20?..qu

2.
I

e... 1.: i. , {”34
z :5... ii 2.5“. Luau.

1.7.... . “A

4
MW

it

a: Mm.

gm

”.5 .9 up.

{PI 1

3-.
3!..-
‘ ‘ . ‘1}!
.1: I. . I, I.
I. .v . I 0
«Z? n. ‘ ‘ humvbflpr ._

N... t:

uuu‘fivuwflunrut .
l

 

5114! f

n;

‘r nlu
Saar.

i, 0‘
¢ ,
1:!
. x.

. Parka“;
mini} a! h.
Jflmfi‘fluvh. Ash

J. z. .t\ (It.

 

I o I J
V nnmmmuqfiqfi “flag.

 

THESIS
Z

 

II IIIIIIII

31293 01555 2338

IIIIIIIW

                                             

This is to certify that the

dissertation entitled

INTERACTIONS BETWEEN LANDSCAPE STRUCTURE AND LADYBIRD
BEETLES (COLEPTERA: COCCINELLIDAE) IN FIELD CROP
AGROECOSYSTEMS .

presented by

MANUEL COLUNGA-GARC IA

has been accepted towards fulfillment
of the requirements for

PhD . ENTOMOLOGY
degree in

 

 

EWIHW

Major professor

Date /?/;2 4f/?6

MS U i: an Affirmative Action/Eq ual Opportunity Institution 0-12771

 

 

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES mum on or baton date duo.

 

 

 

 

 

 

 

 

DALE DUE DATE DUE DATE DUE
- 1313 ,
“M9 —— ’
J’ ‘ Q ~- 0&2
5921120
, ' 1 07 7
AEB_2_2_ZIJIIB __ __
#080410 I"
I. *7 4, 1

 

 

 

 

-—_—__—_—:I

 

INTERACTIONS BETWEEN LANDSCAPE STRUCTURE AND LADYBIRD
BEETLES (COLEOPTERA: COCCINELLIDAE) IN FIELD CROP
AGROECOSYSTEMS

By

Manuel Colunga-Garcia

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Entomology

1996

ABSTRACT

INTERACTIONS BETWEEN LANDSCAPE STRUCTURE AND LADYBIRD
BEETLES (COLEOPTERA: COCCINELLIDAE) IN FIELD CROP
AGROECOSYSTEMS

By

Manuel Colunga-Garcia

Management of agroecosystems to enhance natural regulation of insect pests
requires an understanding of predator ecology and how predatory insects use the
landscape and respond to its structural characteristics. A group of predatory
insects, ladybird beetles, were selected to study patterns of habitat utilization in
response to vegetation type, management practices, and habitat succession in a
complex agricultural landscape.

The field work was conducted at the Long Term Ecological Research (LTER) site
at the Kellogg Biological Station (KBS), in southwest Michigan. First, a life
systems study of Coleomegilla maculata lengi was conducted by sampling
wooded habitats to determine beetle aggregation sites in the landscape, and by
using stable isotopes to determine paths of energy flow in the beetle-crop system.

Secondly, abundance patterns of fourteen species of coccinellids were monitored

weekly during the growing season using yellow-sticky traps. The sampled
landscape consisted of an array of field crops under different management
practices, interspersed with perennial biomass plantations and vegetation in a
state of secondary succession. Seven years of weekly abundance records were
analyzed using Shannon Wiener and richness indices, Kendall's coefficient of
concordance, and principal component analysis. The results of the analysis were
used to produce a spatially explicit population model for comparative analysis of
landscape-predator interactions.

The main finding of the study were:

a) Habitat Succession. Ladybird beetle species diversity peaked during the second
year of secondary succession with a successive decrease in diversity thereafter. In
the Poplar plantation, the assemblage of beetle species showed a succession of
dominance by three beetle species;

b) Management Practices. Reduced chemical inputs (herbicides and fertilizers)
decreased beetle abundance and species diversity in corn fields but these
components increased in wheat. In the corn-soybean rotation sequence, C. m.
Iengi, an important native predator, was most abundant when corn vegetation was
dominant and least abundant when soybean dominated the landscape; and

c) Habitat type. Adults of C. m. Iengi used habitats associated with woodlots,
hedgerows, and rows of trees to aggregate prior to the onset of winter. Large
aggregations occurred near agricultural fields where corn or alfalfa was grown
the previous summer. In early spring, flowers constitute important sources of

food for beetles before they move to field crops to search for prey. After feeding

on flower pollen, adults beetles move to alfalfa or wheat, and finally to corn in
the late summer prior to selecting sites for overwinter.

Within the framework of the KBS LTER theme that ecological knowledge can
replace chemical subsidies, this work contributes to the role that management
practices can have on beneficial insects and identifies landscape characteristics
conducive to maintaining higher numbers of beneficial insect predators within

agroecosystems.

Copyright by
MANUEL COLUNGA-GARCIA
1996

DEDICATION
I dedicate this dissertation primarily to my wife, Maria Guadalupe, for all her
understanding, support, and patience that made possible this work. I also dedicate
it to my children, Sara Gabriela, Maria Elizabeth, and Miguel Angel, because
their presence constantly encouraged me, specially during the difficult moments I

had to go through.

vi

ACKNOWLEDGMENTS

I would like to express my sincere thanks to Dr. Stuart. Gage for all his support
and direction throughout this study and during my personal development as a
scientist. I would like to extend my appreciation to Dr. Mark Whalon, Dr.
Edward Grafius, Dr. Peggy Ostrom, and Dr. Douglas Landis for serving on my
graduate committee and reviewing this manuscript.

Several individuals helped during the field work and I thank them all.

I would also like to thank my friends, Rosemary Faber, Rod Murphy and Dawn
Ciciora for their editorial assistance in the revision of this manuscript.

The Consejo Nacional de Ciencia y Tecnologia, and the Universidad Autonoma
Chapingo of Mexico, as well as the Long Term Ecological Research at the
Kellogg Biological Station are gratefully acknowledged for their financial
support during this study.

Finally, I would like to extend my thanks to the faculty, staff, and graduate
students of the Department of Entomology and the Kellogg Biological Station for

all their encouragement.

vii

TABLE OF CONTENTS

LIST OF FIGURES ................................................................................ xi
LIST OF TABLES ................................................................................. xv
INTRODUCTION ..................................................................................... 1
IMPORTANCE OF LONG TERM ECOLOGICAL RESEARCH ................ 1
General perspective ......................................................................... 1
Agriculture ........................................................................................ 1
THE KBS-LTER ................................................................................... 2
COCCINELLIDS AS THE SUBJECT OF THIS STUDY ........................... 3
GOALS, HYPOTHESIS, AND OBJECTIVES ......................................... 4
Goals ............................................................................................... 4
Hypothesis ....................................................................................... 5
Objectives ........................................................................................ 5
METHODS .......................................................................................... 5

CHAPTER 1: LIFE SYSTEM OF PREDATORY COCCINELLIDAE
(COLEOPTERA) ON AGROECOSYSTEMS: CASE STUDIES ON

Coleomegilla maculata Iengi Timberlake .............................................. 8
BIOLOGICAL CYCLE ........................................................................... 8
SEASONAL HABITAT UTILIZATION .................................................... 8

CASE STUDY I. Toward the Characterization of Landscape Use by
Overwintering Aggregates of Coleomegilla maculata Iengi Timberlake
(Coleoptera: Coccinellidae) in Field Crop Ecosystems in Relation to

Wooded Habitats. .............................................................................. 11
Introduction .................................................................................... 11
General methodology ...................................................................... 11
Results ........................................................................................... 15
Discussion ...................................................................................... 23

CASE STUDY ll. Habitat Use Patterns of Coleomegilla maculata Iengi

Timberlake During Summer in Field Crop Agroecosystems. ................ 25
Methodology ................................................................................... 27
Results and Discussion ................................................................... 28

viii

CHAPTER 2: THE ISSUE OF SCALE IN STUDIES OF PREDATORY
COCCINELLIDAE (COLEOPTERA) IN AGRICULTURAL LANDSCAPES32

FRAMEWORK OF SCALE .................................................................. 32
Importance ..................................................................................... 32
Coccinellids and scales of biodiversity ............................................ 33

LONG TERM MEASUREMENTS OF ARTHROPOD POPULATIONS....35
Hierarchy theory and the concept of sampling ................................. 35
Characteristics of long term sampling .............................................. 36
Sticky traps as sampling tools for coccinellids ................................. 37
Comparison of two sampling methods: sweep-net sampling and sticky
trap in wheat at different time intervals. ........................................... 38
Temporal patterns of coccinellids at the K38 LTER main site ........... 50

CHAPTER 3: ASSEMBLAGE OF PREDATORY COCCINELLIDAE
(COLEOPTERA) IN RESPONSE TO TEMPORAL DIVERSITY WITHIN AN

AGROECOSYSTEM ............................................................................. 51
CASE 1: HABITAT MATURATION IN ALFALFA, POPLAR, AND
SECONDARY SUCCESSION ............................................................. 53

METHODS ..................................................................................... 54
Data analysis. ................................................................................ 55
RESULTS ....................................................................................... 58
DISCUSSION ................................................................................. 69
CASE 2: CORN-SOYBEAN ROTATION UNDER CONVENTIONAL
TILLAGE ........................................................................................... 72
METHODS ..................................................................................... 72
RESULTS ....................................................................................... 74
DISCUSSION ................................................................................. 77

CHAPTER 4: ASSEMBLAGE OF PREDATORY COCCINELLIDAE
(COLEOPTERA) IN RESPONSE TO HABITAT MANAGEMENT (TILLAGE

AND CHEMICAL INPUTS) WITHIN AN AGROECOSYSTEM .................. 80
METHODS ........................................................................................ 82
RESULTS .......................................................................................... 85

Corn 1993 ...................................................................................... 85
Soybean 1994 ................................................................................ 88
Wheat 1995 .................................................................................... 90
Principal Component Analysis .......................................................... 92
DISCUSSION .................................................................................... 94

CHAPTER 5: DEVELOPING A MODEL OF INTERACTIONS BETWEEN
PREDATORY COCCINELLIDS (COLEOPTERA) AND LANDSCAPE

STRUCTURE IN FIELD CROP AGROECOSYSTEMS. ........................... 97
INTRODUCTION ............................................................................... 97
OBJECTIVE .................................................................................... 100
METHODOLOGY ............................................................................. 101
RESULTS ........................................................................................ 103

1. Definition of the problem ........................................................... 103
2. Delimitation of the universe of concern ...................................... 103
3. Object of control and associated environment ............................ 103
4. Determination of Stimuli-Results ............................................... 104
5. Selection of model type ............................................................. 107
6. Selection of computer technology .............................................. 107
7. Mathematical relationships ........................................................ 109
8. Parameter estimation ................................................................ 110
9. Programming ............................................................................ 115
10. Sensitivity analysis ................................................................. 118
DISCUSSION .................................................................................. 125

SUMMARY AND CONCLUSIONS ....................................................... 131

APPENDICES .................................................................................... 136

BIBLIOGRAPHY ................................................................................. 156

LIST OF FIGURES

Figure 1 Phases in the development of the coccinellid-landscape model. .7

Figure 2. Biological cycle of predatory coccinellids and the role of each
biological stage. ............................................................................... 9

Figure 3. Life system of predatory coccinellids in agricultural landscapes.10

Figure 4. Classification of tree-related habitats for C.m. lengi aggregates
inthelandscape. ............................................................................ 15

Figure 5. a) South side of the Elsesser woodlot, MSU campus, East
Lansing, MI. b) Coleomegilla maculata lengi adults feeding on “spring
beauty” in the Elsesser woodlot, c) C.m. lengi feeding on dandelion
flowers at the edges of the Box woodlot, MSU campus, and d) C.m.
lengi adults hidden under leaves from the previous corn crop at the
edges of the Box woodlot ................................................................ 16

Figure 6. Relative abundance of C. maculata lengi (measured as adults per
flower) and dandelion flowers in one mzquadrats along a transect on
the south edge of the Box ll woodlot. MSU, East Lansing MI. 1994. ...17

Figure 7. Relative abundance of C. maculata lengi (measured as adults per
flower) and dandelion flowers in one mzquadrats along a transect on
the south edge of the Elsesser woodlot, MSU, East Lansing MI. 1994.18

Figure 8. Relative abundance of C. maculata lengi (measured as adults per
flower) and dandelion flowers in one mzquadrats along a transect on
the south edge of the K88 woodlot, , Hickory Corners, MI. 1994. ....... 19

Figure 9. Spatial distribution of C. maculata lengi overwintering adults on a
100 m perimeters around the Elsesser woodlot. MSU, East Lansing MI.
1994 ............................................................................................... 20

Figure 10. Distribution of aggregation sites (black squares) of C. maculata
lengi in the Kellogg Biological Station landscape. Large polygon is the
actual area observed. S= single tree, H= hedgerow, C = cluster of
trees, R = row of trees, and W = woodlot. ......................................... 21

Figure 11. Experimental design to study habitat utilization patterns by C.m
lengi at the Kellogg Biological Station, Ml. ....................................... 27

Figure 12. Mean trap captures (1r SE.) of C. maculata lengi in three
habitats, as measured by sticky traps, at the Kellogg Biological
Station, MI. 1992. ........................................................................... 29

xi

Figure 13. Surface maps of the spatial distribution of C. maculata lengi as
measured by sticky traps in four field crops during two times of the
season: May and August. 1992. ...................................................... 31

Figure 14. Expressions and dimensions of diversity at different scales on
time and space in agroecosystems .................................................. 35

Figure 15. Temporal dynamics of coccinellid adults in spring wheat
measured by sticky traps (upper 1.2 m and canopy level) and sweep-
net sampling. Kellogg Biological Station. 1993. ............................... 43

Figure 16. Temporal dynamics of coccinellid adults in winter wheat
measured by sticky traps (upper 1.2 m and canopy level) and sweep-
net sampling. Kellogg Biological Station. 1994. ............................... 45

Figure 17. Correlation between sticky trap sampling at two levels
(upper=1.2m, and canopy) and sweep-net sampling considering whole
season captures of coccinellid adults in wheat. ............................... 47

Figure 18. Patterns of coccinellid captures during a season measured as
frequency of sticky trap catches in 100 degree days intervals using
data collected from 1989 to 1995. Habitats included were corn, wheat,
alfalfa, soybean, poplar and secondary succession. ........................ 52

Figure 19. Patterns of coccinellid captures during the year measured as
frequency of sticky trap catches in 200 degree days intervals from
1989-1995. Habitats included are corn, wheat, alfalfa, soybean, poplar
and secondary succession. ............................................................. 53

Figure 20. Experimental design of the Long Term Ecological Research site
Kellogg Biological Station showing the spatial distribution of
treatments and replicates. Hickory Corners, MI ................................ 55

Figure 21 Relative abundance, species diversity, and richness of a complex
of predatory coccinellids in alfalfa, poplar and secondary succession
during seven years of habitat growth at the Long Term Ecological
Research site at the Kellogg Biological Station. Hickory Corners, MI. .61

Figure 22 Temporal distribution of three dominant coccinellid species in
poplar during the seven years of habitat growth at the Long Term
Ecological Research site at the Kellogg Biological Station. Hickory
Corners, Ml ..................................................................................... 63

Figure 23 Habitat scores and principal components of fourteen species of
coccinellids captured in poplar, alfalfa, and secondary succession
during seven years of habitat growth at the Long Term Ecological
Research site at the Kellogg Biological Station. Hickory Corners, MI. .68

Figure 24. Experimental design of the Long Term Ecological Research site
Kellogg Biological Station showing the spatial distribution of
treatments and replicates. Hickory Corners, Ml ................................ 73

xii

Figure 25. Relative abundance, species diversity, and richness of a
complex of predatory coccinellids in a six year corn-soybean rotation
under high input conventional tillage at the Long Term Ecological
research site at the Kellogg Biological station. Hickory Corners, MI..76

Figure 26. Habitat scores and principal components of eleven species of
coccinellids captured in a six year corn- soybean rotation under high
input conventional tillage at the Long Term Ecological Research site at
the Kellogg Biological Station. Hickory Corners, Ml. .......................... 79

Figure 27. Experimental design of the Long Term Ecological Research site
Kellogg Biological Station showing the spatial distribution of
treatments and replicates. Hickory Corners, Ml. 1993-1995 ............. 82

Figure 28. Relative abundance, species diversity, and richness of a
complex of predatory coccinellids in corn, soybean, and wheat under
four agronomic practices at the Long Term Ecological Research site at
the Kellogg Biological Station. Hickory Corners, MI.1993-1995. CT
=conventional tillage, NT =no tillage, LI =low chemical input, and Zl =
zero chemical input ......................................................................... 86

Figure 29. Habitat scores and principal components of thirteen species of
coccinellids captured in corn, soybean, and wheat under four
management practices at the Long Term Ecological Research site at
the Kellogg Biological Station. Hickory Corners, Ml. 1993-1995. CT
=conventional tillage, NT =no tillage, LCI =low chemical input, and

ZCI = zero chemical input. .............................................................. 93
Figure 30. Interactions of different subsystems in agroecosystems at

different hierarchical spatial scales. ................................................ 98
Figure 31. Life system of ladybird beetles in the landscape .................. 100

Figure 32. Methodological approach followed during the development of
the model on interactions between the landscape and ladybird

beefles. ........................................................................................ 102
Figure 33. Factors involved in the interactions between beneficial insects
and the landscape in agroecosystems. .......................................... 105

Figure 34. Stimuli (inputs) and response (outputs) of a landscape
simulation model involving 13 species of ladybird beetles. ............. 106

Figure 35. Spatially explicit population model in which temporal
abundance of predators is evaluated on a cell by cell basis based on
the habitat status in each cell for each time step (i). ...................... 108

Figure 36. Determination of levels of abundance for ladybird beetle
species based on groups of abundance means of C. septempunctata
after a pairwise comparison (Tukey HSD ). .................................... 111

xiii

Figure 37. Process of estimation of probabilities for levels of abundance
using C. septempunctata in corn as an example. ........................... 113

Figure 38. Modular organization of a model that simulates the response of
thirteen species of ladybird beetles to landscape configurations. 1 16

Figure 39. Comparison between model output and field data in relation to
total relative abundance, species richness, and species diversity of
ladybird beetles. ........................................................................... 119

Figure 40. Changes in abundance, species richness, and diversity of a
complex of ladybird beetle species predicted by a simulation model on
a corn-soybean-wheat rotation scheme in response to increasing from
one initial crop (corn) to two initial crops (corn and wheat). ............ 126

Figure 41. Changes in abundance, species richness, and diversity of a
complex of ladybird beetle species predicted by a simulation model on
a corn-soybean-wheat rotation scheme in response to reduction on
chemical inputs. ........................................................................... 127

Figure 42. Changes in abundance, species richness, and diversity of a
complex of ladybird beetle species predicted by a simulation model on
secondary succession habitats in response to changes from natural
succession to a managed succession (different succession ages, 1 to
3, and plowing after the four year) ................................................. 128

xiv

LIST OF TABLES

Table 1. Characteristics of tree-related aggregation sites of C. maculata lengi in
the Kellogg Biological Station Landscape 1995. ........................................ 22

Table 2. Number of trees with and without C. maculata lengi aggregations on
four selected rows of trees at the Kellogg Biological Station, Hickory
Corners MI. 1994. ..................................................................................... 23

Table 3. Yearly mean (iS.E.) trap captures of eleven species of coccinellids
sampled in alfalfa using yellow sticky traps from 1989-1995. Long Term
Ecological Research- Kellogg Biological Station, Hickory Corners, MI ....... 60

Table 4. Kendall’s coefficients for coccinellid species concordance (species = 9)
between consecutive years in poplar, alfalfa, and secondary succession. Long
Term Ecological Research- Kellogg Biological Station, Hickory Corners, MI.62

Table 5. Yearly mean (iS.E.) trap captures of eleven species of coccinellids
sampled in Populus using yellow sticky traps from 1989-1995. Long Term
Ecological Research- Kellogg Biological Station, Hickory Corners, MI ........ 65

Table 6. Yearly mean (iS.E.) trap captures of eleven species of coccinellids
sampled in early secondary succession using yellow sticky traps from 1989-
1995. Long Term Ecological Research- Kellogg Biological Station, Hickory
Corners, MI ............................................................................................... 66

Table 7. Yearly mean (iS.E.) trap captures of eleven species of coccinellids
sampled using yellow sticky traps in a six year corn soybean rotation under
high input conventional tillage Long Term Ecological Research- Kellogg
Biological Station, Hickory Corners, MI. .................................................... 75

Table 8. Kendall’s coefficients for coccinellid species concordance between
consecutive years in a corn-soybean rotation at the KBS LTER, Hickory
Corners, MI. .............................................................................................. 76

Table 9. Agronomic protocol for each one of the different field crops in the KBS-
LTER. 1993-1995. CT =conventional tillage, NT =no tillage, LCI =low
chemical input, and ZCI = zero chemical input. ......................................... 84

Table 10. Yearly mean trap captures (x 100) iS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in corn under four management
practices. Long Term Ecological Research- Kellogg Biological Station,
Hickory Corners, MI. 1993 ......................................................................... 87

Table 11. Kendall’s coefficients for coccinellid species concordance in corn under
different management practices at the KBS LTER, Hickory Corners, MI. 1993.

XV

CT =conventional tillage, NT =no tillage, LCI =low chemical input, and ZCI
= zero chemical input. ............................................................................... 88

Table 12. Yearly mean trap captures (x 100) iS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in soybeans under four
management practices. Long Term Ecological Research- Kellogg Biological
Station, Hickory Corners, MI. 1994. ........................................................... 89

Table 13. Kendall’s coefficients for coccinellid species concordance in soybean
under different management practices at the KBS LTER, Hickory Corners, MI.
1994. CT =conventional tillage, NT =no tillage, LCI =low chemical input,
and ZCI = zero chemical input ................................................................... 90

Table 14. Yearly mean trap captures (x 100) :tS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in wheat under four
management practices. Long Term Ecological Research- Kellogg Biological
Station, Hickory Corners, MI. 1995. ........................................................... 91

Table 15. Kendall’s coefficients for coccinellid species concordance in wheat
under different management practices at the KBS LTER, Hickory Corners, MI.
1995. CT =conventional tillage, NT =no tillage, LCI =low chemical input,
and ZCI = zero chemical input ................................................................... 92

Table 16. Mean and coefficient of variation (c.v.)(%) of the abundance, species
richness and species diversity of a complex of ladybird beetles described by
a simulation model in response to different landscape configuration
treatments. Number of crops and chemical input treatments were applied on a
rotation scheme of corn-soybean-wheat. .................................................. 125

xvi

INTRODUCTION

IMPORTANCE OF LONG TERM ECOLOGICAL RESEARCH

General perspective

Global, international, worldwide... are some of the words used to describe several of the
current human activities and their repercussions, particularly when referring to
environmental issues (di Castri and Hansen 1992). Our environment is experiencing
global change with regard to loss of habitats, earth warming, water and air pollution, soil
erosion and salinization (de la Court 1992, Mohrrnann 1992, Olson, 1992). Policies in
different parts of the world are addressing issues in conservation, restoration, and
ecosystem management to achieve a sustainable development. In this context, ecologists
are challenged to provide the necessary knowledge to ensure that outcome. Ecological
phenomena occur at different scales in time and space, and there is currently a need to

obtain ecological information that occur on long time scales (Callahan 1984).

Agriculture

Several environmental problems have originated from agricultural practices (Gilpin et a1.
1992, Poincelot 1990). For example, when in the 70’s the Green Revolution changed
agriculture into a high input activity, an era of prosperity was visualized for humans

(Gilpin et al. 1992). For some time, this expectation became reality. However, the long

term environmental impact and repercussions of this high input agriculture counter the
economic benefits. Intensive use of chemical products such as fertilizers and pesticides,
intensive irrigation, and increase of monoculture cropping, have altered the equilibrium
of nature (Altieri 1987, Edwards 1990, Claridge 1991, Pimentel et al. 1992). Ecological
management of agriculture, proposed as a counteraction for the excesses mentioned
above, is still in the process of development. Pest management for example, needs to
evolve toward agroecological management, particularly in relation to scales and strategies
(Levins 1986, Barret 1992, Pimbert 1991). Spatial scales should evolve from single farms
or small regions defined by one pest toward an agro-geographic regional perspective,
while temporal scales will traverse from single season to long-term steady state or
oscillatory dynamics (Levins 1986). Design of appropriate agroecosystems should be the
main strategy for pest management programs, thus minimizing the need for human
interventions (Levins 1986). The transition toward an ecological-managed agriculture
will cause, however, an increase in complexity that has yet to be fully addressed in

ecological theory.

THE KBS-LTER
To promote research on ecological phenomena that occur at large scales, the Long Term
Ecological Research (LTER) network was established as a NSF funded program aimed to
conduct and facilitate ecological research of ecological phenomena that occur over long
temporal and broad spatial scales (Franklin et al. 1990). The only program within the
LTER network which focused on agricultural ecology is located at the Kellogg Biological

Station (KBS). The KBS LTER, established in 1987, shares with the other programs a

commitment to conduct research in five core areas: a) pattern and control of primary
production, b) pattern and control of organic matter accumulation in surface layers and
sediments, c) patterns of inorganic inputs and movements of nutrients through soils,
groundwater and surface waters, (1) patterns and frequency of site disturbances, and e)
spatial and temporal distribution of populations selected to represent trophic structure

(Callahan 1984).

The general hypothesis of the KBS LTER is that “agronomic management based on
ecological concepts can effectively substitute for reliance on chemical subsidies in
production-level cropping systems” (Van Cleve and Martin 1991). Several disciplines

including entomology, are involved in ecological research toward this end.

COCCINELLIDS AS THE SUBJECT OF THIS STUDY
In 1988, Dr. Stuart H. Gage designed a long term program, within the KBS LTER
framework, aimed to monitor the flow of organisms in agricultural landscapes (Gage et
al. 1993). Among the species sampled is a complex of coccinellids which receive special
attention in the sampling program because:
a) they represent an important trophic structure in agroecosystems (predatory insects).
b) their diversity can be an indicator of the integrity of the ecosystem
c) they are easy to identify in the field
(1) they can be present in most habitats in agroecosystems

e) they can be monitored at larger scales.

Early publications showed that these expectations were being met (Maredia et al. 1992
a,b,c).

When I joined the LTER program in 1992, I decided to continue the focus on
coccinellids. They are part of the complex of natural enemies in several agroecosystems
because they prey principally on aphids and scale insects (Hodek 1970). As most
beneficial insects, coccinellids are susceptible to agricultural practices (V anderBosch
1982) an therefore they

can be used as ecological indicator to assess the integrity of agroecosystems. Ground
beetles have been commonly used as ecological indicators in agroecosystems (Desender
et al. 1994, Luff and Woiwod 1995), however they provide insight only at the ground
level scale. A more complete view of the system can be obtained if the above-ground
dimension is incorporated with the study of plant dwelling predators, such as

coccinellids.

GOALS, HYPOTHESIS, AND OBJECTIVES
Goals
My goals in conducting this research were personal and scientific.
My personal interest was to expand the temporal and spatial scope of my knowledge
since my entomological background was focused on working at smaller scales. Usually,
Ph.D. students are constrained (because of time) to conduct short term studies, and
therefore, the future researcher must learn later how to conduct research at larger scales.

Working at the KBS LTER was an excellent opportunity to fulfill my desire by gaining

expertise during my Ph.D. program in the implementation, analysis, and integration of a
long term study in an agroecological context.

The scientific goal of this research was to analyze patterns of seven years of habitat
utilization by a complex of coccinellids in all the different habitats and management
practices that occur in the KBS LTER main site and synthesize that information in a

landscape-coccinellids model.

Hypothesis
This work was conducted with the hypothesis that assemblages of coccinellid predators
with high mobility and a wide range of habitat utilization can be affected by changes in

the temporal and spatial diversity of the landscape.

Objectives

To pursue the scientific goal of my research four specific objectives were proposed:

a) gain insight into the life system of coccinellids using Coleomegilla maculata lengi
Timberlake as a case study

b) determine an adequate scale for the analysis of interaction patterns between
coccinellids and agricultural landscapes.

c) analyze the effect of temporal diversity and management practices in the assemblage
of coccinellids species

d) integrate results in a computer simulation model

METHODS

This thesis was organized into five chapters (Figure 1) and a detailed explanation of the
methodology used is provided within each chapter.

Chapter 1 describes a study of the life system of coccinellids in the landscape using Cm.
lengi case studies. The first case deals with the use of wooded habitats in agricultural
landscapes by overwintering adults, and the second case characterizes patterns of habitat
utilization in field crop agroecosystems during the Summer (May - August).

Chapter 2 is an analysis of the issue of scale to provide a framework for ecological
studies in agroecosystems. It includes the assessment of the use of sticky traps as a
sampling tool for long term studies of coccinellid patterns as well as the analysis of long
term patterns of coccinellids in different habitats in the landscape.

Chapter 3 is an analysis of temporal diversity in the assemblage of coccinellid species
which include two case studies: a) the effect of habitat maturation in alfalfa, poplar, and
secondary plant succession, and b) the effect of a corn-soybean rotation.

Chapter 4 is an assessment of the effect of agricultural management in the assemblage of
coccinellids, and Chapter 5 is the development of a model of interactions between the
landscape and coccinellids to assess the effect of spatial diversity on the assemblages of

coccinellids.

