grasshoppers and grassland health: managing grasshopper outbreaks without risking environmental...

Grasshoppers and Grassland Health

NATO Science Series

A Series presenting the results of activities sponsored by the NATO Science Committee. The Series is published by lOS Press and Kluwer Academic Publishers, in conjunction with the NATO Scientific Affairs Division.

A. Life Sciences B. Physics C. Mathematical and Physical Sciences D. Behavioural and Social Sciences E. Applied Sciences F. Computer and Systems Sciences

1. DisarmamentTechnologies 2. Environmental Security 3. High Technology 4. Science and Technology Policy 5. Computer Networking

NATo-PCO-DATA BASE

lOS Press Kluwer Academic Publishers Kluwer Academic Publishers Kluwer Academic Publishers Kluwer Academic Publishers Ie'S Press

Kluwer Academic Publishers Kluwer Academic Publishers Kluwer Academic Publishers lOS Press lOS Press

The NATO Science Series continues the series of books published formerly in the NATO ASI Series. An electronic index to the NATO ASI Series provides full bibliographical references (with keywords and/or abstracts) to more than 50000 contributions from international scientists published in all sections of the NATO ASI Series. Access to the NATO-PCO-DATA BASE is possible via CD-ROM "NATO-PCO-DATA BASE" with user-friendly retrieval software in English. French and German (WTV GmbH and DATAWARE Technologies Inc. 1989).

The CD-ROM of the NATO ASI Series can be ordered from: PCO, Overijse. Belgium

Series 2. Environment Security - Vol. 73

Grasshoppers and Grassland Health Managing Grasshopper Outbreaks without Risking Environmental Disaster

edited by

Jeffrey A. Lockwood Alexandre V. Latchininsky Entomology Section, Department of Renewable Resources, University of Wyoming, laramie, Wyoming, U.S.A., Association for Applied Acridology International

and

Michael G. Sergeev Department of General Biology, Novosibirsk State University, Novosibirsk, Russia, Association for Applied Acridology International

Springer-Science+Business Media, B.V.

Proceedings of the NATO Advanced Research Workshop on Acridogenic and Anthropogenic Hazards to the Grassland Biome: Managing Grasshopper Outbreaks without Risking Environmental Disaster Estes Park, Colorado, U.S.A. September 11-18, 1999

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 978-0-7923-6530-3 ISBN 978-94-011-4337-0 (eBook) DOI 10.1007/978-94-011-4337-0

Printed an acid-free paper

AII Rights Reserved © 2000 Springer Science+Business Media New York Originally published by Kluwer Academic Publishers in 2000

Softcover reprint of the hardcover 1 st edition 2000

No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording ar by any information storage and retrieval system, without written permission from the copyright owner.

TABLE OF CONTENTS

Preface ............................................................. vii

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. IX

Introduction A Novel Approach to Solving Complex Ecological Problems: An International Polylogue on the Art and Science of Applied Acridology ..... I

J.A. Lockwood, A. V Latchininsky & M. G. Sergeev

Part 1. Grasshoppers as Integral Elements of Grasslands I. Do Grasshoppers Diminish Grassland Productivity?

A New Perspective for Control Based on Conservation. . . . . . . . . . . . . . . . . . . .. 7 G.E. Belovsky

2. What are the Consequences of Ecosystem Disruption on Acridid Diversity and Abundance? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 31

F.A. Gapparov & A. V Latchininsky

3. What is the Role of Grassland Vegetation in Grasshopper Population Dynamics? ............................................. 61

0. Olfert

Part 2. Grasshopper Population Ecology and Management 4. How do Spatial Population Structures Affect Acridid Management? ......... 71

M.G. Sergeev, 0. V Denisova & l.A. Vanjkova

5. Can Micropopulation Management Protect Rare Grasshoppers? ............. 89 M.E. Chernyakhovskiy

6. What Factors Govern Orthopteran Community Structure and Species Prevalence? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 97

T. Kisbenedek & A. Baldi

Part 3. Grasshopper and Locust Control Strategies and Tools 7. How Can Acridid Population Ecology Be Used to Refine Pest Management

Strategies? ..................................................... 109 M. Lecoq

VI

8. What are the Consequences of Non-Linear Ecological Interactions for Grasshopper Control Strategies? ................................. 131

A. Joern

9. What Tools Have Potential for Grasshopper Pest Management? A North American Perspective ..................................... 145

J.A. Onsager & 0. Olfert

Part 4. Grasshopper Control and Grassland Health 10. How Can Acridid Outbreaks Be Managed in Dryland Agricultural

Landscapes? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 157 v.E. Kambulin

11. Can Locust Control Be Compatible with Conserving Biodiversity? ......... 173 MJ. Samways

12. How Does Insecticidal Control of Grasshoppers Affect Non-Target Arthropods? ........................... . . . . . . . . . . . . . . . . . . . . . . . .. 181

I.M Sokolov

Summary The Risks of Grasshoppers and Pest Management to Grassland Agroecosystems: An International Perspective on Human Well-Being and Environmental Health ................................ 193

J.A. Lockwood & A. V. Latchininsky

List of Contributors .................................................. 217

List of Participants .................................................. 219

PREFACE

Acridids (grasshoppers and locusts) can range from being rare curiosities to abundant menaces. Some are threatened with extinction and become subjects of intensive conservation efforts, while others are devastating pests and become the objects of massive control programmes. Even within a species, there are times when the animal is so abundant that its crushed masses cause the wheels of trains to skid (the Rocky Mountain grasshopper, Melanoplus spretus Walsh in western North America in the 1860s and I 870s), while at other times the animal is alarmingly scarce (the Rocky Mountain grasshopper went extinct in the early 1900s).

Why are there these extremes in one insect family, and even in a single species? The NATO workshop examined this paradox and its implications for Environmental Security, which must address both the elements of land use (agricultural production and pest management) and conservation of biodiversity. The reconciliation of these objectives clearly demands a critical assessment of current knowledge and policies, identification of future research, and close working relationships among scientists. Insects can present two clear faces, as well as the intervening gradation. These extremes require us to respond in two ways: conservation of scarce species and suppression of abundant (harmful) species. But perhaps most important, these opposite poles also provide the opportunity for an exchange of information and insight.

]n some NATO and CIS (Commonwealth of Independent States comprising the ex-USSR with the exception ofthe Baltic nations) countries, grasshopper outbreaks are often economic disasters. Anthropogenic changes (e.g., overgrazing and desertification) have effected the duration and frequency ofthese events, and extensive control programs have been developed to protect agricultural resources. However, in temperate grasslands, grasshoppers are an abundant, native and important element of the ecosystem, and there are many rare species. Hence, we must attempt to manage pest outbreaks without disrupting the vital, natural processes that sustain steppic ecosystems.

The classical view of insect pest management and conservation biology is that these two fields are philosophically, methodologically, and politically antagonistic. If there was a single theme developed in the workshop, it was that the field of acridology has a tremendous capacity for the control and conservation perspectives to become mutualistic. Indeed, there appears to be a remarkable lack of enmity between what might be traditionally expected to be two "camps" of biologists. In simple terms, we know of no conservationist calling for the elimination of grasshopper control programmes and no pest manager pursuing eradication of these organisms. Furthermore, there are even a number ofacridologists working on both the issues of managing rarity (conservation) and abundance (control). As such, the opportunities, both realized and potential, for respectful and synergistic discourse are as great in this field as in any other discipline in biology. There are several research and application issues that demand the attention of

Vll

viii

orthopterists engaged in both control and conservation. I) The problem of rarity is of great importance - how do some species

erupt from rarity to outbreak while others languish or decline throughout their range?

2) The diminishment of negative, insecticidal impacts is becoming a matter of unifying concern - how can we further reduce the non-target effects of insecticides so as to optimize both the economic and the environmental outcomes of control programmes?

3) The ecological power of biological control is a matter of great excitement and concern - how can we reasonably assure that these tactic results in the suppression of the pest acridids without negatively impacting the non-target species?

4) The taxonomic impediment would seem to be a matter of concern mostly to the conservationists, but the breakthrough work ofUvarov on locust phase variation serves as a continual reminder that pest management is deeply endebted to a vibrant and vital base of systematic knowledge, for example - how can we use taxonomic and phylogenetic insights to select and use pathogens for control of pest species while minimizing the chances of environmental harm?

5) Conservationists recognize that anthropogenic impacts will continue to increase. Therefore, only by working with those involved in pest management can the goals of conservation and control be truly integrated - how can we manage landscapes via grazing, cultivation, and other practices to protect orthopteran biodiversity without encouraging infestations of pest species?

6) The development of a sustainable agriculture depends on understanding the native species and their place in complex systems - how can we assess the ecological roles and economic values of acridids in agroecosystems and adapt our production systems accordingly?

Thus, it is evident that conservation and control both require, in fundamental terms, an understanding of the population biology and community ecology of the organisms of concern. Whether we manage so as to prevent outbreaks or extinctions, biological insight is required. The workshop opened lines of communication, explored mutualistic research programmes, and began the search for resources necessary to work toward a sustainable relationship between humans and grasshoppers. The goal of the workshop was to foster international understanding of the potential for grasshopper pest management and conservation programmes to cooperate in mutual support of their common goal - keeping good stewards on the land to sustain healthy, diverse and productive ecosystems. And we believe that this goal was achieved, much to the credit ofthe participants, who came with strong opinions, rich experiences, extensive expertise, divergent philosophies, varied economics, and diverse cultures, along with open minds, collegial attitudes, cooperative spirits, and an authentic desire to solve problems.

ACKNOWLEDGEMENTS

We wish to acknowledge the support of the North Atlantic Treaty Organization's Scientific Affairs Division, RhOne-Poulenc's Locust Control and Public Health Unit, the Association for Applied Acridology International, the Orthopterists Society, and the University of Wyoming for support ofthis workshop. We are sincerely grateful for the immense patience, adaptability, and first-rate service provided by Laramie Travel Center. We also thank the proprietors of Creekside Suites and Wildflower Catering for their friendliness, professionalism, and impeccable services. We thank our translators Gregory Rayner and Mikhail Rojavin and their agency, All Language Alliance, for outstanding work on our behalf. Finally, our special appreciation goes to our colleagues for their reviews, criticisms and linguistic improvements.

ix

Par

tici

pant

s o

flh

e N

AT

O A

dvan

ced

Res

earc

h W

orks

hop

"Acr

idog

enic

and

Ant

hrop

ogen

ic H

azar

ds t

o th

e G

rass

land

Bio

me:

Man

agin

g G

rass

hopp

er O

utb

reak

s lI

'itho

llt R

iski

ng E

nvir

onm

enta

l D

isas

ter,

" S

epte

mbe

r 11

-18.

199

9, E

stes

Par

k, C

olor

ado

(US

A)

Fro

m l

eft t

o ri

ght:

M.

Che

rnya

khov

skiy

. G

. B

e1ov

sky,

L S

okol

ov,

T K

isbe

nede

k. S

. S

chel

l. 1

. O

nsag

er,

M.

Lec

oq,

M.

Sam

way

s. D

. O

tte,

A.

Mon

ard,

S. K

rall,

L.

Psh

enit

syna

, V

. K

ambu

lin,

N.

Jago

. F.

Gap

paro

v. T

Joe

rn,

O.

Olf

ert,

1.

Loc

kwoo

d.

Pho

to:

A.

Lat

chin

insk

y (U

nive

rsit

y o

f Wyo

min

g)

INTRODUCTION

A Novel Approach to Solving Complex Ecological Problems: An International Polylogue on the Art and Science of Applied Acridology

1.A. LOCKWOOD I , A.Y. LATCHININSKyl, and M.G. SERGEEy2

I Entomology Section Department of Renewable Resources University of Wyoming Laramie, WY 82071-3354 USA 2 Department of General Biology Novosibirsk State University 2, Pirogova St., Novosibirsk 630090 Russia

For every complex problem, there is a simple solution. And it is wrong. -- H. L. Mencken

1. General Approach to Complexity

A great deal of effort has been directed at the impact of acridids (grasshoppers and locusts, Orthoptera: Acrididae) to crops, where the pest status of these insects is unambiguous. However, the ecology and economics of the world's grasslands make the role of these native organisms much more complex. The NATO Advanced Research Workshop on "Acridogenic and Anthropogenic Hazards to the Grassland Biome: Managing Grasshopper Outbreaks without Risking Environmental Disaster" was developed and conducted using an approach that differed markedly from the format of a standard, scientific conference. This strategy was explicitly designed to capture the diverse views held by acridologists, other scientists, and pest managers working on both rare and abundant acrid ids of temperate grasslands -- and then to synthesize these views into a coherent and compelling interpretation of current conditions and future needs. The approach was much more complex than the standard conference, in which formal presentations consume the majority of the time, with questions and discussions being relegated to the brief periods following each paper. Our goal was to develop an international polylogue (rather than the monologue of the formal presentation or the simple dialogue that arises between two people) from which original, creative, synthetic, and collective insights might emerge.

J. A. Lockwood et al. (eds.), Grasshoppers and Grassland Health, 1-6. © 2000 Kluwer Academic Publishers.

2

2. Rationale and Design of the Workshop

To understand the approach taken in the workshop, and hence to appreciate the nature of the emergent perspectives, it is important to consider the principles and pragmatics that gave rise to this project. It is our hope that the strategy employed in this meeting might be adopted or adapted by others seeking to bring diverse, international perspectives into a common and coherent focus, particularly where the stakes are high (in terms of human suffering and environmental degradation) and we can ill-afford intellectual debates for their own sake.

2.1. THE QUESTION OF "WHY?"

There are five, emerging global trends which provided the impetus for a workshop designed to develop collaborative and consensual understandings and recommendations for the management of grasshopper and locust outbreaks on the world's grasslands.

First, available evidence suggests that acridid outbreaks are severe and perhaps growing worse in grassland ecosystems. At the tum of the millennium, a major outbreak may be developing in the western USA and Canada, severe infestations continue to plague China, and one of the worst plagues in modem history is developing in Russia and Central Asia (including Kazakhstan and Uzbekistan). As such, it is clear that the challenges of acridid outbreaks are not just fabulous material for historical accounts but that these events continue to confront a growing human population. Furthermore, ongoing environmental changes (e.g., warming temperatures, declining rainfall, expanding tillage, and increasing livestock) have the potential to further aggravate the conditions that give rise to acridid outbreaks.

Next, the global understanding of the effects of humans on the environment is steadily increasing. With growing environmental concern, the traditional approaches to managing acridid outbreaks (e.g., blanket applications of broad-spectrum insecticides over immense areas) are being challenged, with their net benefits being called into question. The role of these insects in sustaining the productivity of agroecosystems is being clarified, and the impacts of our livestock grazing and pest management practices are becoming evident. The long-term changes brought about by the use of synthetic chemicals and the introduction of exotic species are now under intense scrutiny. Issues of water quality, bioaccumulation, chemical sensitivity syndromes, and food safety suggest that human health is being increasingly linked to environmental well-being.

Third, governmental and other "external" sources of support for grasshopper and locust control are declining. Both within and between nations, the responsibility for monitoring and treating acridid populations has been dramatically decentralized for a variety of economic and political reasons. This phenomenon may be the single most important factor in defining the future direction of research and management pertaining to the control of outbreaks. With the localization of programmmes to control acridid outbreaks, site-specific, low-input pest management strategies are in high demand.

