed from the trees than adult WPW. The ants had little difficulty
preying on GM adults perhaps because the adult females are not capable of
flight and can only escape predation by moving up the host tree. The ants
caused the adults to drop from trial trees by repeatedly biting them.

F. exsectoides directly reduced the density of WPW on trial trees shortly
after the experiments began. The ants also negatively influenced the weevils

indirectly by causing them to stop feeding and move to new feeding locations

96 -

on the trees. This study was conducted in mid August when adult WPW spend
most of their time feeding on pine shoots before overwintering (USDA, 1983).
Ant predation may also disrupt WPW oviposition in the spring by directly
reducing egg laying and interfering with adult males that are trying to locate
females on the trees.

The ants removed the RHPS larvae more quickly than JPBW and GM
larvae. After 24 h, the ants removed an average of 67 percent of the RHPS
larvae as compared to an average of only 8.6 percent of JPBW larvae.

Although RHPS larvae do have two mechanisms to deter predation, colonial
feeding and the noxious fluid they dispel when approached by predators, they
are relatively naked as compared to JPBW larvae that are protected inside silk
chambers. F. cxsectoidcs did reduce experimental populations of fourth-instar
GM, but the results from this study showed that the setae were not a sufficient
deterrent against ant predation. It appeared that it was the large size of the
caterpillars and their mobility that allowed them to escape predation by

F. exsectoides. Therefore, it seems likely that these ants would not be able to
capture GM caterpillars that were beyond the fourth instar stage. This study
may have overestimated predation on GM larvae because many of the
caterpillars dropped from the trees in response to ants presence. However,
disruption of the feeding of GM larvae should result in decreased defoliation of
trees. The fate of individual caterpillars was not followed, but, instead
predation was measured by the benefit the tree received by counting the
number of caterpillars removed.

The only developmental stage that F. cxsectoidcs showed no interest in
was the egg stage of the gypsy moth. This may have occurred for two reasons:

1) the egg stage is sessile and movement may be necessary for initial

97

recognition of a prey item by ants (Mathews and Mathews, 1978), and 2) the
egg mass of the GM is protected by a thick layer of abdominal hair-like scales
which the adult females deposits along with the eggs. The pupae of both the
JPBW and the GM were readily preyed upon by the ants even though they are
relatively motionless. This may suggest that GM egg masses were not
acceptable prey because of the protective covering on the egg masses rather
than the sessile condition. However, shortly after the JPBW and GM pupate and
shortly before they emerge as adults slight movement does occur and may
serve as a sufficient stimulus to elicit a predation response.

Relative abundance of each prey item in the field may also influence
predation (Way, 1963). Colonies of F. exsectoides are active in the field from
April to September (Haviland, 1947). The insect pests examined in this study
were present at different times of the year, but many of the developmental
stages of the insects overlap (Table 10). From late April to May, GM are in the
early stage of development in Michigan (USDA, 1983), and were readily preyed
upon by F. cxsectoidcs. In the jack pine forest where this study was conducted,
RHPS did not occur but a close relative, the European pine sawfly, Neodiprion
scrtifcr (Geoffroy), were present as larvae during April and May. The
European pine sawfly may offer a potential abundant food resource to the ants
during this time and may be a more acceptable prey item to the ants then GM
larvae if the ants prey on this sawfly as readily as they preyed on the RHPS
larvae.

In May, the JPBW larvae hatch from egg masses in Michigan and may
provide a relatively abundant source of food for the ants. F. cxsectoides will
prey- on the JPBW through the pupal stage which usually lasts until August. In

August WPW adults become available in the environment and may replace the

 

98

.268 2a 308.: 82: :2: can» 2: we 38: 2: 830: 3:: 2:.

 

........... 2.72

 

3&3.

 

20

 

3mm».

 

mama

 

.8380 .Eom 3:332 .23. 2:; .32 :2: moi
“com:—

832888 .K :5 .35 83.5 .8 5:32.:me Enemaom .9 .038.

99

JPBW as prey for the ants. At this time brood production in the ant mounds is
beginning to decrease (Haviland, 1947) and the need for protein foods for the
colony may decline as well (Edwards, 1951).

In conclusion, the JPBW would appear to offer the most stable food
resource to F. cxsectoides in the jack pine forest where this study was
conducted because the ants will prey on small and large JPBW larvae and
pupae. The JPBW also appeared to the most abundant insect pest in these jack
pine forests. This may, in part, explain why the ants can prey on JPBW larvae

and pupae despite their protective silk chamber.

The Influence of Aphids on Ant Predation

One of the most interesting findings of the study was that ants switched
from tending aphids to attacking the insect pests when they wandered near a
tended aphid colony. Green and Sullivan (1950) suggested that ant predation is
often a result of defense of aphid colonies. They observed ants removing 1,800
forest tent caterpillar larvae, Malacosoma dissiria an. When young larvae
moved on trees to feed or moult, they often passed through an aphid colony
which immediately invoked an attack by the ants. During our observations we
found that when an ant encountered a RHPS it would do one of the following:
1) antennate the larva and return to tending aphids, 2) attack and sometimes
remove the larva from the tree, or 3) recruit other ants from the surrounding
area by leaving the aphid it was tending and running up and down the tree
trunk which appeared to excite other ants in the area. Many of the
encounters between ants tending aphids and RHPS resulted in the ant

attacking the larva or removing it from the tree.

100

Bradley and Hinks (1968) observed that ants protecting aphid colonies
would chase away mirids. The mirids would escape further predation by
running back an inch or two and remaining motionless. This is similar to our
study in that an attack on a weevil by an ant that was tending an aphid colony
did not always result in a capture, but would often force the weevil to find a
new feeding location on trees.

In an omnivorous species such as F. exscctoides, individual workers tend
homopterans to collect honeydew as well as forage for insects. However,
Traniello (1989) found that workers of Formica schaufussi forage more
persistently following a reward of sucrose than following a reward of insects.
F. exsectoides preyed on WPW in both the presence of aphids and absence ‘of
aphids on trial trees. Although we recorded more interactions between the
ants and the WPW on trees when aphids were present, the ants were more
persistent in foraging on weevils when aphids were absent. On trees
containing aphids, 52 percent of the interactions between the ants and the
weevils were non-persistent. On trees that did not contain aphids, 38 percent
of the interactions were non-persistent. Persistent foraging on a more
constant food resource such as honeydew produced by the aphids may in part
explain why, at times, the ants simply chase away the predators of the aphids.
However, when individual workers switch from tending aphids to preying on
insects, they may be just simply taking advantage of the prey item which is
more abundant in the environment (Way, 1963).

The effects of the interactions between ants and aphids in relation to
biological control can be either negative or positive depending upon a
number of factors. Ant attendance of certain aphid species interferes with the

control. exerted by the natural enemies (Flanders, 1943; Samways et al., 1982).

101

Ant attendance of certain species of homoptera is known to increase the
population of homoptera to economically injurious levels (Flanders, 1951;
DeBach et al., 1951) and may reduce leaf size and stem wood production of
certain types of trees (Dixon, 1971).

In this study, ant predation was positively influenced for each of the
insect pests examined by the presence of aphids on the jack pines. Ant
densities were always higher and more constant on trees that contained aphids
as compared to trees that did not contain aphids. Whether the ants were
searching for stray aphids or insect pests is not clear, but the results were
always the same, fewer insect pests remained on trees that contained aphids.
The effects of aphid feeding on the growth of jack pines was not examined,
but, the trees did not appear to suffer from aphid feeding. The shoots of the
jack pines that did contain aphid colonies were not discolored or brittle.
However, before a conclusion can be made on the benefits to the trees derived
from ant attendance of aphids in this system, the effects of aphid feeding on
jack pine growth needs to be examined.

This study shows that F. exsectoidcs is a good candidate for biological
control of certain forest insect pests. Finnegan (1974) suggests that ants
combine certain qualities that make them ideal biological control agents.
These characteristics include, that ants: 1) attain a high p0pulation density;
2) forage at all levels in the forest, from in the soil to the tree canopy: 3) have
an extended period of activity: 4) are polygynous, assuring a prolonged nest
life; 5) form colonial nests; 6) forage for and attack all stages in the life cycle
of insect prey; and 7) specialize in foraging for those prey that explosively
increase in total numbers of individuals. F. cxsectoides can form as many as

400 mounds per acre (Allen et al., 1970). F. exsectoides preys on a wide variety

102

of forest insect pests and on several developmental stages. Predation was not
examined on all of the developmental stages of each insect, but the results of
ant predation on the JPBW and GM suggest that their is great potential for this
ant to prey on many developmental stages for similar types of insects. We also
observed F. exscctoidcs bringing arthropods that dwell in the soil back to their
nests. 7F. exsecroides forages over most of the spring and summer, (Haviland,
1947) when many important insect pests are present in jack pine forests. This
species of ant is polygynous and queens from one colony were readily
accepted when introduced to new colonies. F. exsectoides forms colonial nests
that are connected by a series of underground tunnels. This enables the ants
to share food and brood and thereby increases the stability and permanence in
the ant. colony over large areas and over an extended period of time. Although
F. cxsectoides preys on a wide variety of insect prey, JPBW larvae and pupae
represented a large portion of the prey brought back to the nests. JPBW
densities were at very low levels in the jack pine forest where this study was
conducted during 1987 and 1988. F. exsectoides may prey exclusively on JPBW
when it reaches outbreak levels. This would satisfy the last criteria Finnegan
(1974) includes in his list of characteristics for biological control agents. The
results of this study and information from the existing literature suggest that
F. exsectoides would be an excellent biological control agent to incorporate
into an integrated pest management programs in forest ecosystems.