 

Methodological Approach

 

 

 

 

 

 

 

C. maculata Life ‘ C. maculata
475-7 Summer study I 7 . overwintering .
‘ 1992 System ' study 19944995 .

 

 

 

 

CHAPTER1 , ._

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

’ ‘ Sampling . 1 ' Sampling trends ‘- '
1 study - AnalySIs _ KBS LTER
I.» 19934994 1 0f Scale ,_ 1999-1995
A ._.,CHAPTER2 . ‘
cm." Temporal ~ ”3'“.
rotation . . I successmn
1989-1994 D'Ve'S'tY j 1989-1995
CHAPTER 3 . I

 

 

 

JZZZL

' .‘_,.Crsp.ManéQé'fiéhf

 

 

Tillage and chemical inputs
1993-1 995

',fCHAPrEn4"“°‘

JIZZL

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ . Conceptual MOCIGI I V Parameter'
model ‘ Deve|opment estimation
CHAPTERs. ‘

 

 

 

 

 

 

Figure 1 Phases in the development of the coccinellid-landscape
model.

CHAPTER 1: LIFE SYSTEM OF PREDATORY COCCINELLIDAE
(COLEOPTERA) ON AGROECOSYSTEMS: CASE STUDIES ON

Coleomegilla maculata lengi Timberlake

Coleomegilla maculata lengi Timberlake, is a native species of America distributed east
of the Rocky Mountains (Gordon 1985). In Michigan, it is one of the most abundant

native species of coccinellids in agricultural crops (Maredia et al. 1992a).

BIOLOGICAL CYCLE
C. m. lengi, as well as all other species of coccinellids, is a holometabolous insect
with larvae as the active immature stage which have similar habits as the adults.
Adults, as predators, spend most of their active time searching for food and sites
for reproduction, and if a habitat is suitable, oviposition will occur. The subsequent
larval stage will search for food until the completion of its development (Ferran
and Dixon 1993), and after a pupal stage, a new adult emerges (Hodek 1973, Frazer

1987) (Figure 2).

SEASONAL HABITAT UTILIZATION
The life system of C.m. lengi is mostly defined by the interaction of the adult with
the landscape since the dynamics of the immature stages are constrained by the

habitat selected by the adult for reproduction.

 
 
  
     

   

—' m
Ada" \ ..Feeding

Larva
-‘
P” m

  
  

Dispersion

Reproduction

   

Figure 2. Biological cycle of predatory coccinellids and the role of each
biological stage.

From the adult perspective, the life system consists of a cycle that involves the
presence of aggregation sites during the winter and sites for feeding and
reproduction during the summer, with temporal patterns of movement and
dispersion among the different habitats (Figure 3). From the classification of
movement by Wratten and Thomas (1990), three types of movement can be used to
characterize Cm. lengi. First is the colonization of crops from aggregation sites in
the spring, and returning to them during the fall (Hodek et al. 1993), a phenomena
regulated primarily by photoperiod and temperature (Hagen 1962). Second is the
movement among habitats (mainly crops) during the growing season, in which

searching for food is normally the driving factor (Frazer 1987). C.m. lengi, for

lO

instance, has a difficult time searching for prey in polyculture corn (Andow 1991).
In polyculture situation, C. m. lengi adults move rapidly and remain in this habitat
less time than in monoculture corn (Wetzler and Risch 1984). The third type of
movement is the aggregative numerical response to habitats when high prey density

occur (Wright and Laing 1980, Evans and Youssef 1992).

Environmental Components

 

Photoperiod
Temperature
Wind

       
 
   

 
     

   

 

 
   
   
 

 

  
 
 

   

Movement Cultivated
' habitats
Aggregation
A99regatlon .
sites Landscape components Feeding
Reproduction
Overwintering
M t .
03:27:12 Uncultivated -
habitats

   

Figure 3. Life system of predatory coccinellids in agricultural landscapes.

The following is a description of two case studies designed to gain more

understanding of C. m. lengi habitat utilization in agricultural landscapes.

11

CASE STUDY 1. Toward the Characterization of Landscape Use by Overwintering
Aggregates of Coleomegilla maculata lengi Timberlake (Coleoptera: Coccinellidae)

in Field Crop Ecosystems in Relation to Wooded Habitats.

Introduction

C.m. lengi adults aggregate during the winter in or adjacent to wooded areas either
far from or near to crop fields (Roach and Thomas 1991). Specific sites for
overwintering include litter, stones, fallen trees, or grass clumps (Hodek 1973).
Early in the spring, adults become active and start feeding on nearby flowers and
begin their dispersal to agricultural habitats (Benton and Crump 1981). During the
fall adults start to aggregate again in overwintering sites (Benton and Crump 1979).
Discovery of overwintering aggregates of C.m. lengi in woodlots of an MSU
experimental area in 1994, provided an opportunity to observe their dynamics prior
to dispersion and to better understand the landscape features used by these

organisms to select aggregation sites.

General methodology.

Study area. I conducted this work at the MSU research farms (East Lansing, MI)
in spring of 1994 and later at the MSU Kellogg Biological Station (Hickory
Corners, MI) during spring of 1995.

Spring 1994 Aggregation sites at the MSU experimental station were located in two

major woodlots: Elsesser and Box. The first site was surrounded by alfalfa and wheat

12

fields, and the second was 10m adjacent to a corn field. In the north-western corner of the
Box woodlot there was a small woodlot, Box 11, in which I also observed aggregates of
beetles. Each day, woodlots were observed to document when beetles leave aggregation
sites. Spring flowers which were frequented by the adults for pollen feeding were
collected for identification.

When dandelion flowers (T araxacum oflicinale) were abundant in the landscape and
adult beetles were feeding on them (early May), I measured the number of adults per
flower to estimate relative abundance. This sampling method was used because: a) there
were enough flowers at the edges of the woodlots to assess the preference of C. m. lengi
adults for specific locations, and b) the color of the beetles contrasted with the yellow
flowers which make easy to count them.

To conduct this sampling, I established a transect parallel to the southern side of each
woodlot. The length of the transects was 105m in the Elsesser and 80m in the Box 11. I
did not sample the Box woodlot because most dandelion plants were destroyed during the
plowing of the field adjacent to the woodlot. I sampled once in each woodlot recording
the numbers of beetles on flowers and the number of flowers every 1m2 quadrat along
transects. Beetles were “focused” on their activity in the dandelion flowers and slow
movement of the observer seldom disturbed them.

After sampling MSU woodlots and assessing the association of C. m. lengi and dandelion
flowers when beetles were near aggregation sites, I went to KBS to find aggregates of
beetles in woodlots. Searching dandelion flowers concentrations I found a beetle
aggregation in a woodlot located in the north-western comer of the LTER main site.

There, I sampled beetles using the same protocol as above.

 

During that spring I also measured the spatial distribution of C. m. lengi in the
surrounding area in a woodlot that had an aggregation site. I selected the Elsesser
woodlot for this set of observations because it was surrounded by alfalfa which facilitated
the use of the sweep net sampling method. I used a sweep net (25 sweeps per sample) in
transects parallel to the edges of the woodlot. I made a total of 85 samples in a ratio of
100 m2 to the woodlot using wooden stakes as visual cues to determine the location of
sampling points. The interval between sampling points was 25-30 m. I assigned each
sample location to a (X,Y) coordinate. Using the coordinate and the numbers of beetles
captured at each sampling location, I produced a surface map of the spatial distribution of
C. m. lengi using the Kriging method provided in the program SURFER 4 (Golden

Software 1990).

Spring 1995. With the expectation that I could use C. m. lengi adults on dandelion
flowers as indicators of the presence of aggregation sites near woodlots, I designed a
sampling method to determine the distribution of beetle aggregation patches in
agricultural landscapes. I selected KBS for this study because the abundance of tree
related habitats interspersed with agricultural fields allowed me to see patterns of habitat
discrimination by C.m. lengi.

To conduct the sampling, I divided the KBS landscape in 1 km2 quadrats using an aerial
photo and physical features of the landscape as references. I systematically visited
dandelion patches along woodlot edges and those that occurred at the base of individual
trees present in each quadrat. In some parts of KBS the dominant flower was garlic

mustard (Alliaria ofi‘icinalis), and after observing that beetles in aggregate sites used this

l4

flower as source of food, I included the beetle-garlic mustard association in addition to
dandelion as an indicator of potential aggregates. Once I found associations of beetles on
flowers, I searched the area more thoroughly to find evidence of an aggregation site
(beetles under leaves, branches, or in crevices at the base of trees, or adults crawling in
the vegetation near the selected location in the case of woodlots). When I confirmed the
existence of an aggregate, I recorded the location and the type of crop (if any) present in
the vicinity during the previous season. Later in the summer I identified the main species
of trees present in aggregation sites. Due to the limited time available for sampling, (1
could sample only when dandelion was flowering and when days were warm so flowers
were open) no attempt was made to estimate abundance of beetles in the aggregation sites
to enable me to focus on searching for beetle aggregates in the landscape. The
beetle/flower ratio method I used during 1994 was useful only as an indicator of the
presence of aggregation sites. I could not use it to estimate relative abundance during the
season because: a) there were two kind of flowers (dandelion and garlic mustard) with
different plant architecture and b) the number of dandelion flowers at the base of
individual trees was very low compared to woodlot areas.

To characterize the use of tree-related habitats in the landscape by overwintering beetles,
I used the following classification: a) row, b) cluster, c) hedgerow, and d) woodlot

(Figure 4).

 

 

Cluster

 

Hedgerow

 

Figure 4. Classification of tree-related habitats for C.m. lengi
aggregates in the landscape.
Rows were normally associated with field margins and roads, and hedgerows were field
boundaries. Clusters had the same constitution as hedgerows but they were isolated

patches (2-3 trees).

Results

Spring 1994.

The conditions of the two major woodlots at the MSU experimental station,
Elsesser and Box were different. In Elsesser (Figure 5a) there was a variety of
spring flowers on which beetle adults fed: wood anemonae Anemone quinquefolia,

spring beauty Claytonia virginica, and dandelion Taraxacum oflicinale (Figure 5,

c). This woodlot was surrounded by alfalfa, and adult beetles were hidden under

litter.

 

Figure 5. a) South side of the Elsesser woodlot, MSU campus, East Lansing, MI. b)
Coleomegilla maculata lengi adults feeding on “spring beauty” in the Elsesser
woodlot, c) Cm. lengi feeding on dandelion flowers at the edges of the Box
woodlot, MSU campus, and d) Cm. lengi adults hidden under leaves from the
previous corn crop at the edges of the Box woodlot.

In the Box wooodlot, the only spring flower present was the dandelion. The field

across this woodlot was a field in which corn was planted the previous year. Along

17

the south edge of this woodlot there were leaves from the previous corn crop and
adult beetles were hidden under them (Figure 5, d).

I observed that Cm. lengi displayed major activity when dandelions flowered in the
landscape. Despite the fact that dandelions were almost everywhere in the
landscape, beetle adults concentrated on those flowers that were near aggregation
sites. This behavior, mentioned by Benton and Crump (1981) was confirmed after
measuring the beetles-flowers ratio. In the Box 11 woodlot, adults concentrated on
dandelion flowers nearby the hibernation spots with a gradient of diffusion

thereafter (See the peaks in Figure 6).

 

f
U

BeetlesIFlowere Ratio

 

Flowers/m2
.1.

 

T
.3

 

 

 

 

 

Distance (meters)

 

[ . Flowers — BeetleIFlowefl

 

Figure 6. Relative abundance of C. maculata lengi (measured as
adults per flower) and dandelion flowers in one m2 quadrats
along a transect on the south edge of the Box 11 woodlot.
MSU, East Lansing MI. 1994.

18

When the woodlot ended (after 60 m), there was an open field with plenty of
dandelion flowers, but there were almost no adults found on dandelion. In the
Elsesser woodlot a major aggregation site was found at the southwestern corner. In
addition, several aggregations were observed along the woodlot edge as confirmed

by the distribution of beetles on dandelion flowers (Figure 7).

70 1 .4

 

Flowers/m2

 

 

Beetles/Flowers Ratio

 

 

 

 

 

 

Distance (meters)

 

[ . Flowers —BeetlelFlower]

Figure 7. Relative abundance of C. maculata lengi (measured as
adults per flower) and dandelion flowers in one m2 quadrats
along a transect on the south edge of the Elsesser woodlot,
MSU, East Lansing MI. 1994.
In the KBS LTER woodlot the distribution of adults on dandelion flowers
coincided with the presence of beetle aggregates in the woodlot (peaks in the Figure
8). An interesting observation related to this woodlot was that corn was planted in

the fields near the woodlot (corn is an important habitat for Cm. lengi) the

previous season and the number of overwintering beetles in 1994 was high. From

19

50 to 80 m there were 23 quadrats with one to 60 adults and the average beetle-
flower ratio was 0.21 i 0.04(Figure 4). During the summer of 1994 soybeans were
the crop in the landscape next to the woodlot. The following spring (1995) only
three quadrats contained beetles and the average was 0.02 i 0.01 (approximately

only a 10% of that observed the previous year).

 

 

 

 

 

 

 

 

so 1
I
I
45,” ____________ - ................... Tang
8
E .- 2
£30,--- ................ .-_. _________ «— - 45%
, _
o " _ k
i: - 8
33
154———|————-— ————— -—-—- ——————————— I ~0.25§
I I I
o- ' 4' , 0
o 15 30

 

Distance (meters)

 

I - Flowers —BeetleIFlower]

 

Figure 8. Relative abundance of C. maculata lengi (measured as
adults per flower) and dandelion flowers in one In2 quadrats
along a transect on the south edge of the KBS woodlot, ,
Hickory Corners, MI. 1994.
The spatial distribution map around the Elsesser woodlot showed three areas of
concentration of Cm. lengi at the southwest and southeast corners of the woodlot

(Figure 9). Besides those concentrations of beetles next to aggregation sites, the

other concentration was observed near wheat. Although I can assume that beetles in

20

wheat may have come from the aggregation in the woodlots, the lack of previous

sampling made the assumption difficult. Three days later I could not find any adults

in the aggregation sites.

N
* wheat

alfalfa

 
   

Figure 9. Spatial distribution of C. maculata lengi
overwintering adults on a 100 m
perimeters around the Elsesser woodlot.
MSU, East Lansing MI. 1994.

Spring 1995.

Aggregation sites of C.m. lengi were very dispersed across the KBS landscape
(Figure 10), and results showed that any tree-related habitat in the landscape can be
a potential aggregation site if it has a southern exposed to sun light, is located near

agricultural fields, and is not subject to human disturbance (Table 1).

21

 

 

 

Figure 10. Distribution of aggregation sites (black squares) of C. maculata lengi in
the Kellogg Biological Station landscape. Large polygon is the actual
area observed. S= single tree, H= hedgerow, C = cluster of trees, R =
row of trees, and W = woodlot.

The observation that I did not find aggregation sites near trees on private property

which have a clear base (no weeds nor litter) indicated that human disturbance was

an important factor. However for obvious reasons, I could only sample a small

22

portion of those trees (mostly trees close to the road) so further confirmation of this
is necessary. I also observed that rows of trees (which are normally close to a road,
were subject to removal of weeds and debris. In early spring weeds were primarily
composed of dandelions which provide early food for the beetles. Therefore, human
management of the landscape played an important role in the dynamics of C. m.

lengi at this time of the season.

Table 1. Characteristics of tree-related aggregation sites of C. maculata lengi in the
Kellogg Biological Station Landscape 1995.

 

 

Type Number Orientation Dominant tree Neighboring
(Genus) crops
Row 5(8) 4(4) EW Carya, Querqus, and Mostly corn,
1(3) NS Prunus alfalfa and
soybean
Cluster 4(4) South side Rhus, Moms, and corn, soybean
Prunus
Hedgerow 4(5) 3(3) EW Prunus, Moms, and corn, alfalfa,
1(2) NS Juglans soybean
Woodlot 0(3) conifer Picea, Pinus
2(4) deciduous South side Prunus, Moms, Juglans, corn, soybean

UImus, and Robinia

 

Numbers inside parenthesis represent the number of similar elements present in the observed area.

Although no determination of relative abundance was made, my visual assessment
indicated that woodlots and clusters of trees (such as hedgerows) have more beetles

per aggregation than individual trees in a row.

23

Four important concentrations of adult beetles associated with individual trees were
found in what I called rows (Figure 4). Of all the trees in one row, approximately fifty
percent of them were used as aggregation sites by C.m. lengi (Table 2). No special
preference for any tree species was observed. The dominant species seemed to provide

more aggregate sites.

Table 2. Number of trees with and without C. maculata lengi aggregations on four
selected rows of trees at the Kellogg Biological Station, Hickory
Comers MI. 1994.

 

 

Sampling Measurement Tree Genus
site Quercus Carya Prunus

Site 1 All trees 14 28 11
Trees with beetles 7 16 5

Site 2 All trees l3 l3 8
Trees with beetles 3 4 4

Site 3 All trees 2 4
Trees with beetles 1 2

Site 4 All trees 3 4 1
Trees with beetles 3 3 0

Discussion

Tree-related habitats are an important place for C.m. lengi aggregates in the
agricultural landscape. Hedgerows and woodlots seemed to harbor the major

aggregates if neighboring fields were corn or alfalfa rather than soybeans. The

24

smallest aggregates were in areas where soybean was the crop the summer before.
An interesting study to conduct in the future would be to investigate the life system
of Cm. lengi in the landscape under conditions of a corn-soybean rotation (a
common practice in Michigan) near aggregation sites. I recommend sampling
beetles during November to March before the activity of beetles starts, because at
that time flowers or grasses are not grown and beetles are not mobile.

The use of spring flowers that are very abundant in the landscape was useful to
localize tree-related aggregates of C. m. lengi in agricultural landscapes. This
method takes advantage of the beetle’s behavior to feed on those flowers while it
remains close to aggregates during the early spring. Although this method has the
disadvantage of localizing aggregation sites after they have been used for
overwintering, the habit of coccinellids to use the same location for overwintering
for several years (Benton and Crump 1979) may overcome the problem.

This study also allowed me to see the importance of understanding the life system
of beneficial insects in the landscape and the impact that agricultural management
may have on them. In several tree-related habitats near agricultural fields or roads,
dandelions were mowed at a time when beetles were feeding on them, a practice
which could cause adult mortality and can also reduce primary food sources for
adults before dispersion. The objective of eliminate dandelions at field edges is to
reduce the population of these flowers that become weeds within agricultural fields.
However, as I showed in this work, only those flowers near aggregation sites are
used for the beetles. Preserving localized patches of flowers associated with

aggregation sites may enhance the population of beneficial coccinellids.

25

CASE STUDY II. Habitat Use Patterns of Coleomegilla maculata lengi Timberlake

During Summer in Field Crop Agroecosystems.

Forage legumes, small grains, and corn constitute very important habitats for
several species of coccinellids including C.m. lengi (Hodek 1973, Honek 1982,
Kieckhefer and Elliot 1990, Maredia et al. 1992b). Cm. lengi use aphids as a main
source of food (Gordon 1985), and can also feed on eggs and larvae of many
coleopterans and lepidopterans (Conrad 1959, Warren and Tadic 1967, Groden et
al. 1990). Adults also feed on pollen from several sources including com (Hodek
1973). In field crop agroecosystems, winter wheat (planted during previous fall)
and alfalfa (a perennial crop) are habitats available early in the growing season.
Corn and soybean (planted during the spring) become available later on. When
several of those habitat are present in the landscape, patterns of habitat utilization
by a polyphagous species like C.m. lengi may become very complex. Continuous
sampling in different habitats during the growing season can provide information
on the number of individuals of this species at different times. By comparing
population densities in different habitats we can infer a pattern of habitat utilization
by C. m. lengi. This inference could be strengthened if we were able to analyze the
feeding history of C. m. lengi at different times of the season. An approximation to
this approach can be done with the use of stable isotope analysis. The method of
stable isotopes has been used for food web studies (Wada et al. 1991) and it is based on
the observation that the carbon and nitrogen isotope composition (513C and 515N,

respectively) of organic material of a consumer is similar to or deviates by a consistent

26

amount to its diets (Boutton et al. 1983, Ostrom and Fry 1993). The nitrogen isotope
value of an organisms deviates by approximately 3%o from its diet. Consequently, it may
be possible to determine the trophic position of any organism within a food chain (Wada
et al. 1991). It is also possible to determine the relative importance of food sources that
have unique isotope ratios (Fry and Sherr 1984, Harrigan et al.1989). The relative

contribution of different food sources to the consumer is given by:

n
aconsumer = 2 F26;
i

where Fi and Bi are respectively the fractional contribution and isotopic composition of C
or N from an individual source, 1' (Harrigan et al. 1989). Stable isotope analysis thus
enables researchers to determine pathways of energy flow through food webs. This
information is essential if researchers want to know not only where organisms were
present (determined by counting methods) but also which food source contributed to their
diet.

This work was designed to determine habitat use patterns of C.m. lengi in field
crop agroecosystems, by combining analysis from temporal and spatial distribution
of C.m. lengi adults with the analysis of pathways of energy flow in which this

species is involved.

Methodology
Study area. This work was conducted at the Kellogg Biological Station during the
Summer of 1992. The experimental design, consisted of four plots each containing a

different field crop: alfalfa, corn, rye, and wheat is shown in the Figure 11.

27

Sampling method. Six double-sided, yellow cardboard sticky traps (22.5 x 14.0 cm)
suspended 1.2 m above the ground from a metal rod (as described by Maredia et al.

1992b) were placed in each crop (Figure 11).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o e e e
c: O 0 O 0
e o
8 200m 3}
e e e e 9’.
Q
o e e e
e e e e
V , .
Corn Rye Wheat Alfalfa
Corn ' I lAlfalfa

Figure 11. Experimental design to study habitat
utilization patterns by C.m lengi at the Kellogg
Biological Station, MI.

Yellow cardboards were replaced every second week. Each trap was placed at a
permanent geopositioned location, and the experimental area was gridded using X,Y
coordinates measured in meters where the initial point (x=0, y=0) was located at the
south-westem corner.

Sampling was conducted from May to September. Cm. lengi adults on traps were
counted, recorded, detached from traps and frozen. Along with the beetles, plants were

sampled in the different habitats at each of the trap locations and samples were frozen.

28

Using the XY coordinates from each trap surface maps were calculated to obtain the
spatial distribution of C. m. lengi based on sticky trap captures using the Kriging method
provided in Surfer 4 (Golden Software 1990).

The isotopic composition of Cm. lengi was determined at the Environmental
Geochemestry Laboratory in the Department of Geological Sciences at MSU. Beetle
adults and plant samples were dried at 40°C, ground to homogeneous fine powder, and
stored in a freezer. A 20 mg aliquot of plant tissue or a whole insect was combusted in a
sealed quartz tubes in the presence of excess copper and copper oxide (Macko et al.
1987). Carbon and nitrogen gases were obtained by cryogenic gas separation. The
isotopic determination was performed using a PRISM stable isotope ratio mass

spectrometer (VG ISOTECH).

Results and Discussion

The temporal dynamics of C. m. lengi showed that alfalfa, wheat, and corn were used
sequentially (Figure 12). The peak catch of this ladybird beetle species in wheat was
observed the week after the alfalfa was cut, and the peak catch in corn coincided with the

beginning of pollination in this crop.

29

 

25 -
tn
g 20
a ' Alfalfa
N
L) L5-
0.
E
C
(U
0 0.5
2
OJ
120 140 160 180 200 220 240 260 230
25
8
'5 2° Wheat
0:!
%
L5<
O
Q
E 1.0 -
'—
C
‘U 05
o:
2
00
120 140 100 180 200 220 240 260 280
3£

"5 Corn

Mean Trap Captures

 

120 140 160 180 200 220 240 260 230

Julian Days

Figure 12. Mean trap captures (3: SE.) of C. maculata lengi in three
habitats, as measured by sticky traps, at the Kellogg
Biological Station, MI. 1992.

30

The spatial distribution maps of beetle catch are shown for May and August (Figure 13).
These show that the abundance of C. m. lengi is highest in alfalfa and com during May

and August respectively.

In May (Figure 13) I observed a trend of Cm. lengi dispersal coming from the northern
part of the fields. At the north edge of the fields the row of trees (Figure 11) could have
been the source of overwintering beetles that dispersed into the alfalfa in May. In August,
beetles in corn were observed to be dispersing from western and southern edges (Figure
13) which were adjacent to other com fields (Figure 11).

The results of the stable isotopes analysis corresponded with the field observations. C. m.
lengi beetles captured in alfalfa during May revealed a 68% contribution of a com-based
diet and 32% alfalfa (Ostrom et al. in press). The conclusion was that beetles captured in
May were overwintering adults which fed on a com-based diet prior to dispersion to
overwintering sites. This situation was very likely because com is the preferred habitat for

C. m. lengi at the end of the season (Figure 12).

For C. m. lengi adults captured in corn during August, the isotopic analysis showed a
balanced contribution from alfalfa (52%) and com (42%), even though large numbers
were captured in wheat after the alfalfa was cut. No apparent explanation was found for
this situation, and it helps to illustrate the complexity of the habitat-insect interactions

that are occurring in the field.

31

 

corn rye wheat alfalfa corn rye wheat alfalfa

 

Figure 13. Surface maps of the spatial distribution of C. maculata lengi as measured
by sticky traps in four field crops during two times of the season: May and
August. 1992.

This research was able to characterize some elements of the life cycle of Cm. lengi by
including a field crop habitats (alfalfa, wheat, and corn) used by the beetles for feeding
and reproduction as well as tree-related habitats potentially used by the beetles to make
aggregations during the winter. Having field crop habitats near overwintering sites
minimized the probability of “isotopic contamination” during the early season, which
could have occur if beetles coming from distant areas had already fed in habitats outside
the experimental design.

This work showed the importance of stable isotopes to record the feeding history of
predatory beetles in field crop ecosystems, and thus to delineate pathways of material

transfer between agricultural crops and insects.

CHAPTER 2: THE ISSUE OF SCALE IN STUDIES OF PREDATORY
COCCINELLIDAE (COLEOPT ERA) IN AGRICULTURAL LANDSCAPES
Since I was involved in a long term monitoring of a complex of organisms I consider it
necessary to address the issue of scale and its implications on my work. In any ecological
study there must be a correspondence between space and time scales. Either extreme -
long term study in a small spatial scale, or short term study in a large spatial scale- can
produce misleading results (Wiens 1989). A detailed discussion on the subject of scale
can be found in Levin (1992) and Allen and Hoekstra (1992). In this chapter I will focus
on the main considerations of scale I had to weigh with respect to my work: a)
establishing a framework of scale, b) measurement of coccinellid dynamics at larger

scales, and c) establishing an analytical approach for field measurements of coccinellids.

FRAMEWORK OF SCALE

Importance

Allen and Hoekstra (1992), discussing differences among ecological levels of
organization, defined scale as “the grain and extent that are required to see the entities
that characterize the level (of organization)”. Graig can be defined as “the size of the
individual units of observation, the quadrats of a field ecologist or the sample unit of a
statistician”, while _e_xt_er;t is “the overall area encompassed by a study” (Wiens
1989).Determination of the scale for an ecological phenomena can be highly influenced
by the subjectivity of the observer (Allen and Hoekstra 1992). Therefore it is important to

define a framework of scale that a researcher can use as a reference to assess the scope of

32

33

an ecological study and as a means to increase the predictability of ecological

phenomena.

Coccinellids and scales of biodiversity

One aspect that readily emerges from the literature with regard to coccinellids is the issue
of diversity of habitats and food sources that these organisms can use in agricultural
systems (Hodek 1973). Biodiversity is a rather complex concept describing different
features that vary in time and space (Altieri 1994). In referring to arthropods, biodiversity
research has focused on vegetational diversity and the effect it has on arthropod
populations (Sheehan 1986, Andow 1991). Literature dealing with biodiversity
concerning arthropods, and in particular coccinellids, is vast and can become too complex
unless an attempt is made to classify the different expressions of diversity. These include
plant architecture (Kareiva and Perry 1989, Grevstad and Klepetka 1992), plant
phenology (F erran et al. 1989, 1991), intercropping (Andow and Rish 1985, Coderre et al.
1989, Trujillo-Arriaga and Altieri 1990), cover crops (Bugg and Dutcher 1993), and
multiple habitats (Honek 1985a), Galecka 1992, Maredia et al. 1992b). Crop rotation
(Altieri 1994) and plant succession (Price 1991) are also expressions of diversity that may
affect coccinellids.

Since there was no hierarchical classification of diversity in the literature, I developed a
classification system by grouping the most common expressions of diversity in
agroecosystems based on their spatial (individual plant, field, landscape) and temporal

(daily, seasonal, two or more years) dimensions (Figure 14).

34

In this diagram I made a correspondence between temporal and spatial scales (i.e plant-
daily, field-season, landscape-2 or more years). The importance of considering both type
of scales in long term studies has been previously mentioned and is exemplified by the
results obtained by Elliot and Kiechhefer (1990). They sampled a complex of coccinellid
species during 17 years in small grains. After that time, they were able to explain
temporal population trends within single seasons, but not across growing seasons. They
concluded that it is important to consider other habitats in the surrounding landscape to

have a better understanding of long term patterns of coccinellid dynamics.