3

Fourth, as the economic resources for acridid control programmes decline, so too do the human resources in the form of expertise, experience, and knowledge. While several nations had extremely productive and viable centers of acridology just a decade ago, it now appears that no single nation has the intellectual critical mass necessary to independently develop and implement new management strategies. In North America and the CIS, the number of applied acridologists has declined by at least 50% in the last 10 years. Both university and government laboratories have aggressively downsized their research and development programmes, and there appears to be no reprieve in sight.

Finally, the loss of economic resources and human capital related to applied acridology is being offset, to some degree, by increasing international communications among scientists and managers. Where the potential for a critical mass of expertise in a single nation has all but disappeared, the quality of the remaining workers is extremely high and their willingness to share information, exchange ideas, and explore common needs has significantly increased. Organizations such as The Orthopterists' Society and the Association for Applied Acridology International (co-sponsors of the NATO workshop) represent the best efforts of scientists and managers to develop international lines of communication that help assure that the most economically viable and environmentally sound pest management tactics can be developed and implemented. The urgent need for international collaboration in applied acridology is further heightened by the trans-border nature of many acridid outbreaks.

In summary, this workshop was an extension of a growing, grassroots effort by the world's acridologists to build new relationships and programmes in the face of declining resources and increasing acridid outbreaks. The common cause uniting these diverse scientists and managers is an explicit recognition that the interface of pest management with environmental risk must serve as the starting point for new understandings and approaches. It is imperative that grasshopper and locust management systems be derived via simultaneous solutions for the interrelated challenges of economics and ecology on the world's grasslands.

2.2. THE QUESTION OF "WHEN?"

The utopian time to solve a problem is once all of the facts are known, the processes are clarified, and potential approaches have been tested. However, in the "real world", we recognize that it is irresponsible to withhold our best professional judgement until absolute certainty has been obtained. While we wait for "just one more study", acridid outbreaks are decimating food supplies, and misguided control strategies are risking human and environmental health. We suggest that such "paralysis by analysis" will not meet the needs of the human population or the environment, as we cannot stop the world until we have enough data upon which to make unambiguous recommendations. As such, the question becomes not "What is the best answer?", but "What is a better answer?".

It is evident that the world is changing and that acridid outbreaks will occur and will be treated whether or not the scientist-manager offers input or suggests changes. In simple terms, if the applied acridologist does not proactively define the future of pest management, it will be determined by politicians, bureaucrats, and industries - all of who

4

have interests that may contlict with the sustained stewardship of the land and its people. As such, the scientist-manager is in the difficult position of either using inadequate knowledge to develop imperfect approaches which represent better, but less-than-ideal, tactics or abrogating the responsibility for making recommendations until the ideal system can be implemented. We have consciously chosen to move ahead with humility, knowing that we risk error but also being aware that we have an ethical obligation to act on our best judgement in the face of uncertainty.

2.3. THE QUESTION OF "WHERE?"

The setting for a workshop to address issues of international importance is a difficult choice, as one does not want to predispose the nature of the meeting through the selection of a venue. As such, practical considerations largely drove the decision to host the workshop in the United States. Locating the meeting in Estes Park, Colorado provided easy access to transportation and other logistical services (including simultaneous and sequential translation). The other consideration was to choose a venue that was scaled to the size of the workshop, so that participants had a setting that was private and somewhat secluded to create an atmosphere of collegiality and intimacy necessary for honest discussions and sincere resolutions of differences.

2.4. THE QUESTION OF "WHO?"

Selecting the participants was the most difficult challenge in developing the workshop, as the outcome of the project was completely in the hands of the scientists and managers attending the meeting. We were seeking a "real time" creative effort, in which individuals would synthesize -- not simply report -- their ideas. In this context, three essential criteria were used to select the participants.

First, we sought individuals who were internationally respected for their knowledge of applied acridology and related fields. Our participants included many of the world's finest experts in the biology, ecology, taxonomy, conservation, and management of acrid ids. We consciously included both scientists and managers, but excluded agency officials and other administrators who lacked first-hand experience with grasshoppers and locusts.

Next, we looked for highly creative individuals, with reputations for being original thinkers and capable listeners. Because the workshop was participant-driven, it was essential to identify individuals who would truly engage in the process of mutual discovery. It was also critical to assure that the participants would be adaptable to an unusual format, as we chose to minimize formal presentations and maximize structured discussion and problem-solving.

Finally, we wanted a diversity of participants to reflect the full richness of scientific and management experience and perception. We chose people to optimize the range of geographic (within NATO parameters), cultural, political, economic, and philosophical backgrounds. Balancing this diversity against the risks of having too many

5

participants (thereby precluding authentic dialogues) or too few participants (thereby omitting important perspectives) was an enormous task. The goal was to assemble a creative, critical mass of people in order to generate respectful and resolvable conflict. Our intention was that this group should have the potential for finding common ground on complex and controversial issues. But it was essential that the emergent consensus would have to be earned, not assumed.

The list of participants (Table 1) represents our best efforts to address the three criteria. There is no doubt that we could have chosen other participants; exclusion in no way implies a lack of expertise, experience, collegiality, originality or creativity. Rather, this list reflects our best efforts in context of personal experiences, accidents of historical association, and admittedly incomplete knowledge.

2.5. THE QUESTION OF "WHAT?"

To generate the most productive discourse possible, four methods were employed. First, a survey was conducted prior to the workshop in order to assess the participants' views on the risks associated with acridid outbreaks and management strategies. Second, we designed a set of four "problems", the solutions of which were to focus dialogue towards a final set of recommendations. Third, before each of the aforementioned "problems", there was a set of presentations intended to stimulate thought and discussion pertaining to the issue. Fourth, we concluded with the development of a set of consensus recommendations pertaining to the economically viable, environmentally sound acridid pest management.

The methods and results of the "collective" workshop products (i.e., surveys, problems sets, and recommendations) are presented as the concluding chapter of this book. The presentations used as the foundations of the workshop discussions form the "core" of this book. These individual analyses offer valuable insights as to the nature of acridid pest management and conservation in context of ecological and economic principles, and they provide an essential context for the collective products.

6

TABLE 1. Participants in the NATO Advanced Research Workshop on Acridogenic and Anthropogenic Hazards to the Grassland Biome: Managing Grasshopper Outbreaks without Risking Environmental

Disaster

Name Affiliation Country

Gary Belovsky Dept. of Fisheries & Wildlife, Utah State University USA

Mikhail Chemyakhovskiy Dept. of Zoology & Ecology, Moscow State Russia Pedagogical University

Forkat Gapparov Uzbek Institute for Plant Protection (UzNIIZR) Uzbekistan

Nick Jago Natural Resources Institute (NRI) England

Anthony Joem School of Biological Sciences, University of USA Nebraska-Lincoln

Vladimir Kamboulin Kazakhstan Quarantine Inspection (formerly) Kazakhstan

Tibor Kisbenedek Hungarian Natural History Museum Hungary

Stephan Krall Deutsche Gesellschaft fUr Technische Germany Zusammenarbeit (GTZ)

John Larsen United States Department of Agriculture (APHIS) USA

Alexandre Latchininsky Dept. of Renewable Resources, University of USA Wyoming

Michel Lecoq CIRAD-AMISIPRIF AS France

Jeffrey Lockwood Dept. of Renewable Resources, University of USA Wyoming

Annie Monard Plant Production & Protection Division, FAO-UN Italy

OwenOlfert Agriculture & Agri-food Canada Canada

Jerome Onsager United States Department of Agriculture (ARS) USA

DanielOtte Academy of Natural Sciences, Philadelphia USA

Ludmila Pshenitsyna Dept. of General Biology, Novosibirsk State Russia University

Michael Samways Zoology & Entomology Dept, University of Natal South Africa

Scott Schell Dept. of Renewable Resources, University of USA Wyoming

Michael Sergeev Dept. of General Biology, Novosibirsk State Russia University

Igor Sokolov All-Russian Institute for Plant Protection (VIZR) Russia

Part 1. GRASSHOPPERS AS INTEGRAL ELEMENTS OF GRASSLANDS

1. DO GRASSHOPPERS DIMINISH GRASSLAND PRODUCTIVITY?

A New Perspective For Control Based on Conservation

G.E. BELOVSKY Ecology Center and Department of Fisheries and Wildlife Utah State University Logan, Utah 84322 USA

Abstract

Experimental studies of the effects of grasshopper consumption on plant production are presented. The long held claim that grasshopper consumption of plants in some years can reduce forage for livestock and wildlife is supported. However, examining grasshopper consumption over a longer term (multiple years), I find that grasshoppers enhance plant production. This emerges because grasshoppers accelerate nutrient cycling primarily by increasing the proportion of litter provided by faster decomposing plants. The greater availability of nutrients further increases the abundance of faster decomposing plants because they are competitively favored under these conditions and this further enhances nutrient availability and plant production. Therefore, the short-term loss of forage for livestock and wildlife is outweighed by the long-term enhancement offorage production. The rangeland conditions which lead to grasshoppers producing beneficial production effects are reviewed.

1. Problem

1.1. INTRODUCTION

E.O. Wilson [1] claims that biodiversity provides economic, cultural and biological wealth to mankind. Economic wealth arises from the products that biodiversity provides; cultural wealth emerges because biodiversity is part of every nation's history and art, and biological wealth refers to the aesthetic pleasure and environmental health provided by biodiversity.

7


8

As ecologists have become increasingly concerned with the global loss of biodiversity, they have begun to stress the biological wealth provided by biodiversity. In particular, the idea that endemic species and native ecosystems may provide ecological services that benefit humanity has emerged as a major motive for conservation [2, 3]. However, it is a difficult and contentious task to decide how these ecological services should be economically valued for comparison with the economic gains provided by their destruction [4].

Valuation of ecological services is not always problematic. For example, if products from an ecosystem are being exploited and current management practices for the system reduce biodiversity and this diminished biodiversity decreases ecological services that diminish production, then biodiversity needs to be restored to rejuvenate production. I present in this chapter an example based on grasshoppers (Orthoptera: Acrididae) and rangeland production of forage plants in the western USA, where grasshoppers enhance the production of forage for mammalian herbivores.

I will show that grasshoppers can reduce the forage for livestock within a year as commonly assumed; however, I will also show that grasshoppers actually stimulate forage production between years and this long-term enhancement exceeds annual losses. The stimulation of production occurs because grasshoppers can accelerate nutrient cycling in the grassland ecosystem, more effectively than mammalian herbivores can. This means those commonly accepted rangeland management practices that try to control grasshopper consumption may actually reduce the ecosystem's long-term productivity, rather than making more forage available to mammalian herbivores. Consequently, grasshopper control may reduce ecological services provided by grasshoppers and this may be detrimental to the economic wealth provided by rangelands.

To reach the above conclusions, I conducted long term (six years) experiments in grassland that manipulated grasshopper herbivory to assess how this herbivory changes 1) nutrient cycling, 2) plant production and 3) plant species composition within and between years.

1.2. STUDY SYSTEM

The study was conducted at the National Bison Range in Montana, USA. This site is Palouse Prairie, which is the ecosystem of the intermountain region of western North America bounded by the Rocky Mountains on the east, the Cascade Mountains on the west, southern Alberta and British Columbia in Canada on the north, and northern Utah and Nevada in the US on the south. This grassland averages 35 cm/yr of precipitation, which primarily falls as rain in spring (May - June). I chose a 4 ha site at -750 m elevation for its uniformity in slope and aspect (very flat) and homogeneity in vegetation (plant biomass and species composition) for conducting the experiments. More than 90% of the plant biomass at this site was represented by three mono cot species (Elymus smithii, Poa pratensis and P. compresa), with a variety of non-woody dicots comprising the remainder of the plants. During the study period (May, 1994 - July, 1999), aboveground net plant production (NPP) was measured with a radiometer and it varied between 108 and 237 glm2 or approximately 2.2 fold.

9

NPP at this site is a function of precipitation and soil N availability. First, over six years at the National Bison Range, monocot NPP is positively correlated with May -September precipitation (P < 0.02) and soil nitrogen availability during May - September (P < 0.03), not annual nitrogen availability, as indicated by N absorbed by buried resin bags (see below). Over this same period, dicot NPP was not correlated with precipitation and was negatively correlated with annual N absorbed by resin bags (P < 0.02), not nitrogen availability only from May - September. Second, at the 4 ha study site during 1994, 21 resin bags were buried in early May and removed in early October to indicate soil N availability during the plant growing season, and plant biomass was measured by clipping a 0.1 m2 area around each resin bag. Plant biomass was positively correlated with N absorbed by resin bags (r = 0.41, N = 21, P < 0.02). Finally, at the study site, nitrogen fertilization from June - August increased plant biomass [5,6].

The common (>90%) grasshoppers at this site were represented by Gomphocerinae (Chorthippus curtipennis), Oedipodinae (Camnula pellucida) and Melanoplinae (Melanoplus sanguinipes, M femurrubrum, M dawsoni, M bivittatus). M sanguinipes comprised more than 50 - 70% of the grasshoppers in a year. During the study period, grasshopper hatchling densities varied between 11 and 911m2 among the six years or more than 8 fold, and peak adult densities varied between 4 and 361m2 or more than 9 fold. During the study grasshopper density exhibited a build-up and then decline (Fig. 1). High grasshopper densities are typical for Palouse Prairie [7 - 9].

The range of plant and grasshopper abundances at this site is not atypical for arid grasslands in which grasshopper control has been a management concern. Therefore, the patterns reported here may also apply to other areas throughout the world.

~ en 100 z w c 75 0::_

~'E a..~ o-x: U) U)

:i C)

50

25

- HATCHLINGS -- ADULTS

YEAR

Figure I. Field grasshopper populations during the study

10

1.3. ANNUAL (SHORT TERM) STUDIES

Claims that grasshoppers reduce plant abundance so that mammalian herbivores, especially livestock, have less forage are based upon annual measures or calculations of grasshopper consumption. Experiments were conducted to directly measure the impact of grasshopper consumption on forage availability in my study area.

1.3.1. Experimental Design In 1994 and 1996,40 cages were placed in an 8 x 5 grid over vegetation in May at the 4 ha study site with at least 1 m between cages. In 1994, 0.1 m2 cages were used and these cages were replaced with larger ones, 0.35 m2, in 1996. The cages were constructed of aluminum window screens [10]. Half of the cages received natural precipitation, while the other half received supplemental water every other day, which in total equaled 50% of the average amount of precipitation from May - Sept. Cages in each water treatment contained either grasshoppers at the field density (5 cages) or no grasshoppers (15 cages). Grasshopper field density at the site was measured weekly by counting the number of grasshoppers in 24 - 0.1 m2 rings [11] at -1 Ia.m. and-4 p.m .. The cages were stocked in mid-June with M sanguinipes nymphs ~ 3,d instar), because these were the most common grasshoppers at the site in mid-June. The number used to stock the cages approximated the current field density. Individuals in each cage were counted weekly [10] and M sanguinipes individuals were added or removed from each cage weekly to maintain field density. The individuals added were selected from the current most common developmental stage and removed individuals were killed and their corpses were left in the cage.