The ability of colonies of F. exsectoides to prey on more than one type of
insect and switch from preying on insect pests to collecting honeydew from
aphids may aid in maintaining high and constant populations of this ant in
forest ecosystems. Large and constant populations of predaceous ants may

contribute to maintaining insect pest populations at levels below the economic

 

103

thresholds. In addition to directly reducing insect pest populations, the
disruption of feeding and oviposition of insect pests by ants should be
considered by land managers. The ability of predaceous ants to disrupt the
feeding of insect pests and therefore, reduce damage to trees would be
especially beneficial for certain types of crops such as Christmas trees that are

valued for their esthetic appearance as well as growth.

104

Literature Cited

Adlung, K. G. 1966. A Critical Evaluation of the European research on use of
red wood ants (Formica rufa Group) for the protection of forests against
harmful insects. Z. ang. But 57: 167-201.

Allen, D. C., F. B. Knight and J. L. Foltz. 1970. Invertebrate predators of the
jack-pine budworm, Charistoneura pinus, in Michigan. Ann. Entomol.
Soc. Amer. 63: 59-64.

Andrews, E. A. 1925. Growth of ant mounds. Psyche. 32: 75-87.

Andrews, E. A. 1926. Sequential distribution of Formica exsectoidcs Forel.
Psyche. 33: 126-150.

Andrews. E. A. 1929. The mound building ant, Formica exscctoidcs Forel,
associated with tree happers. Ann. Entomol. Soc. Amer. 22: 370-391.

Bradley, G. A., and J. D. Hinks. 1968. Ants, aphids and jack pine in Manitoba.
Can. Entomol. 100: 40-50.

Bruns, H. and A. Schrader. 1955. Diminution of the quantity of N. sertifer
cocoons near nests of the red wood ant. Waldhyg. 1:33-68.

Campbell, R. W., and T. R. Torgersen. 1982. Some effects of predaceous ants
on western spruce budworm pupae in north central Washington.
Environ. Entomol. 11: 111-114.

DeBach, P.. C. A. Fleschner, and E. J. Dietrick. 1951. A biological check
method for evaluating the effectiveness of entomOphagous insects.
J. Econ. Entomol. 44: 763-766.

Dixon, A. F. G. 1971. The role of aphids in wood formation. J. appl. Ecol. 8:
165-179.

Edwards, W. H. 1936. Change in the foraging behavior of the garden ant
Lasius niger L. Entomologist's Monthly Mag. (London) 87: 280.

Eisner, T., I. S. Iohnessee and J. Carrel. 1974. Defensive use by an insect of
plant resin. Science 184: 996-999.

Eskafi, F. M. and M. M. Kolbe. Predation on larval and pupal Certitis capitata
(Diptera: Tephritidae) by the ant Solenopsis gcminara (HymenOptera:
Formicidae) and other predators in Guatemala. Environ. Entomol.

19: 148-153.

Finnegan, R. I. 1971. An appraisal of indigenous ants as limiting agents of
forest pests in Quebec. Can. Entomol. 103: 1489-1493.

105

Finnegan, R. J. 1974. Ants as predators of forest pests. Entomophaga Mem.
Hors. Ser. 7: 53-59.

F'mnegan, R. J. 1975. Introduction of a predaceous red wood ant Formica
lugubris (Hymenoptera: Formicidae) from Italy to eastern Canada. Can.
Entomol. 107: 1271-1274.

Finnegan, R. J. 1978. Predation by Formica lugubris (Hymenoptera:
Formicidae) on Choristoneura fumiferana (LepidOptera: Tortricidae)
Can. For. Serv., Bi-Mon. Res. Note 34: 3-4.

Flanders. S. E. 1943. The Argentine ant versus the parasites of the black scale.
Calif. Citrograph 28: 117-128.

Flanders, S. E. 1951. The role of the ant in the biological control of
hymenopterous insects. Can. But. 83: 93-98.

Goesswald, K. 1940a. Observations on the protection of pine plantations
against Diprion pini. A. f Forst-u. Jagdwesen 72: 370-371.

Green, G. W., and C. R. Sullivan. 1950. Ants attacking larvae of the forest tent
caterpillar, Malacosoma disstria I-Ibn. (Lepidoptera: Lasiocampidae). Can.
Entomol. 82: 194-195.

Haney, P. B., R. F. Luck and D. S. Moreno. 1987. Increases in densities of the
citrus red mite Panonychus cirri (Acarina: Tetranychidae), in
association with the argentine ant, Ridomyrmcx humilis (Hymenoptera:
Formicidae) in southern California citrus. Entomophaga 32: 49-57.

Haviland, E. E. 1938. Mound changes after ten years in colonies of
Formica exsectoides Forel. Ann. Ent. Soc. Amer. 31: 50-57.

Haviland, E. E. 1947. Biology and control of the Allegheny mound ant.
J. Econ. Eat. 40: 413-419.

Mathews, R. W. and J. R. Mathews. 1978. W John Wiley and
Sons. 507 pp.

McNeil, J. N., J. Delisle and R. J. Finnegan. 1978. Seasonal predatory
activity of the introduced red wood and, Formica lugubris
(HymenOptera: Formicidae), at Valcartier, Quebec, in 1976.
Can. Entomol. 110: 85-90.

Montgomery, B. A. 1988. Michigan Forest Pest Report, 1987. Mich. Cooperative
For. Pest Management Program. 81 pp.

Peirson, H. B. 1922. Mound building ants in forest plantations.
J. of Forestry. 20: 325-336.

Samways, M. J., N. Magda and A. J. Prins. 1982. Ants (Hymenoptera:
Formicidae) foraging in citrus trees and attending honeydew-
producing homoptera. Phytophylaciica 14: 155-157.

106

Smith, H. R. and R. A. Lautenschlager. 1978. Predators of the gypsy moth.
USDA Agric. Handb. 534. 72 pp.

Stary, P. 1966. Aphid parasites (HYM., Aphidiidae) and their relationship to
aphid attending ants, with respect to biological control. Insectes
Sociaux 13: 185-202.

Tedders. W. L., C. C. Reilly, B. W. Wood, R. K. Morrison and C. S. Lorgren. 1990.
Behavior of Solenopsis invicta (Hymenoptera: Formicidae) in pecan
orchards. Environ. Entomol. 19: 4-4-53.

Torgersen, T. R. and R. R. Mason. 1987. Predation on egg masses of the
Douglas-fir tussock moth (Lepidoptera: Lymantriidac). Environ.
Entomol. 16: 90-93.

Traniello, J. F. A. 1987. Social and individual responses to environmental
factors in ants. Experientia Supplementum. 54: 63-80.

Traniello, J. F. A. 1989. Foraging strategies of ants. Ann. Rev. Entomol.
34: 191-210.

U.S. Department of Agriculture. Christmas Tree Pest Manual. 1983.North-
central Forest Exp. Sta., St. Paul, MN. 108 pp.

Warrington, S. and J. B. Whittaker. 1985. An experimental field study of
different levels of insect herbivory induced by Formica rufa predation
on sycamore (Acer pseudoplatanus) I. lepidoptera larvae. J. Appl. Ecol.
22: 775-785.

Way, M. J. 1963. Mutualism between ants and honeydew-producing
homoptera. Ann. Rev. But. 8: 307-344.

Wellenstein, G. 1952. On the nutrition and biology of red wood ants. Z.
Pflanzenkr. 59: 430-451.

Wellenstein, G. 1957. The influence of ants of the Formica rufa group on
arthropods of forests. A. ang. But. 4: 368-385.

Weseloh, R. M. 1985. Predation by Calosoma sycophanta (Coleoptera:
Carabidae): evidence for a large impact on gypsy moth. Lymantria
dispar (Lepidoptera: Lymantriidae). pupae. Can. Entomol. 117: 1117-1126.

Weseloh, R. M. 1988. Effects of microhabitat, time of day, and weather on
predation of gypsy math larvae. Oecologia 77: 250-254.

Wheeler, w. M. 1910. MW

Columbia Univer. Press.

Youngs, L. C. and R. W. Campbell. 1984. Ants (Hymenoptera: Formicidae)
preying on pupae of the western spruce budworm (Lepidoptera:
Tortricidae) in eastern Oregon and western Montana. Can. Entomol.
116: 1665-1669.