Diversity in Agroecosystems

Spatial Temporal

   

Plant
succession 2 or more

Landscape

a 8

Crop ye r

rotation
Field Season
Individual .
plant Dally
Dimension Expressions Dimension

Figure 14. Expressions and dimensions of diversity at different scales on time
and space in agroecosystems.

35

LONG TERM MEASUREMENTS OF ARTHROPOD POPULATIONS

Hierarchy theory and the concept of sampling

Hierarchy theory has become an important paradigm to address the issues of scale in
ecology (Urban et al. 1987). It recognizes the existence of hierarchical scales of
organizations (systems) which are constrained in their behavior by intrinsic limitations
(the potential manifestation of its elements) and by the activity of higher levels in the
hierarchy (Allen and Starr 1982). Interactions among different systems are possible due to
the presence of “signals” that transmit back and forth between different levels of
organization in the hierarchy (Allen and Starr 1984). The lower the level, the faster and
more intense is the transmission of signals; the higher the level, the slower the speed of
the transmission. Long term changes in climate, for example, may have an impact on the
population dynamics of specific herbivores in agricultural crops. However, such an
impact may be in the form of gradual change taking several decades to occur. On the
other hand, a pest outbreak in the herbivore subsystems will cause an immediate response
from the management subsystem (the farmer), which in turn will affect the herbivore
subsystem (with a pesticide application). Such response will also have an impact on the
beneficial organisms subsystem (increasing mortality), all of this occurring in the time
span of a few days. The objective of sampling in this hierarchical fiamework is to
intercept the signals among subsystems or levels of organization of interest at the

appropriate scale so that through an analysis we can interpret the system’s function.

36

Characteristics of long term sampling

Sampling is one of the crucial activities in any ecological study. Southwood (1978)

summarized the main principles and techniques of this issue. Long term sampling,

however, requires special considerations that have not been emphasized properly in the
literature, Most of such studies are descriptive by nature since the goal is to discover
temporal and spatial patterns that short term scale studies are unable to address (NSF

1978). Since it is very difficult in these kind of studies to set the appropriate scale of

measurements a priori, researchers tend to increase, as much as they can, the extent of the

information obtained, the frequency of the observations, and the duration of the sampling

(Allen and Starr 1982). This way, researchers try to minimize the possibilities of missing

important information during the sampling process. From studies of this nature with

arthropods, the following considerations can be distinguished for a long term sampling:

a) The use of standard methodology that allow the comparison of information obtained
at different sampling points in time and space (NSF 1978, Gage and Russell 1987).

b) Establishment of fixed sampling locations. Randomized sampling in long term studies
mixes temporal and spatial variances. By spatially fixing sampling points, we can
analyze separately temporal and spatial changes among sampling points (Taylor 1986,
Gage et al. 1990).

c) Use of a trapping method with low maintenance, that can be placed in different
habitats, and that is constantly sampling organisms. Examples of such devices for
long term studies are: light traps (Taylor 1986, Nowinszky and Ekk 1988, Szentkiralyi

1992, Kadar and Loevei 1992), suction traps (Taylor 1986, Quinn et al. 1991),

37

pheromone traps (Nasr et a1. 1984, Gage ct al. 1990, Srivastava et al. 1990), pitfall
traps (Burel 1989), and malaise traps (NSF 1978).
d) Regular time estimates to be able to detect patterns of temporal changes (Gage and

Russell 1987).

Sticky traps as sampling tools for coccinellids

Different techniques have been used to sample coccinellids in agricultural fields: the use
of a sweep-net, "D-VAC" suction sampling, collection via light and sticky traps, and
counting predators observed within a certain time period (Lapchin et al. 1987). Several of
these techniques have been the subject of studies to evaluate their efficiency in measuring
coccinellid population densities in wheat (Lapchin et al. 1987, Iperti et al. 1988, Elliot et
al. 1991), lentil (Schotzko and O'Keeffe 1989), and strawberry (Frazer and Raworth
1985)

Sticky traps have been used for sampling coccinellids in several studies (Ewert and
Chiang 1966; Mack and Smilowitz 1979, Ives 1981, Dean 1982, Neuenschwander 1982,
Ricci 1986, Sirota 1990, Maredia et al. 1992b). The color of the traps was usually white
or yellow as these colors have been found to be attractive for coccinellids (Dowell and
Cherry 1981, Benton and Crump 1981, Maredia et al. 1992a). However, it was not until
the KBS-LTER project (Gage et al. 1993), that sticky traps were used intensively and
extensively as the main sampling tool in a long term study (Maredia et al. 1992b,c). The
reason Gage began using this sampling method was that sticky traps were suitably

installed at permanent locations in a variety of habitats and were attractive to coccinellids.

38

This system was easily maintained and allowed for continuous sampling process during

the entire growing season.

Comparison of two sampling methods: sweep-net sampling and sticky trap in wheat

at different time intervals.

Trap devices successfillly used for long term sampling have had an attractive component
(pheromone or light traps) or have been simple passive (pitfall or suction) traps
(Southwood 1978). Sticky traps are a combination of both type of traps because they are
stationary in the landscape intercepting flying insects and also they have an attractive
component (the color). Sticky traps depend on insect activity, which complicates the
interpretation of the meaning of trap captures. In particular I was interested in
determining to what extent I would be able to use sticky traps catches as a measure of
relative abundance of coccinellids. To that end I designed a two year study to conduct a
comparison between using sticky trap catch and catch using sweep-net sampling. Wheat
was selected because this crop is highly preferred by coccinellids (Honek 1985a), the
number of coccinellids captured using sticky traps is high (Maredia 1992c) and it was
reasonable to expect to have significant numbers of coccinellids via sweep-net sampling.
I used the sweep-net as the method of comparison because it has been evaluated and used
as a measure of relative abundance of coccinellids in wheat (Elliot and Kieckhefer 1990,

Elliot et al. 1991).

39

Methodology.

Study area. I conducted this study during 1993 -1994, using three and four wheat fields
respectively each year. I used spring wheat in 1993, and winter wheat in 1994.

Sampling. In each field I placed six double-sided, yellow cardboard sticky trap (22.5 x
14.0 cm) suspended 1.2 m above the ground from a metal pole (as described by
Maredia et al. 1992a). These traps were designated upper traps. A second yellow
cardboard was attached to the pole at the height of the crop canopy and moved
weekly as the crop grew. Sampling was started in the first week of May and was
continued until the crop was harvested. Coccinellid adults caught on traps were
counted, recorded, and removed every week Traps were replaced every second
week.

For sweep sampling a sweep-net of 37.5cm of diameter was used. Sampling was
conducted following an imaginary diagonal path in each field. In 1993 I made 25
sweeps per sampling period while in 1994 I made two samples of 50 sweeps per
sampling period in each field. Insect captures in both years were converted to
number of adults per sweep.

Nomalitvjssessmentjtnd Taylor’s powerfilaw. I assessed the normality of the data
from the two sampling methods using the Sapiro-Wilks test (SPSS 1993). Then I used
Taylor’s power law to assess the need for a transformation to normalize the data. I first
divided the whole sampling period in each year into intervals of 100 degree days
(50F). Second, I estimated mean and variance for each 100 degree day interval in

each year per each field. Third I used the following equation:

40

log (variance) = log a + b log (mean)
in which log a and b were respectively the Y intercept and the slope in the general
equation for a straight line Y=a + bX.
Dara were transformed by solving the equation:

2 = x”

where z is the transformed value, x is the original (raw) number, and p is equal to
1-b/2 (Southwood 1978).
Due to the low numbers of coccinellids obtained using the sweep sampling method in
1993, I pooled all the coccinellid species together to work with coccinellids as a complex.
Correlation of sampling methods. I conducted a test for correlation between the different
sampling methods using the Pearson product-moment correlation (SPSS 1993).
Correlation was performed at three different time scales: 100, 400, and 800 degree day
interval. Fields from the two years were used together as replicates for correlation

analysis.

Results.

Insect dynamics. Sticky traps at both levels had similar temporal patterns during both
years with more adult beetles captured on the canopy traps than on the 1.2m traps (Figure
15 and Figure 16). In 1993, peak catches on sticky traps occurred after the peak observed
using sweep-net captures. A similar lag in response was observed in 1994 when peak
abundance of beetles caught using sticky traps occurred after a peak observed using
sweep-net sampling around 400 degree days. However, in this year the highest number of

coccinellid caught using sweep-net sampling occurred toward the maturation of the crop.

41

This increase in the population of beetles was detected in the sticky traps although the
difference in numbers of beetles between the two sampling methods was not proportional
to differences observed at other times of the season.

Normalization of data. Data from all sampling methods had a significant departure from
normality (P < 0.01) at 100 degree day intervals. The Taylor’s power law assessment
suggested the following powers (p) to normalize data: upper traps p=0.5, canopy traps

p=0.3, sweep-net p=0.8. Based on these results I selected the square root transformation

(W) for the three methods. The only significant departure from a normal
distribution after data transformation (P <0.01) was observed in the canopy trap sampling
when using data from time intervals of 100 degree days

Samplingmethods correlation. Significant correlation (r) was found between captures
from both levels of sticky traps: upper and canopy at all intervals (r >0.76, P < 0.001).
The correlation between sticky trap sampling and sweep-net was non significant (r = 0.29,
d.f.=40, P > 0.58) at 100 degree day intervals to significant correlation at 400 degree day

intervals (r = 0.8, d.f.=11, P < 0.001) and at 800 degree days (r > 0.9, d.f.=5, P < 0.005).

42

 

 

 

 

  
     

 

 

   

 

 

 

3 ' 1 I
a Upper-trap : J
a ; s
a 2- z a
3 . .
13 1 I
< , .
F. 1 ‘ 3 i
g 3
0 B 1 i l
3 ' ;
n. Canopy-trap
g ;
~ 2 . t
Q :
3 I
'O l
< : r
5 1 ‘ ’ 1
§ : E .
0 i I i i
3 . . . .
23,. Sweep-net
g . : s 2
a 2‘ a 5 E
g : f 3
s ' a
5 : :
g 0 fl -1 l l
0 200 400 600 800 1000

May-26 Jun-16 Jul-3 Jul-20
Degree days (10°C)

Figure 15. Temporal dynamics of coccinellid adults in spring wheat
measured by sticky traps (upper 1.2 m and canopy level) and
sweep-net sampling. Kellogg Biological Station. 1993.

43

 

GD
.1

Upper-trap

Mean Adults l trap

 

 

 

—t—-——-——a——

 

 

  
 
   

j
Canopy-trap

l
I
l
I

Mean Adults I trap

 

—«——- —-—--————--—-——-———-I

q___—___.._-______.

I
I
I
I

1

 

 

Sweep-net

“1 £

 

I
I
l
I
l
l
l
l
l
l
l
l
l
1
1
l

 

 

 

 

a Mean Adults I 25 sweeps

200 400 600 800 1000
May-26 Jun-16 Jul-3 Jul-20

Degree days (10°C)

Figure 16. Temporal dynamics of coccinellid adults in winter wheat measured
by sticky traps (upper 1.2 m and canopy level) and sweep-net
sampling. Kellogg Biological Station. 1994.

44

An analysis of the high correlation observed at the whole year scale interval (800 degree
days) showed the existence of two clusters of data: one of low numbers of insects in 1993
and one of high numbers of insects in 1994 (Figure 17). T tests showed that means

(YiSE) in 1994 for sweep-net (0.08 i 0.02), upper sticky trap (1.14 i 0.44) and

canopy sticky trap (3.25 i 0.64) were significantly higher (P S 0.001) than the
respective means obtained in 1993 for the same methods: sweep-net (0.02 i 0.008),
upper sticky trap (0.09 i 0.33) and canopy sticky trap (0.34 i 0.005). A correlation
analysis conducted at the same scale (800 degree days), but separately for each year,

showed no significant correlation between sweep sampling and sticky traps (P > 0.56).

Discussion.

The results of this analysis of sampling methods allowed me to interpret sticky trap catch
as a sampling method for coccinellids in a long term study. Sticky traps catches, for
example, have been used as a measure of the spatial distribution of the corn rootworm
adults (Midgarden et a1. 1993). The Taylor’s power law for corn rootworms indicated an
aggregated distribution (p~ 0.25), which is similar to what I obtained in this study for
coccinellids on sticky traps at the canopy level (p=0.3). The level of aggregation for
coccinellids was less in the upper sticky trap estimates (p=0.5), and in sweep-net
sampling (p=0.75). Changes in the level of aggregation registered by each of the sticky
trap levels could be the result of the ratio of influence of each trap level. Canopy traps
may have captured beetles near the vicinity of the traps and therefore may be highly

influenced by the local spatial distribution of the coccinellids; while the upper traps may

45

have been visible by beetles from several meters away from the traps and thus reducing
the aggregational influence. This hypothesis was supported by the results of the sweep-net
sampling, since the transect attempts to cover most of the field, which made samples less
susceptible to changes in the spatial distribution of coccinellids within a field. These are
important considerations if a sampling program based on sticky traps is conducted to

determine spatial distribution of coccinellids within a field.

 

Mean AdultsIT rap
N u

A
l

 

 

 

 

0 I I I
0 1 2 3 4

Mean Adults/25 sweeps

Figure 17. Correlation between sticky trap sampling at two levels
(upper=1.2m, and canopy) and sweep-net sampling considering
whole season captures of coccinellid adults in wheat.

Another key consideration before discussing the correlation of sampling methods at
different time intervals is the fact that sweep sampling is an instantaneous measure of

relative abundance, while sticky traps are an accumulative measure over time which may

46

include insects arriving, leaving, and remaining in the field. Other changes due to daily
weather conditions are averaged by the constant exposure of the traps (Tollefson and
Calvin 1994). This may have played a role in the lack of correlation between sweep-net
and sticky trap sampling methods at narrow intervals (100 degree days). However, as time
intervals became broader, trap captures tended to better represent the relative abundance
of coccinellids. This observation was also made for pitfall traps (Baars 1979), in which
year-round catches are more linked to insect abundance while narrow intervals become
measures of a mixture of density-activity (Nentwig 1988, Lester and Morrill 1989). The
fact that sticky traps and sweep-net samples are sensitive to changes in relative abundance
within the same habitat (wheat) was demonstrated by the differences observed by the two
methods when comparing captures between two different years. However subtle
differences like those that occurred among wheat fields in the same year showed no
correlation between methods even at whole season intervals. These results have important
implications for the analysis of sticky trap captures in this thesis. First, it is beyond
question that total counts methods are the best way to determine abundance of
coccinellids in field studies. However, as the scope of the field studies increases in
complexity (more habitats, long term, large extent), total counts techniques become
cumbersome (in time and personnel), making them no longer feasible for such studies.
Second, other sampling techniques commonly used for coccinellids such as sweep-net
sampling cannot be implemented in all kind of habitats and capture low number of
coccinellids. Third, sampling techniques that rely on trapping systems are the best options
currently available for large scale sampling despite their sensitivity to different factors

that affect the interpretation of insect captures (Spence and Niemela 1994). Fourth, the

47

influence of factors that affect the interpretation of trap captures is sensitive to the time
scale used in both: the sampling interval and the analysis. Environmental factors can be
averaged for large exposures of traps in the field, and year-round counts tend to average
for activity effects and relate more to abundance accounts.

As a final note I want to suggest a potential hypothesis regarding the attractiveness of the
color yellow by coccinellids. Observations in wheat showed that mature (and overmature)
wheat can be a suitable habitat for coccinellid reproduction (I observed an increase in the
number of eggs and larva of coccinellids and lacewings). This activity and the abundance
of coccinellids are barely detected by the sticky traps, despite the fact that sweep
sampling showed an important increase in the relative abundance of the coccinellid
population (Figure 16). Although there could be several explanations for this observation,
my hypothesis concerns the lack of attractiveness of the yellow color at this phenological
state of the crop. As was mentioned previously the color yellow is generally attractive for
coccinellids. Why it is attractive has not been a matter of discussion. One explanation
could be that the yellow color reflects the condition of the crop when there is an attack of
sap sucking insects. In this case, when wheat is maturing and changing its color, the
yellow color may have lost its attractiveness for the coccinellid adults. If this hypothesis
is true, then sticky traps may become a passive sampling devices without the power to

attract nearby insects.

Temporal patterns of coccinellids at the KBS LTER main site
An additional asset of the sticky trap sampling method was that insect captures reflected

temporal patterns exhibited by the populations of the different species of ladybird beetles.

48

For example, an analysis conducted on data obtained from seven years (1989-1995) in six
habitats showed that the dominance of C. septempunctata in the landscape reached its
maximum between 400-600 degree days (Figure 18). Peak populations of subdominant
and rare species on the other hand occurred either before or after the one of C.
Septempunctata. In other example, an annual time series showed the existence of patterns
that suggest the influence of special factors on the abundance of ladybird beetle species
(Figure 19). For instance, C. septempunctata had the maximum peak in 1989, while C.
munda and A. bipunctata, had theirs in 1990. C.m. lengi, which had normally two peaks
in the season (Figure 18) had the first one gradually increasing from 1989 to 1992, with
decreases thereafter. B. ursina, was not present from 1989-1992, and then started to

increase its population.

Conclusions on the use of sticky traps as a sampling tool for coccinellids

a) Sticky traps are the best tool currently available to conduct long term sampling of a
complex of coccinellids in multiple habitats.

b) There is evidence that sticky trap captures have a strong abundance component when
averages are scaled up to a year-round basis.

c) Annual time series showed temporal patterns that suggested the possibilities of
diversity effect on coccinellids.

d) Year-round captures from the perspective of diversity in agroecosystems set my the
scale of my thesis study at the higher temporal (2 to more years) and spatial (landscape)

dimensions (Figure 14).

49

 

 

     

C. septempunctata
(scale 1:6)

C.m. lengl

 

———~

H. undulata

H. axyrldls

A. blpunctata

 

c. stigma

 

H. parenthesis

 

300 600 900 1200
Degree Days (10°C)

Figure 18. Patterns of coccinellid captures during a season measured as frequency of

 

sticky trap catches in 100 degree days intervals using data collected from 1989
to 1995. Habitats included were corn, wheat, alfalfa, soybean, poplar and

secondary succession.

 

sooo
C.aumanomumma
2000

1000

900
C.m. lengi

400

 

2 on
C. munda

 

150

100

   

Adults I 210 traps

1 2 o
A. bipunctata
s o

40

 

B. ursina
4 o o

 

 

 

1999 1990 1991 1992 1993 1994 1995
Time (years)

Figure 19. Patterns of coccinellid captures during the year measured as
frequency of sticky trap catches in 200 degree days intervals from
1989-1995. Habitats included are corn, wheat, alfalfa, soybean,
poplar and secondary succession.

CHAPTER 3: ASSEMBLAGE OF PREDATORY COCCINELLIDAE
(COLEOPT ERA) IN RESPONSE TO TEMPORAL DIVERSITY WITHIN

AN AGROECOSYSTEM

Ecological management of agricultural systems relies heavily on the potential for
beneficial organisms to regulate pest populations (Vandermeer 1995). An increase
in the vegetational diversity has been considered as a necessary step to enhance
natural regulation of pests (Pimentel 1960). Diversity, mentioned in the previous
chapter, had several expressions depending on time and space (Figure 14). The
effect of increasing diversity in agroecosystems is explained in terms of two
hypotheses: a) the difficulty of herbivores to find their food, and b) the
enhancement of the action of beneficial natural enemies (Risch et al. 1983,
Sheehan 1986). The rationale behind the second hypothesis, in regard to spatial
diversity, is that beneficial organisms will have alternate sources of food and
shelter necessary to maintain their presence in the landscape (Dover 1989,
vanEmden 1990). Adult parasitoids, for example, that depend on the availability of
nectar sources, will be greatly benefited by the presence of flowers near agricultural
fields (Idris and Grafius 1995). Predators, often seek wooded habitats near
agricultural fields to spend the winter (Roach and Thomas 1991).

While spatial diversity can be seen in terms of landscape heterogeneity, temporal
diversity relates more to landscape change. In agricultural landscapes, most

changes are the result of man-made disturbances (Forman and Godron 1986).

51

52

Agricultural fields, for example, are chronic disturbance patches in which a crop is
planted every year, raised and eliminated (harvested) in the landscape. If the same
crop is grown every year in the same place, the chain of disturbances associated
with such crop become predictable. When a disturbance is predictable, organisms
adapt to it (Resh et al. 1988). Crop rotation could be seen as a way to maintain the
change in the landscape. From a pest management perspective, CI‘Op rotation would
be an effort to delay the adaptability of certain species (pest) to periodical
disturbances associated with always raising the same crop in the same place.

When a previously disturbed patch is left undisturbed (to promote, for example,
uncultivated areas in agroecosystems), then it becomes a “regenerated patch”
(Forman and Godron 1986). In these patches, temporal changes faced by the
arthropod community are now due to successional changes in species composition
and changing plant architecture (Southwood et a1. 1979).

Since the KBS LTER main site represent an excellent scenario to evaluate the
response of coccinellid assemblages to temporal diversity, I focused part of my
analysis on two main sources of temporal diversity: plant succession and crop

rotation.

53

CASE 1: HABITAT MATURATION IN ALF ALFA, POPLAR, AND

SECONDARY SUCCESSION

Price (1991) using parasitoids as an example of secondary consumers, synthesized
the importance of plant succession dynamics in determining communities of
beneficial organisms. Southwood et al. (1979) showed that insect diversity is
correlated with plant diversity during the early stages of secondary succession, but
later the correlation of insect diversity switches to the increasing complexity of
plant architecture.

One of the treatments in the KBS LTER was an early secondary succession which
represented a habitat with heterogeneous and changing plant composition subject to
zero management disturbances. A second treatment was a poplar plantation, which
represented a wooded perennial habitat with uniform plant composition but with a
changing plant architecture due to plant growth from small cuttings to full grown
trees on a six year period. A third habitat used in this study as a control, was
alfalfa. This was a herbaceous perennial habitat with uniform plant composition
subject to periodical agricultural disturbances.

My specific objective was to study the effect of consecutive years of growth and
maturation in alfalfa, poplar plantation and secondary succession, on the

assemblages of predatory coccinellids species.

54

METHODS

Study area. This study was conducted at the Long Term Ecological Research
(LTER) site at the Kellogg Biological Station (KBS), MI from 1989-1995. The
experimental design was established in 1989 consisting of seven treatments: four
wheat-corn-soybean rotations, two perennial biomass plantations: (alfalfa and
poplar) and a secondary succession. All treatments were replicated six times using
one hectare plots as experimental units (Figure 20). Alfalfa Medicago sativa L was
first planted in 1989, and after a five-year cycle, a new stand of alfalfa was planted
in 1994. Due to severe attack by the potato leaf hopper Empoasca fabae (Harris)
pesticide applications (one spray) were used in alfalfa during 1990, 1991, and 1994.
Alfalfa was harvested four times each year. Poplar (Populus x euramericana) tree
cuttings (15 cm) were planted in Spring of 1989, with a cover crop (red fescue
Festuca rubra L.) seeded in 1990. Secondary succession plots were tillage in 1989,

with no more agricultural management conducted during the rest of the study.

Sampling method. In each plot, five double-sided, yellow cardboard sticky trap
(22.5 x 14.0 cm) suspended 1.2 m above the ground from a metal trap (as described
by Maredia et al 1992b), were placed at permanent geopositioned locations. Yellow
cardboards were replaced every second week. Coccinellid adults caught on traps
were counted, recorded, and removed every week. The duration of the sampling

was from May to August of each year.

55

 

 

N '5': -.-:
T'eatments ‘1 E133. '5

Ffi' a“?
”war 59152; 5.ch5;

1W" 1 rm": "
ictI 31112 n513 gIcI4§

sky. 46’. ::-.' :..

conventional tillage \ E ,5 .5... :.5._‘.;5;.;;15
' com S] 1

field succession

no tillage lcr”. zc13 gnt4
low chemical input Wheat 111...-.. .. -..9

zero chemical input / soybean LPI ' i’zc'Zf FT4“
.. l a c

_ ----- .15 1011 a! ‘1014
Replicates II? 11116;
.. , 3:201] ct2. lcr31
1.. 3. mg]... .. 5..» :l
’lc16ct6j:
-Wmm- Ilaam3fil

Figure 20. Experimental design of the Long Term Ecological Research site Kellogg
Biological Station showing the spatial distribution of treatments and replicates.
Hickory Comets, MI.

BEBE EH3

 

 

 

 

 

 

 

 

 

Data analysis.

To conduct inter-year comparisons with selected treatments, I selected a data subset
with a similar scale. Since temperature is an important factor in the development of
insects used degree days as a criteria to determine the rank to use for data analysis.

For each year I selected the data that were included within 300 to 1000 degree days.

56

Yearly mean trap captures per species in each sampling plot were calculated by
averaging each plot trap capture for each sampling week and then averaging all

sampling weeks to obtain the yearly mean.

Three analytical techniques were used to analyze and compare coccinellid species
assemblages: a) multiple comparisons of relative abundance, species richness, and
species diversity (H’), b) the Kendall rank correlation test, and c) principal
component analysis. This analytical approach has been effective in detecting
responses of assemblages of coccinellids to habitat utilization in the landscape

(Colunga-Garcia et al. In press)

Multiple comparisons

Relative abundance is the yearly mean of the number of adult beetles captured per
trap in each treatment. This was estimated for individual species as well as for the
complex of coccinellids (i.e. by summing all individuals regardless of species).
Species richness is the yearly mean of the number of species captured per trap in
each treatment.

Species diversity (H’) was estimated using the Shannon Wiener equation

 

H = -2 [Pi][10gio(Pi)l

where pi is the proportion of the species i in the sample (Magurran 1988).

Data were normalized prior to statistical analysis using a VX+O-001

transformation. Analysis of variance (ANOVA) and the Tukey-HSD test were used

57

to test for statistically significant differences among treatments or years (SPSS

1993).

Kendall rank correlation

Kendall’s coefficient (1) was used as a measure of similarity between species

ranking of different treatments (or years) (Southwood 1978). I estimated ‘1’ based

on the mean trap captures per treatment (or years) using the Kendall’s rank
correlation test (Kendall 1955, SPSS 1993). For this analysis, rare species were

eliminated (Bullock 1971).

Principal component analysis

Multivariate methods are a set of tools applied to ecology that allow researchers to
describe or test differences in community structure (Green 1980). Of these
techniques, principal component analysis (PCA) facilitates the reduction in
dimensions of the original variables to produce a small number of variables
(components) ordered by the amount of variance they explain (James and
McCulloch 1985). Since the PCA procedure recognizes only linear combination of
the original variable (James and McCulloch 1990) I log (x+0.001) transformed the
data before analysis (SPSS 1993). Prior to PCA the data were arranged on a matrix
in which treatments were the headings of the columns and species were the
headings for the rows. Each intersection row (r)-column (c) had a value that

represented the yearly mean of the species “r” in the treatment “c”. Interpretation of

58

components is conducted by constructing plots using component values as
graphical coordinates (Randerson 1993). For each PCA, two plots were produced.
The first plot used the components for coccinellid species (also known as
eigenvalues), and the second plot used the components for treatments (or years)
(also known as scores). Visual interpretation of components is greatly reinforced

with the use of both plots together (Randerson 1993).

RESULTS

The fourteen species of coccinellids sampled in this study were, Coccinella
septempunctata (L.), Coleomegilla maculata lengi Timberlake, Cycloneda munda
(Say), Adalia bipunctata (L.), Hippodamia parenthesis (Say), Brachiacantha ursina
(Fab.), Hippodamia convergens Guerin-Meneville, Chilocorus stigma (Say),
Coccinella trifasciata perplexa Mulsant, Hippodamia tredecimpunctata tibialis
(Say), Hyperaspis undulata (Say), Anatis labiculata (Say) Hippodamia glacialis
glacialis (F.), and Harmonia axyridis (Pallas). H. axyridis (an exotic species) was

first captured in 1994.

59

Alfalfa

The dominant species of coccinellids in this habitat during the seven years of study
was C. septempunctata. The changes in relative abundance, species diversity, and
richness observed in this habitat (Figure 21), were mostly driven by the changes in
relative abundance observed by this species (Error! Reference source not found.)
The lowest species diversity was observed in 1989 and in 1994, which were years
in which alfalfa was newly planted. The relative abundance of coccinellids was
very high during the first year of study (mostly C. septempunctata). Thereafter the
numbers of coccinellids remained relatively stable with no significant differences
(Figure 21). Species richness also showed no significant differences except for the
third (1991) and last (1995) year of the study in which were observed respectively
the lowest and highest number of species (Figure 21). Three species that were never

captured in this habitat were: H. g. glacialis, H. t. tibialis and A. labiculata.