The 15 cages in each water treatment containing no grasshoppers received one of three additional treatments: no further manipulation, grasshopper frass added, or grasshopper corpses added. Frass was collected from M sanguinipes individuals of the current most common developmental stage in the field. These individuals were kept at field temperature and light/dark cycles in 20 liter aquaria and fed an assortment of plant species from the site at the current field plant biomass per grasshopper. Once a week, the frass in each aquarium was collected and per capita frass production was determined by dividing total frass mass by the average number of grasshoppers in the aquarium ([individuals at start + individual at the end]/2). Given the observed field density of grasshoppers during a week, the appropriate mass of frass was added to a cage. Corpses were added by catching M sanguinipes individuals, killing them and placing them in a cage, based on the decline in grasshopper density observed in the field during a week. Five cages represented each treatment. All treatment combinations (water/controVcorpse/frass/grasshoppers) were represented in a row (5 rows), and these treatments were randomly assigned.

Living plant biomass in each cage was measured in early October by clipping plants in half of each cage's area, and litter was collected from the same area. In the following May, this was repeated. The living plant material was separated between grasses and forbs; living plants and litter were dried at 60° C and then were weighed.

11

An index of soil nitrogen availability was measured in each cage using an ion exchange resin bag [12, 13]. Resin bags were made using undyed nylon (pantyhose) containing 109 ofRexyn. The Rexyn absorbs nitrogen from the soil, which can be released using 2N KCI and the extractant can be analyzed colorimetrically for NO]· and NH4 + using an autoanalyzer [14 - 18]. A resin bag was buried at 15 cm in each cage in May and removed in early Oct. to indicate N availability during the plant-growing season; another bag was placed in each cage in early Oct. and removed the following May to indicate N availability for initiation of the next year's plant growth.

1.3.2. Results Whether plant biomass by October was diminished by grasshopper consumption depended upon water availability, significantly decreasing with natural precipitation (Bonferroni contrast: F = 3.99; df= 1,32; P < 0.05), but remaining unchanged with added water (Fig. 2). Both 1994 and 1996 were drought years, experiencing respectively 53% and 68% of average (l8.9cm) precipitation during the plant-growing season (May - Sept), therefore, the added water treatment produced precipitation conditions that were approximately average. Even though grasshoppers decreased plant abundance by October under the drought conditions, the plant abundance in the following spring was unaffected by grasshopper consumption (F = 0.25; df = 3, 64; P < 0.86; data not shown) or by added water during the previous year's plant-growing season (F = 0.78; df= 1,64; P < 0.38; data not shown). Therefore, when grasshopper consumption did reduce plant abundance, this was a transient, not a long-lasting effect.

Under approximately normal precipitation (added water), grasshopper consumption did not decrease plant abundance, because grasshoppers recycled N through their frass that could be readily absorbed by plants, and with the normal precipitation (added water) the plants exhibited compensatory growth (Fig. 2). Even with considerable between and within year variation, this was demonstrated by plants exhibiting the greatest abundance when grasshopper frass was added to cages along with water (Fig. 2: Bonferroni contrast: F = 5.61; df= 1,32; P < 0.02). The N measured in fall-collected ion exchange resin bags indicated that cages containing grasshoppers or with added frass had the greatest N availability (Fig. 3: Bonferroni contrast: F = 2.69; df= 3,60; P < 0.05). In accord with the absence of differences between treatments in the following spring's plant abundance, no differences in spring availability ofN were observed (F = 0.54; df= 3, 48; P < 0.68; data not shown). Therefore, increased availability of N from grasshopper frass caused only short-term effects on plant growth.

1.3.3. Implications/or Long Term Studies These findings are not surprising in light of commonly held ideas that grasshoppers are most likely to reduce forage abundance for livestock in dry years [19, 20]. These results emerged regardless of grasshopper density relative to plant abundance. The years 1994 and 1996 had comparable plant abundances (86.6 ± 11.3 vs. 68.5 ± 6.7 glm2, SE, N = 5), but very different grasshopper densities (7 vs. 35 adults/m2). Furthermore, under these conditions, grasshoppers substantially reduced fall plant abundance, but by comparable amounts (39.3 vs. 30.3%) in the two years. Similar results in other years have been

12

observed at the National Bison Range, where the fraction of plant biomass that is edible (nutritionally useable) to grasshoppers is far less variable than total plant biomass, which

-:§

A)

125 100

75 50 25

1994

s UJ UJ O .......................... ~

~ o B) iii 125 Iz ~ c..

100 75

1994

125 100

75 50 25

1996

O ........................... ~

125 100

75 50 25

1996

OL-II ..... ~-----' O"fb~trI~~~

~~f:~~L.~fl~ c:pv- c:p ... +0

TREATMENT

Figure 2. Annual results for plant biomass in October from 1994 and 1996 with natural precipitation (A) and added water (B)

is largely inedible [10]. With normal precipitation, grasshoppers reduced plant abundance by only 12.7%, which was not significant. Furthermore, grasshopper consumption did not carry over to reduce the next spring's plant abundance.

Regardless of precipitation, grasshoppers enhanced N availability in soil through the deposition of frass. However, even with greater availability of N, plants could not compensate for grasshopper consumption without normal precipitation. However, by next spring, differences in N availability disappeared and next spring's plant abundance was independent of grasshopper consumption.

Therefore, potential grasshopper damage to forage may be less dependent on grasshopper density than precipitation, and may not influence long-term plant abundance. However, this may suggest that droughts lead to grasshoppers creating economic damage that necessitates grasshopper control.

10 8 6 4 2

1996

O...-..... ~ ............

1996

TREATMENT

Figure 3. Annual results for N availability over the plant growing season in 1994 and 1996 with natural precipitation (A) and added water (B)

1.4. MULTIPLE YEAR (LONG TERM) STUDIES

13

The within year studies presented above are typical of how grasshopper economic damage has been assessed. However, grasshopper consumption might have longer lasting impacts on plant abundance. First, long-term grasshopper consumption might eventually weaken plants and lead to their death or reduced production. This effect is well established for grazing by livestock [21, 22]. Second, it has long been hypothesized that herbivory might modify nutrient cycles in ecosystems, which can either reduce or increase plant abundance depending on whether nutrient cycling is slowed down or speeded up [23]. Herbivory might change nutrient cycling by deposition of excrement, changing the quantity and quality (a nutrient content and decomposition rate) of plant litter, and by sequestering nutrients in herbivore bodies. These long-term effects may be far more important than the short-term perspective commonly taken towards grasshopper consumption.

Herbivory's hypothesized effects on nutrient cycling are summarized in Fig. 4.

14

A) B) HERBIVORE .......... .

/, FASTER SLOWER

DECOMPOSED DECOMPOSED p,LANTS PLANTS

I ,~ \ •...•

I NUTRIENT .... :. AVAIL1BILln:......."

...... '.l... T ~ ~ DETRITUS

HERBIVORE ............ :

/,~ FASTER SLOWER:

DECOMPOSED DECOMPOSED: , RLANTS PLANTS: ~/ '\-.....•

l... NUTRIENT .... , AVAIL1BILln: .... ··· ~

~ t ~~" DETRITUS

FAST CYCLE ...••••.. SLOW CYCLE - - -

Figure 4. The conditions for herbivores to decrease nutrient cycling and primary productivity (A) and to increase nutrient cycling and primary productivity (8) are presented. The solid lines reflect the trophic transfers in the ecosystem. The broken lines reflect the pathways by which nutrients are recycled in the ecosystem. Line thickness reflects the relative magnitude ofthe trophic transfer or recycling pathway

The release of nutrients from excrement and the corpses of dead herbivores have been termed the Fast Cycle [24], because these sources of detritus are readily decomposed to make the nutrients rapidly available for plant uptake. The release of nutrients from plant litter has been termed the Slow Cycle [24], because this source of detritus slowly decomposes so that nutrients are slowly released for plant uptake. Furthermore, the Slow Cycle can be further affected by herbivory, if herbivores preferentially feed on different plant species which differ in how rapidly their litter is decomposed [25, 26]. Therefore, if the herbivore preferentially feeds on plants that produce slower decomposing litter, thus removing them from the litter pool, then the Slow Cycle may become faster, while preferential feeding on plants that produce faster decomposing litter may slow down the Slow Cycle.

The hypothesized net effect of this proportional shifting of nutrient cycling between Fast and Slow Cycles is for preferential feeding on slow decomposing plants to increase nutrient cycling, and preferential feeding on fast decomposing plants to decrease nutrient cycling [26]. It has been hypothesized that slow decomposing plants are competitively superior when nutrients are less available, and fast decomposing plants are competitively superior when nutrients are more available [26]. This would lead to herbivores changing the competitive outcome between the two types of plant species as herbivory speeds up or slows down nutrient cycling, which would produce different

15

cycling can lead to ecosystems that have more (faster nutrient cycling) or less (slower nutrient cycling) plant production. However, whether faster nutrient cycling enhances ecosystem primary production depends on whether the effects of accelerated nutrient cycling overshadow the deleterious effects of herbivory on plant growth and survival. In terrestrial ecosystems, this has largely been examined for mammalian herbivores [24 - 29], but may be more important for insect herbivores [30, 31].

1.4.1 Experimental Design Starting in May 1994, 24 - 1 m2 areas were established as replicated mesocosms of the study site's ecosystem and maintained for 6 years. Eighteen areas were covered by cages to contain grasshoppers so their densities could be manipulated and nine areas (6 - 1 m2 and 3 - 9m2 areas) served as controls (field grasshopper densities). Garden edging was buried 15 cm in the soil to delineate areas. Cages were constructed of nylon insect screens, supported with a PVC frame, and attached to garden edging with plastic clips. Each cage had a sleeve that permitted access to the mesocosm.

The mesocosms were manipulated in three ways. Experiment A. In six areas, grasshopper density was manipulated (50% or 125% of field density for each year: 3 replicates of each) and each area received plant litter produced in an area with field grasshopper density (control areas). This manipulation examined the Fast Cycle (variable grasshopper consumption/litter produced by a constant level of grasshopper consumption). Experiment B. In six areas, grasshopper density was held constant (field density for each year) and litter was manipulated (litter produced in an area where grasshopper density was 50% or 125% of field density for each year - Exp. A, above: 3 replicates of each). This manipulation examined the Slow Cycle (constant grasshopper consumption/litter produced by a variable level of grasshopper consumption). Experiment C. In six areas, grasshopper density and litter were manipulated (50% or 125% of field density for each yearllitter produced in each: 3 replicates of each). This manipulation examined net effects of Slow and Fast Cycles.

Litter exchanges (Experiment A and B) were conducted in early October by collecting dead plant material and clipping live plant material at a height of 1.5 cm in each area, removing it, weighing it with a handheld Pescola spring balance and then spreading it on the appropriate area. To keep the litter in the appropriate area from October - May when the cages were removed to prevent winter damage, avian netting (2.54 cm mesh) was staked over each area. The six areas not receiving litter exchanges (Experiment C) had their litter manipulated in the same way to control for any disturbance due to litter collection and clipping.

Grasshopper field density at the site was measured weekly by counting the number of grasshoppers in 24 - 0.1 m2 rings [11] at ~lla.m. and ~ p.m.; individuals also were identified to a developmental stage: early (1 st _3 rd instar), late (4th - 5th instar) and adult. Cages were stocked with M sanguinipes nymphs ~ 3rd instar) in mid-June to 50, 100 or 125% of the current field density. M sanguinipes was the most common grasshopper at the site. Individuals in each cage (Experiments A - C) were counted weekly [10] in 3 -

16

0.05 m2 rings at -lla.m. and -4 p.m .. M sanguinipes individuals were weekly added or removed from each cage (Experiments A - C) to maintain experimental densities. Added individuals were the current most common developmental stage and removed individuals were killed and their corpses were left in the cage.

The 6 - 1m2 control areas (field grasshoppers), where litter was clipped, but not removed, were compared to the 3 - 9m2 uncaged control areas (field grasshoppers), where litter was neither clipped nor removed, to assess the impact of small spatial scale, litter manipulation methods, and cages on ecosystem measurements.

The following measurements were made each year in each of the 27 areas. 1. An index ofN availability to plants was obtained by measuring N absorbed

by an ion exchange resin bag placed in the ground in May and removed in October to reflect the plant-growing season and another placed in the ground in October and removed in May to reflect the initiation of plant growth in the spring.

2. Total soil N content, inorganic soil N content and percent soil water content (1 - dry weight at 100°C for 48 h1fresh wet weight) were measured for soil taken from a core (15 cm) in May and October.

3. In vitro N mineralization was measured as the difference in inorganic N between the above soil core and soil maintained between sampling periods in an incubation tube (capped PVC tube: 2.54 cm diameter x 15 cm) that was buried to 15 cm in May and removed in October and another placed in the ground in October and removed in May [32, 33].

4. PlantN content was measured with a 5 g-dry sample « 2% of plant biomass in area) collected in June.

5. Plant litter content was measured from a 5 g-dry sample « 2% of litter biomass in area) collected in October.

6. Plant species composition and proportion of bare ground were assessed in July by point-sampling [34] at 100 points along 4 transects in each area.

7. Litter decomposition rate for common sources (field) of the two most abundant grasses (P. pratensis and E. smithi!) was measured using decomposition bags. Four ny Ion mesh (1 mm) bags of each plant species (10 g collected in June from the field) were placed on the soil in May and collected in October and May over the next 2 years. Decomposition was measured as percent dry matter (60°C for 48 h) change and percent change in total N.

8. Plant aboveground production and living biomass were measured using radiometer readings taken every 2 weeks from May - October. The radiometer measures the ratio of far-red/infrared reflected radiation and this ratio can be converted into living plant biomass [35, 36]. The conversion into living plant biomass was based on a regression between green plant biomass measured by clipping and the radiometer readings for 10 - 0.10 m2 areas at the study site at the start of the experiment and 3-4 areas (0.1 m2) every 2 weeks from May - October. Annual plant production is computed as the

17

May plant biomass plus all positive differences between consecutive biomass measures made every 2 weeks from May - October.

9. Plant damage by grasshoppers was measured by examining 50 blades of P. pratensis and 50 blades of E. smith;;, and recording whether they had been fed on by grasshoppers.

Plant and litter N measurements were made by extracting N using micro-Kjeldahl methods. Inorganic N in the soil and resin bags was extracted using 2N KCI [14, 15]. The N in extractants (g N/g dry of soil, plant or litter) was assessed colorimetrically with an autoanalyzer [14, 16 - IS].

Monthly precipitation values were available from a USFWS weather station located within 4 km of the study area and at the same elevation.

1.4.2. Results Soil, plant composition and plant production measurements for the 6 - 1m2 control areas did not differ from the 3 - 9 m2 control areas for each year over the six year study (Table I), while values did significantly vary among years. This indicated that the small

TABLE I. A comparison of the ecosystem characteristics for control areas where plant litter is clipped versus unC\ipped

Characteristic Statistic Degrees offreedom P

Primary production F=0.21 1,39 <0.64

Annual soil nitrogen F =0.032 1,28 <0.86

spatial scale of the mesocosms, caging, and litter manipulation methods were unlikely to be responsible for any differences measured in the areas where grasshopper density and litter were experimentally manipulated. Therefore, the mesocosms are likely to reflect the processes of the larger ecosystem.