MANUSCRIPT 2.

A Scots Pine Growth Simulation Model; A Focus on Ecological
Management Strategies

107

108 ‘

Introduction

The long term nature of Christmas tree production as well as the rising
costs and environmental effects associated with conventional practices,
demands that we explore alternative and improved Christmas tree production
technologies. Odum (1984) describes an ”improved" agroecosystem as one that
moves toward the reduction of one or more of the following factors: soil
erosion, loss of biological diversity, and chemical pollution in agricultural or
forest production. An example of such a system is a pine, birdsfoot trefoil and
predaceous ant intercropping. This ”improved" agroecosystem theoretically
provides an ecologically sound alternative to existing practices. A simulation
model was developed to test this hypothesis.
System Description. We have attempted to simulate an agroecosystem
which includes the following components: Scotch pines trees, Pi nus
sylvestris,, birdsfoot trefoil, Lotus corniculatus, and colonies of the Allegheny
mound ant, F. exsectoides Forel. The pines are grown on a short-term rotation
and harvested for Christmas trees. A relatively low variety of birdsfoot trefoil,
a leguminous ground cover, provides weed control and nitrogen through litter
decomposition and nitrogen fixation, and is intercropped with the Christmas
trees. Additionally, the trefoil provides both food and shelter to p0pulations of
aphids that are normally found on trefoil. The aphids in turn provide
honeydew to the ants as an alternative food when insect prey is scarce. When
the pests approach outbreak densities, the trefoil can be mowed forcing the
ants to switch their foraging from honeydew to the insect pests of pines. An
additional benefit of mowing the trefoil is a pulsed input of high nitrogen

content organic matter to the soil. This input of nitrogen has the potential of

109

increasing tree growth and leaf area by increasing soil nitrogen and cation
exchange capacity over a period of time.
Components of the Simulation Model. The model is written in Fortran
and is organized into eight subroutines: 1) a DRIVER to set the initial
conditions; 2) WEATHER to generate typical seasonal temperatures: 3) TREE
which selects leaf area for the trees at a given time, defoliation due to insect
pests, as well as competition between the trees and the trefoil; 4) TREFOIL
which provides an estimate of trefoil biomass; 5) NITROGEN which estimates
rates of nitrogen application and site improvement; 6) PEST which estimates
the number of prey available to the ants; 7) APHID which selects the amount of
honeydew available to the ants; and 8) ANTS which selects a predation rate that
determines the amount of insects removed from the trees. The overall system
flowchart is represented by Fig. 17.
Driver. The driver calls the following subroutines at weekly iterations: tree,
weather, trefoil, nitrogen, aphid, pest, and ant over an entire growing season.
The season consists of a 31 week period. The driver also keeps a yearly account
of the site improvement and the defoliation caused by the insect pests.
Temperature subroutine: The only weather parameter required for this
model was temperature. The design of the weather generator was further
simplified by taking advantage of the fact that temperatures for any
particular climate can be roughly fit to a ‘normal' distribution curve.

In the case of southeastern Michigan, the peak of this curve falls near
July 22 with the lowest temperature falling near the last week in January.
There is a flattening of the curve near the beginning of the year, thus the
peak is shifted slightly towards the end of the cycle.

The equation used to generate the data shown in Fig. 18. is:

 

 

 

 

 

 

 

 

 

110

SYSTEM FLOW CHART

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PESTS

 

 

 

 

 

ANTS

 

 

 

 

 

DRIVER
.1“
BIRDSFOOT TREFOlL APHIDS ‘H
NITROGEN
.1, .r—‘
PINES
OUTPUTS
END

 

 

 

Fig.

 

 

 

 

 

 

17 .System flow chart

 

 

111

64.3*exp[-(Day-203)2/ 19280] =14. [1]

It generates temperature, in degrees Fahrenheit that closely predicts
the peak and valley and agrees with monthly averages obtained from the
National Weather Service within 0.80- This curve is indicated by the solid
green line shown in Fig. 18. The dashed lines signify randomized hot and cold
years generated by the weather subroutine.

The subroutine allows the user to choose from four possible temperature
regimes: hot, cold, normal and random. The normal was taken from a 30 year
average from the First Order Noaa Weather Station. As depicted in the
flowchart (Fig 19.), hot is a season that varies from the normal temperatures
by -20 to +80, cold is a season that inverts hot, and random more closely
simulates normal weather patterns and varies from the normal by +50 .

Since all random number generators on computers are 'pseudo-random'
at best, the subroutine calculates daily temperatures and average weekly
temperature to lessen the problem. The mean weekly temperature is then
passed back to the main driver routine for distribution throughout the
program.

Tree Suhroutine: Scotch pine is capable of growing under almost any site
conditions. On an average site, Scotch pines can be harvested for Christmas
trees in 6 to 8 years. In attempting to model the growth of Scotch pine, we were
looking for what is referred to as a "level 4" model (Landsberg, 1981) which
would simulate growth as a function of radiant energy, water balance, and soil
nutritional status. However, an extensive review of the literature indicated
that previously developed models were lacking in two important criteria.
First, most existing models are static in the sense that they employ the standard

parameters of diameter at breast height (dbh), and site to predict biomass of

112

mm1._.<m3 z<o_Io=2

218${enmomaeervexeeneo u aEee
>00

06%»... 0.0

 

 

3533 Eeocom .. r
35003 200 1!
.8583 «a: 1....

.3583 .2502 ll

 

 

eJnroJadurel Kuog

 

113

 

mifialize Temp
L and Seed
i

 

 

l

 

Increment Week

 

 

Caicu§are first
andlcs?
Dg of The Week

 

 

 

 

J
f—_{ Increment Day
I
, Y

 

i

 

Calculate Temp
0? given Day

 

 

 

WEATHER SUEROUTINE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N J
ls Yr. Cold? Y A23 7533‘?
N g—— J
[ Add -5 to +5
Random Yr? L f0 Temp
\
l——N\n%;/eek>
/
iY
E Caiculare l
; average Temp 1
I J
1
@ Figure 19.

Weather Subroutine

114 —

wood volume- Moreover, these models are more suited for longer
rotations/timber management than for Christmas tree culture. Second, these
models only estimate above ground-biomass or wood volume and do not
measure what Christmas growers "sell" which is leaf area and form (Pregitzer
1988).

The basis for our new TREE subroutine (Fig. 20) stems from what is
referred to as a "pipe” model theory" (Waring et al., 1982). This theory
maintains that a unit weight of tree foliage is serviced by a specific cross-
sectional area of sapwood in the crown. Based on this relationship, Waring et
al. (1982) citing Whitehead, provides an equation to determine leaf area of
Scotch pine given the cross sectional area of sapwood directly below the

crown. The equation is:

Leaf area (m2) = 0.14 * Cross-sectional sapwood area (cmz) [2]

Data were then collected on sapwood area at ground level for different
age classes of Scotch pines at the Tannenbaum Christmas tree farm located
near the Michigan State University campus (Appendix 1). Total leaf area per
tree was then estimated using equation [2] and regression analysis was used to
predict leaf area as a function of cumulative degree days at base 452- The

resulting equation is:

Leaf arealtree(m2) = 0.000104 *Cumulative degree days - 0.042 [3]
Adjusted :2 = 0.838

It was also necessary to incorporate a site factor into this equation in

order to quantify for the contribution of the trefoil nitrogen to increased leaf

115

Figure 20- TREE SUBROUTINE

START_

V

 

 

Call TCDD
Call site
Call spacing

 

 

 

 

 

Calculate trees/A.
NU-(207.7)x (208.7)
BR WR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Calculate Tree DD Cm," Reduce Trefonl
diameter biomass by R
TDD-TCDD+3997 . R-CD. .NU
0.67 43,560
V/ |
11.993-15,99 1.45
15,991-19.99 2.14
Call
- 19389-2338 2.67
Nitrogen
Subroutine 23'986'27-93 2.89
27,985-31.982 3.27
31,983-35.980 3.65
35.981-39,987 3.78
-4 , 7 ,
Calculate leaf 39'9“ 3 9 6 403
area I tree; LAT J
LAT-site+N '.1'TDD R d t f .l
120 .00 ' on .9 ”ce '9 °'
L J . biomass by R
at" ' '
4.78 M
43.560

 

 

Calculate total
leaf area
TLA - LAT‘ NU

 

 

 

Call cumulative
seasonal Deg. D.
CSDD

 

 

 

 

116

Tree Subroutine Canti.