60

.6256 am as: as. 3o 9: a eases $585? on was. “same 3 8323. ea 523 32 e as? am; “a 88:

 

m «m o 9o c He N am e «o o «o o «o semséoefi
C so o «c o «o o no G «o o no c so 338%.:
N “N m an N an o «o o HO o no o no assess
o no o «o o «o o «o 0 9° c He o no 31832.5
o «o o no o «o C so a no o no a so 83833....
f: is so no no no eo He .3 no so no no no esteem
o no o no m «e o no o «o o «o o «o Sausage
em a: a: «m gem .3. me no so no no no no He cease
em «a 3. «e no no em «N eo ac em 48 em .1. assesses
o «o o «o o «o O HO o no men? m an E516
m an oao oao Nam o «o Nam «.94 3335.9
eo ac no Ho a: as axis no He no 9o .3 «a assesses
Bee «3 a: am am “N 024$ 22: a: so “mm Ben «2 Ease

a vm H hmm on ow H mmh an mm H mmv an mm H 5n n mm: H ovm an mm H New 0 ow H mm: Saaonzmfimimu .D

 

32 33 $3 N92 53 8o— awo—
So> 0:82 860mm

 

d2 .EoEoU bod—0mm .cotfim _mflwo—omm wmozox £9833“ 32%—com 88,—. waoq .32 8 33 80¢ man: 303m
323% main «babe E 3383 mum—$588 Co 830% sootsom mo 83% mo .m.m H 83 5 3:5me as: :38 be“; m 033.

61

Rolatlvo Abundance
(Idultcllnp)

 

 

89909192939495 89909192939495 89909192939495
Year:
1.8
Poplar Alfalfa Early Secondary
f 1_5 Succession
T.’
0
c
c
E.
I:
O
C
C
II
.:
tn
89909192939495 89909192939495 89909192939495
Years
8
Poplar Alfalfa Eady Secondary
Succession
A
n.
o
a
3%
.l: .2
2
x i
9.

 

89909192939495 89909192939495 89909192939495

Your:

Means (1' 8.13.) within a habitat which are followed by difi‘ercnt letters are significantly
different at the 0.05 level (T ukey HSD method).

Figure 21 Relative abundance, species diversity, and richness of a
complex of predatory coccinellids in alfalfa, poplar and
secondary succession during seven years of habitat growth at
the Long Term Ecological Research site at the Kellogg
Biological Station. Hickory Corners, MI.

62

Significant changes (P > 0.05) in species concordance in this habitat were observed

in 1989-1990, 1992-1993, and 1993-1994 (Table 4).

Table 1. Kendall’s coefficients for coccinellid species concordance
(species = 9) between consecutive years in poplar, alfalfa, and
secondary succession. Long Term Ecological Research- Kellogg
Biological Station, Hickory Corners, MI.

 

 

Interval Poplar Alfalfa Secondary

(years) succession
1989-1990 0.23 * 0.55 * 0.68
1990-1991 0.77 0.61 0.51 *
1991-1992 0.89 0.71 0.78
1992-1993 0.24 * 0.50 "‘ 0.56
1993-1994 0.84 0.46 * 0.79
1994-1995 0.57 0.57 0.49 *

 

Coefficients with an asterisks (*) indicate no concordance in species ranking
(P > 0.05) (Kendall’s rank correlation test).

Poplar

The most important event observed in this habitat was the “succession” of
dominance by three species of coccinellids: C. septempunctata, C .m. lengi, and B.
ursina. In 1989, when poplar reached a height of 1.5 meters, C. septempunctata

was the dominant species in the habitat. From 1990 to 1992, C. m. lengi was

63

captured in high numbers with a peak in 1991. Starting in 1993, B. ursina became

the dominant species in the habitat (Figure 22).

 

 

 

 

21 ‘ O Cocaine/Ia septempunctata
18 D Coleomegilla maculata lengi
m . Brachiacantha ursina
Q}
g a 15
Q
N _
a. 9
E '5 9 ‘
'- '3
C v
«I 5 1
3 _
. IA
0 L: o—-- .0 o. 'o‘
1989 1990 1991 1992 1993 1994 1995
Year

Figure 22 Temporal distribution of three dominant coccinellid species in poplar

during the seven years of habitat growth at the Long Term Ecological
Research site at the Kellogg Biological Station. Hickory Corners, MI.

Species richness was similar (P > 0.05) during the seven years of study (Figure 21).

Two species, H. t. tibialis and A. labiculata were not observed in this habitat

during the seven years. Species diversity was significantly lower during 1989,

1991,

and 1995 (the peak years of abundance for the three dominant coccinellids

mentioned above). The relative abundance of the complex of coccinellids was

significantly higher in 1991 and 1995 influenced by the high relative abundance of

64

C. m. lengi and B. ursina respectively for each year (Table 5).
The Kendall’s coefficient showed that species concordance changed significantly

(P > 0.05) from 1989 to 1990, and from 1992 to 1993 which represents the

transition years of the dominant species (Table 4).

Secondary succession

The most significant trend observed in this habitat was the temporal change in
species diversity which was low during the first year, reached a peak the following
year, and showed a gradual decline thereafter. Finally in 1995 species diversity
reached a maximum (Figure 21). Changes in relative abundance were driven mostly
by the abundance of C. septempunctata (the dominant species in this habitat) (Table
6) which followed a cycle that peaked in 1989 and 1994. Richness was higher the
second and last years of study while it was lowest during the third and fifth years
(Figure 21). H. axyridis, was first detected in this habitat in 1994. By 1995 this

species ranked second in relative abundance behind C. septempunctata.

65

.9268 am: 93: as. 8.0 2e a 829% 39:89st as 8:2 eases E 832E as 523 38 a £5? 3m 8 88:

 

 

no so so «o so He no so so so so as f:«: ”88289:
N Hm o Ho o Ho o Ho 0 Ho o Ho o Ho ”ascend
C so mam on; Nam C so oao one assuage
o «o a so o «o o “no C so a «o o «o Essa:
C so o ...o o «o O HO o «o c He c He assess
m u: o HO o as C so o «o o «o o so auras:
o Ho o «o O HO o «o m an m «N c so 383830
sesame: nemaomm sinuses so «o no «a an Hm so so Ease
so Ho so He 3 am £2 “mm £2 «S ..S “I: a: a: acacugsgs<
mm am pm a: no a» an am so «o .3 so an an gamete
93. in: in a»: in «a: £2 «3. so» «am 88...? mm a» 8:90
o «o a. j. m j. N am N «N O HO m an 339228.:
am «2 a; «mm am so some; 32 «mm: 38“ch am «a 93.5.9
so «o «mm «mm 3. as an «2 a.» «: poms: 08.28 39839290
mao_ voo_ m¢a_ maa_ ~¢o_ oom_ mwa_
58> 25286on

 

dz £8500 boo—om: .couflm _moBo—omm wwo=ov~ -ngmomom _momwfioom 85H wqoq .33 3 $2 88% mama“ .035
3020» mama: “SEEK E 3388 350538 .«o 83on 58.58 me 333 we .m.m H 83 5 3533 an: :38 3.3»; m 2an

66

.8505 am: 9:6 .22 2.0 2e a sauce 58%? as 8:2 eases B 832.8 as 523 so. a as? new i as»:

 

 

so «o no no so so so so so Ho so «o as“: nnmmtsésm
N ..nN o HO O «c c «o c He o so C so 358%.:
a no N «N o Ho N an c so a «o o «o assesses
o «o o go o as o «o o no o «o o «a “eases:
o a o o a o o a o o a o c a o o a o o a c 833339
m an on; oao o so o no o «o one asks:
c HO o go o Ho o Ho m an N «N o «o 38363.6
USNHSE fauna? nmNuSN so He no so «N “N me «o Essa
a o q. a a o a o a N a N as E a NN 9 mm a No a 2 a E. as N a : assesses .w
mm «m an a: no u; «N am no «o so «c an an assess
n S a a: a F. a NE a E a. e: as 2 a 3 a E a cNN as on a 2. a m a w 8:90
o a o a. a a. m a v N a N N a N o a o m ..n w assemiamm
am ”I: 9; ”IN 3.. so .ZNHS :2 «mm: oases. an «m $8550
a o a o a mN a mN a m a e as N a N_ as N a S 9 ON a z. o 9. a 8m assessmssma u
32 .32 82 No2 32 82 32

80> 25286on

 

.HE .3380 boo—oi .eoufim
_momwofizm wwozoM -noceomom Gamma—com Eco... wood .33 8 $3 88m an: 303m 30:?» wfim: :ommmoooam
bauaooom 380 E @6383 men—0533 mo 38on 5858 mo 33% .«o .m.m H 82 xv 3:596 met :38 330.». o 033.

67

Principal Component Analysis

Graphical assessment of the principal components of species and the site scores
showed that the first component, which accounted for 84% of the variance, was
related to intrinsic differences among coccinellid species. In this first component,
the outcome of the analysis arranged the species from the most abundant, C.
septempunctata on the positive side of the axis, to the least abundant, H. tibialis
and A. labiculata on the negative side of the axis (Figure 23). The lack of habitat
effect in this component was evident from the fact that the three habitats retained
the same position along the first score axis (Figure 23). Therefore, the first
component described the assemblage of coccinellid species at a scale that included
all habitats. The second principal component, which accounted for 11% of the
variance, was a result of site differences. B. ursina, C. munda, and C. m. lengi,
which were more abundant in poplar were positioned on the positive side of the
axis, while species such as C. septempunctata, H. axyridis, and H. parenthesis,
which were more abundant in alfalfa and succession, were positioned on the
negative side of the second component axis as a result of the analysis. The PCA
scores grouped alfalfa and succession separately from poplar, which confirmed our
observations that more noticeable changes in species assemblages occurred in

poplar.

68

 

 

 

 

 

 

 

 

 

 

 

Pcom 1
-2.5 O 2.5
2.5
.B. ursina
C" munga . C.m. lengi
N . . H. undulata A. b' ctt
E H.g.glaCIalIs . C. stigma. 'PUH aa
8 0 . C.t. perplexa
0_ A. labiculata ‘ H. convergens
H't' tibialis H. parenthesis
O C
H. axyridis C. septempunctata
-2.5
0.5 .
Populus
N
(D
L.
o O
0
a) Alfalfa
Field succession
-0.5
0.5 1 15

Score 1

Figure 23 Habitat scores and principal components of fourteen species of
coccinellids captured in poplar, alfalfa, and secondary succession
during seven years of habitat growth at the Long Term Ecological
Research site at the Kellogg Biological Station. Hickory Corners, MI.

69

DISCUSSION

Coccinellid species assemblages had two interesting responses in this study. The first, in
plant succession, showed that maximum coccinellid species diversity occurred in second
and third years. Thereafter, an increase in the dominance by C. septempunctata gradually

reduced the diversity of coccinellids.

Although I did not measure vegetation, I observed that the increase in the abundance
levels of C. septempunctata coincided with an increase in the population of goldenrod
(Solidago spp.). From the perspective of biological control, it seems that managing
secondary succession would be a realistic approach to take to promote predator diversity
(for example, mulching the field every three years). Before considering any such
managerial decisions regarding this habitat, it will be necessary to study more organisms
representing a broad gamut of beneficial insects (parasitoids included). There is
additional evidence, indicating that this may be the correct approach. Nentwig (1988),
after his work in meadows suggested that in order to maximize diversity and density of
beneficial insects, succession should be stopped or “held” by the third year. Moreover,
current theories on disturbance predict a maximization of diversity in habitats where
disturbances occurred frequently enough that resident and colonizing species can live in
the same area (Resch et al. 1988). If disturbances are too frequent, resident species may
be eliminated, and if disturbances are not frequent enough then competitive exclusion

may eliminate pioneer colonizing species.

70

In summary “abandoned fields” in the landscape will require some sort of management if

they are going to enhance the presence of beneficial insects within agroecosystems.

A second important response observed in this study was the “succession” of dominant
species observed in the poplar plantations. Here plant composition was uniform, but with
a different architecture (due to tree growth). The two dominant species during the first
years of popar growth: C. septempunctata and C. m. lengi were associated with crops in
agroecosystems. During the initial stages of grth poplar acted as an additional habitat
in the landscape for these species of coccinellids. However, toward the end of this cycle,
the dominant coccinellid species was that characteristic of deciduous habitats: B. ursina
(Maredia 1992a). By including a wooded habitat within the landscape, the richness of
coccinellids in the landscape increased, but such species may not have an impact on pests
in agricultural crops. This work showed how, in a relatively short time (seven years),
there is a shift in species at different stages of deciduous habitats. In fact, it has been
reported that species of predators in wooded habitats are different from species living in
agricultural crops (Usher 1995). Deciduous habitats, from a pest management
perspective, are considered important not because they provide more beneficial insects
species but because of the refuge they provide to beneficial insects living in agricultural
crops in agroecosystems. This involves all the vegetation associated with deciduous trees.
In this study, poplars were treated as a crop, and except for the presence of some grass,

there was not additional vegetation.

71

Finally, the response of coccinellids in alfalfa, showed no patterns that can be attributable
to the maturation of this habitat. This observation, coincides with Kieckhefer et al. (1992)
and suggests that the variation observed in alfalfa may be due to other factors such as

changes in temperature and/or food availability (Honek and Kocourek 1986).

72

CASE 2: CORN-SOYBEAN ROTATION UNDER CONVENTIONAL TILLAGE

Crop rotation is an agricultural practice designed to improve the quality of the soil
(Copeland and Crookston 1992), but it is also an important feature in pest
management (Flint and Roberts 1988). Crop rotation has been more successful
against pests with limited mobility and a narrow host range such as soil insects
(Brust and Stinner 1991). Since highly mobile polyphagous herbivores are not
affected by this practice, I hypothesized that a similar criteria could be applied to
coccinellids (generalist predators with high mobility).

Since one of the treatments at the KBS LTER was a corn-soybean rotation under
conventional tillage, I analyzed the response of the assemblage of coccinellids to

such management practice conducted during 1989-1994.

METHODS

Study area. This aspect of the study was conducted at the Long Term Ecological
Research (LTER) site at the Kellogg Biological Station (KBS), Ml from 1993-
1994. The experimental design was established in 1989 consisting of seven
treatments: four wheat-corn-soybean rotations, two perennial biomass plantations:
(alfalfa and poplar) and a secondary succession. All treatments were replicated six
times using one hectare plots as experimental units (Figure 24). A rotation of Com

Zea mays and Soybean Glycine max under high input conventional tillage was

studied from 1989 to 1994.

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Treatments 1 .35 ”‘5. “"5
CONVENTIONAL TILLAGE {p53 crs 55 m5-
232:2; CTl _Emz m3 1ci4
B field sumssion ' 1 L:.......:: gnu-u.“ Theo-nhli:
no tillage sl 1ci21§ gzci3 m4
low chemical input ..... ----- ‘ ...._.....: = --------- “—— ‘“
zero chemical Input pl 2ch g 53 CT 4

. ..... .9": i... ..... 5 T... ..... 3' lCll . . p2 a3 ‘ZCI4N

Replicates p6 a6 N62 f. .........

1-6 "m: is“; W, ézml CT2 lc13 a4
56 41°“) ‘°°"‘ ‘1 ..... .......... " “it?“

... ...... ‘ :ntl, 32.§p3,'s4
lci6fi' CT6 ..... -..
. .. _.. f ,

a1 52 i CT3 §p4

 

 

 

 

 

 

 

 

Figure 24. Experimental design of the Long Term Ecological
Research site Kellogg Biological Station showing the
spatial distribution of treatments and replicates. Hickory
Corners, MI.

Sampling method. In each plot, five double-sided, yellow cardboard sticky trap
(22.5 x 14.0 cm) suspended 1.2 m above the ground from a metal trap (as described
by Maredia et al 1992b), were placed at permanent geopositioned locations. Yellow
cardboards were replaced every second week. Coccinellid adults caught on traps
were counted, recorded, and removed every week. The duration of the sampling

was from May to August of each year.

Data analysis. Data analysis was conducted using the same procedure described in

page 55 of this chapter.

74

RESULTS

Species diversity

The eleven species of coccinellids sampled in this study were, Coccinella
septempunctata (L.), Coleomegilla maculata lengi Timberlake, Cycloneda munda
(Say), Adalia bipunctata (L.), Hippodamia parenthesis (Say), Brachiacantha
ursina (F ab.), Hippodamia convergens Guerin-Meneville, Chilocorus stigma (Say),
Coccinella trifasciata perplexa Mulsant, Anatis labiculata (Say), and Harmonia
axyridis (Pallas). H. axyridis (an exotic species) was first captured in 1994.

C.m. lengi was the only species that significantly responded to the crop rotation
process, since it was more abundant during the corn phase (Table 7). C.
septempunctata, the other dominant species, showed a pattern of variation in
abundance that did not relate to rotation. No pattern of relative abundance was
observed with the rest of the species.

Significant differences were found on total relative abundance during 1989, and
species diversity during 1994 (Figure 25). Species richness remained the same

during the whole rotation scheme.

Coefficient of similarity

Regarding species concordance, a significant similarity was found between 1990

and 1992 (soybean years) and 1991 and 1993 (corn years) (Table 8).

75

.85? no: 93b 35. So 2o .s eases 3.888% .5 sons. 859% 3 3328 2s 523 32 s :8? So so 88:

 

 

so so so so so so so so so so so so oes: 38828.:
N ..sN o so o so o «o o «o o «o o so flotsam
o so N sN o so N so o «o o so o «o assess:
o so o «o o so o «o o so o so o so 3....er
o so o so o «o o so o so o so o so 83835....
o so o so o so o so o so o so o so 3.58.:
o so o so o so o .o o «o N “N o so sssoesssu
sostooo: ososooo ooososN so so so «o N “N so so Essa
so so so so sN «N oso: sNN osoo «No n2o. soo oso s: 38:83....
so so oo «2 oo so sN so so so so so. so so season
oos as: no so: so so: oso_ sos ooo ..SNN osoosoo so so 8.52.0
o so s as o as N sN N «N o so o so 35:28.:
so so. oso soN so so ooNsoo 3N. soo: soosoos so so 88550
so so soNsoN so so so «a oso Us: ooNs: sossooo 32838859
ooo_ soo_ oooo Noo_ _oo_ ooo_ soo_
88> 2.8288on

 

A2 £8500 >8on 528% 188205 wmozom -nocoooom _mommfioom 88H wcoq
.owozz 12838250 was @8838: E8 m8N=fl8b 39: 82:88 :wE 835 .8938 83.3w :80 89a xmo m 5
one: .228 328% memo: 838% 328583 mo 88on 888 me 3:68 we .m.m H A2: 5 8.533 8.5 888 38on h 833.

76

Shannon-Wiener (H) R9'1gzzAbungrnee

Rlchness
(specieshrapl

 

n

Com Soybean Com Soybean Com
1989 1990 1991 1992 1993 1994

Means (15.5.) which are followed by different letters are significantly different
at the 0.05 level (T ukey HSD method).

Figure 25. Relative abundance, species diversity, and richness of a complex of
predatory coccinellids in a six year corn-soybean rotation under high input
conventional tillage at the Long Term Ecological research site at the
Kellogg Biological station. Hickory Comers, M].

Principal Component Analysis
Graphical assessment of the principal components of species and the site scores

showed that the first component, which accounted for 72% of the variance, was

77

related to intrinsic differences among coccinellid species. In this first component,
the analysis arranged the species from the most abundant, C. septempunctata on the

positive side of the axis, to the least abundant ones on the negative side of the axis

(Figure 26).

Table 8. Kendall’s coefficients for coccinellid species concordance between
consecutive years in a corn-soybean rotation at the KBS LTER, Hickory
Corners, MI. (P= statistical significance level).

 

 

1989 1990 1991 1992 1993
corn soybean corn soybean com
1990 0.4
(P=0.33)
1991 0.4 0.6
(P=O.33) (P=O. 14)
1992 0.6 0.8 0.4
(P=0.14) (P=0.05) (P=O.33)
1993 0.6 0.4 0.8 0.2
(P=0.l4) (P=0.33) (P=0.05) (P=O.62)
1994 0.2 0.4 O 0.6 -0.2
(P=O.62) (P=O.33) (P=1.00) (P=O. 14) (P=O.62)

 

The second principal component, which accounted for 14% of the variance, was
primarily a result of habitat preferences. Species which preferred corn positioned
on the positive side of the axis: B. ursina, Cm. lengi, C. t. perplexa, and C. munda.
Species such as H. parenthesis, and A. labiculata which preferred soybean, were

positioned on the negative side of the second component axis.

DISCUSSION
The most important finding of this work is the change in dominant species

observed from year to year during the rotation process, in which C.m. lengi seemed

78

more affected than the others. C. septempunctata was the dominant species during
the soybean phase, while C.m. lengi was dominant (except in 1989) in the corn
phase. The potential implications of this rotation on dominance lies in the fact that
it will never be possible to take advantage of the population of coccinellids
produced the previous year. For example, in a corn year, the population of C.m.
lengi increases in the landscape. That will potentially produce great numbers of
overwintering Cm. lengi ready to populate the landscape during the following year.
The following year, however, the crop rotation scheme dictates that it will be
soybean, which is a habitat that promotes the population build-up of C.
septempunctata, and probably deters C.m. lengi. Therefore, in some sense all the
C. m. lengi population generated during a corn year are wasted during the soybean
year. Probably the same process occurs with C. septempunctata, although
potentially, as shown during 1989, this species can occur in large numbers during
corn years. In this study, however, it was not possible to evaluate the full impact of
rotation on coccinellid populations because the surrounding landscape always
provided an alternate food resource. For example, during soybean years, corn was
present in other fields nearby the experimental area. However, it is evident that this
agricultural practice, besides disrupting the population dynamics of several pests,
can also produce the same effect on beneficial insects. Presence of alternate crop
habitats in early spring (wheat, alfalfa) could provide a survival opportunity for

coccinellid populations that were produced during a previous season.

79

 

 

 

 

 

 

 

 

Pcom 1
-1 0 1 2
2
B. ursina
' o
1 . . C.m. lengi
C.t. perplexa
N C. munda
E
o 0
0 H. convergens
D. I. O 0
CC. stigma A. bipunctata .
H. axyrldls
-1 O C. septempunctat;
A. labiculata
H. parenthesis
O
-2
0.8
Corn-93
0.6 o
0'4 Corn-91
O
N 0.2
2
o 0 Corn-89
o O
(D
-0.2 .
. Soy-90
-0_6 sigma

0.5 0.7 0.9 1.1 1.3 1.5

Score 1

Figure 26. Habitat scores and principal components of eleven species of
coccinellids captured in a six year corn- soybean rotation under high
input conventional tillage at the Long Term Ecological Research site
at the Kellogg Biological Station. Hickory Corners, MI.

CHAPTER 4: ASSEMBLAGE OF PREDATORY COCCINELLIDAE
(COLEOPTERA) IN RESPONSE TO HABITAT MANAGEMENT

(T ILLAGE AND CHEMICAL INPUTS) WITHIN AN AGROECOSYSTEM

The influence of humans on agroecosystems through all of the activities conducted to
produce crops has long been recognized (Risser 1986). Many of these activities, aimed to
address increased crop production (such as plant nutrients, pest control, etc.) had
unexpected consequences over the long run (soil erosion, water pollution, pest resistance,
etc.) (de la Court 1992, Wheeler 1992). This situation motivated the search for ecological
alternatives to make agriculture sustainable (Hetch 1987). The goal would be to achieve
an equilibrium of energy flow in agroecosystems which eventually will allow for the
reduction of elimination of the need for artificial inputs (Crews et al. 1991). In this
process, a systems perspective will be needed to evaluate the impact of recommended
ecological practices in the context of the whole agroecosystem. Conservation tillage, for
example, recommended for ecological management of the soil (Hilner 1985), requires the
reconsideration of pest management approaches (Kells and Meggitt 1985, Kirby 1985,
Ruppel and Sharp 1985). From a pest management point of view low input agricultural
practices are considered to diversify agroecosystems which makes them less susceptible

to pest outbreaks (Reijntjes et al. 1992, Altieri 1994).

The KBS LTER principle hypothesis is that “agronomic management based on

ecological concepts can effectively substitute for reliance on chemical subsidies in

80

81

production-level cropping systems” (VanCleve and Martin 1991). To test this
hypothesis, the comparison of crop production under different tillage and chemical
input treatments using a corn-soybean-wheat rotation was included in its
experimental design. Conservation tillage practices change the ground
microenvironment affecting soil and litter dwelling insects (Musick and Petty 1973,
Steffey et a1. 1992). These practices encourage the development of organisms that
benefit from increased humidity provided by crop residues or from the presence of
weeds (Musick and Petty 1973, Steffey et al. 1992). However, there is a higher
proportion of pests whose population density and subsequent damage decreases as
tillage decreases (Stinner and House 1990). Such reduction in pest populations may
be due to the increase in beneficial insect fauna. Reduction of soil disturbance and
a higher presence of weeds favor important predators such as ground beetles, rove
beetles, spiders, and ants (Stinner and House 1990, Clark et al. 1993).

Studies on coccinellids to evaluate the effect of conservation tillage and reduced
chemical input practices on beneficial arthropods showed different results
depending on the crop and on the insect species. C.m. lengi, for example, had large
populations of adults in conventional tillage corn (Tonhasca and Stinner 1991),

while H. convergens had higher density in no tillage wheat (Rice and Wilde 1991).

The present work was conducted to evaluate the effect of tillage and chemical input

practices on the species assemblages of predatory coccinellids.

82

METHODS

Study area. This study was conducted at the Long Term Ecological Research
(LTER) site at the Kellogg Biological Station (KBS), MI from 1993-1995. The
experimental design was established in 1989 consisting of seven treatments: four
wheat-corn-soybean rotations, two perennial biomass plantations: (alfalfa and
poplar) and a secondary succession. All treatments were replicated six times using

one hectare plots as experimental units (Figure 27).

 

 

 

1v 5
Treatments 1 35..
conventional tillage \oom p'f. 5555”;
no tillage "
low chemical input ’ 3:“
zero chemical input / soybean I E
3 " alfalfa
field succession if"‘b‘1?Z';s3
. ”,7, w“: pz, ..
Replicates p6 . 153
:_ ...... sad in; .m.-s; ‘.
a an
1., _.. al.... I

(- i
‘ la Eli“ : 82‘: . . : S4 _
l ;. .135 54' fl ‘ ,7, 7 ,- . .

Figure 27. Experimental design of the Long Term Ecological Research
site Kellogg Biological Station showing the spatial distribution
of treatments and replicates. Hickory Corners, M1. 1993-1995

 

 

 

 

 

 

 

 

Swpgr

[BEBE

 

 

 

 

33

 

v‘.
Ll
.
‘0‘.

 

 

..-..
:3 I
511':
L.

{ciao-o.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

83

Treatments. Four management practices were tested: High Input Conventional
Tillage, High Input No Tillage, Low Chemical Input, and Zero Chemical Input.
Field crops used were: Corn Zea mays L. In 1993, Soybean Glycine max L. In 1994,
and Wheat Triticum aestivum L. A description of the agronomy protocols are

described in the Table 9.

Sampling method. In each plot, five double-sided, yellow cardboard sticky trap
(22.5 x 14.0 cm) suspended 1.2 m above the ground from a metal trap (as described
by Maredia et al. 1992b), were placed at permanent geopositioned locations.
Yellow cardboards were replaced every second week. Coccinellid adults caught on
traps were counted, recorded, and removed every week. The duration of the
sampling was from May to August of each year, except during the wheat year in

which sampling was interrupted when the crop was harvested.

Data analysis. Data analysis was conducted using the same procedure described in

page 55 of this chapter.

84

Table 9. Agronomic protocol for each one of the different field crops in
the KBS-LTER. 1993-1995. CT =conventional tillage, NT =no
tillage, LCI =low chemical input, and ZCI = zero chemical

 

 

 

 

 

 

input.
AGRONOMIC CORN 1993
PRACTICE CT NT LCI ZI
Tillage Y N N N
Herbicide (No) Y (1) Y (2) N N
Fertilizer Y Y Y N
Cultivation (No.) N N Y (5) Y (5)
Cover crop (before and after N N Y Y
the crop)
SOYBEAN (1994)
CT NT LCI ZI
Tillage Y N Y Y
Herbicide Y (1) Y (2) Y (1) N
Fertilizer N N N N
Cultivation Y (1) N Y (3) Y (3)
Cover crop N N N N
WHEAT (1995)
CT NT LCI ZI
Tillage Y N Y Y
Herbicide (No.) Y (2) Y (2) N N
Fertilizer Y Y Y N
Cultivation (No.) N N Y (5) Y (5)
Cover crop (before and after N N Y Y

the crop)

 

85

RESULTS
The thirteen species of coccinellids found in this study were: Coccinella
septempunctata (L.), Coleomegilla maculata lengi Timberlake, Cycloneda munda
(Say), Adalia bipunctata (L.), Hippodamia parenthesis (Say), Brachiacantha ursina
(Fab.), Hippodamia convergens Guerin-Meneville, Chilocorus stigma (Say),
Coccinella trifasciata perplexa Mulsant, Hippodamia tredecimpunctata tibialis
(Say), Anatis labiculata (Say) and Hippodamia glacialis glacialis (F.), and
Harmonia axyridis (Pallas). H. axyridis (an exotic species) was first captured in

1994.

Corn 1993

Total relative abundance, Shannon-Wiener, and Richness indices were significantly
different between high input treatments (conventional and no tillage) and low and
zero chemical input treatments (Figure 28). C. maculata and C. septempunctata, the
two dominant species in this habitat, were most abundant in the high input
treatments (Table 10).

The only significant similarity in species concordance was found between No till
and Low Chemical Input (Table 11). High dissimilarities in species concordance

were found between the Zero Chemical Input and the rest of the treatments.