Annual effects of grasshopper consumption on forage availability measured in long-term experiments (Experiments A - C) were similar to annual effects in short-term experiments. The proportional reduction in fall (Oct.) plant biomass was positively correlated with early instar (hatchling) grasshopper density (P < 0.000 I ) and spring (May) plant biomass (P < 0.0001), but was negatively correlated with precipitation during the plant growing-season (May - Oct.: P < 0.0001) (Fig. 5a). Grasshoppers decreased fall (Oct.) plant biomass by up to 63%, averaging 26.1 ± 20.9% (SD). Grasshopper consumption had no effect on plant biomass for the next spring (P < 0.20) and spring plant biomass was positively correlated with the precipitation during the current (0.0001) and past growing-seasons (P < O.OOOI)(Fig. 5b). Therefore, grasshopper consumption can reduce forage abundance within a year, especially in dry years, but this short-term reduction does not depress the next year's forage abundance.

18

Nonetheless, long-tenn effects of grasshopper consumption on forage availability were observed in the experimental mesocosms. Measurements for manipulated mesocosms

A)

0.7

C 0.6 WZ ..... 00 =,.:: ..... Co.. w:; .,.,. 1I::::l ZI/I ~ oz FO 00

". .. iHi IL

,o~G'I'o (t'C),,o~"'Q

-1~\S'~ o",.s'o

B) I"~"" .-II: -W I-W ::E -0 0 ., .... ~ (!) Z 1>: "-II)

L£.~;t1;f1':t77 .. . i~ACTION REDUCED = ~.25 + 0.002 SPRING + 0.003 DENSITY .0.036 PRECIPITA nON

r2 = 0.62, N = 90, P < 0.00001

DENSITY (#/m')

, , 150 .. 100

50 o

SPRING RADIOMETER = 11.26 + 11.93 CURRENT PRECIPITA nON + 10.21 PREVIOUS PRECIPITAnON

r2 = 0.51, N = 72, P < 0.00001

Figure 5. The within-year results from the long-term experimental mesocosms are presented. A) examines the factors controlling the fraction of plant biomass removed by grasshoppers by October,

and B) examines the factors controlling the following year's spring (May) plant biomass

were expressed as percent change relative to control areas since the start of the experiment:

O~ Ch 100 S/So /0 ange = x --------C/Co '

where Sj is the mesocosm's spring plant biomass in a year i; So is the mesocosm's spring plant biomass at the start of the experiment (May, 1994), C j is the average spring plant biomass in a year i for the 6 - 1 m2 control areas, and Co is the average spring plant biomass at the start of the experiment (May, 1994) for the 6 - 1 m2 control areas.

19

With these measures of percent change, the following emerged: Experiment A - Fast Cycle Experiments (varied grasshopper density/control-level litted. While values varied among years, spring (May) plant biomass increased relative to controls (field density) with increased grasshopper density (125% of field) and decreased compared with controls with decreased grasshopper density (50% offield) (Fig. 6a; ANDVA: F = 7.0; df= 1,28; P < 0.01). This is expected because a greater grasshopper density produces more frass, which makes more N available to plants (increased Fast Cycle). Experiment B - Slow Cycle Experiments (control grasshopper density/varied litter production). While values varied among years, spring (May) plant biomass increased relative to controls (field density) with litter produced by increased grasshopper densities (125% of field) and decreased compared with controls with litter produced by decreased grasshopper densities (50% of field) (Fig. 6b; ANDV A: F = 22.05; df = I, 28; P < 0.000 I). This means that grasshoppers are speeding up the Slow Cycle, so that more N is released by decomposition of plant litter. Experiment C - Net Effect of Fast and Slow Cycles (varied grasshopper density and in situ produced litter). As would be expected from the above results, the net effect was that spring (May) plant biomass increased relative to controls (field density) with increased grasshopper densities and decreased compared with controls with decreased grasshopper densities (Fig. 7; ANDV A: t = 4.25, df = 4, P < .02). Furthermore, factoring in the measured reduction of forage by grasshopper consumption, forage availability in the fall was still greater with greater grasshopper densities. Therefore, grasshoppers enhanced the ecosystem's overall plant production.

Each year, experimentally increased grasshopper density led to greater plant production than experimentally reduced grasshopper density; however, the degree of change in productivity varied between years and will be discussed later. How increased grasshopper densities impacted the ecosystem's plant production is complicated and has many facets.

Fast and Slow Cycle effects were not additive, but antagonistic (Fig. 8), because Fast Cycle effects, rather than increasing the net effect, led to its decrease. Furthermore, Slow Cycle effects were twice as strong as Fast Cycle effects. Therefore, the Slow Cycle effects of increased grasshopper densities explained the increase in plant abundance relative to areas with control and decreased grasshopper densities. This is expected, because a greater relative reduction in plant biomass by grasshoppers will decrease the relative abundance of litter, which will make the Slow Cycle contributions to N availability for plant growth less important. Nonetheless, the net effects ofthe Fast and Slow cycles with increased grasshopper densities were to increase the availability ofN for plant growth as measured by ion exchange resin bags (Fig. 9), which is expected given that plant growth at this site is N-limited.

20

A) 15

10

W o DECREASE

(!) 5 • INCREASE

Z I!!I!RATIO

« 0

:J: -5 U ~ -10 Z W B) u 0:= 25 W 20 D..

15

10 o DECREASE

5 • INCREASE 121 RATIO

0

-5

-10

YEAR

Figure 6. The effects of the Fast (A) and Slow (B) Cycles relative to control areas <± SE) by year in the long-term experimental mesocosms when grasshopper densities are decreased by 50% or increased by

25%. The ratio <± SE) refers to the net effect between the decreased and increased grasshopper densities

The possible mechanisms by which the Slow Cycle effects emerge are presented in Fig. 9. Litter quantity was lower in each year with increased grasshopper density as

w C) z ~ o .... z w o a= w a.

20

15

10

5 .

o

.1 1 ..... ~ .....

....

rL

.....

I I .

1995 1996 1997 1998 1999

YEAR

DFASTCYCLE

DSLOWCYCLE

• TOTAL

21

Figure 7. The cumulative effect of the Fast and Slow Cycles relative to control areas (.± SE) in the long-term mesoscosms when grasshopper densities are decreased by 50% or increased by 25%. The total refers to the

net effect (± SE) between the decreased and increased grasshopper densities

would be expected with greater consumption. However, the N content of the litter, which reflects how easily it can be decomposed, was greater with increased grasshopper

r2 = 0.99, N = 5, P < 0.003

20

I-15 (J

III U. U. III ..J

~ 0 I-

12 8 SLOV4 CyCLE

Figure 8. The combination of Fast and Slow Cycles to produce a cumulative or total effect

22

Iii' ~u d~ yO

a..ri II

:t::. :t::. :t::. :t::. :t::. 25

20

~ 15

5 10 DDECREASE

ffi 5 DINCREASE

0 M -5 D..

-10

Figure 9. The average effects of decreasing grasshopper density by 50% and increasing grasshopper density by 25% in the long-term experimental mesocosms

densities. The increase in litter N content would be expected if the grasshoppers preferentially fed upon plant species oflower N content that decompose more slowly. This was observed because 1) a greater proportion of blades of E. smithii than P. pratensis had been fed on by grasshoppers by October in each year (X2 = 9.45, df= 1, P < 0.002), E. smithii decomposed 11% more slowly than P. pratensis in litter bags (t = 5.10, df = 2, P < 0.03), and 3) E. smithii contained 28% less N than P. pratensis (t = 13.76, df = 8, P < 0.0001). Therefore, the conditions for herbivory to speed up the slow cycle and thereby, make more N available for plant growth and enhance ecosystem primary productivity (Fig. 4b) were observed. Why these grasshoppers should preferentially feed on a plant that is lower in N content is not known.

Concomitant with the observed changes in the Slow Cycle, P. pratensis and P. compresa were observed to increase in abundance relative to E. smithii as grasshopper density increased (Fig. 10). This would further speed up the Slow Cycle and enhance the ecosystem's primary production, because P. pratensis decomposes faster than E. smithii. P. pratensis is a better competitor at higher N availabilities [37, 38] and is less preferred by grashoppers, which should lead to its increase in abundance compared with E. smithii. Therefore, the grasshoppers change plant composition and create a more productive ecosystem.

23

3

w F = 48.18; df = 2,6; P < 0.0005 U 2.5 z i:§

2 z ::l m

1.5 < ~

1 ~ iii

l 0.5

0

DECREASED CONTROL INCREASED

TREATMENT

Figure 10. The average effect of decreasing grasshopper density by 50% and increasing grasshopper density by 25% in the long-term experimental mesocosms

The grasshoppers' influence on the Slow Cycle might counter the enhancement of ecosystem productivity in two ways. First, the decrease in litter quantity might lead to greater soil evaporation so that plants have less moisture available to them and become water-limited, but this was not observed (Fig. 9). Second, the decrease in litter quantity and increase in litter and soil N might make the soil microbes responsible for decomposition Climited rather than N-limited and decrease decomposition rates [39, 40]. While the decomposition of a common litter source was slower with increased grasshopper densities, the difference was very small and statistically insignificant. Therefore, none of the countermechanisms to herbivory increasing the Slow Cycle and increasing ecosystem productivity were observed.

2. Tools

Very simply, grasshoppers enhance ecosystem productivity and forage availability to mammalian herbivores at this site. Therefore, consumption of plants by grasshoppers cannot be viewed as detrimental and would not warrant control, i.e., there is no economic damage even though grasshoppers decrease fall plant abundance in a given year. This means that the standard measure of grasshopper economic damage is not valid at this site.

Enhancement of plant productivity by grasshoppers changes with adult grasshopper density (Fig. 11), even though enhancement was observed at all adult grasshopper densities observed in the study. At low adult densities, enhancement of primary production is not great, because low densities do not consume enough plant biomass to appreciably affect the Slow Cycle. At very high adult grasshopper densities, enhancement of primary production

24

is not great, because high densities consume so much plant biomass that plants are damaged and cannot take advantage of increased nitrogen availability. Therefore, greatest benefits emerge at an intermediate adult grasshopper density.

W (!) Z c( :t: 0 ...J c( I-0 I-::::S! 0

3. Solutions

25

20

15

10

5

0 0

Y = -16.91 + 3.52X - 0.08X2 N = 5, r2 = 0.99, P < 0.001

10 20 30

ADULT DENSITY (#/m2)

40

Figure II. The effect of grasshopper density on primary productivity

3.1. THIS ECOSYSTEM

One might argue that the grasshoppers' enhancement of ecosystem productivity might not emerge at lower or higher grasshopper densities than observed in the study and grasshoppers might need to be controlled at densities outside this range. Let us consider this. First, an enhancement was observed at densities of7 adults/m2 (1994), but would grasshopper control be warranted at lower densities, even if lower densities decreased ecosystem productivity? Second, an enhancement was observed at densities of36 adults/m2 (1997), but how often do densities become greater than this and perhaps diminish ecosystem productivity? Therefore, when would control be warranted?

These findings raise a new perspective on grasshopper control: might grasshoppers be controlled at an optimum level to maximize their enhancement of ecosystem

25

productivity? Before addressing this question, a caveat needs to be made: this is a very complex question to address and the limited range of grasshopper densities and precipitation observed in this study do not provide sufficient detail to adequately address this management possibility. Nonetheless, this study suggests that there may be an optimum grasshopper density that maximizes ecosystem productivity: 21.5 adults/m2 increasing ecosystem production by 20.9% or 19.6% after grasshopper consumption (Fig. 11). Furthermore, grasshoppers would diminish ecosystem productivity when their densities fall below 5.5 adults/m2 or increase above 37.4 adultslm2. The above predictions would change with precipitation; drought would diminish the enhancement of ecosystem productivity by grasshoppers and reduce the range of grasshopper densities that lead to increased primary productivity, while greater precipitation would increase the enhancement of primary productivity and increase the range of grasshopper densities producing it. This is particularly important for the Slow Cycle, which is the dominant factor for enhancing the ecosystem's productivity.

To achieve this optimization, much more detailed monitoring of grasshopper densities and precipitation than is usually available would be needed and this may not be economically feasible. Managers might have to be content with letting grasshoppers provide whatever enhancement naturally emerges. This might not be so bad, because grasshoppers provided enhancement in every year over the six years of the study.

3.2. OTHER GRASSLAND ECOSYSTEMS

One might argue that the ecosystem studied here is not typical of grasshopper/rangeland conditions and grasshoppers might not enhance productivity in many other ecosystems. There are a number of ways that an ecosystem might differ from that studied here.

First, other ecosystems might not provide the conditions whereby grasshopper consumption increases ecosystem productivity, i.e., preferential feeding on the slower decomposing plant species. Interestingly, grasshoppers at another site within 4 km preferentially fed on P. pratensis over E. smithii. At this second site, P. pratensis decomposed faster and was higher in N content. Therefore, the grasshoppers at this other site preferentially feed on the faster decomposing plant, which would tend to diminish ecosystem productivity. The second site has similar rainfall and elevation, but plant production is only 75% of my experimental study site. This raises several issues. Why would grasshoppers change their feeding preferences between the two sites? How often do grasshoppers preferentially feed on slower decomposing plants, and thereby, tend to enhance ecosystem productivity? What is the proper spatial scale to differentiate between areas where grasshoppers enhance and diminish ecosystem productivity? Obviously, the spatial scale employed in grasshopper control programs (thousands of hectares) is too large given my observations to discriminate between areas where grasshoppers potentially enhance vs. diminish ecosystem productivity.

Second, how similar is the study ecosystem to other rangeland ecosystems? 1. The range of plant productivity and precipitation values observed at my study

site includes the values reported for many grasslands [41, 42].

26

2. Large non-domestic mammalian herbivores are very abundant at my study site (-2.5 g/m2 or the equivalent of -0.1 cattle/ha). The greatest density of large mammalian herbivores reported for the Great Plains of North America by early European explorers was -5-9 glm2 [43, 44]; therefore, this site supports an abundance of large mammalian herbivores by North American standards. Furthermore, cattle grazing is one of the reasons why unprotected Palouse Prairie has been destroyed and the National Bison Range is the largest tract of undisturbed Palouse Prairie remaining.

3. Palouse Prairie sites generally are known to have historically supported large grasshopper densities [7 - 9]; therefore, grasshopper densities reported in my study are not unusual. In comparison to large mammals, the grasshoppers annually consume 1.25 - 2.25 times more plant biomass. However, grasshopper densities in other grasslands may not be as great, e.g., North American tall grass prairie [45, 46].

4. The incidence of lightning-caused wildfires is much lower in Palouse Prairie than other North American grasslands [47].

Consequently, without repeating my experiments in other grasslands, one cannot ascertain how general the results are, and ifthey do not hold in other ecosystems, why this is the case.

Third, could large mammalian herbivores be substituted for grasshoppers to provide the same enhancement of ecosystem productivity? This is not a simple substitution of consumption by one herbivore for another, because large mammalian herbivores impose damage to plants beyond consumption through trampling, soil compaction and uprooting of plants. These additional impacts will tend to diminish ecosystem productivity [48,49]. Furthermore, at my study site, I added large mammal feces to cages in an amount (g-dry/m2) comparable to the field deposition by mammals and monitored plant production. Even though the large mammals annually deposit a comparable amount ofN in feces and urine as grasshoppers do in frass (0.47 glm2/yr vs. 0.5 glm2/yr), plant production was greater with frass than with mammalian feces in both fall (109%) and spring (28%) (repeated measure ANOVA: F = 7.28; df= 1,8; P < 0.03). The main reason for this is that mammalian feces decompose slower than frass.