 

 

Calculate new leaf area

NLA-TLA-site+N'.1'TDD-CSDD
120 1000

-0.042

 

 

 

Calculate old leaf
area; OLA
OLA. TLA - NLA

 

 

 

Current leaf area
Total leaf area
CLA - TLA

l
l

 

 

 

OUTPUTS
Current leaf area
Reduced trefoil

 

 

 

117

area. Unfortunately. time constraints did not permit sampling trees at
representative sites. As a second best alternative, a site factor coefficient was
developed for the leaf area equation. This coefficient is based on potential
corn yield (corn equivalent yield) for various Michigan soil types. The soil
type where the sapwood data was collected is one of the best in Michigan with
a corn yield potential of 120 bushels per acre. The coefficient is based on the
assumption that maximum leaf area is obtained on the best soil types. This

coefficient is:

Original Site «I» Site Improvement Factor(from NITROGEN)/120 [3]

We assumed that any particular site could not be improved more than its
maximum potential. Thus, in the model, this factor is always 1.0 or smaller.
The third major component of the tree subroutine attempts to account
for reduced trefoil biomass as a result of competition for radiant energy
between the trees and the trefoil. In order to incorporate competition, reduced
trefoil planting density/biomass was considered to be a function of trees
crown diameter. Data on tree crown diameter for different age classes were
collected at the Tannenbaum farm and used to determine a trefoil reduction

factor based on the following equation:

Trefoil Reduction Factor == (Crown Diameter‘pi*No.Trees/Acre)l43560

[4]

The user of the simulation model is able to specify both between-row and
within-row spacing as well as the original site factor (in corn yield

equivalents) from the driver. Outputs of this subroutine are leaf area, adjusted

118

by the site improvement coefficient and the pests; and the trefoil reduction
factor.

Other assumptions in the development of this subroutine are:

1) there is a linear relationship between leaf area growth and time from 0 to
10 years. However. when actual data were plotted. same inflection is noted
between 1-2 years and between 9-10 years.

2) crown diameter remains constant during any one growing season.

3) trefoil would eliminate all weeds and that any competition between trefoil
and the trees is only for radiant energy.

4) total leaf area is directly related to site quality and that corn yield can be
used as an indicator of site quality for Scotch pine.

TREFOIL SUBROUTINE: Birdsfoot trefoil. a perennial forage legume, has
become the most frequently used plant in soil and water conservation projects.
In addition to being a competitive plant. it provides weed control through the
release of allelopathic chemicals. This in conjunction with its potential for
nitrogen input, made it a prime candidate for the intercrop.

We had originally intended to model the growth of trefoil as a function
of radiant energy, soil moisture and soil nutritional status. similar to the TREE
subroutine. However, as development of the other subroutines progressed we
found that only two points in the trefoil's development cycle were needed. The
first is when the plant reaches a dry weight of approximately 10 g/plant and
thus has sufficient leaf area and sap production to sustain aphid populations.
This point is estimated to be at approximately 200 degree-days at base 45. The
second point is when the trefoil reaches its maximum biomass potential (i.e.
the top of the growth curve). This point is estimated to be at approximately 800

degree-days.

119

The flowchart describing this subroutine is presented in Fig. 21. The
main input in this subroutine is planting density specified from the driver.
The driver limits planting density between 10 and 30 plants m2 and provides
the user with an indication of the appropriate planting density which is
dependent upon site quality. The system provides for harvesting the trefoil
when the pest densities reach destructive levels. In addition, this subroutine
allows for one mid-season management mow. If the ants adequately control
the pests, then the mowing command again becomes an option. The output of
this subroutine is biomass which is directed to the aphid and nitrogen
subroutines.

Actual trefoil above-ground biomass yield is based on an equation
develOped by McGraw et al. (1986), employing planting density as the driving

variable. The equation is:

Biomass Yield (31ml) = (654‘Planting Density)l3.54 + Planting Density)
[6]

This figure was then converted to biomass yield in pounds/acre.

Other major assumptions in this subroutine are:

1) that there may be a maximum of three mowings per growing season and at
least 800 degree days are required between mowings for trefoil to reach
maximum biomass levels.

2) that biomass levels remain constant after 800 degree days until the trefoil
is mowed.

3) that in the event that the trefoil is not mowed during the season (by the
user or by the aphids), standing biomass at the end of the season is mowed and

enters into the NITROGEN subroutine the following growing season.

120

Figure 21. TREFOIL SUBROUTINE

(SIMPLIFIED)

INPUTS FROM DRIVER
1) PLANTING DENSITY
2) SEASONAL DDAYS

 

 

 

 

IF
$002200

 

YES
CALL COMPETITION

 

 

 

 

 

PLANT DENSITY/ACR
-4046 x DENS xCOMP.

 

    

  

DDZBOOANT
PEST - MOW

 

  

 

(LBS BIOMASS / ACRE

22045 W\
.54 + DEM

4046]

J

I
(RESET SDD -Q
I /

BIOMASS DECOMPOSITION - \I f
REMAINING BIOM + NEw BIOM RETURN

f

 

 

k

 

 

 

 

121

NITROGEN SUBROUTINE: Nitrogen is a limiting factor in Scotch pine
culture. While many tree-soil systems are deficient in nitrogen, there have
been no attempts to model the cycling of the nitrogen in the tree-soil system.
Instead the approach has been to model the behavior of the system based on
expected performance using empirical and non-mechanistic equations.

The purpose of the nitrogen subroutine is to determine the nitrogen
application rate due to the decomposition of the mowed trefoil in pounds/acre
of organic material, and to calculate the subsequent site improvement due to
the addition of nitrogen in bushels/acre. The site improvement value is a
critical variable in the Scotch pine leaf area growth equation indicating the
effect of site on leaf area growth.

The only input variable is the biomass of mowed trefoil. The output
variable is the site improvement in terms of corn yield equivalent. Thus,
when no mowing occurs and the biomass is zero, there is no subsequent site
improvement. If the trefoil biomass is more than zero. mowing has occurred,
and the site improvement due to the decomposing trefoil is calculated.

A negative exponential equation determined the rate of decomposition

of organic matter as a pulsed input (Coleman et al. 1984).

X/XO = e‘l‘t XO= the amount of trefoil at time 0 [7]
X = the amount of trefoil at time t
k = fractional loss rate

t = time interval

Equation [7] was modified to calculate the amount of mowed trefoil

decomposed per week by solving for 'X' and subtracting this value from the

122

total amount of trefoil mowed. The time interval 't' was set to be one week. The

equation is:

Xdecomp = x0 - xo* e-kt [8]
where Xdecomp = the amount of decomposed trefoil biomass

in one week.

The value of the decomposition rate constant 'k' has been examined in
many ecosystems with low and high production rates and over a wide range of
environmental conditions. A decomposition rate constant of 1.5 for a
temperate grassland was used in the model. To obtain a rate per week. this
value was divided by the number of weeks in our growing season.

The next step was to determine the amount of nitrogen applied to the
system by the decomposed trefoil. Assuming an available nitrogen content of
3.6% of total trefoil biomass (Fribourg et al. 1956), the actual nitrogen

application rate in pounds/acre/week is:
N apprate = Xdecomp/27 [9]

The site improvement factor in bu com/acre is then calculated using a
logarithmic equation that determines yield response of corn to the application
of. nitrogen in the form of alfalfa taps (Fribourg et al.. 1956). It was assumed
that this response is similar for trefoil litter application. because trefoil is
closely related to alfalfa. Trefoil and alfalfa are both nitrogen fixing legumes

(Leguminacea).

Site Improvement = {105.2(1-10 - 0.01484(N+35.2)] - 73.6}/100

123

where N = nitrogen application rate. [10]

We modified the above equation in several ways. First. upon closer
examination of the equation is was evident that the response of corn yield to
nitrogen application in the form of alfalfa tops was unreasonably high. For
example, the equation predicts that an application of only 6 pounds/acre of
alfalfa nitrogen improved corn yields by 73 bu/acre to 80 bu/acre. To obtain a
more realistic value a long-term availability of nitrogen of 1% was assumed
for site improvements.

Finally. we assumed that all the nitrogen contained in the trefoil
biomass is available to the Scotch pine for uptake after decomposition.
Furthermore, nitrogen fixation and immobilization by microbes and leaching
of available nitrogen from the soil were not incorporated in this subroutine
because there was not enough time during the systems science course and it
would be a very complex subroutine to develop. The difference in availability
of nitrogen in sand. loam or clay was also not taken into consideration. The
flowchart for the NITROGEN subroutine is represented in Fig. 22.
REDHEADED PINE SAWFLY SUBROUTINE: The redheaded pine sawfly is an
important defoliator of young hard pines in eastern North America. The
sawfly occurs in colonies of a few to more than 100 larvae, and just one or a
few colonies can readily strip a small tree (l-Sm) of its foliage (Averill et al.
1982; MacAloney & Wilson 1964). Even lightly defoliated Christmas trees are
unfit for sale (USDA. 1983). Adult females prefer to lay their eggs on trees that
are under stress.

The redheaded pine sawfly overwinters as pupae in the topsoil or duff.
When degree-day accumulation surpasses 298, then the adults emerge. Below

298, ants are actively preying on the pupae. Between 298 and 500 degree days.