86

14 -- Corn Soybean Wheat
b

Relative Abundance
(adults/trap)

 

 

CT NT Ll Zl CT NT Ll 21 CT NT Ll Zl

Soybean Wheat

Shannon-Wiener (H)

 

CT NT Ll ZI CT NT LI Zl CT NT Ll Zl

Corn Soybean Wheat

Richness
(species/trap)

 

 

CT NT Ll ZI CT NT Ll Zl CT NT L| Zl

Means (i S.E.) within a crop which are followed by the same letter are not
significantly different at the 0.05 level (Tukey HSD method).

Figure 28. Relative abundance, species diversity, and richness of a complex of
predatory coccinellids in corn, soybean, and wheat under four
agronomic practices at the Long Term Ecological Research site at the
Kellogg Biological Station. Hickory Comers, MI.1993-1995. CT
=conventional tillage, NT =no tillage, LI =low chemical input, and Z1 =

zero chemical input.

87

Table 10. Yearly mean trap captures (x 100) iS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in corn under four
management practices. Long Term Ecological Research- Kellogg
Biological Station, Hickory Corners, MI. 1993.

 

 

Treatment
Species name Conventional No tillage Low chemical Zero chemical
tillage input input
C. septempunctata 171 + 38 bc 231 + 43 c 33 + 19 a 65 + 21 ab
C.m.lengi 339 + 72 c 119 + 47 b 36 + 18 ab 4 + 4 a
H. parenthesis 0 + O 5 + 3 3 + 3 O + O
C. munda 24 + 10 17 + 8 4 + 4 0 + O
H. axyridis O + O 0 + O O + O O + 0
H. glacialis 0 + 0 O + O O + 0 O + O
B. ursina 10 + 10 2 + 2 0 + O 6 + 6
A. bipunctata 2 + 2 2 + 2 O + 0 0 + 0
A. labiculata 0 + O 0 + 0 0 + 0 O + 0
Ct. perplexa 2 + 2 O + 0 0 + 0 O + 0
H. convergens 0 + O O + 0 O + 0 O + O
H.t.tibia1is O + O 0 + 0 0 + 0 O + 0
C. stigma 0 + 0 0 + 0 0 + 0 0 + 0

 

Means (:1: S.E.) within a row which are followed by different letters are significantly
different at the 0.05 level (Tukey HSD method).

Table 11. Kendall’s coefficients for coccinellid species concordance

in corn under different management practices at the KBS
LTER, Hickory Corners, MI. 1993. CT =conventional tillage,
NT =no tillage, LCI =low chemical input, and ZCI = zero
chemical input.

 

 

CT NT LCI
NT 0.59
(P=0.068)
LCI 0.62 0.84
(P=0.06) (P=0.01 1)
ZCI 0.51 0.46 0.30
(P=0.l32) (P=0.l76) (P=0.384)

 

Soybean 1994

The most important observation in this habitat was the dominance of C.
septempunctata in all treatments and the low relative abundance of this species in
the low chemical input treatment (Table 12). This, in turn, caused a significantly
low relative abundance for the total of coccinellids (Figure 28). No significant

differences were found in either diversity (Shannon index) or species richness

(Figure 28).

A highly significant similarity was found between no tillage and zero chemical

input treatments (Table 13). The rest of the treatments were dissimilar.

89

Table 12. Yearly mean trap captures (x 100) iS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in soybeans under four
management practices. Long Term Ecological Research- Kellogg
Biological Station, Hickory Corners, MI. 1994.

 

Treatment

Species name Conventional No tillage Low chemical Zero chemical
tillage input input

 

C. septempunctata 710 + 93 b 1063 + 121 b 228 + 69 a 611 + 104 b

Cm. lengi 6 + 3 22 + 4 29 + 12 23 + 7
H. parenthesis 15 + 6 23 + 7 5 + 3 11 + 7
C. munda 4 + 3 4 + 3 10 + 7 2 + 2
H axyridis 2 + 2 4 + 3 6 + 6 2 + 2
H. glacialis 0 + 0 2 + 2 0 + 0 0 + 0
B. ursina 0 + 0 0 + 0 0 + 0 7 + 3
A. bipunctata 8 + 5 6 + 3 0 + 0 12 + 5
A. labiculata 10 + 2 O + 0 0 + 0 2 + 2
Ct. perplexa 0 + 0 0 + 0 0 + 0 0 + 0
H. convergens O + 0 2 + 2 0 + 0 3 + 3
H.t.tibialis 0 + 0 0 + 0 0 + 0 0 + O
C. stigma 2 + 2 0 + 0 0 + 0 0 + 0

 

Means (:1: S.E.) within a row which are followed by different letters are significantly
different at the 0.05 level (Tukey HSD method).

90

Table 13. Kendall’s coefficients for coccinellid species concordance
in soybean under different management practices at the
KBS LTER, Hickory Corners, MI. 1994. CT =conventional
tillage, NT =no tillage, LCI =low chemical input, and ZCI =
zero chemical input

 

 

 

CT NT LCI

NT 0.59

(P=0.032)
LCI 0.34 0.57

(P=0.226) (P=0.044)
ZCI 0.41 0.67 0.41

(P=0.l34) (P=0.017) (P=0.l47)

Wheat 1995

In this habitat, the highest species diversity was found under conventional tillage,
while the highest total relative abundance was in the zero chemical input treatment
(Figure 28). Cm. lengi and C. septempunctata, the dominant species, preferred the
zero chemical input treatment , although Cm. lengi was also attracted to

conventional tillage treatment (Table 14).

The Kendall’s coefficient of concordance showed a significant similarity between

all treatments (Table 15).

91

Table 14. Yearly mean trap captures (x 100) iS.E. of adults of thirteen species of
coccinellids sampled using yellow sticky traps in wheat under four
management practices. Long Term Ecological Research- Kellogg
Biological Station, Hickory Corners, MI. 1995.

 

Treatment

Species name Conventional No tillage Low chemical Zero chemical
tillage input input

 

C. septempunctata 361 + 52 a 519 + 66 ab 450 + 45 ab 719 + 100 b

C.m. lengi 72 + 16 c 6 + 6 a 22 + 16 ab 72 + 20 be
Hparenthesis 39 + 10 28 + 10 22 + 16 50 + 11
C. munda 6 + 6 11+ 7 19 + 14 6 + 6
Haxyridis 22 + 16 22 + 7 ll + ll 17 + 17
Hglacialis 17 + 11 6 + 6 11+ 7 11+ 7
B. ursina 0 + 0 0 + 0 0 + 0 6 + 6
A. bipunctata 0 + O 0 + 0 0 + 0 0 + 0
A. labiculata 0 + 0 11 + 7 0 + 0 0 + 0
Ct. perplexa 6 + 6 0 + 0 6 + 6 0 + 0
H. convergens 6 + 6 0 + 0 0 + 0 0 + 0
Httibialis 6 + 6 0 + 0 0 + 0 0 + 0
C. stigma 0 + 0 0 + 0 0 + 0 0 + 0

 

Means (2h S.E.) within a row which are followed by different letters are significantly
different at the 0.05 level (Tukey HSD method).

92

Table 15. Kendall’s coefficients for coccinellid species concordance
in wheat under different management practices at the KBS
LTER, Hickory Corners, MI. 1995. CT =conventional tillage,
NT =no tillage, LCI =low chemical input, and ZCI = zero
chemical input.

 

 

CT NT LCI
NT 0.61
(P=0.016)
LCI 0.85 0.70
(P=0.001) (P=0.007)
ZCI 0.88 0.611 0.85
(P=0.001) (P=0.016) (P=0.0017)

 

Principal Component Analysis

Graphical assessment of the principal components of species and the site scores
showed that the first component, which accounted for 63% of the variance, was
related to intrinsic differences among coccinellid species. In this first component,
the analysis arranged the species from the most abundant, C. septempunctata on the
positive side of the axis, to the least abundant, H. tibialis and C. stigma on the

negative side of the axis (Figure 29).

93

 

 

 

 

 

 

 

 

Pcom 1
-1.5 .005 0.5 1.5 205
2 B. ursina1
A. bipunctata Q
1 C.m. lengi
C. stigma .
N
E O O A. labiculata .
0 ' C- munda C. septempunctataw
8 O H. convergens .
0‘ . Ct perplexa
H.t. tibialis ' ' H, parenthesis
'1 H. axyridis
. O
H.g.glacialis
-2
0-6 iCom-CT
Com-Zl C NT
om-
o.4 0 '
Soy-Zl
0.2 Soy-CT
~ 0
e 0 .Corn-LI
(:30) Soy-NT 3.01“”
O
'02 Wheat-2|
.0 4 Whea.t-NT
' O
Wheat-CT Wheat-Ll
-0.6 0
0.5 0.7 0.9 1 .1 1 I3 1'5

Score 1

Figure 29. Habitat scores and principal components of thirteen species
of coccinellids captured in corn, soybean, and wheat under
four management practices at the Long Term Ecological
Research site at the Kellogg Biological Station. Hickory
Corners, M1. 1993-1995. CT =conventional tillage, NT =no
tillage, LCI =low chemical input, and ZCI = zero chemical input.

94

The second principal component, which accounted for 14% of the variance, was
mostly the result of habitat preferences. Species which preferred corn such as B.
ursina and A. bipunctata, were positioned at the extreme end of the positive side of
the axis, while species such as H. parenthesis, H. axyridis, and H. g. glacialis which
preferred wheat, were positioned at the extreme end of the negative side of the axis
(Figure 29). An additional treatment grouping was observed between corn and

soybean under both low and zero chemical input treatments.

DISCUSSION

The most important response by coccinellids to management treatments occurred in
corn. Total abundance, species diversity, and species richness of predatory
coccinellids were higher in the high chemical input systems (conventional and no
till) and significantly lower in low and zero chemical input treatments. C. m. lengi
and C. septempunctata, the dominant species in this habitat, were significantly
more abundant in high chemical input systems. In the LTER experiment,
differences in chemical inputs caused differences in weed management, since up to
five cultivations in low and zero chemical input treatments were necessary for weed
control. Presence of weeds increases both plant diversity and plant density in
agroecosystems (Koskinen and McWhorter 1986) and C.m. lengi is negatively

affected when plant density increases (Risch et al. 1982).

95

The result that conventional tillage corn harbored higher numbers of C. m. lengi
coincides with observations by Tonhasca and Stinner (1991), although they found
significantly smaller populations in no tillage corn. No tillage practices promote the
increase of weed populations (Kells and Meggitt 1985). However in the KBS LTER
weeds were rapidly controlled by the use of herbicides, which may have been the
reason for the similarity on Cm lengi abundance in both tillage treatments
(conventional and no tillage).

The observation that C. septempunctata was also more abundant in high input
treatments was unexpected because this species has not been reported to be affected
by plant density or plant diversity. A possible cause was the intense disturbance
due to weed management that low and zero chemical input treatments were subject
to during the season. The peak of abundance of C. septempunctata in corn
corresponded with early phenological stages of the plant, and most of the
cultivations were made at that time. This constant disturbance may have rendered

the habitat unsuitable for this species.

Another interesting pattern in predator abundance was observed in wheat. This crop
did not require any cultivation for weed control in low and zero chemical input
treatments. This situation meant that disturbance was not an issue in these
treatments. Under these circumstances, the zero chemical input treatment registered
a high relative abundance of coccinellid, including the two dominant species: Cm.

lengi and C. septempunctata.

96

The results on coccinellid abundance obtained in corn and wheat illustrates the
complexity that exists within sustainable systems. It also shows the importance of
not making generalizations from what occurs in one crop nor what occurs with one
group of beneficial insects. If it is true that the presence of weeds may deter some
beneficial insects such as Cm. lengi, it may also enhance the activity of other
beneficial insects such as the ground beetles. In wheat, small population of weeds
may diminish the population density of some predators but it also promotes the

activity of other predators such as coccinellids.

CHAPTER 5: DEVELOPING A MODEL OF INTERACTIONS BETWEEN
PREDATORY COCCINELLIDS (COLEOPTERA) AND LANDSCAPE

STRUCTURE IN FIELD CROP AGROECOSYSTEMS.

INTRODUCTION

Agricultural systems are considered to have lower complexity when compared to natural
systems (Risch 1987, Stinner and Stinner 1989). This distinction has been made based on
the agricultural model of the 70’s and 80’s where monoculture fields with high-yield
hybrids grow under intense chemical inputs provided by the farmer (Bird 1994). Concerns
for the need for adequate knowledge to manage agroecosystems ecologically have made
researchers aware of the potential complexity that agroecosystems have (Kogan 1986,
Vandermeer 1995). For example, interactions between vegetation (cultivated and natural),
herbivores, and beneficial insects occur at different spatial scales that transcend the single
plant to the landscape (Levins and Wilson 1980, Risser 1986) (Figure 30). Moreover,
these interactions can be highly influenced by weather (Wellington and Trimble 1984) and
human management (Stinner and Stinner 1989). When issues increase in complexity, a
systems approach becomes necessary to provide the holistic View required to obtain a
thorough understanding of the interactions that occur among the different components or

subsystems in agricultural landscapes (Jeffers 1978, Bird et al. 1990).

97

98

Interactions of Subsystems in Agroecosystems

       
 

  
          
       
   
  
  
   

management asrolandseape

sub-system

 

producers

plant
sub-system

successob n

   
 

  

III ‘
Q I I I

interaction

, . levels
microclrmate

herbivores

herbivores
sub-system

weather

sub-system
parasitoids
V

ntomophagous _

beneficial organisms sub-system

Figure 30. Interactions of different subsystems in agroecosystems at different
hierarchical spatial scales.

Modeling is the step within the systems approach in which the interactions that occur
between different elements of the system are expressed mathematically (Tait 1987, Bird et
al. 1990). This allows researchers to evaluate their degree of understanding of the system
of interest and propose further or alternative studies (Getz and Gutierrez 1982).

Modeling has been extensively used to study interactions between organisms and the
landscape (Saarenmaa et al. 1988, Folse et al. 1989, Hyman et al. 1991, Turner and Dale
1991, Doak et al. 1992, Milne et al. 1992). In landscape ecology, modeling has become
particularly useful because of the difficulty in conducting experiments at larger spatial or

temporal scales (Baker 1989, Turner et al. 1989, Hargrove and Pickering 1992).

99

Simulation models allow researchers to explore scenarios difficult to reproduce in the
landscape, and as a result of several simulations and analyses, hypotheses can be
formulated and tested with an appropriate experimental design (Fahrig 1991, Pickett et al.
1994)

Different aspects of the interactions between ladybird beetles and the landscape were
described in previous chapters. This chapter will describe the development of a model
designed to integrate the results obtained in those chapters. Ladybird beetles have been the
subject of multitrophic simulation models (Carter et al. 1982, Gutierrez and Baumgaentner
1984). In these studies however, only one habitat was included. A model including a
group of ladybird beetle species and a landscape may become very complex because the
large number of potential interactions that exist among environmental conditions,
landscape elements, and ladybird beetles (Figure 31). For example, environmental
conditions including temperature, rain, wind, and photoperiod change through the season
and drive the population dynamics of ladybird beetles. In addition, the structure (type,
number, shape, size, and configuration of patches), the fimction (flow of energy), and
anthropogenic disturbances in the landscape have been shown to affect the population
dynamics of ladybird beetles (Chapters 3 and 4).

Thus, the challenge for a researcher is to determine the amount of knowledge sufficient to
include in the model (Berryman 1991a,b, Onstad 1991). In this process, it is fiindamental
to assess the appropriate scale needed to evaluate the interactions that occur in the system

to be modeled (see Chapter 2).

100

 

Environmental Conditions

Weather

 

 

Figure 31. Life system of ladybird beetles in the landscape

OBJECTIVE
The objective of this part of the research was to develop a simulation model to synthesize
the observations and results obtained as part of the KBS LTER project (described in
chapters 3 and 4). This model will be used to study the response of assemblages of
ladybird beetles to changes in the structural configuration of the landscape. The ultimate
goal of this model with regard to the KBS LTER program is to assess the implications of

modifying the agricultural landscape design in the composition of the beneficial insects

community.

101

METHODOLOGY

A systems approach (Tummala and Haynes 1977, Kitching 1983) was used to develop the

model in this study. This consisted of several steps beginning with the definition of the

problem and culminating with the validation of the model (Figure 32). The last step

however, was not accomplished at the present stage of development of the model. Since

most of the steps described in the methodological diagram (Figure 32) are self descriptive,

I will provide a detailed explanation of each step in the Results.

There are some terms, however, that require a definition in advance to understand the

diagram. They are:

a) Universe of Conpem. The system under study.

b) OQ'ect of control. The part of the system (universe of concern) that we want to
manage.

0) Associate environment. The part of the system that influences the object of control but
that will not be managed.

d) Stimuli or inputs. Factors of the object of control that affect the response of the
system.

e) Response or outputs. The responses to inputs/stimuli factors.

f) Sensitivity analysis. Test of the different variables of the model to see how changes in
those variables affect the output of the model.

9) Validation. Test of the model to assess the degree of accuracy and precision of its

predictions compared to real observations.

102

 

 

 

 

 

Delimitation of Pattern Analysis Chapters
Universe of Concern of Field Data 1 to 4

ass mama:

   

 

   
          
     
   
   
 

Delimitation of
Object of Control and .
Associated Environment

(24‘!!!
‘ .‘(4' flffi

 

   
 

.-z$$$$.’-.’-$’.’-:'- :5”? .-:-’ 559”" "3’13

 

Determination of
Input/Stimuli
Output/Results

 

 

Selection of j:

 

 

 

Statistical Analysis Chapters
of Field Data 3 and 4

 

 

     
   
   
 

    

Parameter
Estimation

 

 

 

 

Sensitivity g3?
analysis

 

 

 

 

Figure 32. Methodological approach followed during the development of the model
on interactions between the landscape and ladybird beetles.

103

RESULTS

1. Definition of the problem

As part of the KBS LTER project, a seven years of study (1989-1995) has revealed
patterns of habitat utilization by a complex of coccinellids in a variety of field crops (under
different management practices), perennial habitats, and secondary succession vegetation.
This information needed to be synthesized so it could be studied the response of

assemblages of coccinellid species to different landscape scenarios in agroecosystems.

2. Delimitation of the universe of concern
The universe of concern was the agricultural landscape comprised of field crops and
uncultivated habitats in which plants, herbivores, and beneficial insects interact among

themselves under the influence of weather and human management (Figure 30).

3. Object of control and associated environment

The objects of control were a complex of 13 ladybird beetle species and a group of six
habitat types (corn, wheat, alfalfa, soybean, poplar, and secondary succession). The
associated environment was defined by weather and human management. However, after
analyzing the universe of concern (Figure 30), an important consideration had to be made
with respect to the herbivore subsystem. It seems that by considering vegetation in the
landscape (producers) and a complex of predators (secondary consumers), the model
could have included the herbivore subsystem (primary consumers) as has been the case in

some tri-trophic models reported in the literature (Carter et al. 1982, Gutierrez and

104

Baumgaertner 1984). This approach, however, was not followed in this study because of

the following reasons.

a) The potential complexity that would be added to the model. Ladybird beetles are
generalist predators which have a wide range of prey preferences (Hodek 1967).
Moreover, some species can alternatively feed on plant products such as pollen and
extrafloral nectars (Pemberton et al. 1993).

b) The assumption made that, at the scale of this study, the habitat effect on ladybird
beetles includes the effect of the herbivore subsystem. This assumption is based on the
numerical response of coccinellids to insect outbreaks (Frazer and Raworth 1985).
Therefore if a habitat registers a high number of coccinellids, it is reasonable to expect

that prey availability is also high.

4. Determination of Stimuli-Results

The decision on the type of results expected from the model was made during earlier
stages in the model’s development. It was decided that results would include the temporal
and spatial distribution for each species within the ladybird beetle complex. The selection
of the stimuli that could determine the above results was further complicated because
there were multiple factors related to both the landscape and to the ladybird beetles
(Figure 33)

An important step toward the selection of the appropriate stimuli was the analysis of the
temporal and spatial scales in relation to population dynamics of coccinellids described in
Chapter 2. This analysis determined that a “year” was an appropriate temporal scale for

the model. This determination had important implications in the selection of the stimuli

105

since several of the potential factors that can affect the life system of coccinellids vary
within a season. Thus considering yearly abundance of coccinellids, by averaging within-

season factors, will reduce the complexity of the model.

Adult
mobility

Weather
thresholds

'Landscape

 

Figure 33. Factors involved in the interactions between beneficial insects and the landscape
in agroecosystems.

A model that focuses on across season variability can be affected by four factors: a)
habitat type, b) habitat age, c) habitat management, and d) habitat influence. These factors
were used as the stimulus inputs. They were provided to the model based on information
derived from the spatial configuration of the landscape, and the results were the temporal

and spatial distribution of the ladybird beetles (Figure 34).

106

 

 

  
 
   
    

   
 
  

 
 
   

Ladyblrd
beetle

 

 

Ladybird
beefles

 

 
    

Specles 13

 
 
 
  

apaaq
Pllq

K931 5:25:52;

 

 

 

 

 

 

 

Figure 34. Stimuli (inputs) and response (outputs) of a landscape simulation
model involving 13 species of ladybird beetles.

Habitat type was based on the six habitats chosen and manipulated in the KBS LTER
design: corn, wheat, soybean, alfalfa, poplar, and secondary succession. Habitat age was
based on habitat aging of perennial habitats and secondary succession. 3mg;
management was due to tillage and chemical input practices utilized on corn, soybean, and
wheat. Finally, habitat influence was based on the interaction of different habitats within

the landscape.

107

5. Selection of model type

A spatially explicit type of population model, which combines a population simulator with
a landscape map that characterizes the spatial distribution of the landscape features, best
fit the proposed objectives(Dunning et al. 1995). In this type of model, also know as
spatial mosaic model (Baker 1989), the landscape is divided into equal-sized grid cells
(Figure 35). Change in the abundance of each ladybird beetle species was modeled on an
individual-cell basis. Results (outputs) were in the form of computer files containing digital
maps of the spatial distribution of ladybird beetles. These files were accessible by
Geographic Information Systems (GIS) technology for further spatial analysis. An
important initial step in the implementation of this model is the selection of the grain or
resolution of the model i.e. “the size of the individual units of observation” (Wiens 1989).
Since the information needed to estimate parameters of population abundance was
obtained from one hectare field plots (chapters 3 and 4), this was chosen (1 ha.) as the

grain of this model.

6. Selection of computer technology

The computer model was coded using Quick Basic version 4.5. This language permitted a
structured program based on modules called by a main module. The model was
documented to facilitate an eventual translation to other computer language, such as C or
C++. Output maps from the model were produced in ASCII format ready to be accessed

by GIS software such as ARC INFO for additional spatial analyses.

108

 

Predator
Abundance
Species 1

        
  
    

 

 

 

TIME (i)

Species 9

  
 

2 3 4 5 6 7 8 91011 1213 14 15 161718 19 20212223242526 27
a, wwwmmmfim.

“mummmmwfiwt
$EEN§§§E§§1
afimflfiflfimfifl‘
Emmflflfifimfifi'
gflfimfimfimmm‘

 

 

 

 

 

 

 

 

Old Field

 

 

I _Woodtoot . i _ Woodloot

 

Figure 35. Spatially explicit population model in which temporal abundance of
predators is evaluated on a cell by cell basis based on the habitat status
in each cell for each time step (i).

109

7. Mathematical relationships
The abundance of each ladybird beetle species as a function of the stimuli (inputs) used in

the model is given by the following equation:

Abundance (m, Gen) = 10,000 x Habitat type (time, cell) x Habitat age (time, cc") x Habitat management (time, cc", x

Habitat influence (time, cell).

where time is in years, and each cell is one hectare with row and column coordinates (see
Figure 35)

Each of the habitat effects provide a value between 0-1 in which “1” is the best condition.
The model currently assumes that at the beginning of each year, there is a potential
population density of 10,000 beetles of each species per hectare (i.e. one beetle of each
species per m2). This constant supply of beetles each year was made by assuming that the
KBS LTER experimental area was surrounded by habitats that potentially could have
supplied habitats with beetles at the beginning of each year. While this assumption may be
acceptable for small landscapes, its validity will decrease as the size of the landscape in a
simulation increases. It is expected that fiirther analysis of field data will allow me to
account for population dynamics within a season. This information will provide a
mathematical relationship which in addition to the effect of the habitat factors, will
account for the density of ladybird beetles during the previous year to predict the

population density for the following year.

110

8. Parameter estimation

a) Habitat type

Determination of species abundance as a function of habitat type was designed with a
stochastic process based on probabilities of levels of abundance. The level of abundance
was coded as a range of values named with terms such as “very low” or “high”. The
determination of levels of abundance and the estimation of their respective probabilities for
each species in each habitat will be described in the following paragraphs and shown in

Figure 36, and Figure 37 .

Determination of levels of abundm.

C. septempunctata was selected as a model species because it had the broadest range of
abundance in the KBS LTER study. All the annual trap catches per treatment per year (49
combinations) were used for pairwise comparisons to obtain groups of means that were

not statistically different (P > 0.05). Prior to statistical analysis data were normalized

using a Vx +0001 transformation. Analysis of variance (ANOVA) and the Tukey-
HSD test were used to test for statistically significant differences among

treatments or years (SPSS 1993).

As a result of this analysis, 18 groups of means were obtained. Five of them were
selected to cover the possible range in values of abundance for the dominant
species: C. septempunctata and C.m. lengi. These five groups were classified as

very low, low, medium, high and very high (Figure 36).

111

Species model C. septempunctata

 

 

7 treatments 7 years
(com-soybean-wheat)
Conventional tilage Alalfa X
No tillage pom, 1989 - 1995
LOW chem. WM Sec. succession
Zero chem. hput

 

 

 

 

 

 

 

49
yearly means

Pairwise comparison

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Tukey-HSD

‘I.

1..

171
to 10‘ E
9 151 :
2 14‘ fi- : :
(D 1:1 1 1
§ 11< . :
. 11 : 1
5 101 2
Q g, t :
E at : 5
a .. _. : g

r :
g It ——: : 2
_ m ———.: : E
8 3‘ —I z

3'_:1 I: 3 2

t-—:: :: : 5

O 1:3:w....wrev.w:::.;.:.:w:.:v;vc.3.:.;.:.cficrrt

Ii'isz'p: Iiiétéfiiufltufli‘ua!

 

 

Mean Abundance

Selection of
Levels of Abundance

A‘A‘AA.A

 

 

IIIIIII
012 3 45 6
'oh

{3'}, Low Medium High v-ry High ‘33 Low mam H.
C. septempunctata Other
C. maculata lengi ladybird beetles

Figure 36. Determination of levels of abundance for ladybird beetle species based on

groups of abundance means of C. septempunctata after a pairwise
comparison (Tukey HSD ).

112

A different set of groups was selected for the other ladybird beetle species because their
range in abundance was lower than the dominant species. The values for all the 18 groups
that resulted fi'om the Tukey-HSD analysis, as well as the values selected for the different

levels of abundance, are shown in the Appendix A.

Estimation of probabilities for levels of abundance .

Maximum coccinellid abundance was selected from each of the following habitats: com
(1993), soybean (1994), wheat (1995), alfalfa, poplar, and secondary succession (1989-
1995). Then, I determined the frequency of each level of abundance. Frequencies were
converted to percentages, and the original values used for levels of abundance were
normalized to a scale of 0-100 (Figure 37). The resulting probabilities for abundance levels

per species in each habitat are shown in the Appendix B.

i b) Habitat age

The effect of habitat age was provided by perennial habitats (poplar and alfalfa) and
secondary succession in accordance with the analysis conducted in Chapter 3. For each
ladybird beetle species all annual means in each of the habitats were normalized to a scale
of 0 to 1; where l corresponded to the highest mean. The resulting estimations (expressed
in percentages) are shown in the Appendix C. For agronomic treatments (corn, soybean,

and wheat) values were always 1 since they did not provide any age effect (annual crops).

113

Species: C. septempunctata
Habitat: corn 1993

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Conventional No Low chemical Zero chemical staffing;
tillage tillage input input 0 ‘9
mean- 1.7 rrieant 2.3 mean-0.3 mean- 0,7 mean
I
Highest mean

// \\

Original replicates

7 V

 

1.4 .2

1 .6 1 .9 2.1 2. 7 4
Assignation
of values from
replicates to
levels of
abundance

A. A A A A A A L

 

 

 

 

 

 

 

 

 

 

 

 

 

     

A A A

00518.522.533.544.55.51503.577.58l50t510fls111t512

Very Low Medium High Very High

Q

Low

Detennlnatlon
of percentages
for levels of
‘ abundance

00.518522.533.544.555!8&577.5OI.500.91010.51111.512

    

 

Very Low Low Medium High Very High

Q

  

 

Normalization
of values (0400)
A A A A A A A A A A A A A A A A for levels 0'
osiotsnasouabasaounumnuunumabundance
Very Low Low Medium High VetyHigh
Abundance Levels

Figure 37 . Process of estimation of probabilities for levels of abundance using C.

septempunctata in corn as an example.

114

c) Habitat management

Habitat management is the influence of agronomic practices on corn, soybean, and wheat,
based on the analysis conducted in Chapter 4. Practices considered were conventional
tillage (also considered as high chemical input), no tillage, low chemical input, and zero
chemical input.

For each species all treatment means in each habitat were normalized to a scale of 0 to 1;
where 1 corresponded to the highest mean. The resulting estimations (expressed in
percentages) are shown in the Appendix D. Alfalfa, poplar, and secondary succession were

considered without agronomic management and were assigned a value of 0.

d) Habitat influence.