Fourth, grasshoppers at other locations may be more likely to enhance nutrient cycling by feeding on plant litter, which would speed up the Slow Cycle. The dominant grasshopper at my study site and the species employed in the experiments, M sanguinipes, does not consume litter to any great extent, nor do any of the other Melanoplines [50]. However, other grasshopper species, especially Gomphocerines, often consume large quantities of plant litter [51, 52], which would lead to a more rapid release of nutrients from plant litter by shifting the mechanism from the Slow Cycle (litter decomposition) to the Fast Cycle (grasshopper frass).

Therefore, it is unclear how often grasshoppers increase ecosystem productivity as I observed in the experiments. Nonetheless, what is clear is that a reduction in plant biomass by grasshopper consumption does not necessarily imply that grasshoppers are making forage less abundant for mammalian herbivores and does not necessarily indicate the need for grasshopper control. I suspect that grasshoppers may frequently enhance primary production in grassland ecosystems, especially at locations where grasshoppers

27

consume considerable quantities of plant litter, and grasshoppers are certainly more likely to enhance primary production than large mammals with their uprooting of plants and soil compaction.

Finally, these results indicate that we cannot assess the impact of grasshoppers on rangeland ecosystems without a better understanding of grasshopper population dynamics [53, 54]. For example, if higher grasshopper densities enhance nutrient cycling and grasshopper populations are unable to attain these densities, the ecosystem will not exhibit higher productivity. This means that grassland ecosystem functioning may be dependent not only on the presence of grasshoppers, but their abundance.

3.3. CONCLUSION

The observations from the ecosystem studied here and the implications for grasshopper control typify the "state of the art" in our management of nature. As we conduct careful experimental studies of ecosystems and try to apply this knowledge, we find that what we firmly held as truths about how nature operates and how we should manage it may be wrong. Therefore, we should be humbled by how little we know and be very cautious in the application of our knowledge to management.

As we gain more knowledge about ecosystems, we are continually discovering that many of the component species are playing crucial roles in maintaining the ecosystem's sustained operation (i.e., health). The functioning of ecosystems depends upon the component species and species may not easily be substituted and still maintain these functions [55]. Therefore, long-term sustainability of ecosystem productivity may be the essence of the biological wealth provided by biodiversity and we need to understand this value better and protect it.

4. Acknowledgements

I wish to thank The Utah State Agricultural Experiment Station and The National Science Foundation (DEB-9317984, DEB-9707654) for financial support. J.B. Slade, D. Branson, and J. Chase aided in establishing the long-term experiments. J.B. Slade read the manuscript and provided criticism.

5. References

l. Wilson, E. O. (1989) Threats to biodiversity, Sci. Amer. Sept., 108-116. 2. Daily, G. c., Matson, P. A., and Vitousek, P. M. (1997) Ecosystem services supplied by soil, in G. C.

Daly (ed.), Nature's Services: Societal Dependence on Natural Ecosystems, Island Press, Washington, D.C., pp. 113-132.

3. Costanza, R. and Folke, C. (1997) Valuing ecosystem services and efficiency, fairness, and sustainability as goals, in G. C. Daly (ed.), Nature's Services: Societal Dependence on Natural Ecosystems, Island Press, Washington, D.C., pp. 49-68.

4. van Kooten, G. C. (1998) Economics of conservation biology: a critical review, Env. Sci. Pol. 1. 13-25.

28

S. Schmitz, O. J. (1993) Trophic exploitation in grassland food chains: simple models and a field experiment, Oecologia 93, 327-33S.

6. Branson, D. (1999) Grasshopper life history strategies in a variable environment. Ph.D. dissertation, Utah State University, Logan.

7. Shelford, V. E. (1963) The Ecology of North America, University of Illinois Press, Urbana. 8. Rogers, L. E. and Uresk, D. W. (1974) Food plant selection by the migratory grasshopper (Melanoplus

sanguinipes) within a cheatgrass community, Northwest Science 48, 230 - 234. 9. Sheldon, J. K. and Rogers, L. E. (1978) Grasshopper food habits within a shrub-steppe community,

Oecologia 32, 8S-92. 10. Belovsky, G. E. and Slade, J. B. (199S) Dynamics of some Montana grasshopper populations:

relationships among weather, food abundance and intraspecific competition, Oecologia 101,383-396. II. Onsager, J. A. and Henry, J. E. (1977) A method for estimating the density of rangeland grasshoppers

(Ortboptera, Acrididae) in experimental plots, Acridida 6, 231-237. 12. Binkley, D. and Matson, P. (1983) Ion exchange resin bag method for assessing forest soil nitrogen

availability, Soil Sci. Soc. Am. J. 47, 1050-IOS2. 13. Binkley, D. (1984) Ion exchange resin bags: factors affecting estimates of nitrogen availability, Soil

Sci. Soc. Am. J. 48, 1181-1184. 14. Page, A. L., Miller, R. H., and Keeney, D. R. (1982) Methods of Soil Analysis. Part 2, American

Society of Agronomy, Inc., and Soil Science Society of America, Inc., Madison. 15. Adamsen, F. J., Bigelow, D. S., and Scott, G. R. (1985) Automated methods for ammonium, nitrate and

nitrite in 2M KCI-phenylmercuric acetate extracts of soil, Communications Soil Science and Plant Analysis 16, 883-898.

16. Adamski, J. M. (1976) Simplified Kjeldahl determination for sea water by a semi-automated digestion method, Anal. Chem. 48, 1194-1197.

17. Nelson, D. W. and Sommers, L. E. (1980) Total nitrogen analysis ofsoi! and plant tissues, J. Assoc. Offic. Agric. Chem. 63, 770-778.

18. Association of Official Agricultural Chemists. (1984) Official Methods of Analysis, Association of Official Agricultural Chemists, Washington, D.C.

19. Pfadt, R. E. and Hardy, D. M. (1987) A historical look at rangeland grasshoppers and the value of grasshopper control programs, in J. L. Capinera (ed.), Integrated Pest Management on Rangeland: a shortgrass prairie perspective, Westview Press, Boulder, pp. 183-195.

20. Wang, T. and Walgenbacj, D. D. (1989) Economic thresholds of grasshoppers on cool season grasses, in J. Fowler (ed.), GHIPM Annual Report (FY 1989), USDAIAPHISIPPQ, pp. 124-131.

21. Milchunas, D. G. and Lauenroth, W. K. (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments, Ecol. Monogr. 63,327-366.

22. Wallace, L. L. and Dyer, M.1. (1995) Grassland management: ecosystem maintenance and grazing, in A. Joem and K. H. Keeler (eds.), The Changing Prairie: North American Grasslands, Oxford University Press, New York, pp. 177-198.

23. Hutchinson, G. E. and Deevey, E. S. (1949) Ecological studies on populations, Biological Progress I, 325-359.

24. McNaughton, S. J., Ruess, R. W., and Seagle, S. W. (1988) Large mammals and process dynamics in African ecosystems, BioScience 38, 794-800.

2S. Pastor, 1., Naiman, R. J., Dewey, B., and Mcinnes, P. (1988) Moose, microbes, and the boreal forest, BioScience 38, 770-777.

26. Pastor, J. and Naiman, R. J. (1992) Selective foraging and ecosystem processes in boreal forests, American Naturalist 139, 690-70S.

27. Frank, D. A. and McNaughton, S. J. (1993) Evidence for the promotion of aboveground grassland production by native large herbivores in Yellowstone National Park, Oecologia 96,157-161.

28. Pastor, 1., Dewey, B., Naiman, R. J., McInnes, P. F., and Cohen, Y. (1993) Moose browsing and soil fertility in the boreal forests ofisle Royale National Park, Ecology 74, 467-480.

29. Holland, E. A., Parton, W. J., Detling, J. K., and Coppock, D. L. (1992) Physiological responses of plant populations to herbivory and their consequences for ecosystem nutrient flow, American Naturalist 140, 68S-706.

30. Seastedt, T. R. and Crossley, D. A. (1984) The influence of arthropods on ecosystems, BioScience 34, 157-161.

29

31. Lovett, G. M. and Ruesink, A. E. (1996) Carbon and nitrogen mineralization from decomposing gypsy moth frass, Gecologia 104, 133-138.

32. Binkley, D. and Hart, S. C. (1989) The components of nitrogen availability assessments in forest soils, Adv. Soil Science 10,57-112.

33. Wedin, D. and Tilman, D. (1990) Species effects on nitrogen cycling: a test with perennial grasses, Gecologia 84, 433-441.

34. Daubenmire, R. F. (1947) Plants and Environment, John Wiley and Sons, New York. 35. Pearson, R. L., Miller, L. D., and Tucker, C. J. (1976) Handheld spectral radiometer to estimate

gramineous biomass, Applied Gptics 15, 416-418. 36. Milton, E. J. (1987) Principles offield spectroscopy, Int. J Remote Sensing 8, 1807-1827. 37. Smika, D. E., Hass, H. J., and Power, J. F. (1965) Effects of moisture and nitrogen fertilizer on growth

and water use by native grass, Agronomy Journal 57, 483-486. 38. Tilman, D. (1988) Plant Strategies and the Dynamics and Structure 0/ Plant Communities, Princeton

University Press, Princeton, New Jersey. 39. Seastadt, T. R. (1995) Soil systems and nutrient cycles ofthe North American prairie, in A. Joern and

K. H. Keeler (eds.), The Changing Prairie: North American Grasslands, Oxford University Press, New York, pp. 157-174.

40. Zheng, D. W., Agren, G. I., and Bengtsson, J. (1999) How do soil organisms affect total organic nitrogen storage and substrate nitrogen to carbon ratios in soils? A theoretical analysis, Gikos 86, 430-442.

41. Lauenroth, W. K. (1979) Grassland primary production: North American grasslands in perspective, in N. French (ed.), Perspectives in Grassland Ecology, Springer-Verlag, New York, pp. 3-24.

42. Bourliere, F. and Hadley, M. (1970) The ecology of tropical savannas, Ann Rev. Ecol. Syst. I: 125-152. 43. Watts, L. R., Stewart, G., Connaughton, c., Palmer, L. J. and Talbot, M. W. (1936) The management

of range lands, u.s. Senate Doc. t 99, 501-522. 44. French, N. R., Steinhorst, R. K. and Swift, D. M. (1979) Grassland biomass trophic pyramids, in N.

French (ed.), Perspectives in Grassland Ecology, Springer-Verlag, New York, pp. 59-87. 45. Evans, E. W. (1989) Interspecific interactions among phytophagous insects of tall grass prairie: an

experimental approach, Ecology 70, 435-444. 46. Mitchell, J. E. and Pfadt, R. E. (1974) A role of grasshoppers in a shortgrass prairie ecosystem,

Environ. Entomot. 3,358-360. 47. Bragg, T. B. (1995) The physical environment of Great Plains grasslands, in A. Joem and K. H. Keeler

(eds.), The Changing Prairie: North American Grasslands, Oxford University Press, New York, pp. 49-81.

48. Rietkerk, M. and van de Koppel, J. (1997) Alternate stable states and threshold effects in semi-arid grazing systems, Gikos 79, 69-76.

49. Wallace, L.L. (1987) Effects of clipping and soil compaction on growth, morphology and mycorrhizal colonization of Schizachyrium scoparium, a C. bunchgrass, Gecologia 72, 423-428.

50. Mulkern, G. B., Pruess, K. P., Knutson, H., Hagen, A. F., Campbell, J. B., and Lambley, J. D. (1969) Food habits and preferences of grassland grasshoppers of the north central Great Plains, North Dakota Agr. Exp. Sta. Bull. #481.

51. Pfadt, R. E. and Lavigne, R.I. (1982) Food habits of grasshoppers inhabiting the Pawnee site, Agr. Exp. Sta. Univ. Wyo. Sci. Mono #42.

52. Pfadt, R. E. (1994) Field guide to common western grasshoppers, Wyo. Agr. Exp Sta. Bull. #942. 53. Joern, A. and Gaines, S. B. (1990) Population dynamics and regulation in grasshoppers, in R. F.

Chapman and A. Joern (eds.), Biology o/Grasshoppers, John Wiley and Sons, New York, pp. 415-482. 54. Belovsky, G. E. and Joern, A. (1995) The dominance of different regulating factors for rangeland

grasshoppers, in N. Cappuccino and P. Price (eds.), Population Dynamics: New Approaches and Synthesis, Academic Press, New York, pp. 359-386.

55. Tilman, D. (1997) Biodiversity and ecosystem functioning, in G. C. Daly (ed.), Nature's Services: Societal Dependence on Natural Ecosystems, Island Press, Washington, D.C., pp. 685-700.

2. WHAT ARE THE CONSEQUENCES OF ECOSYSTEM DISRUPTION ON ACRIDID DIVERSITY AND ABUNDANCE?

Abstract

F.A. GAPPAROV1 and A.V. LATCHININSKy2 J Uzbek Institutefor Plant Protection (UzNIIZR) 700140 Tashkent, Uzbekistan

2 Entomology Section, Department of Renewable Resources University of Wyoming Laramie, WY 82071-3354, USA

Since late 1950s, the water of the two tributaries of the Aral Sea, the Amu Darya and Syr Darya rivers, has been contaminated by agrochemicals and increasingly diverted for irrigation of cotton and rice monocultures. This resulted in the rapid desiccation of the Aral Sea which lost >50% of its area and >80% of its volume by 1999. The shoreline receded by > 100 km, and the exposed sea bottom became a source of toxic dust that feeds frequent dust storms and further contaminates the environment. The vast areas of the rivers' deltas occupied by reeds shrunk by 70% losing a substantial part of their once abundant fauna. The desertification of the milieu led to the simplification of plant communities and the predominance of halophilous, weedy vegetation. Grasshopper communities have also changed accordingly, with mesic species being substituted by xeric forms, invading new habitats from the adjacent deserts. The once predominant Migratory Locust in the Aral Sea basin lost its economic importance to the Italian Locust. This species readily colonizes roadsides, fallows and weedy patches which constitute its most favorable breeding areas. As a result, some grasshopper species now demand continuous attention inflicting damage to cereal crops, and the Italian Locust became a permanent menace to the dwindling local agriculture, requiring extensive chemical treatments regularly. Recurrent insecticide interventions contribute to further deterioration of the ecological conditions in the Aral Sea basin.

31


32

1. Introduction

In even the smallest drop of water there is a grain of gold

Uzbek proverb

The 20th century witnessed many ecological catastrophes of different scales. The most severe of these, such as the radiation disaster at Chemobyl, the deforestation of the Amazon, and the drying of the Aral Sea were of anthropogenic origin. Indeed, humans have become the single most important factor jeopardizing the well-being, even the existence, of ourselves and our neighbours on this planet - individual animals and plants species, and entire ecosystems.