124

NITROGEN SUBROUTINE

INPUT
Trefoil biomass

 

 

 

 

 

 

 

Update biomass for the

 

week

 

 

 

Calculate decomposed biomass
decomp - biomass'exp(-0.29)
biomass(NH)-Biom(N)-decomp

 

I

 

 

Calculate N applied for 1 week
NAPPRATE .- decomp/27

 

 

 

Cale. pot. corn yield/week
SlTElND-105.2‘(1-10“(—.02
'(NAPPRATE+35.2) ) )

 

I

 

 

Calc. site imprav. factor per
week: siteimpr-siteind-73.60

 

 

 

125

the females oviposit. If the trees are grown on a poor site or if the trees are
older, it is assumed that the number of eggs laid reaches its maximum
potential. Between 750 and 850 degree days the larvae hatch and start feeding
immediately. By knowing the number of larvae. the amount of defoliation can
be estimated. Studies showed that from the first to fifth instars, each larva
can consume about 1.1 inches of foliated branch (Averill et al. 1982). The
number of larvae used in the defoliation calculation equals the number of
larvae available to the ants as food.

After 850 degree days pupation begins and defoliation tapers rapidly.
The number of larvae that reach this stage and survive the winter. plus the
number of pupae from the beginning of the season are used both for the
number of sawflies available to the ants as prey and to determine the
population density for the next year. This value is used to calculate the new
season‘s pupae and reinitiates the cycle. The outputs of this subroutine are: 1)
the number of larvae available to the ants as prey. and 2) the amount of

defoliation. The subroutine is represented in Figure 23.

EUROPEAN PINE SAWFLY SUBROUTINE: The European pine sawfly,
Neodiprion sertifer Geoff, larvae feed only on the older needles of pine trees.
The trees usually recover and outgrow the injury in 2-3 years (USDA, 1983).

In the model, the season begins with calculating the number of eggs
available. After the larvae hatch in mid- to late April, the number of larvae
available to the ants as food is calculated and the defoliation caused by the
larvae. The number of adults available to lay eggs is determined from the
number of pupae. The total number of adults is the number that emerged from
their cocoon in the current year and a small percentage of adults that did not

eclose during the previous season. The outputs from this subroutine are: l)

126

Figure 23. REDHEADED PINE SAWFLY SUBROUTINE

  
  
  
 
 

 

YES CALL GIVEN #
OF PUPAE

 

 

 

 

 

 

 
 
 
 

 

IF m CALCULA
>293 DEG #PUPAE
AY AVAIL. As
CALCULATE NEW . YES (EMERGE’
SEASONS #PUPAE (CALCULATE #
OF ADULTS

 

  
  

 

ALCULATE # EGGS LAID:
o IERTREES-MORE EGGS

IF

50<DDA
<850

   
   
   
 
 

 

   

 

YES
TE#

CAL ULA
PUP/SE FOR NEW CALCULATE I"
LARVAE AS FOO
CALCULATE
END l DEFOLIATION la
:1-

 

 

 

 

 

127

the number of larvae and pupae available to the ants as prey. and 2) the
amount of defoliation caused by the larvae. The subroutine is represented in

Figure 24.

APHID SUBROUTINE: Honeydew constitutes the most abundant type of food
collected by many species of ants (Carol and Janzen. 1973: Skinner. 1980).
However. most ants also require protein usually in the form of insect prey to
complete their diet. The ant species used in this model. F. exsectoides. tends
aphids for honeydew and preys on a wide variety of insect prey (Andrews.
1929: Headley, 1943; Haviland, 1947).

Biomass of trefoil is the only input for the APHID subroutine. If the
average biomass of trefoil is less than 10 glplant. the variables are reset to 0.
If the average trefoil biomass is greater than 10 g/plant, we assumed that
there was sufficient plant materials to support aphids and the subroutine is
initiated. A population density for the aphids is randomly selected from one of
two aphid population density series developed by Banks (1954). The density of
aphids is determined at weekly intervals according to Taylor‘s method for
estimating aphid densities on beans. Total honeydew produced (g/week) is
determined based on the population density of aphids. This figure is input into

the ANT subroutine. The subroutine is represented in Figure 25.

ANT SUBROUTINE: The ant species we've chosen to incorporate in the model
is the Allegheny mound ant, Formica exsectoides Forel. This ant occurs
primarily in eastern United States but is also scattered throughout other parts
of the country as well (Andrews 1926). F. exsectoides is a predator of many
forest insect pests, but also feeds on honeydew which is produced by aphids

and other Homoptera. Because of this broad feeding habit, we assumed that F.

128

Figure 24. EUROPEAN PINE SAWFLY SUBROUTINE

  
  
   

 

CALCULATE

AVAIL. EGGS

CALCULATE ADEFOLIATION
NEW SEASONS

CALCULATE CALCULATE
DEFOLIATION (N EW)TOTAL
BY LARVAE POPULATION

 

 

 

 

 

 

 

 

 

CALCULATE

#AVAILABLE @
PORFOOD

YES %DEFOLIATION CALCULATE '
NEW
'0 POPULATION

CALC. #
AD”: T gfifigéfigy UPDATE ADUL LAID
”'- POPULATIONT

PUPAE
%DEFO. I]
-0

 

 

 
  
  
   

 

 

 

 

 

 

 

 

 

 

 

 

129

APHID SUBROUTINE

 

   

RESET
VARIABLES-0 I

 

 

  

BIOMASS
Zlogan

 

   

 

«DELECT POPULN.
ENSITTSF)ACTOR

I

STEP WEEK
COUNTER (W)

 

 

 

 

 

 

CALCULATE
POP. DENSITY
AD - S + W

I

CALCULATE
TOTAL APHID

 

 

 

 

130

exsectoides can maintain high and active populations in the absence of
abundant insect defoliators by switching their foraging towards collecting
more honeydew or perhaps soil dwelling arthropods.

In the ANT subroutine. manipulating this switching behavior is the key
element in our biological pest management program. This is done by mowing
the trefoil, thus significantly reducing the honeydew available to the ants.
This obliges the ants to switch their foraging toward the insect pests of on
pines. Mowing is done during the night when the number of ants foraging is
reduced.

The amount of honeydew from the aphids subroutine is input into the
ANT subroutine. If degree-day accumulation is less than 64, it is assumed that
the ants are not actively foraging on either insect prey or honeydew. When
degree day accumulation is greater than 64, a predation rate is selected. The
predation rate is derived from published figures on the amount of prey] day
brought back to a mature nest of a closely related Formica spp. (Skinner, 1980),
because values were not available for F. exsectoides . The maximum number of
insects brought back to the nest in June is 3.600. This value correlates with the
greatest need for protein to meet maximum brood production at this time. The
minimum value of insects brought back to the nest is in September with a
value of 500 insects/day and correlates with the decrease in protein needed
because brood production has significantly declined (Haviland, 1947). When
the ant colonies are younger then 4 years old, the predation rate is reduced by
0.9 due to the smaller foraging force that would be associated with smaller ant
colonies. Although the demand for food increases during this phase of colony
development, there is only a maximum amount of prey an individual forager

can retrieve.

131

The subroutine is presented in Fig. 26. Next, the subroutine calculates
the amount of prey eaten. This equation depends upon the amount of available
honeydew and utilizes the predation rate previously selected.

The output from the ANT subroutine is: P = the number of prey eaten by
the ants.

The assumptions of the subroutine are:

1) F. exsectoides will increase predation for a specific type of prey as it
becomes available in the environment..

2) F. exsectoides can maintain and function as a colony in the absence of their
main honeydew resource for relatively short periods of time (1-2 weeks), after
harvesting the trefoil.

3) the ants will prey more heavily on insect prey in the absence of honeydew.

Analysis of Output

Simulations from this model addressed the more basic questions about
the ”improved” system including, trefoil, leaf area and defoliation, site
improvement, and variations of defoliations as a functions of tree age, number
of ant colonies and site quality.
Trefoil vs. Leaf Area and Defoliation: The simulations showed that
regardless of the original site classification, the improvement in terms of
increased leaf area depends is influenced more by the number of mowings
than the density of trefoil. To show the effect of trefoil density on leaf area
for a given site classification, the leaf area generated by the site with no
trefoil input was subtracted from all curves. The resulting curves are relative
to this ”constant" site (Figs. 27-30).