Habitat influence is considered a key factor in determining the spatial distribution of
ladybird beetles in the landscape. For example, the presence of deciduous habitats in the
landscape promotes the presence of species such as B. ursina and C. munda (Colunga-
Garcia et al. in review). Also, preliminary analyses showed that wheat may influence the
abundance of ladybird beetles in corn or soybean. The estimation of this parameter,
however, requires the determination of the ratio of influence for each habitat. For
example, it is necessary to know how far into the landscape a hectare of poplar can
influence the presence of B. ursina and C. munda in other habitats. Answers to this
question, will require elaborate initial assumptions which can be tested with subsequent

analysis based on simulation tests. This approach will be implemented in the future

115

development of the program after conducting sensitivity analyses of the model with

respect to the other three factors.

d) Probability of being in a crop.

This parameter was incorporated to account for factors not included in the model at its
current stage because some species are rarely present in annual crops, but when present,
they may occur in high numbers. This parameter was calculated by determining the
frequency (%) for species presence in a field crop habitat across the seven years of the

study. The resulting estimates are shown in Appendix E.

9. Programming

The structure of the computer code for the model consists of a main module which
includes a time controller and calls eight sub-modules (Figure 38). The computer code is
shown in Appendix F.

The first three modules set the initial condition of the model before proceeding with the
actual simulation. The Mble definition module defines the types of variables and sets the
dimensions of the arrays used in the model. The mmeterflccpiisition module accesses a
digital file that contains the numerical values of the parameters of the model (habitat type,

habitat age, habitat management, habitat influence, and habitat probability).

116

 

 

 

 

 

 

 

 

 

 

 

 

A HIM“!
,8,
U effect

 

   

 

 

 

 

«ls

 

 

 

   

initializationééé

 

 

 

 

 

 

 

 

 

A Time
.;, Edith! m
V controller .,,. m.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 38. Modular organization of a model that simulates the response of thirteen
species of ladybird beetles to landscape configurations.

117

The landscape initialization module creates temporary files for landscape structure to be
used in the simulation process.

The following modules analyze information about landscape structure, estimate insect
abundance and produce the output files. The time controller allows the model to run the
simulation for variable number of years. The landscape configpration module reads an
external digital file (provided by the user) which contains information on X,Y coordinates,
habitat type, habitat age, and habitat management for each cell in the landscape. The
coccinellid abundance module estimates the abundance of each species in each cell based
on the information obtained in the landscape configuration module. The species divers_ity
apalysis module computes means, variances, and standard errors for the total abundance of
ladybird beetles, species richness, and the Shannon Wiener index in the whole landscape.
The species spatial distribution module produces ASCII files that contain information on
the spatial distribution of each species per time unit (year) and that can be accessed by a
Geographic Information System software such as ARC INFO. Finally, the landscape
gauge module, “updates” the landscape configuration for the following year of the
simulation based on management criteria. For example, this module can be manipulated so
it follows a corn-soybean-wheat rotation and alfalfa is replanted every four years. By
modifying the input file used in the landscape configuration module, or changing the
management criteria in the landscape change module the user can produce different

landscape configurations to be used as “treatments” in the simulation.

118

10. Sensitivity analysis

Three approaches were used to conduct the sensitivity analysis of the model. The first
approach was to compare species assemblages between observed (field observations) and
expected (model output) data. Yearly ladybird beetle total abundance, species richness,
and species diversity for the entire KBS LTER landscape was estimated for the years
1989-1994. Then a computer simulation was conducted using the KBS LTER landscape
configuration based on the same habitats and planting sequence. The simulation was
replicated 10 times for each year to obtain yearly means of total abundance, species
richness and species diversity. Two tests were used to compare between the observed data
and the model output: a) the Kolmogorov-Smirnov goodness of fit test, which is a more
powerful test than the chi-square test when n is small (Zar 1984), and b) the Pearson

product-moment correlation test (SPSS 1993).

There was no statistical difference (P > 0.1) between the distribution of the observed data
and the output of the model for the three variables measured (Figure 39). High correlation
was found with species diversity (r=0.8, P = 0.056), and total abundance (r=0.77, P =

0.075). No correlation was detected with species richness.

119

 

‘LO
8 ‘E .
c 0.11: “Si: x”
(U eiies .‘ef’Q 4’
U Tigoc‘ \\ g
c: \ ,’
3 0.6 -l \\ I,
:2 \\‘{a l
a, 0.4 -
.2
iii
6 0.2 " Model ""‘0'

Fielddata _.—

n:

0.0 I T I I

 

 

 

1989 1990 1991 1992 1993 1994

 

 

 

 

 

 

 

(D
a — a
t: % 1‘3
5 3 2
E G)
2 4 “=
+ 1
1 - Model ......0
Field date _.—
o I I I I o
1989 1990 1991 1992 1993 1994
‘12 (L8
.é‘ *3
a '5 U
(D 1’ 3;
.2 g a»
D I:
0'2 ‘ 2:11am :2;
011 t r t 1 0

 

 

 

1989 1990 1991 1992 1993 1994

Time

Figure 39. Comparison between model output and field data in relation to total relative
abundance, species richness, and species diversity of ladybird beetles.

120

The results of the simulation and subsequent analysis in relation to total abundance and
species diversity were considered acceptable because the major disagreement between the
model output and the field data was with respect to observations made in 1989.
Coccinellid abundance during 1989 was highly influenced by the abundance of C.
septempunctata in the whole landscape and in particular that in wheat. Subsequent
observations during the following years have lead me to the hypothesis that the high
abundance observed in 1989 was the exception rather than the rule.

Another observation of this analysis, was that both richness and diversity were
overestimated in all years of simulation. An analysis of species abundance as a result of the
simulation showed that this overestimation occurred because the “very low interval” for
the non-dominant species (Appendix A) may have been too broad, increasing the
probability for a species to be present in the landscape. Ifthat probability is too high, then
species richness and species diversity will be overestimated, although the trend would

remain the same.

The second approach was an analysis of variance of species abundance under different
agronomic treatments (in field crops) and habitat aging (in perennial and succession
habitats). For the agronomic treatments, I designed an “experimental landscape” for each
of the field crops (corn, soybean, and wheat) consisting of 25 replicates of each of the
following agronomic treatments: conventional tillage, no tillage, low chemical input, and
zero chemical input. This information was used to conduct a simulation and the mean

abundance of each ladybird beetle species per crop per treatment was estimated in each

121

simulation. Data were normalized using a “(+0001 transformation for statistical
analysis. The simulation in each crop was replicated 10 times and the means of the
10 simulations were used for the analysis of variance (ANOVA) and the Tukey-
HSD test to test for statistically significant differences among agronomic
treatments (SPSS 1993). For treatments on habitat aging in perennial and successional
habitats, a same approach that was implemented for agronomic treatments was used. The
difference was that instead of agronomic treatments I included different ages (from 1 to 6
years) as treatments. The number of simulations, data normalization, and statistical tests
were also the same.

The results of the statistical analysis conducted on field data (see Table 3, p. 63 for habitat
aging, and Tables 8, 10, and 12, p. 89-94 for agronomic treatments), were compared with
the results of the statistical analyses conducted on the outputs of the model.

The combination of species and treatments produced 234 pairwise comparisons for
agronomic treatments and 182 for aging treatments. Of these pairwise comparisons, 15%
in agronomic treatments and 35% in perennial treatments produced significant statistical
differences when the same comparisons conducted on field data did not produce
significant statistical differences. A revision of the structure of the model indicated that
those cases in which there was a statistical analysis conflict between data from the model
and field data, were cases in which pair of treatments had field data with means that
differed in 50% or more, but were not significantly different (P > 0.05). For example, field
data on C.m. lengi in soybeans under conventional and no tillage had a mean (i S.E.) of
0.06 i 0.03 adults per trap and 0.22 i 0.04 adults per trap respectively. In this example,

the abundance in conventional tillage is 1/3 less than the abundance with no tillage but

122

differences were not statistically significant because of the high degree of variability in
conventional tillage. However, after the proportional differences between treatments were
included in the model, the estimates of abundance for Cm. lengi as a result of the
simulation showed statistically significance. An easy solution would be to assign the same
value to the parameters of both treatments since they are not statistical different. This way,
the simulation would produce similar values and there would not be conflict with the
statistical analysis. I think however that the solution to this problem will require a more
detailed analysis of the variance that occur within each treatment.
The third approach was to hypothesize, based on analysis of field data, the potential
response of ladybird beetle assemblages to changes in complexity in the landscape, and
observe if such predictions could be supported by the model.
Three ways were selected to increase the complexity of the landscape such that the
abundance and diversity of ladybird beetles might be affected: a) increasing the numbers of
habitats in the landscape (Chapter 1), b) changing management practices to reduce
chemical inputs (Chapter 4), and c) causing periodical disturbances to revert secondary
succession fields to earlier stages of succession (Chapter 3).
The “experimental design” implemented in the model used a landscape whose dimensions
were 10*10 hectares, and a simulation length of 9 years. Specifics of treatments were as
follows.
1) Modifying complexity by increasing the number of habitats. The model consisted of a
corn-soybean-wheat rotation scheme under high chemical input with two treatments

a) The entire area planted to corn at the beginning of the simulation

123

b) Half of the area planted to corn, and the other half planted to wheat at the
beginning of the simulation.
1) Modifying complexity by changing management practices. It consisted of a corn-
soybean-wheat rotation scheme with two treatments:
a) The entire area planted to corn under high chemical input at the beginning of
the simulation.
b) The entire area planted to corn under zero chemical input at the beginning of
the simulation.
1) Modifying complexity by managing disturbance in succession. It consisted of a
secondary succession habitat with two treatments
a) The entire area started with a succession field of age 1 and succession
continued undisturbed.
b) One third of the area had secondary succession of age 1,one third of age 2,
and one third of age 3. After a succession field reached four years of age, fields
were “plowed” and reverted to age 1.
All simulations were replicated ten times.
Based on the data analyses conducted in previous chapters, it was predicted that the
increase in complexity would cause:
0 a stabilization of diversity and number of beetles in 1) and 3),
0 an increase in the fluctuation in number of beetles and a decrease of diversity in

2)

124

As a measure of stabilization or fluctuation the coefficient of variation (CV) was used
such that the lower the CV. the greater the stability, while the higher the CV. the greater

the change the change in abundance existed.

The results of these simulations are shown in the Figure 40, Figure 41, and Figure 42.
The observed trends occurred as predicted. An increase in landscape complexity occurred
by increasing the number of crops, showed a reduction in the coefficient of variation while
maintaining the same mean with respect to the one-crop treatment (Table 16). Changes in
landscape complexity by reductions in chemical inputs increased the coefficient of
variation of abundance and richness, and reduced the diversity. Finally, The change in
landscape complexity by managing secondary succession habitats reduced the coeflicient
of variation and increased the diversity of coccinellids in the landscape. At the current
state of the model I could not test the statistical significance of some the outcomes (see
discussion on the second approach for sensitivity analysis), however, I expect to be able to
do that after tuning some of the parameters in the model. Nevertheless, I considered the

fact that trends in model simulation followed patterns observed in the field.

125

Table 16. Mean and coefiicient of variation (c.v.)(%) of the abundance, species richness
and species diversity of a complex of ladybird beetles described by a simulation
model in response to different landscape configuration treatments. Number of
crops and chemical input treatments were applied on a rotation scheme of corn-
soybean-wheat.

 

Treatment Abundance Richness Diversity

mean C.V. mean C.V. mean C.V.

 

 

Number of initial crops

one (com) 5070 10 4.3 12 0.9 27
two (corn, wheat) 5090 5 4.3 6 0.9 14
Chemical inputs

high 5070 10 4.3 12 0.9 27
zero 4400 63 3.8 30 0.6 33

Secondary succession

 

 

 

natural 5410 36 6.7 7 0.8 47
managed 4780 16 ' 6.6 3 1.1 10
DISCUSSION

This resulting model during this first iteration produced trends in species diversity and
abundance of ladybird beetles that followed the temporal patterns observed in the field.
Simulation results are promising because of the potential for comparative analysis of
studies that involve the interactions between species assemblages of ladybird beetles and

the landscape.

126

 

Relative Abundance

 

 

 

 

Richness

 

 

 

 

Diversity

 

0.2 -t Onguop ...ou.
Twocrops+

0.0 I I I H I l I
1 2 3 4 5 6 7 8 9

 

 

 

Time (years)

Figure 40. Changes in abundance, species richness, and diversity of a complex of ladybird
beetle species predicted by a simulation model on a corn-soybean—wheat
rotation scheme in response to increasing from one initial crop (corn) to two

initial crops (corn and wheat).

127

 

 

 

 

 

 

 

 

 

 

 

 

 

Q)
U
C
(U
U
C
55
.3
E
0
DC [I Zorohput---O---
D Hidihput+
1 2 3 4 5 6 7 8 9
(D
U)
Q)
C
.5
CK
1 q Zqohput...on.
Highinput+
o I I I f l I I
1 2 3 4 5 6 7 8 9
:E‘
f
.2
D
0.2 - Zeroinputu-On-
ngnlnput+
0.0 I I I I l I r

 

 

1 2 3 4 5 6 7 8 9
Time (years)
Figure 41. Changes in abundance, species richness, and diversity of a complex of

ladybird beetle species predicted by a simulation model on a corn-soybean—
wheat rotation scheme in response to reduction on chemical inputs.

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8
C
(U
'U
C
é
Q1
.2
E
0)
CE
1 2 3 4 5 s 7 a 9
U)
(D
a)
C
.C
.9
I
1~ 3:23;”;2;
0 I I I I I I '
1 2 3 4 5 6 7 8 9
1.4 ”xii
1'2M
3.2: 1.0 ," \
(D ’1 \
E 0.87,] \G
._ D \‘x
D 3.29 so---a---e.---a
0.2 _ 3.31:“. "'0'”
0.0 I I I I ' I I

Time (years)

Figure 42. Changes in abundance, species richness, and diversity of a complex of ladybird
beetle species predicted by a simulation model on secondary succession habitats
in response to changes from natural succession to a managed succession
(different succession ages, 1 to 3, and plowing after the four year).

129

Some discrepancies were observed between the results of statistical tests applied to field
data and the results of tests applied to estimates of abundance of beetles produced by the
model. This shows the need for conducting more detailed analyses on the variance in
abundance observed by some species of ladybird beetles in response to treatments.

Although the trends in species diversity and species richness produced by the model were
similar to those observed in the field data, they were overestimated. This indicates the

2

need to fine tune the “very low” and “low’ intervals of abundance.
Once the sensitivity analysis of the habitat type, habitat age, and habitat management
parameters produces satisfactory results, it will be possible to start the next iteration in the
development of the model which includes estimates of the habitat influence parameter.
This will be very important because the model currently simulates only temporal patterns,
even though it has also been designed to produce spatial patterns.

The next step will be to validate the model. This process can be achieved using primarily
new data collected at the KBS LTER. Alternatively, the model can also be validated by
conducting experiments that test the predictions of the model. Because, the design of the
model is directed toward comparative analysis, predictions of the model for field
validations should be comparative by nature.

Another important step in the near future will be to expand the temporal scope of the
model to include intra-seasonal variability. Important information in the simulation process
will be gained by dividing a season into early, middle and late season. The implications that
such subdivisions will have on the abundance and species diversity of ladybird beetles will

be important since C. septempunctata, the dominant species, is a mid-season species while

the rest of the species are more abundant early or late in the season.

130

In the future the model could be adapted to account for other groups of predators such as
lacewings or ground beetles. These groups form part of the KBS LTER project (Gage et
al. 1993) and several years of data have been collected following the same general
approach as that used for the ladybird beetles. A model including these groups will

produce a more holistic view on use of the landscape by generalist predators.

SUMMARY AND CONCLUSIONS

The Long Term Ecological Research (LTER) program for agricultural ecology located at
the Kellogg Biological Station (KBS) stated as a general hypothesis that “agronomic
management based on ecological concepts can effectively substitute for reliance on
chemical subsidies in production-level cropping systems (VanCleve and Martin 1991). In
the experimental design an array of field crops (corn, wheat, soybean) under different
management practices were interspersed with perennial biomass plantations (alfalfa and

poplar) and secondary succession habitats making a complex agricultural landscape.

A long term program designed to characterize the flow of organisms within the KBS
LTER landscape was established in 1988 (Gage et al. 1993). This program focused
primarily on insect predators because of their role in the natural regulation of pest
populations in agroecosystems. Among the predator groups, ladybird beetles were
selected to measure the effect that agroecological management of field crop systems could
have on the population dynamics of beneficial insects. Ladybird beetles were ideal for this
study because they have a large number of species with a wide range of habitat utilization

and they are very mobile organisms.

In 1992, I joined the KBS LTER program and decided to conduct my research within the

context of the existing KBS LTER ladybird beetles project. The goal of my research was

to obtain and analyze habitat utilization patterns by a complex of 14 species of

131

132

coccinellids, and to synthesize the results in a landscape-coccinellid model. To achieve this

goal I conducted

a)

b)

d)

a study of the life system of Coleomegilla maculata lengi to characterize their use of
tree related habitats during the winter, and their patterns of field crops use during the
summer.

an assessment of the temporal and spatial scales required to conduct long term studies
of coccinellids, as well as an evaluation of sticky traps as a sampling tool for ladybird
beetles.

an analysis of the response of coccinellid species assemblages to two types of temporal
diversity in the landscape. One was a six-year corn-soybean rotation, and the second
was the habitat aging of alfalfa, poplar, and early secondary succession.

an analysis of the response of coccinellid species assemblages to agricultural
management in a corn-soybean-wheat rotation with practices that included
conventional and no tillage under high chemical inputs, as well as low and zero
chemical inputs.

the first iteration of a spatially explicit population model that will be used as a tool for

comparative analysis in studies involving landscape - ladybird beetle interactions.

133

The major findings in this study were:

A

. Aging of deciduous habitats such as poplar plantations causes a “succession” of

dominant species within the ladybird beetle complex. Dominant beetle species in young
plantations (1-4 years) are also important inhabitants of field crops. Thereafter, the
dominant ladybird beetle species are species typical of only deciduous habitats.
Ladybird beetles in secondary succession reach their peak abundance and species
diversity between the second and fourth years of plant succession. Ecological
management of agroecosystems that involve succession habitats may require some
practices aimed to maintain those fields within earlier stages of succession.

Reducing chemical inputs in field crops have different effects on ladybird beetle species
depending on the crop. Abundance and species diversity are significantly reduced in
corn when chemical inputs (herbicides and fertilizers) are reduced, but in wheat the
reduction of chemical inputs increases abundance and diversity of beetles.
Corn-soybean rotation schemes can affect the abundance of ladybird beetle species.
Cm. lengi populations become abundant during “corn years” but decrease their
abundance during “soybean years”.

Ladybird beetles can use woodlots, hedgerows, and rows of trees as aggregation sites
during the winter as long as they have a southern exposure to sunlight, and are not
subject to disturbance. Large aggregates of beetles may be located in woodlots or
hedgerows near agricultural fields where corn or alfalfa was grown the previous

summer.

134

6. Spring flowers are important sources of food for ladybird beetles that aggregate during
the winter in woodlots, hedgerows, or individual trees. Therefore, weeds such as
dandelions could be controlled when they are within agricultural fields but should be
preserved when they are near tree related habitats.

7. Habitat preference by ladybird beetles in field crop agroecosystems vary through the
summer. Cm. lengi populations prefer alfalfa or wheat during spring but switch its
preference to corn during late summer.

8. Sticky traps prove to be the appropriate sampling method for large temporal and
spatial scale measurements of adult ladybird beetles in diversified landscapes.

9. The applicability of the systems analysis used during the development of the model
provided a framework to organize the approach to landscape level research on insect

dynamics.

Within the framework of the KBS LTER general hypothesis, this work contributed to the
understanding of the potential impact that agroecological management can have on
beneficial insects. It also completed an important phase of a long term project initiated in
1988 on insect predators in the landscape. Finally, this work provides a framework for
future research on beneficial insect and their interactions with the landscape.

Future research can be scaled down to address specific issues related with the fluctuation
in abundance and diversity of beneficial insects in some field crop management practices.
More effort is also needed to determine why ladybird beetles are attracted to the early

years of secondary succession. Research can also scale up to study patterns of response by

135

ladybird beetles to difl°erent landscape configurations by comparing diversified landscapes
versus simple landscapes.

The model in this research could be adapted to account for other groups of predators such
as lacewings or ground beetles to produce a more holistic view of the interactions between

beneficial insects and agricultural landscapes.

APPENDICES

136

Appendix A. Minimum and maximum values of 18 groups of abundance means of C.
septempunctata that resulted from a pairwise comparison (Tukey-HSD) of 49
treatments. Numbers in bold are the intervals selected for levels of abundance of
ladybird beetles during the development of a simulation model.

 

 

Interval Level of abundance
Group Minimum Maximum C. septempunctata Rest of ladybird
C.m. lengi beetles
1 0 0.83 very low very low
2 0.06 1.06 low
3 0.25 1.32
4 0.41 2.28 low
5 0.65 2.37
6 0.71 2.79
7 0.83 3.04 medium
8 1 .32 3.81
9 1 .71 4.55 medium
10 2.28 5.02
1 1 2.31 5.19
12 2.37 5.66
13 2.79 6.35 high
14 3.61 7.35
15 3.81 7.69 high
16 4.03 8.47
17 6.1 1 10.63
18 7.10 11.85 very high

 

137

 

 

 

 

 

 

 

 

 

 

o o o o o o o o o o o o 8 :9: 25>
o o o o o o o o o 9 o 0 so :2...
o o o o o o o 2 c on 2 o o 528.2 88808
cm 8 cm om om om cm 8 om mu me 9 o 26.. 628$
om om cm 8 cm cm 8 9. 8 mm 2. mm o 33 bm>
o o o o o o o o o o o o 8. :9... a;
o o o o o o o o 2: o o o o :9:
o o o o o o o o o o 9 as o 528.2 $.32
om o 8 o o om om om o 8 3 8 o 26..
om o om o 0 cm om on 0 cm 3. cm o 26.. 30>
o o o o o o o o o o o E o E: 28>
o o o o o o 2: o o mm o 9 8 :9:
o o o o o o 0 mm o 8 o o t 52.8.2 .280.
8 8 cm o o om o om on mm on o o 26.. .
om om om o o om o mm om : 8 o o 3.: am>
o o o o o o o o o o o o mm :9: 95
o o o o o o o o o o o o S :9:
o o o o o o o o 0 mm o o 3 528.2 $22,
on cm on S 8 cm. om cm on cm 8 on o 26..
om om om om om om om om om mm om om o 26.. a;
o o o o o o o o o o o o B 5% a;
o o o o o o o o o o o o mm :9:
o o o o o o o o o o o o o 528.2 88.3w
om om o o om om om om om cm 8 cm o 25..
8 cm 0 o 8 om om. om om 8 om on o 33 a;
o o o o o o o o o o o o o :9: has
o o o o o o o o o o o 3 we 5.:
o o o o o o o o 2: o o a. 3 228.2 58
8 o 8 o 0 cm S on o om om mu 3 26..
on o om o o 8 cm om o 8 om o o 33 28>
:81 3.9... :25: m r: 5m? “:50 _&:m :92 97.0 9.an 5qu 0680 so oocmncanm “83¢...
Co
$6on 03an 2333 .05..
6.69968

.Iu>=ooI new .m.:mBmB dinosmx 6636:: .Iuaucsr £6.28» «rump: 663036. .<n_nm_< 9.283 €08.50 65%:
.mufiSm 66623.35 .<u:a_n< .6593 0.1628 6ng .0"?an 6.66526ch .IuEmaI 6:2 .EdnomEo .mwmtequoKom

duno .Eefi>wooooaow no.5 2o... s $352 E8226 5 88on 2.03 2563 .8 3.695% Co m_o>m_ Co 33 32.338.“ .m 5969?.

 

 

 

138

 

 

 

 

on 2: 0 SF 8F o 8F 2 o mF 2 cm mF F
on o o o o o F S o E mm m 2: o
F o o o o O on S 0 cm mm m on m
E o 02 o o o o mF ow m 8 mm mm v 83883
F o o o 0 8F o S o F om 2: m m 5288
8F c o o o o 0 8F 0 SF SF 8 E m
on o o o o o o mF 2: F on m 3 F
F o 2 o 0 SF 8 F o F mm m 3. m
2 o o o o o o S o 8 SF SF on F.
o o o o o o o F o F F om om m «:32
o o o o o o 0 SF 2: mm F 8 8 m
o o o o o o o E m 8 cm on 2: F
o 2: o o o o 2: o 9. mm F F o F
o o 8 o o 0 cm o 2: 8 on F m o
o o F o o o m m mm mm 8 F F m
o o 2: o o o 0 on ma om om 2 F v 538
o o o o 0 8F c mm F 2: on 2: m m
o o o o o 2 o 2: F mm o 8 2 m
2: o o o o o o mm 8 m 8F F 2: F
:8: 85: :25: 2: 3m? :5 7.5m 5:? 38 2520 ES... 850 B 3.83 3ng
883m 283 23.63 mm<

 

..m:m9o>:8 .Iu>coo_._ 95 £383 diuowaz

.8932: Ensues: £33... fitumFI .mFszoBmF .<u_nm_< 6onqu .quEFo .35: .mu_w5m .396:an .<u:a_n< .258

.ouazmo .mbtaE .OuucsEo 6.60522mq .IuEaaI <92 .E.o uowEo .mFmFozaququm .Ouno 62980355 no.6 Em: E
:o_mmooo:m raucoomm new $352 .3523 Fe £265 8 moan 228% 8 8.0me 253 253m. 3 $$ 8593.3 @2581 .0 zucoaq<

 

139

 

 

 

 

 

 

 

o 8 o o o o 2: o 0 on SF SF SF .2. .520 23
0 mm o o 0 2: o o o 2: 3 on no .2. .520 26..
o mm o o 09 o o o o mm mm 9 on 092:. 02 69.3
00F 09 0 SF 0 09 o o c on cm 09 om mug; .250
mm o o o mm o 8F 8.. o mm om om oo .9: .620 EmN
o o o o o o o o 0 2: cm 02 om .2. .520 25..
mm 03 o o o o o om o ov 09 mu 8.. mm_m__._. oz cmmn>om
o o o o 09 o o on 09 ov mm om mm 32.... .250
o o o o o o ow o o o o F on .9: .szo SmN
o o o o o o o o o om E oF mF .2. .520 25..
o o o o o o mm 09 c on 09 mm 2: 36E. 02 500
o o o o o SF SF 03 o 03 o 03 mm $3.... .280
cooI 02m: :95... m _.I 526. ES .35 .554. azwo 2.an Ema: 02.0 no chEmmmcms. Fauna:
8.0QO mzmmn 25.63
5:09.958

.Iu>cooI new 6.5.83 diuomax .8332: .quucax 5.3.5.: ...IumFI 6330.59 .<u.nm_< .3283 ...ouFEo 62.5:
.mu_m.:m .Smtzaai .<u:a_a< .mEaFFm duazmo .mbcaE .0"?an 6.352....ng .IuEmaI .523 .EduowEo .mFmFocqumiom
.Onso .62.; new ...mogom .58 E 30:35 2.22.89 EEoEc .695 3.0on 253 £3.62 Fo @ov 8:353 3:201 .0 x659?

 

140

 

 

om om on or or on 0.. CV om om cor 009 Dow «005$
cm or o O on or 0.. om 0.. co co cow cow cmmn>ow
or o 0.. o c or on on or on cm 00—. 00—. 500

 

 

:8... 3.2.. 225x 2.. .22 2.0 .930 8.2 97.0 22.0 52.: 8&0 F0 5.2...

00.00% 0:000 0.3.63

 

620902.00 .Iu>cooI 0:0 6.00.003. 6.1505:

65:02: .Iu:0:=I 6.003.... 0.IumFI 6.03030. .<u_00_< 63.800 ...ouFEo 62.6.: .mu_m._:m 600.0:an

.<n=0_n< 650.00 .Oumzmo 60:05 dun—SEQ 6.6052800 .InEwaI .500. .EOuomEo 6.0.0::0E0i00
6qu 68056000050 00.0 0.0: .0356 5 E0005 00 2 00.0000 0:000 2.0.60. .8 3:332... .m 50:009..

 

141

Appendix F. Computer code of the computer simulation model to simulate the effect of different
habitat configurations on the relative abundance of 13 species of coccinellids

 

MAIN PROGRAM

REM $INCLUDE: 'c:\LANDYBUG\LANDYBUG.DCL'

'LANDYBUG: A Landscape - Coccinellids model to simulate the effect of
' the landscape configuration in the species assemblages of
' predatory coccinellids.

' Developed by Manuel Colunga-Garcia

' This model simulates the effect of difi‘erent habitat configurations

' on the relative abundance of 13 species of coccinellids.

' Field crop habitats (corn, soybean, and wheat) are subject to different

' agronomic practices (conventional tillage, no till, low chemical input,

' zero chemical input). Poplar, alfalfa, and and secondary succession are

subject to difierent years of maturation.
'**************************************************************

' The statement [$INCLUDE: 'LANDYBUGDCL'] at the beginning of this program
' declares subprograms, functions, variables, and arrays to be used during
' the simulation.