The example ofthe Aral Sea is the classic scenario of a man-made environmental disaster. In the initial phase, there is a poor management decision and consequent action, which is often politically driven. Then, there is a phase of short-term economic benefits from this action, which hide the long-term negative impact and, in fact, aggravate management mistakes. During this phase, the signs of deterioration of environmental quality, although obvious and apparent, are ignored or even deliberately concealed by the managers. Then, the situation begins to degenerate rapidly. Public awareness grows, but the management inertia is strong, so there is an effort to maintain an appearance of economic gains and justify satisfaction with the status quo. Eventually, it becomes clear, even to the managers, that the course of events ought to be reversed, but at this point it is usually too late. The small snowball has developed into a devastating avalanche. It is not possible to stop it, let alone reverse it. Enormous resources are now involved in trying to slow down the process. In case of the Aral Sea, the situation is now in this phase.

Our article does not have for its goal a detailed description of the Aral Sea's tragedy. We would like to investigate only one, very particular (and apparently secondary) problem: The impact of this ecological catastrophe on the locust and grasshopper situation in the region. We think that this glaring example of mismanagement and environmental neglect, reflected in the mirror of the acridid turmoil, can teach us a valuable lesson. When modifying the environment (purposefully or not), we should be ready to face new, unforeseen and unpredictable problems everywhere and with everything. Short-term economic gains may turn into economic disasters and severe losses of environmental and human health in the long run. And improving a dramatically deteriorating situation is much more difficult (if possible at all) and costly than avoiding the initial mistake or stopping the negative process of the environmental interference near the beginning.

It would be a misinterpretation to assert that the acridid pest problem was born with the shrinking of the Aral Sea. The Asiatic Migratory Locust Locusta migratoria migratoria L. and, to a lesser extent, the Italian Locust Calliptamus italicus (L.) frequently produced outbreaks in the first half of the 20th century, requiring the implementation of broad-scale control interventions [38, 52, 63]. However, the desiccation of the Aral Sea has altered and aggravated the situation. The deteriorating ecological conditions were

33

superimposed on the natural fluctuations of locust numbers, producing new and more severe economic, environmental and socio-political problems. But let us return to the Aral Sea and its surroundings. To understand what is going on today, it is necessary to look back several decades.

2. Background: a historical perspective of the problem

2.1. GEOGRAPHICAL SETTING: HYDROLOGY

The Aral Sea is actually a saltwater lake. It is located in Central Asia, in southwestern Kazakhstan and northwestern Uzbekistan, about 450 km east of the Caspian Sea (Fig. I).

Figure 1. Geographic location of the Aral Sea

The Aral Sea has no outlet, being surrounded by three vast deserts, Plateau Ustyurt, Karakum, and Kyzylkum. In the Turks' languages, the word "Aral" means "island". According to one hypothesis, the name reflects the "island" location of this water body in

34

the "sea" of the desert. Another hypothesis connects this name with the fact that the Aral Sea contained more than 300 islands. It the recent past, only some 40 years ago, the Aral Sea was the 4th largest lake on Earth - bigger than any of the North American Great Lakes except Superior. Its area exceeded 69 500 km2 with a volume of 1 064 km3. It was a shallow reservoir, with a maximum depth of 68 m and an average depth of only 15.5 m. Its eastern part, where its only tributaries, the Amu Darya and Syr Darya rivers, bring their water, was the most shallow: the isobath of 10m passed about 60 km from the littoral zone. The water in the Aral Sea was only slightly saline - ten parts of dissolved salts per thousand parts water - which is less than a third of the ocean's salinity. With 24 indigenous species of fish, the sea supported a major fishery averaging 44 000 tons per year in the 1950s [14].

For centuries, the boundaries of this sea, landlocked between three deserts, remained almost unchanged. The water level experienced gradual fluctuations of only about 3 m. According to one hypothesis, the Aral Sea and the Caspian Sea are connected by a subterranean passage and therefore act like communicating vessels. Thus, when the level increases in one of them, it decreases in the other. Some support for this hypothesis can be drawn from the fact that during the last quarter of the 20th century, when the level of water in the Aral Sea dropped dramatically, the Caspian Sea experienced a 2-m rise of its water level.

The Aral Sea's hydrological regime was more or less stable until the end ofthe 1950s: rains (3-9 km3 per year) and inflow from the Amu Darya and Syr Darya rivers (55-56 km3 per year) compensated for the high evaporation losses (58-65 kml) [29]. However, in the early 1960s, this dynamic equilibrium was broken by humans.

2.2. AGROECOLOGICAL SETTING: IRRIGA nON

In the late 1950s, the Soviet government decided to promote extensive agricultural development, particularly in Kazakhstan and Central Asia. In Kazakhstan, more than 25 million ha of virgin and fallow lands were converted into cropland between 1955 and 1966, mostly for grain production. In the countries of Central Asia, especially Uzbekistan, the focus was on cotton production to gain independence from cotton imports. Cotton, the "white gold" of Central Asia, rapidly became an overwhelming monoculture. The area of its cultivation increased three-fold in 20 years, occupying almost 70% of cropland area. By 1983, Uzbekistan alone produced almost as much cotton as the United States [48].

It is necessary to acknowledge that climatic conditions in Uzbekistan in general, and the Aral Sea zone in particular, are very suitable for growing cotton, rice, com, sorghum and other thermophilous, irrigated crops. The climate in this region is intensely continental and arid, with short but cold winters and long, hot and very dry summers. The growing season lasts from 204 to 288 days, which is the longest of anyplace in the former Soviet Union [56]. The summer, noon temperatures fluctuate from 28 to 33°C; the annual precipitation is only 100-160 mm, with the majority (up to 80%) falling in spring [32]. Given the agronomic aridity limit of 200-300 mm of annual precipitation, practically all agriculture in the region needs irrigation. With the exception of some heavily salinized soils (solontchak and solonets), the thin but mostly fertile local soils (calcisoils, sandy soils,

35

alluvial soils and some clay-pan soils) are able to support irrigated agriculture. However, such soils are concentrated along the riverbeds in the Amu Darya and Syr Darya deltas, restricting the possibilities of agricultural production to this zone.

2.3. BOTANICAL SETTING: NATIVE VEGETATION

Although the Aral Sea is encircled by the deserts, the presence of the two powerful rivers created a very particular type of a wetland ecosystem in its basin. Both the Amu Darya and Syr Darya formed vast deltas, which comprised an intricate system of channels, ponds, temporary lakes, large and small tributaries and islands. The Amu Darya delta was much bigger, more heavily exploited, and had a much higher population (about 5 million people) and economic importance than the Syr Darya delta. Therefore, we will limit our description to this ecosystem, unless specified otherwise.

The Amu Darya delta stretched from the northern limits of the city ofNukus, the capital of the Autonomous Republic of Karakalpakia (part of Uzbekistan), to the littoral zone (see Fig. 3). It occupied an enormous area of14 000 km2• The relief of the delta was formed by the Amu Darya deposits; its water had an extremely high (11 gil) content of alluvium. Every year, the river brought about 182 kmJ of the alluvium to its delta, creating banks and islands. There were more than 2 000 islands in the Amu Darya delta [29]. This ecosystem was very dynamic, experiencing several floods and dryings per year, depending on the water level of the river, which, in tum, depended on the amount of snow melting in the distant mountains.

Basically, there were two types of vegetation communities in the delta. A very particular type of fluvial forest called "tugai" occupied the immediate vicinity of the water bodies. The canopy was composed of several native species of poplars: Populus diversifolia, P. pruinosa, and P. arian, as well as the Russian olive tree (Elaeagnus angustifolia), silverberry elaeagnus (E. argentea), and eastern elaeagnus (E. orientalis). The shrubs consisted of the tamarisk (Tamarix pentandra), willow (Salix songarica) and saxaul species (Halimodendron halodendron, Haloxilon sp.). The undergrowth included several reedgrass species (Calamagrostis pseudophragmites, C. dubia), with the British yellowheadlnula britannica in the dryer habitats and the cattail (Typha minima) in the wet patches. Such fluvial forests were very dense and practically impenetrable for people, which made them excellent breeding refuges for numerous species of mammals and birds.

The fluvial forests occupied only 2 out of 14 thousand km2 of the Amu Darya delta. The rest was covered by a particular type of moist grassland, practically a virtual monoculture of the common reed (Phragmites australis). The reeds formed dense stands of different heights (up to 10m) depending on the soil moisture. Other plant species in this thicket were far less abundant: the chee reedgrass (Calamagrostis epigeios), other grasses (Aeluropus littoralis, Alopecurus sp.), several sedges (Carex spp.), and some halophilous species (Salsola spp. and others).

The reeds were of immense economic importance to the local people. First, they were used as native pastures for livestock during the warm part of the season. The accessible reed-fields were cut for hay, which constituted practically the only source of fodder during winter. In addition, the reeds were used as a cheap building material for

36

making fences, cabins, roofs, mats etc. Also, they were used for cellulose production and papermaking. The major part of the reed ecosystem was situated on the islands and in remote corners of the delta where it remained intact and provided food and shelter for a diverse local fauna. Characteristically, one of the biggest islands in the Aral Sea is called "BarsakeImes", which means "Place of no return", alluding to these dense, impenetrable reed areas. Such reed ecosystems are very typical for Central Asia and Kazakhstan. They occupy the deltas of all major rivers (IIi, Ural, Karatal etc.) and surround the main lakes (Balkhash, Alakol, Zaisan). These habitats were home to the last populations of Kazakhstan tigers, which existed near lake Balkhash until the 1930s.

Beside the fluvial forests and wetland grasslands of reeds, other vegetative communities were of a secondary importance in the Amu Darya delta. The proportion of the cultivated lands declined toward the littoral zone. Only a few small patches were cultivated for alfalfa. Weedy vegetation and other ruderal plant communities grew mostly around villages, on heavily salinized soils unsuitable for cultivation. Plant species in such areas were represented by halophilous forbs and grasses: the wild licorice Glycyrrhjza glabra, Russian knapweed Acroptilon repens, Johnsongrass Sorghum hale pense, camelthornAlhagi pseudalhagi, Russian thistle Sa/sola rigid a, and Bermudagrass Cynodon dactylon. The zone of croplands was located outside of the reed area, along the main course of the Amu Darya. Rice, cotton, melon, and potatoes were the principal crops; near villages, other cereals, vegetables, vineyards and orchards were also cultivated. Vegetation in the Syr Darya delta (NE of the Aral Sea) was similar. The difference was in the size, with the Syr Darya delta being one-fourth of the area of the Amu Darya delta.

2.4. ACRIDOLOGICAL SETTING: A LOCUST PARADISE

2.4.1. The Asiatic Migratory Locust The acridofauna of the Aral Sea basin is comparatively well studied. The reason for this is the fact that historically the Aral Sea basin supported two immense permanent breeding areas of the Asiatic Migratory Locust, Locusta migratoria migratoria L. [63]. They occupied the reedbeds in the Amu Darya and Syr Darya deltas. During outbreaks, which occurred at irregular intervals (3 to 15 years), migratory flights of huge swarms stretched for several hundreds ofkm, menacing the irrigated croplands in the nearby oases [71]. The outbreaks were usually preceded by abnormally dry years. During such years, larger areas become free from flooding and expanded the habitat suitable for oviposition. Thus, after a severe drought in 1945, the Migratory Locust infested more than 1 million ha in 1946, and chemical treatments to control the plague were implemented on >500000 ha [52]. The dynamic, ever-changing ecosystem of the Amu Darya delta provided very favorable conditions for survival and reproduction of the Migratory Locust. The sandy river banks served as oviposition sites, and the reeds furnished food and shelter for hoppers and adults. Hopper bands of the Migratory Locust concentrated in the depressions where some green vegetation (reeds) survived periods of drought. This tendency to track the temporarily mesic habitats in an overall arid milieu is a typical life strategy of many locusts [65, 67]. It allows the species to exploit the vegetative resources of the habitat better.

Some ofthe hopper bands were enormous - up to 36 km long in the 4th instar and

37

reaching 49 km in the 5th instar; the hoppers climbed on one another forming a layer up to 50 cm thick [38]. After fledging, the adult swarms dispersed in different directions, often reaching 120 km S and 200 km SW. Tsyplenkov [63] described one of the longest migrations of L. migratoria from the Amu Darya delta in which a swarm flew westwards, crossed the Caspian Sea and landed in Azerbaidjan (Transcaucasia), covering a distance of>1 000 km. According to Novitzky [39], the size ofa landed swarm could reach 120 by 10 km (120 000 ha). Ordinarily, the density of the adults in such swarms was 10 to 50 individuals per m2, but in the exceptional cases, it could reach 2 000-3 000 per m2 [38]. During the period of sexual maturation and mating, the locusts were extremely voracious, destroying all available vegetation down to the soil level. Poplars and willows in the fluvial forests would often collapse under the enormous weight of the landed swarms. The average egg-pod density was about 5 per m2, with a maximum of 140 perm2. Each female laid three to four egg-pods containing 32 to 120 eggs with an average ranging from 60 to 80 eggs [63].

Chemical treatments against the Migratory Locust were implemented mostly to prevent gregarisation and assure crop protection [54]. Irrigated fields of cash crops, especially cotton and rice, were jeopardized by the swarms flying out ofthe reeds. Inside the vast areas of reeds, the Migratory Locust, even in very high numbers, seldom damaged the ecosystem. Perceptible economic damage to the reeds occurred only during excessively dry seasons [52]. The essentially graminivorous Migratory Locust was admirably adapted to the reed monoculture and played a major role in the cycling of nutrients [53]. The solitarious phase of the locust is a narrow oligophage (grass-feeder) compared to the gregarious phase which exhibits broad polyphagy - a typical picture for many locust species [23]. Therefore, the principal goal ofthe Migratory Locust control in the Aral Sea basin was to prevent phase transformation and development of the dense swarms of locusts capable of inflicting serious damage to crops.

One of the most interesting elements in the biology ofthe Migratory Locust in the Aral Sea basin is the number of generations per year. The Amu Darya outbreak area of the species is the southernmost in the former USSR. Populations from the breeding areas situated to the north (for example, Lake Balkhash in Southeastern Kazakhstan and even in the nearby Syr Darya delta) are strictly univoltine. The more southern [Xinjiang (Sinkiang), Iran] or less continental (Southern France) popUlations can produce offspring continuously, without diapause, and thus have two or even three, annual generations [69]. In the Amu Darya delta, cases of late (September instead of May) hatching of the Migratory Locust were frequently reported. However, it was not clear if such autumnal hatching was a result of the late drying of an egg-bed where the oviposition occurred in the previous fall, or the result of an emergence by the second, consecutive generation from the eggs laid the same year.

Our laboratory studies of the Amu Darya population showed that some eggs hatched 12 to 14 d after they were laid ifkept at 30°C and 12112 h LID conditions [19]. The eggs from the more northern geographic populations (Lake Balkhash) never developed continuously. They exhibited an obligatory diapause and needed at least 15 d of "cold reactivation" at 0 to 3°C to initiate the process of embryonic development. Such eggs simply did not develop and perished when kept at a constant 30°C after oviposition. For

38

the Amu Darya population, the proportion of eggs that developed continuously ranged from 12 to 82% at 30°C, dropping to zero when the temperature was raised to 35°C [19]. It is interesting to note that the females raised from either diapausing or non-diapausing eggs could produce both types of offspring. It is possible to suppose the existence of a "genetically-imprinted" capacity for alternative types of development in this locust population, which occupies an intennediate geographic position between strictly univoltine (northern) and bi- or trivoltine (southern) populations. These studies revealed the possibility of the Migratory Locust producing two consecutive annual generations in the Amu Darya delta, weather conditions pennitting. Such adaptation could be a useful life strategy under conditions of an unusually long growing season, contributing to the species fitness.