While all of the relationships are linear, the greatest site improvement

occurs when the trefoil density is 10 plants/m2. Only slight improvements

132

Figur8~ 26. ANT SUBROUTINE

9

INPUT
Honeydew

 

 

 

 

IF N

New seas

Y
AntDD-O
P-O

Ealculate Degree Dag

 

 

 

 

 

 

 

 

 

Month Ant DD

 

 

 

1930

 

2304

May 64<DD<298
June 299<DD<831
Ju|y831<DD<1415I
Aug 1416<OD <

Sep 1931<DD<

 
  

  

month PR
W
June 3600
July 1400
Aug 1000
Sep 500

 
 

P

 

 

 

 

 

 

133

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANT SUBROUTINE CONTINUED
N
Y
Pr - (0.9) Pr
May 3360<HD<10000 IF\ N
June 7.42xto <HD< Month HD P - PR (0.75)'5
2.23x10 -
July 2.38x10 <HD<
7.15x10 Y
Aug 13400<HD<
40300 P - Pr(1.5)'5
Sep 3350<HD<10000

 

 

 

 

 

134

 

 

A
t 20001
6‘ e—e Treefoil Density - 0.0
m . ra—a Treefoil Density - 10
V ‘ H Treefoil Density - 20 _
a)? 1500; H Treefoil Density - 30 "
+1 «I
C e
o
4.; ..
0‘)
g '1
o "I
+1 1 -
$3 I /"
u _
,2 500- /=
0": -I /.

J c
g : /;/
.._ O-J .e,4fie.qr$.ejese.
8 0.0 2.5 5.0 7.5 10.0
.4

Season #
( Trees = 1209 Soil = 60 Ants = 1 Mow = 0 )

27- EFFECT OF TREEFOIL ON TREE GROWTH

135

e—e Treefail Density = 0.0
H Treefoil Density =- 10
H Treefoil Density - 20
H Treefoil Density - 30

L‘

“

 

 

 

0 T & fl 17W 1 3 r‘S T 6 I G t q—r r
0.0 2.5 5.0 7.5 10.0

Leaf Area Relative to Constant Site ( Sq. Ft. )
B
O
C)
Le

Season #
(Trees = 1209 Soil = 60 Ants = l Mow
Figure 28. EFFECT OF TREEFOIL ON TREE GROWTH

1 )

 

136

 

 

A

fi-J 2000-

5. _ e e—e Treefoil Density - 0.0

U) -I a—a Treefoil Density - 10

j; I H Treefoil Density - 20 _

a): 1500_ H Treefoil Density - 30

4.: -I

C -I

o

*" 'l

tn

8 d

.9 I .

0 'I /':

.9. 500- /=

é’ 'l _

O -l /

0 d L.

L. 4 .,/

<

0‘. 0 r rfirfirarfiairHfi—e rfififir
8 0.0 2.5 5.0 7.5 10.0
_I .

Season #

(Trees = 1209 Soil == 100 Ants = 1 Mow = 0 )
“8“" 29: EFFECT OF TREEFOIL ON TREE GROWTH

Leaf Area Relative to Constant Site ( Sq. Ft. )

.3.
N
O
‘P

137

 

 

 

. J H Treefoil Density - 0.0
-I H Treefoil Density - 10
1 H Treefoil Density - 20
.I
-I
I
1
Ojr—fi via T$f$fi4 r? rte—I—
0.0 2.5 5.0 7.5 10.0
Season #

(Trees = 1209 Soil == 100 Ants = 1 Mow = 1 )
“8“" 30EFFECT OF TREEFOIL ON TREE GROWTH

138

occur when higher densities of trefoil are used. One mowing per season
results in a maximum improvement of 4000 ft2 over a 10 year period. If the
trefoil is not mowed at least once per year, (allowing only decomposition at the
end of the season) this figure is reduced by 50%.

If the primary objective is to increase leaf area and produce a high
grade Christmas tree, then planting the trefoil at the highest density would
seem. to be the best strategy. The simulations showed that defoliation does not
decrease proportionately to increases in the density of trefoil planted. In fact
there is a slight trend toward a decrease in defoliation with lower densities of
trefoil (Figs. 31 & 32). These results can be best understood by remembering
that as the density of trefoil increases, the number of aphids and therefore
available honeydew to the ants also increases. The model assumes that the ants
will forage most readily on the food resource that is available in the greatest
quantity, and, colony demands relative to brood production are also accounted
for. Theses simulations suggest that the density of trefoil planted should be
optimized, not maximized.

Site Improvement: One function of the trefoil is to supply nitrogen to the
soil in an attempt to improve the site quality. This is especially relevant to
Michigan Christmas tree production because pine plantations are usually
established on marginally sandy soils. Site improvement is measured in terms
of additional corn yields (corn equivalents) per acre. Thus, only site
improvement for a site with a base rating of 60 com equivalents was analyzed.
The analyses was also limited to sites with 1 ant colony and mowing regimes of
0 (end of the season decomposition) and 1 (end of the year decomposition plus
one management mow).

Although there is a carry-over each year of decomposing biomass in

the soil, the decomposition rate and quantities of trefoil biomass are such that

139

700-

L

. e—e Treefoil Density - 0.0
‘ a—a Treefoil Density - 10
" a—e Treefoil Density - 20
55010—0 Treefoil Density - 30
I

   
 

Cumulative Defoliation (sq. ft.)

 

 

r"'r"_r"r'r""rj'
0.0 2.5 5.0 7.5 10.0
Season #
( Trees = 907 Soil =' 90 Ants = 2 )

Figure 31- EFFECT OF TREEFOIL DENSITY ON DEFOLIATION

140

600-

0—0 Treefoil Density - 0.0
H Treefoil Density - 10
H Treefoil Density - 20
H Treefoil Density - 30

 

Cumulative Defoliation (sq. ft.)

 

d
O r I I I I ‘r I

 

0.0 2.5 ' ' 5:0 1 7.5 10.0
Season #

( Trees = 907 Soil = 90 Ants = 4 )
“8“” 32- EFFECT OF TREEFOIL DENSITY ON DEFOLIATION

T I I I I I r I I

141

it is not possible to attain a maximum site improvement (to 120 com
equivalents) within the 10 year growing cycle used in the model. However,
the simulations suggest that the 10 year time frame is enough time to
significantly increase the quality of the site. This is best achieved by mowing
the trefoil during the growing season as opposed to increasing the density of
trefoil per acre (Figs. 33 & 34). Increasing the trefoil density per acre on
sandy soil contributes only marginally to site improvement under either
mowing regime. If the trefoil is not mowed once during the growing season,
the site only realizes a maximum improvement of 15 com equivalents over 10
years.

Mowing more than once during the growing season could significantly
increase soil improvement, but because the model is based a low input
sustainable system, only one mowing was incorporated in this model. Finally,
tree density had little affect on site improvement. The competition factor is
such that it did not significantly reduce site quality.

Variation in Defoliation as a Function of Tree Age, Number of Ant
Colonies, and Soil Quality: The density of insect pests directly influences
the quality and esthetic value of a Christmas trees. The main factor to
influence the amount of defoliation due to insect pests was site quality. The
differences in cumulative defoliation due to soil type and the number of ant
colonies is shown in Figs. 35-37. As the soil type improves, the maximum
defoliation decreases independent of the number of ant colonies. As the
number of ant colonies per acre increases, there is also a decrease in
defoliation. Potential defoliation may be due to the fact that female redheaded
pine sawflies prefer to lay eggs on trees that are under stress, such as would be

the case if the trees were growing on a poor site.

142

30—
-I H Treefoil Density - 0.0
I a—a Treefoil Density - 10
‘ H Treefoil Density - 20
23: H Treefoil Density - 30

 

 

 

0.0 2.5 5.0 7.5 10.0
Season #

( Trees = 1037 Soil = 60 Ants = 1 Mow =
“8"" 33«EFFECT OF TREEFOIL ON SITE CLASSIFICATION

Site Improvement ( Corn Equivalent Yield )
"E

0 . ee—T—e—I—e—r—e—i—e—r—e—I—e—y—efqr,

0)

143

30-
« e—e Treefoil Density - 0.0
. a—a Treefoil Density - 10
‘ H Treefoil Density - 20

23: H Treefoil Density - 30

15~ /

.2 /;
8: //

 

 

Site Improvement ( Corn Equivalent Yield )

e/l
Dr'.e.e rere.esefierer
0.0 2.5 5.0 7.5 10.0
Season #
(Trees = 1037 Soil = 60 Ants = 1 Mow = 1 )

W3 EFFECT OF TREEFOIL ON SITE CLASSIFICATION

144

 

 

 

 

1700-

A J

t‘ -l
.4 T

D.— 4 r. - ’- 3 a 3

£1 13501 i" ; :- : 5 '7 a

o J N

2:3 l ‘ ‘ ' . .

o-o- 1 5"“ ' I

3 10004 4 4L— 4:

Q) -l

C] a

a, «II

.3 a H 5 Ant Colonies

_o_ 650- H 4 Ant Colonies

3 T H 3 Ant Colonies

:EJ ‘ H 2 Ant Colonies

O 1 a—a 1 Ant Colony
1 H No Ant Colonies

300 T 1 I T I 1 I ~ Y I l j 1 I ‘l 1+7 f fit 17 l‘- l *l

0.0 2.5 5.0 7.5 10.0

Season #
( Treefoil = 20. Trees = 907 Soil = 60 )
Figure.‘35. EFFECT OF SOIL TYPE AND ANTS ON DEFOLIATION