' Accessing the module that contain the parameters to acquire them.
CALL Parameters

' Accessing the module that initializes temporal landscape files.
CALL Landini

' Turning the random generator on. It uses the computer clock as a seed.
RANDOMIZE TIMER:

' Begins the simulation using a [Foerext] loop as a time controller.
FOR year = 1 TO NoYears

' Determination of the landscape configuration.
CALL Landiconf

' Estimation of the abundance of coccinellids.
CALL CoccAbund

' Analysis of coccinellid species diversity.
CALL Analysis

' Generation of species distribution output maps.
'CALL MapOutput

' Changeing the landscape for the following year.
CALL Landichange

NEXT year

'**************************************************************

END

142

1. VARIABLES DEFINITION

DECLARE SUB MapOutput ()

DECLARE SUB CoccAbund O

DECLARE SUB Landichange 0

DECLARE SUB Analysis 0

DECLARE SUB Landini 0

DECLARE FUNCTION Degree! (hab%, spp%)
DECLARE SUB Landiconf 0

DECLARE SUB Parameters ()

'*********************************************************************
' Declaration of and Initialization of Variables and Arrays:

' Directories and data files (Variables)

DIM SHARED DataProg AS STRING: ' directory for the landybug program
DIM SHARED DataDir AS STRING: ' directory for data files (*.dat)

DIM SHARED SimuFile AS STRING: ' name of temporal simulation file

'directory for the landybug program
LET DataProg = "c:\landybug\"

'directory for data files (*.dat)
LET DataDir = "c:\landybug\mydata\"

' Related to habitat parameters

' (Variables)

DIM SHARED NoHab AS INTEGER: ' habitats considered in the model

DIM SHARED NoAge AS INTEGER: ' maximum age for perennial habitats

DIM SHARED NoLevel AS INTEGER: ' levels of species abundance

DIM SHARED NoMngt AS INTEGER: ' number of management practices for crops
DIM SHARED NoSpp AS INTEGER: ' species considered in the model

' habitats considered in the model
LET NoHab = 7

' maximum age for perennial habitats
LET NoAge = 7

' levels of species abundance
LET NoLevel = 5

' number of management practices for crops
LET NoMngt = 4

' species considered in the model
LET NoSpp = 13

' (Array5)

DIM SHARED HabType(0 TO NoHab, 1 TO NoSpp, 1 TO NoLevel) AS INTEGER
DIM SHARED HabAge(0 TO NoHab, 1 TO NoSpp, 1 TO NoAge) AS SINGLE
DIM SHARED Haangt(O TO NoHab, 1 TO NoSpp, 0 TO NoMngt) AS SINGLE
DIM SHARED HabInf(O TO NoHab, 1 TO NoSpp, 1 TO NoHab) AS SINGLE

143

DIM SHARED HabProb(0 TO NoHab, I TO NoSpp) AS SINGLE

' Landscape map dimensions (Variables)

DIM SHARED FirstRow AS INTEGER: 'first row in the landscape map
DIM SHARED LastRow AS INTEGER: 'last row in the landscape map
DIM SHARED FirstCol AS INTEGER: 'first column in the landscape map
DIM SHARED LastCol AS INTEGER: ' last column in the landscape map

' first row in the landscape map
LET FirstRow = l

' last row in the landscape map
LET LastRow = 10

' first colurrm in the landscape map
LET FirstCol = l

' last column in the landscape map
LET LastCol = 10

' Time controller (Variables)

DIM SHARED NoYears AS INTEGER: ' number of years for the simulation
DIM SHARED year AS INTEGER: ' year counter

LET NoYears = 2

' For Diversity Analysis (Arrays)

DIM SHARED SppDensity(FirstRow TO LastRow, F irstCol TO LastCol, 1 TO NoSpp) AS INTEGER
DIM SHARED Total(FirstRow TO LastRow, FirstCol TO LastCol) AS INTEGER: ' Total of coccinellids
DIM SHARED Richness(FirstRow TO LastRow, FirstCol TO LastCol) AS INTEGER: 'Richness (No.
Speci65)

DIM SHARED Diversity(FirstRow TO LastRow, FirstCol TO LastCol) AS SINGLE: ' Diversity index
(Shannon)

DIM SHARED Mtotal(l TO NoYears) AS SINGLE: ' Mean

DIM SHARED Vtotal(l TO NoYears) AS SINGLE: ' Variance

DIM SHARED SEtotal(1 TO NoYears) AS SINGLE: ' Standard error

DIM SHARED Mrichness(l TO NoYears) AS SINGLE: ' Mean

DIM SHARED Vrichness(l TO NoYears) AS SINGLE: ' Variance

DIM SHARED SErichness(l TO NoYears) AS SINGLE: 'Standard error

DIM SHARED Mdiversity(l TO NoYears) AS SINGLE: ' Mean

DIM SHARED Vdiversity(l TO NoYears) AS SINGLE: 'Variance

DIM SHARED SEdiversity(l TO NoYears) AS SINGLE: 'Standard error

' Others
' (Variable)
DIM SHARED Ratio AS INTEGER: ' ratio of influence in the landscape

' ratio of influence in the landscape
LET Ratio = 1000

' (Amy-a)

DIM SHARED HabCell(FirstRow TO LastRow, FirstCol TO LastCol) AS INTEGER

DIM SHARED MngtCell(FirstRow TO LastRow, FirstCol TO LastCol) AS INTEGER

DIM SHARED AgeCell(FirstRow TO LastRow, FirstCol TO LastCol) AS INTEGER

DIM SHARED HabAround(FirstRow TO LastRow, FirstCol TO LastCol, 1 TO NoHab) AS SINGLE

144

DIM SHARED HabTotal(0 TO NoHab) AS INTEGER
DIM SHARED HabProp(O TO NoHab) AS SINGLE
DIM SHARED Around(0 TO NoHab) AS INTEGER

2. PARAMETERS ACQUISITION

SUB Parameters
'This module reads parameters from tables in the file LANDYBUGPAR.

'Open the file to input parameters
OPEN DataProg + "landybugpar" FOR INPUT AS #1

'*************************************************

' Description of habitat type codes

'corn = l

' soybean = 2

' wheat = 3

' orchard = 4 (information not determined yet)
' poplar = 5

' alfalfa = 6

' succession = 7

'lnput parameters for level abundance according to habitat type.
LINE INPUT #1, Title$
LINE INPUT #1, LabelsS
FOR hab% = 1 TO NoHab
FOR level% = 1 TO NoLevel
INPUT #1, Valuel, Value2
IF Valuel <> hab% OR Value2 <> level% THEN
PRINT "Error: Table "; Title$2 PRINT "Check the sequence of parameters"
END
END IF
FOR spp% = 1 TO NoSpp
INPUT #1, HabType(hab%, spp%, level%)
NEXT spp%
NEXT level%
NEXT hab%

'************************************************

'lnput parameters for habitat age effect on abundance (years)
LINE INPUT #1, Title$
LINE INPUT #1, Labels$
FOR hab% = 1 TO NoHab
FOR age% = 1 TO NoAge
INPUT #1, Valuel, Value2
IF Valuel <> hab% OR Value2 <> age% THEN
PRINT "Error: Table "; Title$z PRINT "Check the sequence of parameters"
END
END IF
FOR spp% = 1 TO NoSpp

145

INPUT #1, HabAge(hab%, spp%, age%)
NEXT spp%
NEXT age%
NEXT hab%

'************************************************
' Description of management type codes

' no management = 0 (or management effect not determined)
' conventional tillage = 1 (also high chemical input)
' no tillage = 2

' low chemical input = 3
' zero chemical input = 4

'lnput parameters for habitat management effect on abundance.
LINE INPUT #1, Title$
LINE INPUT #1, Labels$
FOR hab% = 1 TO NoHab
FOR mngt% = 0 TO NoMngt
INPUT #1, Valuel, Value2
IF Valuel <> hab% OR Value2 <> mngt% THEN
PRINT "Error: Table "; Title$z PRINT "Check the sequence of parameters"
END
END IF
FOR spp% = 1 TO NoSpp
INPUT #1, Haangt(hab%, spp%, mngt%)
NEXT spp%
NEXT mngt%
NEXT hab%

'************************************************

'lnput parameters for inter-habitat influence on abundance.
LINE INPUT #1, Title$
LINE INPUT #1, Labels$
FOR hab% = 1 TO NoHab
FOR hab2% = 1 TO NoHab
INPUT #1, Valuel, Value2
IF Valuel <> hab% OR Value2 <> hab2% THEN
PRINT "Error: Table "; Title$z PRINT "Check the sequence of parameters"
END
END IF
FOR spp% = 1 TO NoSpp
INPUT #1, Hablnf(hab%, spp%, hab2%)
NEXT spp%
NEXT hab2%
NEXT hab%

'**************************************#*********

'lnput parameters of probability for a species to be in a field crop habitat.
LINE INPUT #1, Title$
LINE INPUT #1, Labels$
FOR hab% = 0 TO NoHab
INPUT #1, Valuel
IF Valuel <> hab% THEN
PRINT "Error: Table "; Title$z PRINT "Check the sequence of parameters"

146

END
END IF
FOR spp% = 1 TO NoSpp
INPUT #1, HabProb(hab%, spp%)
NEXT spp%
NEXT hab%

'************************************************

CLOSE #1: ' closes the file

END SUB

3. LANDSCAPE INITIALIZATION

SUB Landini

'************************************************

'Checking the data directory.
CLS
LOCATE 5, l: PRINT "Is < "; DataDir; " > the directory for your data files? <Y> <N>";
DO
INPUT Response$
IF UCASEMResponseS) = "N" THEN
LOCATE 7, 1: PRINT "You have to correct the directory address of the program"
LOCATE 8, 1: PRINT " in the file < landybugdcl >"
LOCATE 10, l: PRINT "The program has ended"
END
ELSEIF UCASE$(Response$) = "Y" THEN
EXIT DO
ELSE
END IF
LOOP

'************************************************

'lnput of the data file for the simulation
CLS
DO
LOCATE 3, l: FILES DataDir + "*.dat"
PRINT : PRINT "Input the data file name for this simulation (without the extension '.dat')"
INPUT FileS
PRINT : PRINT "Is "; Files; ".dat"; " the correct name? <Y> <N>";
INPUT Responses
IF UCASE$(Response$) = "Y" THEN
EXIT DO
ELSEIF UCASE$(Response$) = "N" THEN
CLS : LOCATE l, l: PRINT "TRY AGAIN"
ELSE
END IF
LOOP

147

'************************************************

' Creating the temporal file <Lanbug01> for the year =1 with information on
' landscape structure contained in the original data file.
CLS
SimuFile = "LanbugOI"
OPEN DataDir + FileS + ".dat" FOR INPUT AS #2
OPEN DataDir + SimuFile + ".dat" FOR OUTPUT AS #3
DO
LINE INPUT #2, DataS
PRINT #3, DataS
LOOP UNTIL EOF(2)
CLOSE #2
CLOSE #3

'************************************************

END SUB

4. LANDSCAPE CONFIGURATION (INPUT)

SUB Landiconf

'************************************************

’lnput of habitat data in the temporal file Lanbug??.dat.
OPEN DataDir + SimuFile + ".dat" FOR INPUT AS #4

LINE INPUT #4, Title$
LINE INPUT #4, Label$
FOR row% = FirstRow TO LastRow
FOR col% = F irstCol TO LastCol
INPUT #4, Valuel: 'row coordenate in the file
IF Valuel <> row% THEN PRINT "Error row"
INPUT #4, Value2: 'column coordinate in the file
IF Value2 <> col% THEN PRINT "Error column"
INPUT #4, HabCell(row%, col%): 'habitat type
INPUT #4, MngtCell(row%, col%): 'management type
INPUT #4, AgeCell(row%, col%): ' habitat age
NEXT col%
NEXT row%
CLOSE

'************************************************

'Habitat proportion in the landscape

'Reseting values to 0

FOR habitat% = 1 TO NoHab
HabTotal(habitat%) = 0

NEXT habitat%

AllCells% = 0

148

'Calculating the amount of cells for each habitat type (absolute)
FOR row% = FirstRow TO LastRow
FOR col% = FirstCol TO LastCol
habitat% = HabCell(row%, col%)
HabTotal(habitat%) = HabTotal(habitat%) + l
AllCells% = AllCells% + 1
NEXT col%
NEXT row%

'Calculating the proportion of cell for each habitat type (relative)
FOR habitat% = 1 TO NoHab

HabProp(habitat%) = HabTotal(habitat%) / AllCells% * 100
NEXT habitat%

EXIT SUB: ' This instruction skips the next part of the module. It should

be eliminated once habitat influence is ready for this program.

'***************************************llHIIIIIIIHIHIHIIIIHII
'Habitat influence

FOR row% = FirstRow TO LastRow
FOR col% = F irstCol TO LastCol
left% = row% - Ratio: right% = row% + Ratio
upper% = col% - Ratio: lower% = col% + Ratio

'Reseting values to 0

FOR habitat% = 1 TO NoHab
Around(habitat%) = 0

NEXT habitat%

AllCells% = 0

FOR windrow% = left% TO right%
FOR windcol% = upper% TO lower%
IF windrow% >= FirstRow AND windrow% <= LastRow THEN
IF windcol% >= FirstCol AND windcol% <= LastCol THEN
'Calculating the amount of cells for
' each habitat type (absolute)
habitat% = HabCell(windrow%, windcol%)
Around(habitat%) = Around(habitat%) + 1
AllCells% = AllCells% + 1
END IF
END IF
NEXT windcol%
NEXT windrow%
habitat% = HabCell(row%, col%)
Around(habitat%) = Around(habitat%) - l
AllCells% = AllCells% + 1

'Calculating the proportion of cell for each habitat type (relative)
FOR habitat% = 1 TO NoHab

HabAround(row%, col%, habitat%) = Around(habitat%) / AllCells%
NEXT habitat%

NEXT col%
NEXT row%

149

'************************************************

END SUB

5. COCCINELLID ABUNDANCE ESTIMATION
SUB CoccAbund
' Description of species code
'C. septempunctata = l C. maculata = 2
'H. parenthesis = 3 C. munda = 4
'C. stigma = 5 A. bipunctata = 6
'B.ursina = 7 C. trifasciata = 8
'A. labiculata = 9 H. tredecimpunctata =10
'H. undulata = 11 H. glacialis = 12
'H. convergens = 13

'************************************************

FOR row% = FirstRow TO LastRow
FOR col% = FirstCol TO LastCol
hab% = HabCell(row%, col%)
FOR spp% = I TO NoSpp
'*#********************************
'Determine if the species is in a habitat based on its
'probability of being in a habitat.
Value% = INT((100 - 0 +1)* RND + 0)
IF Value% > HabProb(hab%, spp%) THEN
Abundance4 = 0
ELSEIF Value% <= HabProb(hab%, spp%) THEN
' ***********************************
'Initial value of a cell (1 ha) considering one beetle per n12.
Abundance = 10000

'Times habitat type efl'ect
Degree(hab%, spp%) is a module located in other part of the program
Abundance] = Abundance * Degree(hab%, spp%)

'Times management effect
mngt% = MngtCell(row%, col%)
Abundance2 = Abundance] * Haangt(hab%, spp%, mngt%)

'Times age effect
age% = AgeCell(row%, col%)
Abundance3 = Abundance2 * HabAge(hab%, spp%, age%)

'Times habitat influence effect. (N 0 determined yet)
Abundance4 = Abundance3

' *********************************

ENDIF

' Final asignation of abundance to each species

150

SppDensity(row°/o, col%, spp%) = Abundance4
NEXT spp%
NEXT col%
NEXT row%

'************************************************

END SUB

FUNCTION Degree (hab%, spp%)

'Probability for different levels of abundance for the habitat (hab%)
' and the species (spp%)

Levell% = HabType(hab%, spp%, l)

Leve12% = HabType(hab%, spp%, 2)

Level3% = HabType(hab%, spp%, 3)

Level4% = HabType(hab%, spp%, 4)

Leve15% = HabType(hab%, spp%, 5)

'Determines if the sum of levels of abundance is > 0. Otherwise

'the abundance is O and exits this function

IF Levell% + Leve12% + Level3% + Level4% + Leve15% = 0 THEN
Degree = 0
EXIT FUNCTION

END IF

'********************************#*************

'Generates a random number between 0 and 100
Value% = INT((IOO - 0 +1)* RND + 0)

IFspp%>= l ANDspp%<=2TI-IEN
' For species C. septempunctata and C. maculata
'Lower and Upper values of abundance per each abundance level
L1% = 0: U1% = 7
L2% = 3: U2% =19
L3% = 14: U3% = 38
L4% = 32: U4% = 64
L5% = 59: U5% = 100

SELECT CASE Value%
CASE IS <= Levell%
Abundance = INT((U1% - Ll% +1)* RND + Ll%)
CASE (Levell% + 1) TO Leve12%
Abundance = INT((U2% - L2% + l) * RND + L2%)
CASE (Leve12% + 1)TO Level3%
Abundance = INT((U3% - L3% + l) * RND + L3%)
CASE (Level3% + 1) TO Level4%
Abundance = INT((U4% - L4% + l) * RND + L4%)
CASE (Level4% + 1) TO Leve15%
Abundance = INT((U5% - L5% + 1) * RND + L5%)
CASE ELSE
END SELECT

151

ELSEIF spp% >= 3 AND spp% <= 13 THEN
' For the rest of the species
'Lower and Upper values of abundance per each abundance level
L1% = 0: U1% = 7
L2% = 1: U2% = 9
L3% = 7: U3% = 25
L4% = 23: U4% = 53

SELECT CASE Value%
CASE IS <= Levell%
Abundance = INT((U1% - Ll% + 1) * RND + Ll%)
CASE (Levell% + 1) TO Leve12%
Abundance = INT((U2% - L2% + 1) * RND + L2%)
CASE (Leve12% + 1) TO Level3%
Abundance = INT((U3% - L3% + 1) * RND + L3%)
CASE (Level3% + I) TO Level4%
Abundance = INT((U4% - L4% + 1) * RND + L4%)
CASE (Level4% + 1) TO Leve15%
Abundance = INT((U5% - L5% + 1) * RND + L5%)
CASE ELSE
END SELECT
END IF

Degree = Abundance / 100

END FUNCTION

6. SPECIES DIVERSITY ANALYSIS

SUB Analysis

Determination of the number of cells
Num = (LastRow - FirstRow + 1) * (LastCol - FirstCol + 1)

'III*****III*IIHII**************************************

'Calculation of total adults per cell (all species)
FOR row% = F irstRow TO LastRow
FOR col% = F irstCol TO LastCol
Tot = 0
FOR spp% = 1 TO NoSpp
Tot = Tot + SppDensity(row%, col%, spp%)
NEXT spp%
Total(row%, col%) = Tot
NEXT col%
NEXT row%

'Estimation of mean, variance, and standard error.
T = 0: T2 = 0
FOR row% = FirstRow TO LastRow
FOR col% = FirstCol TO LastCol
T = T + Total(row%, col%)

152

T2 = T2 + (T otal(row%, col%) " 2)
NEXT col%
NEXT row%
Mtotal(year) = T / Num
Vtotal(year) = (T2 - ((T A 2) / Num)) / (Num - l)
SEtotal(year) = SQR(Vtotal(year)) / Num

'************************************************

'Calculation of no. species (richness).
FOR row% = FirstRow TO LastRow
FOR col% = F irstCol TO LastCol
rich = 0
FOR spp% = 1 TO NoSpp
IF SppDensity(row%, col%, spp%) > 0 THEN
rich = rich + 1
END IF
NEXT spp%
Richness(row%, col%) = rich
NEXT col%
NEXT row%

'Estimation of mean, variance, and standard error.
R = 0: R2 = 0
FOR row% = FirstRow TO LastRow
FOR col% = FirstCol TO LastCol
R = R + Richness(row%, col%)
R2 = R2 + (Richness(row%, col%) " 2)
NEXT col%
NEXT row%
Mrichness(year) = R / Num
Vrichness(year) = (R2 - ((R " 2) /Num)) / (Num - 1)
SErichness(year) = SQR(Vrichness(year)) / Num

'************************************************

'Calculation of diversity index
FOR row% = FirstRow TO LastRow
FOR col% = FirstCol TO LastCol
div = 0
FOR spp% = 1 TO NoSpp
IF Total(row%, col%) > 0 THEN
prop = SppDensity(row%, col%, spp%) / Total(row%, col%)
ELSE
prop = 0
END IF
IF prop > 0 THEN
logprop = LOG(prop)
ELSE
logprop = 0
END IF
div = div + (prop * logprop)
NEXT spp%
Diversity(row%, col%) = (—l) * div
NEXT col%

153

NEXT row%

'Estimation of mean, variance, and standard error.
D = 0: D2 = 0
FOR row% = FirstRow TO LastRow
FOR col% = F irstCol TO LastCol
D = D + Diversity(row%, col%)
D2 = D2 + (Diversity(row%, col%) A 2)
NEXT col%
NEXT row%
Mdiversity(year) = D / Num
Vdiversity(year) = (D2 - ((D A 2) / Num)) / (Num - 1)
SEdiversity(year) = SQR(Vdiversity(year)) / Num

'************************************************

' Print results in the screen.

FormatyS = "Year (##)": Formatt$ = " Total = #######"
Formatr$ = " Richness = #.#": Formatd$ = " Diversity = #.##”
PRINT USING FormatyS; year;

PRINT USING FormattS; Mtotal(year);

PRINT USING Formatr$; Mn'chness(year);

PRINT USING Formatds; Mdiversity(year)

'********t************#**************************

END SUB

7. SPECIES SPATIAL DISTRIBUTION (OUTPUT)
SUB MapOutput

FOR spp% = 1 TO NoSpp

' Creating the file for the species/year
IFspp%<10ANDyear<10THEN

MapFile$ = "spO" + LTRIM$(STR$(spp%)) + "_0" + LTRIM$(STR$(year))
ELSEIF spp% < 10 AND year >= 10 THEN

MapFileS = "sp0" + LTRIM$(STR$(spp%)) + "_" + LTRIM$(STR$(year))
ELSEIF spp% >= 10 AND year < 10 THEN

MapFile$ = "sp" + LTRIM$(STR$(spp%)) + "_0" + LTRIM$(STR$(year))
ELSEIF spp% >= 10 AND year >= 10 THEN

MapFileS = "sp" + LTRIM$(STR$(spp%)) + "_" + LTRIM$(STR$(year))
ELSE
PRINT "Error"
END IF
OPEN DataDir + MapFileS + ".txt" FOR OUTPUT AS #5

'Header for an ascii file to be imported by Arclnfo
PRINT #5, "NCOLS "; (LastCol - FirstCol + l)
PRINT #5, "NROWS "; (LastRow - FirstRow + 1)
PRINT #5, "XLLCORNER "; l

PRINT #5, "YLLCORNER "; l

PRINT #5, "CELLSIZE "; 30

154

'Print ASCII data by row and columns
FOR row% = F irstRow TO LastRow
FOR col% = FirstCol TO LastCol
PRINT #5, SppDensity(row%, col%, spp%);
NEXT col%
PRINT #5, ""
NEXT row%

CLOSE #5
NEXT spp%

END SUB

8. LANDSCAPE CHANGE

SUB Landichange

'Skips subrutine if the years for simulation are complete
IF year = NoYears THEN

KILL DataDir + SimuFile + ".dat"

EXIT SUB
END IF

'4!ill**********IIHIHF*********************************

' Creating the temporal file for the following years
' with information on landscape structure
IF (year + 1) < 10 THEN

NewFileS = "LanbugO" + LTRIM$(STR$(year + 1))
ELSE

NewFileS = "Lanbug" + LTRIM$(STR$(year + 1))
END IF
OldFile$ = SimuFile

' Read landscape information from OItFile and updates to a NewFile.
OPEN DataDir + OldFile$ + ".dat" FOR INPUT AS #6
OPEN DataDir + NewFile$ + ".dat" FOR OUTPUT AS #7

LINE INPUT #6, Title$z PRINT #7, Title$
LINE INPUT #6, Label$z PRINT #7, LabelS

FOR row% = F irstRow TO LastRow
FOR col% = FirstCol TO LastCol

INPUT #6, RowCell%: 'row coordenate in the file
IF RowCell% 0 row% THEN PRINT "Error row"
INPUT #6, ColCell%: 'column coordinate in the file
IF ColCell% 0 col% THEN PRINT "Error column"
INPUT #6, HabCell%: 'habitat type
INPUT #6, MngtCell%: 'management type
INPUT #6, AgeCell%: ' habitat age

155

' Updating information according with the current habitat/age
SELECT CASE HabCell%
CASE IS = 1: ' Corn
HabCell% = 2: ' rotates to soybean
CASE IS = 2: ' Soybean
HabCell% = 3: 'rotates to wheat
CASE IS = 3: 'Wheat
HabCell% = 1: ' rotates to corn
CASE IS = 5: ' Poplar
IF AgeCell% < 7 THEN
AgeCell% = AgeCell% + 1
ELSE
AgeCell% = 7
END IF
CASE IS = 6: 'Alfalfa
IF AgeCell% < 5 THEN
AgeCell% = AgeCell% + l
ELSE
AgeCell% = 1
END IF
CASE IS = 7: 'Succession
IF AgeCell% < 6 THEN
AgeCell% = AgeCell% + l
ELSE
AgeCell% = 6
END IF
CASE ELSE
END SELECT

'Writing in new file the updated information
WRITE #7, RowCell%, ColCell%, HabCell%, MngtCell%, AgeCell%
NEXT col%
NEXT row%

CLOSE #6
CLOSE #7

"I"!*IIUIHII*****1H"?III********II"!***********************

'Renarning the temporal file with the new name
SimuFile = NewFile$

KILL DataDir + OldFile$ + ".da "

END SUB

BIBLIOGRAPHY

Allen, T.F.H. and TB. Starr. 1982. Hierarchy: perspectives for ecological diversity.
University of Chicago Press, Illinois.

Allen, T.F.H. and T.W. Hoekstra. 1992. Toward a unified ecology. Columbia University
Press. New York.

Altieri, M.A. 1987. Agroecology: the scientific basis of alternative agriculture. Westview
press, London.

1994. Biodiversity and pest management in agroecosystems. Food Products Press, New
York. .

Andow, D.A. and SJ. Risch. 1985. Predation in diversified agroecosystems: relations
between a coccinellid predator Coleomegilla maculata and its food. J. Appl. Ecol.
22: 357-372.

Andow, D.A. 1991. Vegetational diversity and arthropod population response. Annu.
Rev. Entomol. 36: 561-586.

Baars, M.A. 1979. Catches in pitfall traps in relation to mean densities of carabid beetles.
Oecologia (Bed) 41: 25.46.

Baker, W.L. 1989. A review of models of landscape change. Landscape Ecology 2: 111-
133.

Barret, G.W. 1992. Landscape ecology: designing sustainable agricultural landscapes. J.
Sust. Agriculture 2:83-103

Benton, A.H. and A.J. Crump. 1979. Observations on aggregation and overwintering in
the coccinellid beetle Coleomegilla maculata (DeGeer). J .N.Y. Entomol. Soc. 87:
154-159.

1981. Observations on the spring and summer behavior of the 12-spotted ladybird beetle,
Coleomegilla maculata (DeGeer) (Coleoptera: Coccinellidae). J.N.Y. Entomol.
Soc. 89: 102-108.

Berryman, A.A. 1991a. Population theory: an essential ingredient in pest prediction,
management, and policy-making. Am. Entomol. 37 : 138-142.

1991b. Good theories and premature models. Am. Entomol. 37: 202-204.

Bird, G.W., T.C. Edens, F. Drummond, and E. Groden. 1990. Design of pest
management systems for sustainable agriculture, pp 55-110. In C. Francis, C.B.
Flora and LB. King [eds], Sustainable Agriculture in Temperate Zones. Wiley
InterScience, New York.

Bird, G.W. 1994. Integrated pest management in a sustainable agriculture. Proceedings of
the second national integrated pest management symposium/workshop. Las Vegas,
Nevada. April 19-22, pp 35-41.

Boutton, T.W., M.A. Arshad, and LL. Tieszen. 1983. Stable isotope analysis of
termite food habits in East African grasslands. Oecologia (Berlin) 59: 1-6

Brust, GB. and B.R. Stinner. 1991. Crop rotation for insects, plant pathogen, and weed
control. pp 217-236. In Pimentel, D. [ed.] Handbook of pest management in
agriculture, CRC Press, Boca Raton, Florida.

Bugg, RL. and Dutcher, J.D. 1993. Sesbania exaltata (rafinesque-Schmaltz) Cory
(Fabaceae) as a warm-season cover crop in pecan orchards: effects on
aphidophagous Coccinellidae and pecan ahids. Biol. Agric. Hortic. 9: 215-229.

156

157

Bullock, J.A. 1971. The investigation of samples containing many species. 11. Sample
comparison. Biol. J. Linn. Soc. 3: 23-56.

Burel F. 1989. Landscape structure effects on carabid beetles spatial patterns in western
France. Landscape Ecology 2: 215-226.

Callahan, J.T. 1984. Long-term ecological research. Bioscience 34: 363-3 67.

Carter, N., A.F.G. Dixon, and R. Rabbinger. 1982. Cereal aphid populations: biology,
simulation and prediction. PUDOC, Wageningen.