2.4.2. The Italian Locust Beside the Migratory Locust, the Amu Darya delta was home to a number of other acridid species. Among them, the Italian Locust, Ca/liptamus italic us (L.), had significant economic importance, second only to the Migratory Locust. It colonized anthropogenic habitats: roadsides, weedy patches and old, overgrown alfalfa fields [73]. In fact, the Amu Darya delta represents the southernmost extremity of the distribution range of this species [11]. Farther to the south, it is replaced by another species from the same genus, Calliptamus barbarus Costa, which is better adapted to the arid conditions of the Central Asian deserts. The zone of economic damage of the Italian Locust is located far to the north of the Aral Sea, in the steppes of north-central Kazakhstan and West Siberia. There, C. italicus colonizes very arid, steppe or semi-desert habitats [34, 50, 60, 68]. In the southern part of its range, the shortage of moisture becomes the limiting factor, and the species finds suitable environments only in the wet habitats, such as irrigated agroecosystems [61]. That is why the Italian Locust was given the common name "Oasis Locust" in Central Asia in the 1930s [24].

As mentioned before, in the Amu Darya delta C. italic us historically occupied habitats of anthropogenic origin. There the species found both oviposition and forage sites. Proximity of these sites to croplands and very broad polyphagy made the Italian Locust an important pest. However, because the cultivated land occupied only a small proportion of the delta until early 1960s, the economic importance of the Italian Locust remained far behind the ever-threatening Migratory Locust. Between 1930 and 1960, C. italicus annually infested only 29 000 ha on average, compared to almost 300 000 ha by L. migratoria [22] (Table 3).

2.4.3. Other Acridid Species According to Stolyarov [59], the acridofauna of the fluvial forests' borders included 12 species: Calliptamus italicus (L.), Pyrgomorpha bispinosa deserti Bei-Bienko, Chrotogonus turanicus Kuthy, Tetrix tartara tartara (I. Bolivar), Acrida oxycephala (Pallas), Aiolopus thalassinus (F.), Acrotylus insubricus inficitus (Walker), Mesasippus kozhevnikovi (Serg. Tarbinsky), Duroniella gracilis Uvarov, D. kalmyka (Adelung), Gonista sagitta (Uvarov), and Anacridium aegyptium (L.).

In the reedbeds of phragmites, Locusta migratoria migratoria L. was

39

overwhelmingly predominant. The accompanying species - Tropidopola turanica turanica Uvarov, Aiolopus oxianus Uvarov, and D. kalmyka - were scarce.

The dry patches with bare, salinized soils within the reedbeds were inhabited by several species of the genus Sphingonotus Fieber, Calliptamus barbarus cephalotes Fischer von Waldheim, and P. bispinosa deserti. These species were most common outside the reed area, in the semi-desert. Finally, Oxyafuscovittata (Marschall), T. tartara tartara, Heteracris adspersa (Redtenbacher), C. italicus, Truxalis eximia (Eichw.) and a few others inhabited the banks of the irrigation canals and the borders of the rice fields. Prior to the 1980s, grasshoppers were considered as pests of minor economic importance, and practically were never subject to control [53].

2.5. THE CATASTROPHE

The only available sources of water for irrigation in the region are the two tributaries of the Aral Sea, the Amu Darya and Syr Darya rivers. They account for 90% (110 km3 per year) of the river flow in all of Central Asia [32]. Their water has been used for irrigated agriculture for centuries. The "Mesopotamia" of the Central Asia, the fertile irrigated land between the two legendary rivers of Oxus (Amu Darya) and Jaxartes (Syr Darya), is an ancient settlement area with a history of approximately 3 500 yr. Archeological research has revealed sophisticated irrigation systems that provided water for millions of hectares [14].

Starting from the late 1950s, the water of these two rivers was increasingly diverted for irrigation. Whereas atthe beginning of the century, only about 5% ofthe total river flow was used for agricultural purposes, the rapid expansion of the irrigated area led to an almost total extraction of the average annual flow by the end of 1980s [14]. During the 1980s, several years passed in which little or no water reached the Aral Sea from the Amu Darya. The Karakum canal, the longest (1 300 km) canal in the former Soviet Union, diverted a large proportion of its water to Turkmenistan (15 km 3/year). The Syr Darya river also ceased to reach the Aral Sea in the mid-1980s. On average, the annual flow of both the Amu Darya and Syr Darya into the sea dropped to only 4 to 5 km3 - less than onetenth of that in the 1950s. Consequently, the level, area, and volume of water in the Aral Sea decreased dramatically (Table I). In 1987, a strip of the emerged dry sea bottom divided the Aral Sea into two unequal parts, the "Small Sea" in the north fed by the Syr Darya (3 000 km2, 24 km 3) being separated from the "Large Sea" in the south fed by the Amu Darya (37000 km2, 324 km3) [29].

The area of irrigated land in the Aral Sea basin (s. I.) expanded from 5.6 to 8.5 million ha from the late 1950s till 1990 [31]. At the same time, water use rates doubled. This disproportionate increase in water use resulted mainly from the development of newly irrigated areas on marginal land, where huge amounts of water had to saturate dry soils before cultivation [6]. By 1998, the sea level had dropped 18 m, the Aral Sea had lost 60% of its former area and 80% of its volume (see Table I), and the sea had moved down from the 4th to the 8th place in the list of the world's largest lakes. The shoreline in the most shallow, southeastern part of the Aral Sea had receded 80-120 km from its original littoral limit (Fig. 2) and exposed more 36 000 km2 of the sandy, former sea bottom. At the same

40

time, salinity tripled, and water in the Aral Sea became comparable to ocean water.

TABLE I. Degradation of the Aral Sea 1960-1998 (compiled from different sources)

AREA VOLUME SEALEVEL* SALINITY YEAR

km2 % decrease km3 % decrease m m decrease gil % increase

1960 69500 1,064 53 10

1985 45713 34 468 56 41.5 11.5 23 230

1986 43630 37 380 64 40.5 12.5

1987 42650 39 354 67 40 13

1988 41 134 41 339 68 39.5 13.5

1989 40680 41.5 320 70 39 14 30 300

1990 38817 44 282 73 38.5 14.5

1991 37159 46.5 248 77 38 15

1992 36087 48 231 78 37.5 15.5

1993 35654 49 248 77 37 16

1994 35215 49 248 77 37 16

1995 35374 49 248 77 37 16

1996 31516 55 212 80 36 17 35 350

1997 29632 57 190 82 35 18

1998 28687 59 181 83 34.8 . 18.2 45 450

2010*- 21058 70 124 88 32.4 20.6 70 700

• the average sea level in the Large Aral Sea. By 1998, the level of the Small Sea near the Syr Darya delta dropped 13 m .

•• projected

2.6. CLIMATIC CONSEQUENCES

For centuries, the Aral Sea functioned as a natural "thermoregulator" for the entire basin. As a barrier against the cold winds from the north, the Aral Sea originally had a moderating function in the regional climate system. It served as a catalyst for cloud formation because of enormous masses of vapor (58 km3 per year, equivalent to a 90-cm water layer! [14]) rising from its surface. This moisture replenished the ice and snow caps of distant mountains, thus completing the region's water cycle.

The climate became more continental with a shrinking water surface. The summers became hotter and the winters colder. The growing season shortened, and heatloving cotton fields were turned into rice paddies. In the Amu Darya delta, the temperature extremes reached 47°C in summer and -17°C in winter. The quantity of rainfall dropped nearly 3-fold, from an average of 156 mm to only 56 mm per year [45]. In most regions of the Aral Sea basin, annual precipitation now does not exceed 70 to 80 mm. From 1950 to 1980, the average July temperature in the Amu Darya delta rose from 25.7 to 28.3 °C,

41

whereas the air humidity dropped from 44 to 32% in the same period [5]. Given the regional role of the Aral Sea, it is assumed that even the precipitation patterns in the distant mountains of Central Asia will change with the sea's desiccation [14].

Figure 2. The recession of the Aral Sea (1960 - 1998)

Sand and dust storms became one of the most distinct features of the regional climate. Salts and toxic chemicals had deposited and accumulated in the Aral Sea for decades. Now the sea has changed its role from a sink to a major source of salt, mainly sodium chloride and sodium sulphate. From the desertified seabed, which represents a mixture of extremely fine sand and salt, an estimated 75 million tons of toxic dust are carried away by winds every year [45]. This windborn salty dust is deposited directly on the soil surface or incorporated as aerosols in acid rains, thereby becoming a serious danger to plants and soils in agricultural areas. An average of 500 kg of salt and sand fall annually on each hectare of the Amu Darya delta [7]. The saltstorms carry their toxic load far beyond the Aral region. Dust from the Aral Sea has been found on top of the Himalayas, in the Ganges and Brahmaputra rivers, and even in the Pacific and Arctic oceans as far as 10 000 km away [45]. Between 1966 and 1985, the Amu Darya delta recorded 965 days - an average of 48 days per year - with salt-dust storms [32]. The frequency and magnitude of such storms increased further in the 1990s.

42

2.7. AGROECOLOGICAL CONSEQUENCES

The contraction of the Aral Sea had a devastating impact on local agriculture. The increasing use of highly saline and polluted water for irrigation caused salinization and water logging of soils and their contamination with pesticides and fertilizers. Irrigation erosion (flushing of a thin fertile soil layer by irrigation water) was also common and soil productivity decreased. Cotton yields dropped by 20 to 50% from a peak in 1980 mainly due to increasingly salinized soils [58]. Rice production in the Amu Darya delta declined by two-thirds within 12 yr [1]. In the 1980s, this decrease in agricultural production triggered further expansion of the irrigated areas, which rapidly became salinized. The ever-increasing water diversion for irrigation linked to a rapidly deteriorating soil quality formed a vicious circle of environmental and agricultural degeneration in the Aral Sea basin. Dust storms became a permanent problem aggravating the severe soil erosion.

Another consequence of agricultural monocultures was the contamination of soils by agrochemicals. By the mid-1980s, as much as 300 kg of mineral fertilizers and 50 kg of toxic chemicals were used per hectare. Pesticide use in the Amu Darya delta exceeded the Soviet average by ten-fold [45]. Uninterrupted cotton monoculture also violated traditional crop rotation practices (which mainly used alfalfa) and exhausted the soil nutrients [48].

The native ecosystems of the Aral Sea basin have undergone dramatic changes. Of course, the fishing industry died out. Because of the severely increased water salinity, all 24 native species offish have been lost [14]. The fluvial forests along the rivers have disappeared almost entirely. Only very narrow strips still remain along the main watercourses. Lakes and wetlands have been reduced by 85%. More than 30 000 ha of lakes and bogs have dried up completely [45]. Their desiccation and subsequent desertification was accompanied by simplification of native plant and animal communities due to extinction of some species. A complex, multi-story vegetative community, which sustained a large number of animal species, was transformed into a simple, meager stand composed mostly of halophilous forbs and grasses, with a depauperate fauna [22]. Consequently, the biodiversity has decreased significantly. In the Syr Darya delta, the number of nesting bird species fell from 173 to 38 [29]. About 50 plant species apparently went extinct in the Amu Darya delta [7]. Once highly abundant, populations of muskrat, boar, and deer have totally disappeared. The number of species of mammals dropped from 30 to 10 [29]. Apparently, the entomofauna of the Aral Sea basin also suffered from desiccation. Unfortunately, the relevant, quantitative data remain largely unavailable.

2.8. ACRIDOLOGICAL CONSEQUENCES

In this section, we will extend and assess the current inventory of the acridofauna of the Amu Darya delta and evaluate the trends in its transformation since the beginning of the Aral Sea's desiccation. Specific attention will be given to the changes in the pest acridid situation.

In 1990-1998, during extensive acridological surveys in the Amu Darya delta, we investigated the faunistic composition of locust and grasshopper communities, including

43

analyses of the distribution and abundance of species within different types of habitats. The results of this work were partially reported in our previous publication [22]. We do not pretend that our data are exhaustive because some remote areas of the delta remain unexplored. Yet we believe that our results adequately reflect the current locust and grasshopper situation and allow us to identify trends in this regard.

Species composition and distribution within different types of habitats in the AInu Darya delta are summarized in Table 2. These data suggest several important tendencies. The total number of species found (41) was close to the numbers reported by previous investigators [52, 59, 71], which ranged from 36 to 45. However, species distributions within different biotopes have undergone significant modifications. During the process of desiccation of the Aral Sea, trees and shrubs of the fluvial forests disappeared first. In the very narrow strips that remained, instead of 12 grasshopper species reported by Stolyarov [59], we collected only 6. Mesic species such as T. tartara tartara, A. oxycephala, D. gracilis, D. kalmyka, Mesasippus kozhevnikovi (Serg. Tarbinsky) and H. adspersa were not found. The acridofauna of the fluvial forests' borders appeared to be the least diverse of all studied habitats.

In the reeds, only 8 grasshopper species were found. The inhabitants of this biotope were not numerous. Hygrophilous and mesophilous species, such as T. eximia and D. gracilis were extirpated from the reeds. Together with L. migratoria migratoria, G. sagitta, and H. adspersa, these species now occupy banks ofthe irrigation canals and rice paddies. This grasshopper community is very heterogeneous. It contains 17 species, including some steppic and even semi-desert grasshoppers.

The depletion of the reed grasshopper community is an expression of a general principle of the invertebrate fauna succession following the drainage of wetlands [70]. In general, xerophilous species replace the hygrophilous ones in 2 to 3 yr [35]. These changes track the climatic alterations and modifications in plant communities induced by the drainage.

Cereal crops (rice, wheat, and barley) were home to nine species of grasshoppers, and the economic importance of the acridids as pests of grain crops has increased. After gaining political independence in 1992, the Uzbek government adopted the policy of striving for food security and independence from grain imports. Significant areas of cotton cultivation were re-allocated to grain production, in particular, wheat. The grasshoppers reacted by moving from the adjacent fallows or weedy patches into the wheatfields. Some species, such as C. italicus, C. barbarus cephalotes Fischer von Waldheim, T. turanica turanica and D. kalmyka colonized the cereal cropfields in high numbers, inflicting serious damage. Oxya foscovittata (Marschall) and A. insubricus inficitus were abundant in the rice paddies. The latter species has been recently recorded as a pest of rice for the first time in the Amu Darya delta [13].

Deserts outside the reedbeds contained the highest number of grasshopper species - 23. The predominant species were C. italicus, C. barbarus cepha/otes, Trinchus turcmenus Bei-Bienko, Dericorys annulata roseipennis (Redtenbacher), Sphingonotus halocnemi Uvarov and S. salinus (Pallas). The members of the desert and semi-desert acridofauna started to penetrate into the habitats which were never occupied by these species before. From an acridological standpoint, this process is probably the most

44

apparent consequence of desertification in the Aral Sea basin. For example, Sphingonotus maculatus Uvarov, Sphingoderus carinatus (Saussure), Hyalorrhipis turcmena Uvarov, T. turcmenus and Strumiger desertorum desertorum Zubovsky are typical inhabitants of dry, open spaces with sandy or solontchak soils. These desert species are now found in hayfields, formerly colonized only by mesophilous species like Ochrilidia hebetala hebetata (Uvarov) and D. kalmyka.