145

     

 

 

 

l lDD-I

A "l
1.5 -I

. -I
U- u
3 8251

c:
.9. ”
y I
.9.
“g 550-:
o I _"’"'—“

a) d
.5 . e e 5 Ant Colonies
2 275- H 4 Ant Colonies
:3 d H 3 Ant Colonies
g ‘ . ; . - .- H 2 Ant Colonies
O 1 ‘ ";-3____g7 H 1 Ant Colony

o—e No Ant Colonies
O *r 1' T T j j I I F‘Y U j j r—' j fi— ‘r I f W
0.0 2.5 5.0 7.5 10.0

Season #
( Treefoil = 20 Trees = 907 Soil = 90 )
“3"“ 36' EFFECT OF SOIL TYPE AND ANTS ON DEFOLIATION

 

146

 
  
    
 

 

 

 

500-

A - H No Ant Colonies

33 ‘ H 1 Ant Colony

, 'l s—s 2 Ant Colonies

‘7 1 H 3 Ant Colonies

3 Tbo'j H 4 Ant Colonies

g. -l H 5 Ant Colonies

15 -I .,

.g 4 a

8 .300-

a) -I

c: 1 _ -

a) 4 - '

02 '1 ' ‘

:cj 150~ — _ ‘

:3 " C 7

o l
O 1'— wf T T— r—T' _T— T ‘— T j t r f 1' T r T T _‘l" I
0.0 2.5 5.0 7.5 10.0

Season #
( Treefoil = 20 Trees = 907 Soil = 110 )
37. EFFECT OF SOIL TYPE AND ANTS 0N DEFOLIATION

147

Figure 35 shows that 1200 ft2 of defoliation per acre occurs by the third
year even if there are five colonies of ants placed on each acre. However, the
authors were very conservative for the values used in the predation rates of
the ants. When ants are not employed as biological control agents. the
simulations showed that potentially 1350 to 1490 ft2 of foliage is defoliated.
When ants are not present, a 37 percent decrease in defoliation can be
expected if the trees are planted on a good site. When as many as five ant
colonies are used per acre, defoliation on an ideal vs. a stressed site is reduced
by 8 percent, indicating that less favorable sites potentially can produce good
Christmas trees provided a form of pest management such as predaceous ants

are employed.

Further Model Development

Based on the results of the analyses, we believe that the model warrants
further development and fine tuning combined with additional research in
order to better approximate the biology of the real system. Time did not permit
a sensitivity analysis of all of the model's parameters to determine which ones
had the greatest impact on the overall output. A first step in the fine tuning of
the model would be to test the sensitivity of all parameters.

Similarly, the model only incorporated three of the several pests of
Scotch pine in the Great Lakes region. While a subroutine and code for pales
weevil was developed for this model, time did not permit it to he added to the
model.

One additional area where the model requires some fine tuning is in the
time horizon for long-term site improvement. As indicated in Fig. 38, the rate
of leaf area production is fairly linear until about the year 8 at which time the

rate begins to increase. This is more than likely due to the cumulative effects

148

of the site improvement. Further work with this model should include a
longer time horizon over several rotations in order to ascertain longer-term
site improvement potential.

Finally, any new or revised versions of the model should begin to
address the economics of the conventional and improved systems by
attempting to quantify net returns. Financial and economic analysis of the
system could be built into the model or addressed in a typical partial budget or
benefit-cost framework.

Additional Research

While the model requires some additional sensitivity analyses.
simulation results did point to certain priorities that should be addressed
through research regarding the improved system. In particular, research is
needed on ant predation switching values as leaf area lost is extremely
sensitive to these values. However. literature on these values is virtually non-
existent.

We also believe that a better leaf area growth model could be developed
by incorporating site differences either by ”com equivalent" or by actual soil
nitrogen content into the TREE regression equation. This would involve
collecting data on a wider variety of sites where Scotch pines are known to
grow. It would also be useful to validate any leaf area model by harvesting
trees from representative age classes and removing and drying the leaves to
determine total ovendry weight of leaf biomass. Conversion of leaf biomass to
leaf area can be achieved though existing conversion factors.

While the NITROGEN subroutine provides a relative estimate of site
improvement, additional research on trefoil decomposition rates and nitrogen
contributions through organic matter decay and nitrogen fixation is needed to

develop a more realistic model. Unfortunately there is currently very little

149

information available on this subject. Moreover. to our collective knowledge.
there has been no research on Scotch pine response rates to nitrogen.
Currently. nitrogen fertilizer is applied to Christmas tree plantations only
when the trees show signs on nitrogen deficiency. Research on the entire
nitrogen cycling process in Scotch pine (organic and inorganic) appears to be
sorely needed.

We assumed that due to the aggressive behavior and allelopathic effects
of the trefoil. all of the weeds in the plantation would be eliminated. Further
research is needed to determine if this in fact is true. Additionally. it was
assumed that trefoil would make only a positive contribution to leaf area
growth and the only competition between the trefoil and the trees would be
for radiant energy. Below ground competition for nutrients and moisture as
well as the allelOpathic effects from the trefoil on the trees were assumed to be
negligible. Additional research is required in order to determine whether this
is indeed the case.

Finally. we assumed that the only biological control agent present in
the plantations was one species of a predaceous ant. In most situations. there
are many other natural enemies that negatively influence the pest complex;
including predators other than ants. parasitoids. small mammals and disease.
The addition of the trefoil to the system increases the potential number of food

and shelter resources available for other biological control agents to utilize.

Conclusions
The model did predict that leaf area under the improved system would
increase by 6037 m2 over a lO-year period compared to a conventional system.
A conventional system in the model is assumed to be one where there were no

ants. no pests (due to the use of pesticides) and not long-term site improvement

150

(due to the use of inorganic fertilizer). This amounts to a ”savings" of
approximately 80 degree days for a 10 year rotation (Fig. 38).

While these results do not seem spectacular. it should be noted that we
were very conservative in many of the assumptions. believing that in this
first approximation of the model. it was better to err on the side of caution. For
example. published ant predation rates (Skinner. 1980). were reduced by a
factor of 5 for the pests. Initially the published predation rates were used in
the model and resulted in complete elimination of the pests. Similarly. the
equation used to predict corn yields using nitrogen input from trefoil was
reduced by a factor of 100. .We believe the published figures for both ant
predation and corn yields are very optimistic and possibly do not indicate the
biology of the proposed system very well. This suggests that while models can
be used to summarize existing information. this information may not always be
correct.

Perhaps more importantly. the model predicted an increase in leaf area
with very little inputs by the grower. While no attempts have been made to
determine net margins of the conventional vs. the improved system. in the
economist partial budget format. increased cost of the the improved system
would be the one time trefoil planting and the cost of harvesting the trefoil.
The cost incurred from mowing the trefoil could be offset by selling a portion
of the trefoil for forage if more than one mowing per season occurred. The
decreased costs associated with the improved system are less fertilizer and
pesticides (if any) and fewer labor hours necessary for weed control. Benefits
under the improved system (leaf area) would also increase slightly. In
summary. the net benefits under the improved system appear to be greater

than the benefits under the conventional system. While difficult to quantify.

Difference in Leaf Surface Area
Controllea — Uncontrolled Leaf Surface Area

5750§
52505
4750€
4250;
3750i
szsog
2750§
2250§
1750§
1250?

7503

250

151

 

 

Figure 38.

04-1

Year

Comparing conventional and 'improved system.

-1

10

152

the improved system provides the additional social benefit of reduced ground

water contamination and other chemical related problems.

 

152b
Literature Cited

Andrews. E. A. 1926. Sequential distribution of Formica cxsectoides Forel.
Psyche 33: 127-150.

Andrews. E. A. 1929. The mound building ant. Formica exsecroides Forel.
associated with tree hoppers. Ann. Entomol. Soc. Amer. 22: 370-391.

Averill. R. D., L. F. Wilson and R. F. Fowler. 1982. Impact of the redheaded pine
sawfly (HymenopterazDiprionidae) on young red pine plantations.
Great Lakes Entomol. 15: 65-91.

Banks. C. J. 1954. A method for estimating papulations and counting large
numbers of Aphis fabae Scop. Bull. Entomol. Res. 45: 751-756.

Banks. C. J. and E. D. M. Macaulay. I964. The feeding. growth and reproduction
of Aphis fabae on Vicia faba under experimental conditions.
Ann. Applied Biol. 53: 229-242.

Blackman. T. 1974. Aphids. London and Aylesbury. Gin and Company Ltd.
175 pp.

Carol. C. R. and D. H. Janzen. 1973. Ecology of foraging by ants. Annual
Review of Ecology and Systematics. 4: 231-257.

Coleman. Cole. Elliot.1984. Decomposition. organic matter turnover. and
nutrient dynamics in agroecosystems. In. “W
WWiley lnterscience publication. New York. 233 pp.