Cartwright, B., R.D. Eikenbary, and G.W. Angalet. 1982. Parasitism by Perilitus
coccinellae (Hym.: Braconidae) of indigenous coccinellid hosts and the introduced
Coccinella septempunctata (Col: Coccinellidae), with notes on winter mortality.
Entomophaga 27: 237-244.

Claridge, M.F. 1991. Genetic and biological diversity of insect pests and their natural
enemies, pp. 183-194. In Hawksworth, D.L. [ed.], The Biodiversity of
microorganisms and invertebrates: its role in sustainable agriculture. CAB
International, Oxon UK.

Clark, M. S., J. M. Luna, N. D. Stone and R. R. Youngman. 1993. Habitat
preferences of generalist predators in reduced-tillage corn. J. Entomol. Sci. 28:
404-416.

Coderre, D. L. Provencher, and J. Champagne. 1989. Effect of intercropping maize-
beans on aphids and aphidophagous insects in corn fields of southern Quebec,
Canada. Acta Phytopathol. Entomol. Hung. 24: 59-63.

Colunga-Garcia, M., S.H. Gage, and D.A. Landis. (in press). The response of an
assemblage of Coccinellidae (Coleoptera) to a diverse agricultural landscape.
Environ. Entomol.

Conrad, M.S. 1959. The spotted lady beetle, Coleomegilla maculata (De Geer), as a
predator of European corn borer eggs. J. Econ. Entomol. 52: 843-847.

Copeland, PJ. and RK. Crookston. 1992. Crop sequence affects nutrient composition
of corn and soybean grown under high fertility. Agronomy Journal 84: 503-509.

Crews, T.E., C.L. Mohler, and A.G. Power. 1991. Energetics and ecosystem integrity:
the defining principles of sustainable agriculture. Am. J. Alternative Agric. 6: 146-
149.

De la Court, T. 1992. Critique of the dominant development paradigm. Development 2:
42-45.

Dean, G.J. 1982. Phenology of aphidophagous predators. Ann. Appl. Biol. 101: 182-184.

Desender, K., M. Dufrene, M. Loreau, M.L. Luff, and J.P. Maelfait [eds.]. 1994.
Carabid beetles, ecology and evolution. Kluwer Academic Publishers, Netherlands.

Di Castri, F. and A.J. Hansen. 1992. The environment and the development crises as
determinants of landscape dynamics, pp. 3-18. In Hansen, A.J. and F. di Castri
[eds], Landscape boundaries: consequences for biotic diversity and ecological
flows. Springer-Verlag, New York.

Doak, D.F., P.C. Marina, and RM. Kareiva. 1992. Spatial scale mediates the influence
of habitat fragmentation on dispersal success: implications for conservation. Theor.
Pop. Biol. 41: 315-336.

158

Dover, J.W. 1989. The conservation of insects on arable farmland, pp. 294-318 In
Collins, NM. and J .A. Thomas [eds], The conservation of insects and their
habitats, Academic Press, London, England.

Dowell, RV. and RH. Cherry. 1981. Survey traps for parasitoids, and coccinellid
predators of the citrus blackfly, Aleurocanthus woglumi. Entomologia
Experimentalis et Applicata 29: 356-361.

Dunning, J.B. Jr., D.J. Stewart, B.J. Danielson, B.R. Noon, T.L. Root, R.H.
Lamberson. 1995. Spatially explicit population models: current forms and future
uses. Ecological Applications 5: 3-11.

Edwards, CA. 1990. The importance of integration in sustainable agricultural systems,
pp. 249-264. In Edwards, C.A., L.Rattan, P. Madden, R.H. Miller, and G. House
[eds], Sustainable agricultural systems. Soil and Water Conservation Soc.
Ankeny, IA.

Elliot, N.C. and RW. Kieckhefer. 1990. Dynamics of aphidophagous coccinellid
assemblages in small grain fields in eastern South Dakota. Environ. Entomol. 19:
1320-1329.

Elliot, N.C., R.W. Kieckhefer, and W.C. Kauffman. 1991. Estimating adult coccinellid
populations in wheat fields by removal, sweepnet, and visual count sampling. Can.
Entomol. 123: 13-22.

Evans, E.W. and N.N. Youssef. 1992. Numerical response of aphid predators to varying
prey density among Utah alfalfa fields. J. Kan. Entomol. Soc. 65: 30-38.

Ewert, M.A. and H.C. Chiang. 1966. Dispersal of three species of coccinellids in corn
fields. Can. Entomol. 98: 999-1003.

Fahrig, L. 1991. Simulation methods for developing general landscape-level hypotheses
of single-species dynamics, pp. 417-442. In Turner, MG. and RH. Gardner [eds],
Quantitative methods in landscape ecology: the analysis and interpretation of
landscape heterogeneity. Springer-Verlag, New York.

Ferran, A., P. Gubanti, G. Iperti, A. Migeon, and J. Onillon. 1989. La repartition
spatiale des différents stades de Coccinella septempunctata dans un champ de blé: '
variation au cours de la saison. Entomol. Exp. Appl. 53: 229-236.

Ferran, A., G. Iperti, L. Lapchin, and J.M. Rabasse. 1991. La localisation, 1e
comportement et les relations <<proie-prédateur>> chez Coccinella
septempunctata dans un champ de blé. Entomophaga, 36: 213-225.

Ferran, A. and F.G. Dixon. 1993. Foraging behavior of ladybird larvae (Coleoptera:
Coccinellidae) Eur. J. Entomol. 90: 383-402.

Flint, M.L. and RA. Roberts. 1988. Using crop diversity to manage pest problems:
some California examples. Am. J. Alternative Agriculture 3: 163-167.

Folse, L.J., J.M. Packard, and WE. Grant. 1989. AI modeling of animal movements in
a heterogeneous habitat. Ecol. Modeling 46: 57-72.

Forman, R.T.T. and Godron, M. 1986. Landscape Ecology. John Wiley & Sons. New
York.

Franklin, J.F., C.S. Bledsoe, and J.T. Callahan. 1990. Contributions of the Long-Term
Ecological Research program. Bioscience 40: 509-523.

159

Frazer, B.D. and D.A. Raworth. 1985. Sampling for adult coccinellids and their
numerical response to strawberry aphids (Coleoptera: Coccinellidae; Homoptera:
Aphididae). Can. Entomol. 117: 153-161.

Frazer, B.D. 1987. Predators, pp. 83-93. In Minks, AK. and P. Harrewijn [eds], Aphids,
their biology, natural enemies and control. Vol. B. Elsevier Sci. Pub., Amsterdam.

Fry, B. and BB. Sherr. 1984. d13 C Measurements as indicators of carbon flow in marine
and freshwater ecosystems. Contrib. Marine Sci. 27 : 45-47.

Gage, S.H. and EL. Russell.l987. Pest surveillance systems in the USA--a case study
using the Michigan crop monitoring system (CCMS), pp. 209-224. In Teng, P.S.
[ed], Crop loss assessment and pest management St. Paul, Minn. APS Press.

Gage, S.H., T. Wirth and G.A. Simmons. 1990. Predicting regional gypsy moth
population trend and risk using pheromone trap catch and geographical
information systems. Environ. Entomol. 19: 370-377.

Gage, S.H., M. Colunga, and P.0strom. 1993. Flow of insects through a landscape.
Abstracts, LTER all scientist meeting, Estes Park, Colorado.

Galecka, B. 1992. Importance of forest islands for inhabiting of agricultural landscape by
coccinellids (Coccinellidae, Coleoptera). EKOL. POL. POLISH J .-ECOL. 39:
561-577.

Getz, W.M. and A.P. Gutierrez. 1982. A perspective on systems analysis in crop
production and insect pest management. Annu. Rev. Entomol. 27: 447-466.
Gilpin, M., G.A.E. Gall, and D.S. Woodruff. 1992. Ecological dynamics and

agricultural landscapes. Agric. Ecosystems Environ. 42: 27-52.

Golden Software, Inc. 1990. Surfer version 4. Reference manual. Colorado, USA.

Gordon, RD. 1985. The Coccinellidae (Coleoptera) of America North of Mexico. J .N.Y.
Entomol. Soc. 93: 1-912.

Green, RH. 1980. Multivariate approaches in ecology: the assessment of ecologic
similarity. Ann. Rev. Ecol. Syst. 11: 1-14.

Grevstad, ES. and B.W. Klepetka. 1992. The influence of plant architecture on the
foraging efficiencies of a suite of ladybird beetles feeding on aphids. Oecologia
(Berl.) 92: 399-404.

Groden, E., F.A. Drummond, R.A. Casagrande, and D.L. Haynes. 1990.
Coleomegilla maculata (Coleoptera: Coccinellidae): its predation upon the
colorado potato beetle (Coleoptera: Chrysomelidae) and its incidence in potatoes
and surrounding crops. J. Econ. Entomol. 83: 1306-1315.

Gutierrez, A.P. and J.U. Baumgaertner. 1984. A realistic model of plant-herbivore-
parasitoid-predator interactions. Can. Entomol. 116: 933-949.

Hagen, KS. 1962. Biology and ecology of predaceous coccinellidae. Annu. Rev.
Entomol. 62289-326.

Hargrove, W.W. and J. Pickering. 1992. Pseudoreplication: a sine qua non for regional
ecology. Landscape Ecology 6: 251-258.

Harrigan, P., J.C. Zieman, and S.A. Macko. 1989. The base of nutritional support for
the gray snapper (Lutjanus griseus): an evaluation based on a combined stomach
content and stable isotope analysis. Bull. Marine Sci. 44: 65-77.

160

Hecht, S.B. 1987. The evolution of agroecological thought. In Altieri, M.A. [ed.],
Agroecology: the scientific basis of alternative agriculture. Westview press,
London.

Hilner, HR. 1985. Leadership in conservation tillage, pp. 15-20. In D’Itri, F .M. [ed.] A
systems approach to conservation tillage, Lewis Pub. Chelsea, MI. USA.

Hodek, I. 1967. Bionomics and ecology of predaceous coccinellidae. Annu. Rev.
Entomol. 6: 289-326.

1970. Coccinellids and the modern pest management. Bioscience 20: 543-552.

1973. Biology of Coccinellidae. Junk, The Hague/ Academia, Prague.

Hodek, I., G. Iperti, and M. Hodkova. 1993. Long-distance flights in Coccinellidae
(Coleoptera). Eur. J. Entomology. 90: 403-441.

Honek, A. 1982. The distribution of overwintered Coccinella septempunctata L. (Col,
Coccinellidae) adults in agricultural crops. Z. Ang. But. 94: 311-319.

1985a. Habitat preferences of aphidophagous coccinellids (Coleoptera). Entomophaga 30:
253-264.

Honek, A. and F. Kocourek. 1986. The flight of aphid predators to a light trap: possible
interpretations, pp. 333-344. In Hodek, 1. [ed.] Ecology of aphidophaga,
Academica Prague, Junk Pub. Czechoslovakia

Hyman, J.B., J.B. McAninch, and D.L. DeAngelis. 1991. An individual-based
simulation model of herbivore in a heterogeneous landscape. In Turner, MG. and
RH. Gardner [eds], Quantitative methods in landscape ecology: the analysis and
interpretation of landscape heterogeneity. Springer-Verlag, New York, pp. 417-
442.

Idris, AB. and E. Grafius. 1995. Wildflowers as nectar sources for Diadegma insulare
(I-Iymenoptera: Ichneumonidae), a parasitoid of diamondback moth (Lepidoptera:
Yponomeutidae).Environ. Entomol. 24: 1726-1735.

Iperti, G., L. Lapchin, A. Ferran, J.-M. Rabasse, and J.-P. Lyon. 1988. Sequential
sampling of adult Coccinella septempunctata L. in wheat fields. Can. Entomol.
120: 773-778.

Ives, RM. 1981. Estimation of coccinellid numbers and movement in the field. Can.
Entomol. 113: 981-997.

James, EC. and CE. McCulloch. 1985. Data analysis and the design of experiments in
ornithology. In Johnston, RF. [ed], Current Ornithology 2: 1-63. New York
Plenum.

1990. Multivariate analysis in ecology and systematics: panacea or pandora box? Annu.
Rev. Ecol. Syst. 21: 129-166.

Jeffers, J.N.R. 1978. An introduction to system analysis: with ecological applications.
University Park Press, London.

Kadar, F. and G.L. Loevei. 1992. Light trapping of carabids (Coleoptera: Carabidae) in
an apple orchard in Hungary. Acta Phytopathol. Entomol. Hung. 37: 343-348.

Kareiva, P. and R. Perry. 1989. Leaf overlap and the ability of ladybird beetles to search
among plants. Ecol. Entomol. 14: 127-129.

Kells, J.J. and W.T. Meggitt. 1985. Conservation tillage and weed control. pp. 123-130.
In D’Itri, FM. [ed] A systems approach to conservation tillage, Lewis Pub.
Chelsea, MI. USA.

161

Kendall, M.G. 1955. Rank correlation methods. Hafirer Pub. Co. New York.

Kieckhefer, RW. and N.C. Elliot. 1990. A 13-year survey of the aphidophagous
coccinellidae in maize fields in eastern South Dakota. Can. Entomol. 122: 579-
581.

Kieckhefer, R.W., N.C. Elliot, and D.A. Beck. 1992. Aphidophagous coccinellids in
alfalfa, small grains, and maize in eastern South Dakota. Great Lakes Entomol. 25:
15-23.

Kirby, H.W. 1985. Conservation tillage and plant disease. pp. 131-136. In D’Itri, F.M.
[ed] A systems approach to conservation tillage, Lewis Pub. Chelsea, MI. USA.

Kitching, RL. 1983. An introduction to ecological modelling. University Queensland
Press, Melbourne.

Kogan, M. [ed.]. 1986. Ecological theory and integrated pest management practice. John
Wiley & Sons. New York.

Koskinen, W.C. and CG. McWhorter. 1986. Weed control in conservation tillage.
Journal of Soil and Water Conservation 4]: 365-3 70.

Lapchin, L., A. Ferran, G. Iperti, J.M. Rabase, and J.P. Lyon. 1987. Coccinellids
(Coleoptera: Coccinellidae) and syrphids (Diptera: Syrphidae) as predators of
aphids in cereal crops: a comparison of sampling methods. Can. Entomol. 119:
815-822.

Lester, D.G. and W.L. Morrill. 1989. Activity density of ground beetles (Coleoptera:
Carabidae) in alfalfa and sainfoin. J. Agric. Entomol. 6: 71-76.

Levin, S.A. 1992. The problem of pattern and scale in ecology. Ecology 73: 1943-1967.

Levins, R. 1986. Perspectives in integrated pest management: from an industrial to an
ecological model of pest management, pp. 1-18. In Kogan, M. [ed.]. Ecological
theory and integrated pest management practice. John Wiley & Sons New York.

Levins, R. And M. Wilson. 1980. Ecological theory and pest management. Annu. Rev.
Entomol. 25: 287-308.

Luff, M.L. and I.P.Woiwod. 1995. Insects as indicators of land-use change:a european
perspective, focusing on moth and ground beetles, pp. 399-422. In Harrington, R.
and NE. Stork [eds] Insects in a changing environment, Academic Press, London.

Mack, T.P. and Z. Smilowitz. 1979. Diel activity of green peach aphid predators as
indexed by sticky traps. Environ. Entomol. 8: 7 99-801.

Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University
Press. New Jersey.

Maredia, K.M., S.H. Gage, D.A. Landis and T.M. Wirth. 1992a. Visual response of
Coccinella septempunctata (L.), Hippodamia parenthesis (Say), (Coleoptera:
Coccinellidae), and Chrysoperla camea (Stephens), (Neuroptera: Chrysopidae) to
colors. Biological Control 2: 253-256.

Maredia, K.M., S.H. Gage, D.A. Landis, and T.M. Wirth. 1992b. Ecological
observations on predatory Coccinellidae (Coleoptera) in southwestern Michigan.
Great Lakes Entomol. 25:265-270.

Maredia, K.M., S.H. Gage, D.A. Landis, and J.M. Scriber. 1992c. Habitat use
patterns by the seven-spotted lady beetle (Coleoptera: Coccinellidae) in a diverse
agricultural landscape. Biological Control 2: 159-165.

162

Midgarden, D.G., R.R. Youngman, and SJ. Fleischer. 1993. Spatial analysis of counts
of western corn rootworm (Coleoptera: Chrysomelidae) adults on yellow sticky
traps in corn: geostatistics and dispersion indices. Environ. Entomol. 22: 1124-
1133.

Milne, B.T., M.G. Turner, J.A. Wiens, and A.R. Johnson. 1992. Interactions between
the fractal geometry of landscapes and allometric herbivore. Theor. Pop. Biol. 41:
337-253.

Mohrmann, J.C.J. 1992. New approaches for environmental management. Development
2: 17-21.

Musick, G.J. and H.B. Petty. 1973. Insect control in conservation tillage systems. In
Soil Conservation Society of America [ed.] Conservation tillage; the proceedings
of the national conservation tillage conference, Des Moines, Ankeny, Iowa.

Nasr, E.A. M.R. Tucker, and D.G. Campion. 1984. Distribution of moths of the
Egyptian cotton leafworm, Spodoptera littoralis (Boisduval) (Lepidoptera:
Noctuidae), in the Nile Delta interpreted from catches in a pheromone trap
network in relation to meteorological factors. Bull. Entomol. Res. 74: 487-494.

Nentwig, W. 1988. Augmentation of beneficial arthopods by strip-management. l.
Succession of predacious arthropods and long-term change in the ratio of
phytophagous and predacious arthropods in a meadow. Oecologia (Ber). 76: 597-
606.

Neuenschwander, P. 1982. Beneficial insects caught by yellow traps used in mass-
trapping of the olive fly, Dacus oleae. Entomologia Experimentalis et Applicata
32: 286-296.

N owinszky, L. and I. Ekk. 1988. The swarming outset of Macrolepidopterae observed
by light-trap as correlated with the moon phases. Acta Phytopathol. Entomol.
Hung. 23: 157-166.

NSF. 1978. A pilot program for long-term observation and study of ecosystems in the
United States : report of a Second Conference on Long-Term Ecological
Measurements, Woods Hole, Massachusetts, February 6-10, 1978. Washington,
DC.

Olson, R.K. 1992. The future context of sustainable agriculture: planning for uncertainty.
J. Sust. Agric. 2: 9-20.

Onstad, B.W. 1991. Good models and immature theories. Am. Entomol. 37: 202-204.

Ostrom, HP and B. Fry. 1993. Sources and cycling of organic matter and prehistoric
food webs. pp 785-798 In Engel M.H. and S.A. Macko [eds] Organic
Geochemestry. Plenum Press.

Ostrom, H.P., Colunga—Garcia, M. and S.H. Gage. In press. Establishing pathways of
energy flow for insect predators using stable isotope ratios: field and laboratory
evidence. Oecologia (Berl.).

Pemberton, R.W. and NJ. Vandenberg. 1993. Extrafloral nectar feeding by ladybird
beetles (Coleoptera: Coccinellidae). Proc. Entomol. Soc. Wash. 95: 139-151.

Pickett, S.T., J. Kolasa, and CG. Jones. 1994. Ecological understanding. Academic
Press, San Diego, USA.

Pimbert, M.P. 1991. Designing integrated pest management for sustainable and
productive fiitures. IIED, London. Gatekeeper Series No. 29.

163

Pimentel, D. 1960. Species diversity and insect population outbreaks. Ann. Entomol. Soc.
Am. 54: 76-86.

Pimentel, D., H. Acquay, M. Biltonen, P. Rice, M. Silva, J. Nelson, V. Lipner, S.
Giordano, A. Horowitz, and M. D'Amore. 1992. Environmental and economic
costs of pesticide use. Bioscience 42: 750-760.

Poincelot, RP. 1990. Agriculture in transition. J. Sust. Agric. 1: 9-40.

Price, P.W. 1991. Evolutionary theory of host and parasitoid interactions. Biological
Control 1: 83-93.

Quinn, M.A., S.E. Halbert, and L HI Williams. 1991. Spatial and temporal changes in
aphid (Homoptera: Aphididae) species assemblages collected with suction traps in
Idaho. J. Econ. Entomol. 84: 1710-1716.

Randerson, P.F. 1993. Ordination. pp. 173-218. In Fry, J.C. [ed.] Biological data
analysis: a practical approach. Oxford University Press.

Reijntjes, C., B. Haverkort, and A.Waters—Bayer. 1992. Farming for the future: an
introduction to low-extemal input and sustainable agriculture. Macmillan; ILEIA,
Leusden, Netherlands.

Resh, V.H., A.V. Brown, A.P. Covich, M.E. Gurtz, H.W. Li, G.W. Minshall, S.R.
Reice, A.L. Sheldon, J.B. Wallace, and RC. Wissmar. 1988. The role of
disturbance in stream ecology. J. North Am. Benthological Soc. 7: 433-455.

Rice, M.E. and G.E. Wilde. 1991. Aphid predators associated with conventional- and
conservation-tillage winter wheat. J. Kansas Entomol. Soc. 64: 245-250.

Ricci, C. 1986. Beneficial Coccinellidae caught in yellow traps I some Italian regions. pp.
441-447. In Hodek, 1. [ed] Ecology of aphidophaga, Academica Prague, Junk
Pub. Czechoslovakia.

Risch, S.J., R. Wrubel, and D. Andow. 1982. Foraging by a predaceous beetle,
Coleomegilla maculata (Coleoptera: Coccinellidae), in a polyculture: effects of
plant density and diversity. Environ. Entomol. 11: 949-950.

Risch, S.J., D. Andow, and M.A. Altieri. 1983. Agroecosystem diversity and pest
control: data, tentative conclusions, and new research directions. Environ.
Entomol 12: 625-629

Risch, SJ. 1987. Agricultural ecology and insect outbreaks, pp. 217-238. In Barbosa, P.
and J .C. Shultz [eds], Insect outbreaks. Academic Press, San Diego.

Risser, P.G. 1986. Agroecosystems-structure, analysis and modeling, pp. BAG—ES. In
Kogan, M. [ed.]. Ecological theory and integrated pest management practice. John
Wiley & Sons. New York.

Roach, S.H. and W.M. Thomas. 1991. Overwintering and spring emergence of three
coccinellid species in the coastal plain of South Carolina. Environ. Entomol. 20:
540-544.

Ruppel, RF. and K. Sharp. 1985. Conservation tillage and insect control, pp. 137-144.
In D’ltri, F.M. [ed.] A systems approach to conservation tillage, Lewis Pub.
Chelsea, MI. USA.

Saarenmaa, H., Stone, N.D., L.J. Folse, J.M. Packard, W.E. Grant, M.E. Makela,
and, RN. Coulson. 1988. An artificial intelligence modelling approach to
simulating animal/ habitat interactions. Ecol. Modelling, 44: 125-141.

164

Schotzko, B.J. and E. O'Keeffe. 1989. Comparison of sweep net, D-Vac, and absolute
sampling, an die] variation of sweep net sampling estimates in lentils for pea aphid
(Homoptera: Aphididae), nabids (Hemiptera: Nabidae), lady beetles (Coleoptera:
Coccinellidae), and lacewings (Neuroptera: Chrysopidae). J. Econ. Entomol. 82:
491-506.

Sheehan, W. 1986. Response by specialist and generalist natural enemies to
agroecosystem diversification: a selective review. Environmental Entomology 15:
456-461.

Sirota, J.M. 1990. The temporal and spatial distribution and sex ratio of summer adults of
the seven-spotted ladybird beetle, Coccinella septempunctata, in alfalfa. Master
thesis, Michigan State University, East Lansing.

Southwood, T.R.E. 1978. Ecological Methods. Chapman and Hall, London.

Southwood, T.R.E., V.K. Brown, and RM. Reader. 1979. The relationships of plant
and insect diversities in succession. Biological J. Linnean Soc. 12: 327-348.
Spence, J.R. and J.K. Niemela. 1994. Sampling carabid assemblages with pitfall traps:

the madness and the method. Can. Entomol. 126: 881-894.

SPSS Inc. 1993. SPSS for windows. Release 6.0. Chicago, USA.

Srivastava, C.P., M.P. Pimbert, and W. Reed. 1990. Spatial and temporal changes in
the abundance of Helicoverpa (= Heliothis) armigera (Hubner) in India. Insect
Science and its Application 11: 869-876.

Steffey, K. M. Gray, R. Weinzierl. 1992. Insect management. Pp. 67-74. In Midwest
Plan Service [ed.] Conservation tillage systems and management: crop residue
management with no-till, ridge-till, mulch-till. MidWest Plan Service, Agricultural
and Biosystematics Engineering. Dept., Iowa State University, Ames, IA.

Stinner, B.R. and G.J. House. 1990. Arthropods and other invertebrates in
conservation-tillage agriculture. Annu. Rev. Entomol. 35: 299-318.

Stinner, B.R. and D.H. Stinner. 1989. Plant-animal interactions in agricultural
ecosystems, pp. 355-393. In Abrahamson, W.G. [ed], Plant-animal interactions.
McGraw-Hill Inc. USA.

Szentkiralyi, F. 1992. Spatio-temporal patterns of brown lacewing based on the
Hungarian light trap network (Insects: Neuroptera: Hemerobbidae), pp: 349-357.
In Canard, M. , H. Aspoeck, M.W. Mansell. [eds], Current research in
Neuropterology, Proceedings of the fourth international symposium on
Neuropterology, Tolouse France.

Tait, E.J. 1987. Planning an integrated pest management system, pp. 189-207. In Burn,
A.J., TH. Coaker, and PC. jepson [eds], Integrated pest management. Academic
Press, London. '

Taylor, L.R. 1986. Synoptic dynamics, migration and the Rothamsted insect survey. J.
Animal Ecology 55: 1-38.

Tollefson, J.J. and DD. Calvin. 1994. Sampling arthropods pests in field corn, pp. 433-
473. In Pedigo, LP. and GD. Buntin [eds], Handbook of sampling methods for
arthropods in agriculture, CRC press, Boca Raton Florida, USA.

Tonhasca J r., A. and B.R. Stinner. 1991. Effects of strip intercropping and no-tillage on
some pest and beneficial invertebrates of corn in Ohio. Environ. Entomol. 20:
1251-1258.

165

Tummala, RL. and D.H. Haynes. 1977. On-line pest management systems. Environ.
Entomol. 6: 339-349.

Turner, M.G. and V.H. Dale. 1991. Modeling landscape disturbance, pp. 323-351. In
Turner, M.G. and RH. Gardner [eds], Quantitative methods in landscape
ecology: the analysis and interpretation of landscape heterogeneity. Springer-
Verlag, New York.

Turner, M.G., R. Constanza and EH. Sklar. 1989. Methods to evaluate the
performance of spatial simulation models. Ecol. Modeling 48: 1-18.

Trujillo-Arriaga, J. and M.A. Altieri. 1990. A comparison of aphidophagous
arthropods on maize polycultures and monocultures , in Central Mexico. Agric.
Ecosyst. Environ. 31: 337-349.

Urban, D.L., R.V. O'Neill, and H.H. Shugart Jr. 1987. Landscape ecology. Bioscience
37 : 119-127.

Usher, MB. 1995. A world of change: land-use patterns and arthropod communitiespp.
371-397. In Harrington, R. and NE. Stork [eds] Insects in a changing
environment, Academic Press, London.

Van Cleve, K. And S. Martin. 1991. Long-Term Ecological Research in the United
States: a Network of research Sites. Seattle, Washington.

VandenBosch, R., P. S. Messenger, and A. P. Gutierrez. 1982. An introduction to
biological control. Plenum Press, NY.

Vandermeer, J. 1995. The ecological basis of alternative agriculture Annual Review of
Ecology and Systematics 26: 201-224.

vanEmden, HE 1990. Plant diversity and natural enemy efficiency in agroecosystems,
pp. 63-77. In M. Mackauer, L. Ehler, and J. Roland [eds], Critical issues in
biocontrol. Intercept Ltd., Andover, England.

Wada, E., H. Mizutani, and M. Minagawa. 1991. The use of stable isotopes for food
web analysis. Crit. Rev. Food Sci. Nut. 30: 361-371.

Warren, LG. and M. Tadic. 1967. Biological observations on Coleomegilla maculata
and its role as a predator of the fall webworm. J. Econ. Entomol. 60: 1492-1496.

Wellington, W.C. and RM. Trimble. 1984. Weather, pp. 399-425. In Hufi'aker, CB.
and RL. Rabb [eds], Ecological entomology. John Wiley and Sons, NY, USA.

Wetzler, RE. and SJ. Risch. 1984. Experimental studies of beetle diffirsion in simple
and complex crop habitats. J. Animal Ecol. 53: 1-19.

Wheeler, J.C. 1992. Population and environment: thinking of the next generations.
Development 2: 7-10.

Wiens, J.A. 1989. Spatial scaling in ecology. Functional Ecology 3: 385-3 97.

Wratten, SD. and C.F.G. Thomas. 1990. Farm-scale spatial dynamics of predators and
parasitoids in agricultural landscapes, pp. 219-237. In Bunce, R.G.H. and DC.
Howard [eds] Species dispersal in agricultural habitats. Belhaven Press, London.

Wright, E.J. and J.B. Laing. 1980. Numerical response of coccinellids to aphids in corn
in southern Ontario. Canadian Entomologist 112: 977-988.

Zar, J .H. 1984. Biostatistical analysis. Prentice Hall, New Jersey.

"IIIIIIIIIIIIIIIII