Comeil. I. A. 1981. Unpublished Master's thesis. The behavior and impact of
the pales weevil on Christmas tree plantations. Michigan State
University.

Fagerstrom. T. and U. Lohm. 1977. Growth of Scots pine (Pinus sylvestris L.)
mechanisms of response to nitrogen. Oecologia 26:305-315.

Fribourg. H. A. and W. V. Bartholomew. 1956. Availability of nitrogen from
cr0p residues during the first and second seasons after application. Soil
Sci. Soc. Am. Proc. 20: 505-508.

Haviland. E. E. 1947. Biology and control of the Allegheny mound ant. J. Econ.
Ent. 40: 413-419.

Headley. A. E. 1943. The ants of Ashtabula County. Ohio (Hymenoptera:
Formicidae). Ohio 1. Science 43: 22-31.

Lyons. L. A. 1964. The European pine sawfly. Neodiprion serrifer
(HymenOptera: Diprionidae). Proc. Entomol. Soc. Ontario.
94: 22-35.

153

MacAloney, H. J. and L. F. Wilson. 1964. The redheaded pine sawfly. Forest
Service Leaflet no. 14. USDA Forest Service.

McGraw. R. L.. P. R. Beuselenck and K. T. Ingram. 1986. Plant population
density effects on seed yield of birdsfoot trefoil. Agronomy Journal.
78: 201-205.

McGraw. R. L. and P. R. Beuselenck. 1983. Growth and seed yield of
birdsfoot trefoil. Agronomy Journal. 75: 443-446.

Odum. E. P. 1984. Properties of Agroecosystems. In. W
W. Wiley Interscience Publication. 233 pp..

Skinner. G. J. 1980. The feeding habits of the wood-ant Formica
rufa (Hymenoptera: Formicidae). in a limestone woodlot in
north-east England. J. Animal Ecol. 49: 417-433.

Taylor. L. R. 1970. Aggregation and the transformation of counts of Aphis
fabae Scop. on beans. Ann. Appl. Biol. 65: 181-189.

United States Department of Agriculture. 1983. Christmas tree pest
manual. North Central Forest Experiment Station. St. Paul. MN. 108 pp.

United States Department of Agriculture. 1979. Soil Survey of Ingham County.
Michigan. Washington D.C. 142 pp.

Waring. R. H., P. E. Schroeder and R. Oren. 1983 Applications of the pipe model
theory used to predict canopy leaf area.

Way. M. J. 1963. Mutualisms between ants and honeydew producing
Homoptera. Ann. Rev. Entomol. 8: 307-344.

Wilson. L. F. 1966. Effects of different populations levels of the European pine
sawfly on young Scatch pine trees. J. Econ. Entomol. 59:1043-1047.

Woodmansee. R. G. 1984. Comparative nutrient cycles of natural and
agricultural ecosystems. In:
Com. Wiley Interscience Publication. 233 pp.

Wright. J. W. 1976. Scotch pine varieties for Christmas tree and for forest
planting in Michigan. MSU Research Report 293.

MANUSCRIPT 3.

Designing and Field Testing Portable Colonies of the Predaceous Ant.
Formica exsectoides Forel

_ 154

155

Introduction

This paper introduces a new technique for manipulating ants as
biological control agents. Managing predaceous ants in portable nests could
increase the potential for using a combination of pest management strategies
in forest agroecosystems. The colonies could temporarily be removed or placed
in optimal locations to meet specific management objectives such as chemical
applications or harvesting. Many of the ideas that contributed toward the
development of this technique are based on technologies successfully utilized
by beekeepers.

The objectives of this study were to design. construct and test different
portable ant nests and conditions that favoured colony survival. An additional
objective was to determine how quickly the ants would begin to forage for

insect prey after the transplant.

156

Material and Methods

Twelve portable nests of Formica exsectoides Forel. a native Michigan
mound building ant. were constructed and transported to a 6-year old jack pine
plantation in Alpena County. Michigan. On each of three dates beginning in
late July. a cohort of four ant nests were transplanted into the jack pine
plantation. The number of workers in the colonies increased and the nest
design progressively changed over the 3-week period in accordance with our
observations. Each cohort of nests was taken from the same parental colony.
This was done for two reasons: 1) to have a larger force of nest siblings in one
area (to eliminate intraspecific competition). and 2) to form a super-colony. if

the individual colony size proved to be too small.

Portable Nest Design. The queens were placed in 15 x 8 x 5 cm plastic
containers with 20 holes drilled in the side. top and bottom. large enough for
the workers to utilize but small enough to restrain the queen. The workers and
nest material from the parental colony were placed in 31 x 15 x 45 cm
rectangular plastic waste baskets (portable nests). Small holes were made in
the bottom of the nests for water drainage. The portable nests were placed in
snug-fitting holes in the plantation with the lip of the nest approximately 3
cm above the surface of the ground. Small holes were drilled on the sides of
the nests at ground level to allow the ants to enter and leave the portable nests.
The four colonies of the first cohort each consisted of 500 workers and
one mated queen. The depth of the soil was 30 cm and the queen excluder was
placed 10 cm below the soil surface. The ants were allowed to acclimate to the
nests for 24 h prior to opening up the colonies in the new location. The
colonies were fed honey-water and two crickets/day throughout this time. In

addition, each colony was fed honey-water for 3 days following their release.

157 -

A small colony of 25 redheaded pine sawflies was placed on a tree
approximately 1.5 m away from each nest to provide the ants with insect prey
to improve the opportunity for the colonies to become established.
Observations on ant predation on the sawfly colonies were made for two days
following the transplant and thereafter weekly.

The second cohort consisted of four colonies each containing 1000
workers and one mated queen. The acclimation period was increased to 48 h
and only two nests were left open. A flat lid was fitted over the top of the other
two nests to protect them from possible rain damage. However. fifteen small
holes were made in each lid to allow some moisture to enter the nests. One
small colony of 25 sawflies was placed on each of three trees approximately 1.5
m from the nest.

In the third cohort. the number of workers per nest was increased to
1500 and one queen and each colony was covered with a lid. More nest
material (soil and plant debris). was added. making the soil depth 40 cm and the
queen excluder only 7 cm from the bottom. The ants were allowed to acclimate
to their new nests for 5 days prior to release. The nests were placed 15 m apart
in a row and each cohort was separated by 30 m. Each nest was provided with
small colonies of sawflies. as described above.

We inspected each nest at weekly intervals for the following: 1)
presence of a queen. 2) approximate number of workers. and 3) number of
foraging trails established. 'Also. observations were made on the interactions
between the ants and the sawflies. The queen excluders were removed from
the portable nests with as little disturbance to the soil as possible. However.
the soil around the queen excluder did cave in when the excluders were
removed and had to be repacked around the excluder after returning the

excluder to the nest.

158

Results and Discussion

Two days following the transplant of the first cohort of nests (each nest
contained 500 workers). ants were seen moving nest material and nest mates to
new locations within a 2 m radius of the nests. After 1 week no queens were
found in any of the colonies and little activity was detected in the nests. The
contents of the nests were sorted and only 5 to 19 workers were found. The
sand throughout the nests was damp and and very few galleries were found.

Two days following the transplant of nests from the second cohort
(each nest contained 1000 workers). ants from the two colonies without lids
were seen moving both nest material and nest mates to new locations despite
the noticeable increase in the number of galleries constructed during their
acclimation period. Very few workers and no queens were found in the two
colonies without lids after 1 week. A queen and approximately 50 workers
were found inside of the excluder of the third nest. while the queen and many
workers were near the excluder in the fourth nest. The surviving two
colonies had established two foraging trails each. A foraging trail is defined
as a steady flow of ants both entering and leaving the nests.

When the third cohort of nests (each nest contained 1,500 workers)
were inspected after 1 week. many galleries were established in each of the
four nests. In each excluder. a queen was found along with approximately 50
workers. The queen excluders were then carefully replaced and the colonies
were monitored for 5 more weeks.

Three of the four colonies had established two foraging trails and the
remaining colony established one. The ants were seen performing typical
colony functions such as restructuring nest materials and carrying pupae to

various chambers in the nest. A flow of ants was seen entering and leaving

159

the two exit holes in each of the colonies. The colony with only one foraging
trail exhibited the least activity overall. We observed four ants from the
fourth cohort removing sawflies on nearby trees within 1 week following
their release.

The survival time for two of the nests from the second cohort ranged
from 3 to 7 weeks. The last cohort of nests survived in the portable nests for a
period of 6 weeks. At this time the experiment was terminated.

Manipulating ants in portable nests should be useful when using an
integrated approach to reducing pest populations. This study showed that the
following factors positively influenced colony establishment: 1) the nest lid.
2) the queen excluder. and 3) the acclimation period. The next step in the
development of the portable ant nests is to determine if the ants will remain in
the portable colonies for more than 7 weeks and to increase the scale of

introductions.

160

Figure 39. Portable nest: of Formica exseccoides Feral

 

 

 

 

 

"I71111111111111.1111