Another tendency in the species re-distribution after the desiccation of the Aral Sea manifested in the occupation of a very broad range of habitats by ecologically plastic species. Indeed, such species as the Italian Locust and another representative of the same genus, C. barbarus cephalotes, together with Anacridium aegyptium (1.), Pyrgomorpha bispinosa deserti Bei-Bienko and T turanica turanica were found in six to seven out of nine studied habitats (see Table 2). These grasshoppers combine broad ecological temperament with polyphagy and/or good flight capacities. On the contrary, grasshoppers specialized for particular conditions (e.g., inhabitants of dense, mesic grass stands), like 0. hebetata hebetata, G. sagitta and Pyrgodera armata Fischer von Waldheim were found in just 2 to 3 different habitats and in very low numbers.

With respect to abundance, the Italian Locust became undoubtedly the most abundant acridid species in the Amu Darya delta. Not only was it nearly ubiquitous, but it also occurred almost everywhere in very high densities. This species is extremely opportunistic, and it readily colonizes roadsides, field borders, weedy patches and similar habitats of anthropogenic origin. We have found the species even in sandy dunes and excessively salinized areas where it has never been previously reported. However, the highest densities were always observed in the agricultural area. Vegetation communities composed of weeds (especially Artemisia spp.) and overgrown alfalfa plots provided ideal oviposition sites for the Italian Locust. As the season progressed, hopper bands migrated to the adjacent crop fields in search of fresh, succulent food. The Italian Locust is infamous for its extreme polyphagy [34], and practically not a single crop can withstand the attacks of its voracious hopper bands or swarms of adults.

At the same time, the pest status of the Asiatic Migratory Locust has significantly decreased with the shrinking of the Aral Sea. The once continuous, vast and dense reedbeds have been transformed into fragmented, sparse patches of short and frail reeds. During the initial phases, this process probably even contributed to the expansion of the Migratory Locust, as dry seasons create more areas free from flooding and thus suitable for breeding. However, as the contraction of the Aral Sea and the reeds in the Amu Darya delta progressed, the "living space" of the species was severely reduced (Figure 3). Practice shows that the most efficient means to control the Migratory Locust is total destruction of its habitats (i.e., reedbeds). Several "permanent" breeding areas of this species ceased to exist in the former Soviet Union because of the dramatic changes in the hydrological regime of such rivers as the Kouma, Terek, Kuban and others. Their deltas have been drained and converted to cropland, thus eliminating the conditions for successful reproduction of the Migratory Locust [63, 72]. Something similar is currently happening in the Aral Sea basin. The initial area of reeds was enormous (12,000 km2), and about 25% of it still exists [13]. However, the process of their degradation seems irreversible.

45

TABLE 2. Specific composition and habitat distribution of grasshoppers of the Amu Darya delta (I990-1998)

HABITAT* N SPECIES

2 3 4 S 6 7 8 9

Family Tetrigidae

Tetrix bolivari Saulcy + +

2 T. tartara tartara (I. Bolivar) + + +

3 Paratettix uvarovi Semenov + + Family Pamphagidae

4 Trinchus turcmenus Bey-Bienko + +

5 Strumiger desertorum desertorum Zubovsky + + Family Pyrgomorphidae

6 Pyrgomorpha bispinosa deserti Bei-Bienko + + + + + +

7 Chrotogonus turanicus Kuthy + + + Family Acrididae Subfamily Catantopioae

8 Dericorys tibialis (Pallas) + +

9 D. annulata roseipennis (Redtenbacher) + + 10 Diexis chivensis Umnov + +

Il Egnatius apicalis Stal + +

12 Oxyafuscovittata (Marschall) + + + +

13 Tropidopola turanica turanica Uvarov + + + +

14 Anacridium aegyptium (L.) + + + + + + + 15 Calliptamus italicus italicus (L.) + + + + + + +

16 C. barbarus cephalotes Fischer de Waldheim + + + + + +

17 Heteracris adspersa (Redtenbacher) + + + + Subfamily Acridinae

18 Truxalis eximia (Eichw.) + + + + + +

19 Gonista sagitta (Uvarov) + + +

20 Ochrilidia hebetata hebetata (Uvarov) +

21 Duroniella gracilis Uvarov + + +

22 D. kalmyka (Adelung) + + + + + 23 Dociostaurus tartarus (Stshelkanovtzev) + + + + 24 D. kraussi nigrogeniculatus Sergo Tarbinsky + + +

Subfamily Oedipodinae

25 Aiolopus thalassinus (F.) + + + + +

26 A. oxianus Uvarov + +

27 Hilethera turanica Uvarov + +

28 Locusta migratoria migratoria L. + + + +

29 Pyrgodera armata Fischer de Waldheim + +

46

TABLE 2 continued

N SPECIES 2 3 4 5 6 7 8 9

30 Mioscirtus wagneri (Kittary) + + 31 Oedipoda miniata atripes Bei-Bienko + 32 Acrotylus insubricus injicitus (Walker) + + + + 33 Sphingonotus maculatus Uvarov + + 34 S. halocnemi Uvarov + +

35 S. nebulosus discolor Uvarov + + 36 S. salinus (Pallas) +

37 Sphingoderus carinatus (Saussure) + + +

38 Helioscirtus moseri Saussure +

39 Hyalorrhipis c1ausi (Kittary) + +

40 H turcmena Uvarov + + +

41 Leptopternis gracilis (Eversmann) +

TOTAL 17 8 13 19 13 9 IS 14 6

• Habitats: 1 - banks of irrigation channels; 2 - reeds; 3 - deserts; 4 - hayfields; 5 - alfalfa fields; 6 - cereal crops; 7 - cropfield borders; 8 - steppes; 9 - fluvial forests' borders.

That is why the economic importance of the Migratory Locust has declined steadily over the last four decades. The statistics on the areas infested by this species and the Italian Locust are summarized in Table 3. Habitat destruction is the single, most powerful cause of extinction of acridid species [49]. In the case of locusts, habitat destruction often leads to elimination oflocal populations. Several cases of such extirpations are documented for the Moroccan Locust, Dociostaurus maroccanus (Thunberg), which cannot withstand plowing of its breeding areas [21]. Species with narrow ecological requirements (e.g., the Migratory and Moroccan Locusts) are usually more prone to suffer from habitat destruction compared with more ecologically plastic species (e.g., the Italian Locust).

TABLE 3. Areas infested annually (thousand hal by the two locust species in the Amu Darya delta (compiled from different sources)

Locusta migratoria migratoria L. CaUiptalnUs italicus (L.) PERIODS

average maximum average maximum

1930-1950 282 1031 29 52

1951-1960 257 430 55 86

1961-1970 114 570 170 420

1971-1980 103 334 120 176

1981-1990 34 121 lOS· 180"

1991-1999 51 92 lll" ISO·

• together with grasshopper species

47

An important characteristic of the dynamics of a locust species is the phase status of the population, i.e., its ranking within the solitarious, gregarious, or transitory phase [20].

Quantitatively, this trait for the Migratory Locust can be assessed using several morphometric ratios of the individual insects [4] and then averaging them [17]. In this species, the gregarious individuals are larger, possessing longer wings than the adults of the solitarious phase [66]. Consequently, the most commonly used morphometric index, a ratio between tegmen length and hind femur length (ElF), exceeds 2.0 in the gregarious locusts and rarely reaches 1.9 in the solitarious individuals [3]. Morphometric studies of the Migratory Locust population from the Amu Darya delta from 1980 to 1987 showed that the average ElF index ranged from 1.79 to 1.88 [18]. Apparently, a great majority of the measured locusts belonged to the solitarious phase, suggesting that over the study period, the population dynamics of the species entered a deep recession. Due to the adverse anthropogenic impact, the periods of recession of the Migratory Locust exhibit a strong tendency to occur more frequently and persist longer in the last two decades of the 20th

century than before (compare with data from Shamuratov [51, 52]). It is interesting to note that in the Amu Darya delta, the climatic factor limiting

the abundance of the Italian Locust is the shortage of rain in spring [61]. The optimum lies in the range from 30 to 50 mm, and if the precipitation is below these limits, the eggs die of desiccation [55, 59]. With the loss of the Aral Sea, the climate became even dryer than it was before. However, reductions of the Italian Locust populations did not take place. On the contrary, due to the opportunities to colonize new types of habitats, this species reproduced successfully and expanded its distribution range [55]. Comparison of the breeding areas of L. migratoria and C. italic us in 1960 and 1990 shows that while the Migratory Locust's habitats became fragmented and reduced, the Italian Locust colonized new, vaster areas (Fig. 3). Apparently, the very high ecological plasticity and adaptability of the Italian Locust and comparatively narrow ecological requirements of the Migratory Locust are responsible for reversing the pest status of these two species. It is necessary to add that due to the climatic changes induced by the desiccation of the Aral Sea (higher summer temperatures and earlier autumn frosts), the duration of the postembryonic development of the Migratory Locust decreased from 160-170 days [36, 51] to 150 days or even less [13]. The reduced longevity and the earlier offset of the population due to the precocious decrease of the temperature in the fall limit the possibilities of exploiting the thermoresources of the habitat by the species. It also practically eliminates the possibility of the development of the second consecutive generation during the growing season.

To summarize, the desiccation ofthe Aral Sea has almost "resolved the problem" of the Migratory Locust, by destroying a major part of its habitat. At the same time, the drying had the opposite effect on the Italian Locust, which became an even more serious pest than the Migratory Locust in the past. Several grasshopper species, particularly 0. fuscovittata and A. insubricus inficitus, now also became pests of economic importance damaging cereal crops [13]. As such, one more problem - that of the Italian Locust and grasshoppers - has been added to the plethora of disasters that are looming on the horizon with the shrinking of the Aral Sea.

48

Migratory locust

Italian locust

Migratory locust

~ Italian iocust {) 20 4Gltro Immmm~

Figure 3. Breeding areas of two locust species in the Amu Darya delta in 1960 (a) and 1990 (b) (modified from Khudanov [13])

49

3. Tools

Monitoring of the locust populations in the Aral Sea basin started in the beginning of the 20th century. The first control operations were carried out using various mechanical devices to destroy hopper bands [57]. In the 1920s, chemical control was introduced using baits poisoned with arsenates; later, the arsenate powder was applied from horse-drawn dusters [51]. Until the mid-1980s, dusting was the predominant method of pesticide application while the control agent changed over time from arsenates to organochlorines -DDT (since late 1940s) and BCH (since early 1960s). In the 1980s, the organophosphates and synthetic pyrethroids replaced BHC. Monitoring and control of pest acridids were carried out by special organizations, "Anti-locust expeditions", belonging to the Federal Plant Protection Service. One of these was based in the Amu Darya delta, another in the Syr Darya delta. The expeditions conducted surveys, identified the infested areas and implemented control operations. The surveys took place four times a year. The spring survey began at the onset of mass hatching of hoppers and was carried out until the majority of population reached the 3rd instar. During this survey, the sites of mass hatching were mapped and the infested areas were designated, if necessary, for control treatments.

The threshold for chemical intervention was an average hopper density exceeding 200-300 individuals per ha [64]. The summer survey was carried out when locusts reached the adult stage. It was intended to locate areas of concentrated adults. The results of the summer survey served as a basis for planning control campaigns the following year. The autumn survey of egg-pods took place late in the season, at the end of the adult period. Its purpose was to establish the areas where the egg-pods were laid and to evaluate their densities. An additional early spring survey was carried out before hatching to assess the overwintering survival rate of eggs. The percentage of damaged or dead eggs was used to adjust the locust forecast for the oncoming season. As a rule, this survey was executed only after years of heavy outbreaks.

Insecticides for anti-locust treatments were purchased through the federal budget and applied by the anti-locust expeditions using ground or aerial equipment. In mid-1980s, the list of authorized insecticides included dust powder of an organochlorine (BCH) and liquid formulations of three organophosphates (diazinon, malathion, and methylparathion) and two pyrethroids (deltamethrin and esfenvalerate). Although some of these formulations were designated for use at ultra-low volumes (diazinon and malathion), the lack of the special equipment reduced their use to a minimum. In practice, full-volume (150-300 l/ha from ground and 50 l/ha from air) spraying of EC formulations was the predominant type of insecticide application.

The main objective of the control campaigns was not to let the swarms of the Migratory Locust fly out of the reeds and devastate the adjacent crop areas. In a sense, it was a preventive strategy. During the outbreaks, chemical control was applied to extensive areas, up to 500 000 ha annually [51]. The Italian Locust was controlled only on limited areas until the early 1970s. However, the contraction of the reed area and the emergence of the new habitats suitable for the Italian Locust reversed this proportion. Starting from

50

1993, the areas treated against C. italicus exceeded those for L. migratoria (Table 4). It is to be noted that the limited areas treated in the 1990s are due mostly to the shortage of economic means to conduct a full-scale control campaign. The areas in need of control were generally much higher (see Table 3).

In the beginning of the 1990s, two countries of the Aral Sea basin, Kazakhstan and Uzbekistan, gained political independence. They inherited all environmental and agricultural problems resulting from the desiccation of the sea. Locust monitoring became the responsibility of the National Services of Plant Protection. In the Syr Darya delta (Kazakhstan) the anti-locust expedition ceased to exist because of almost entire extermination of the habitats (reedbeds) suitable for the Migratory Locust. In the Amu Darya delta (Uzbekistan), about 3 000 km2 of reeds still remain, and the anti-locust expedition continues to survey and treat the areas infested by different species of locusts and grasshoppers on a regular basis. In the late 1990s, the list of authorized acridicides included about 20 formulations of products of different chemical groups -organophosphates, pyrethroids, benzoy I ureas (insect growth regulators, I G Rs) and phenylpyrazols. However, in practice, pyrethroid formulations are used most often [13]. Control operations are essentially terrestrial because of the shortage of funds to carry out an aerial campaign. Preventive strategy has been replaced by curative crop protection: the treatments are executed mostly in the crop areas infested by the Italian Locust and grasshopper species. Such treatments have become regular, and their area has increased, demanding more economic and human resources [13]. Reedbeds occupied by the Migratory Locust are treated only in the nearest proximity to cash crops nowadays.

4. Solutions

4.1. IMPROVEMENT OF THE ARAL BASIN ECOLOGY

The solution to the acridid pest problem in the Amu Darya delta is inconceivable if taken out of the context of ecological and agricultural changes in the entire Aral Sea basin. Although the process of the sea's desiccation exhibited a tendency to slow down in the 1990s (see Table 1) it has not reversed. At recent (1990-1995) inflow levels from the Amu Darya river (an average of 9 km3 per year), the Large Aral will continue to recede, separating into several lakes in the next century, with a total projected area of 13000 km2

[62]. The Small Ara1, with an inflow of3.3 kmJ per year, will rise, stabilizing with an area near 3772 km2• To stabilize the entire Aral Sea at its 1996 size (although less than 50% of the original, 1960 size) would require an inflow of about 23 km3 per year, or about double that of recent years. To restore the sea to its 1960 dimensions would require many decades of inflows amounting to an estimated 55 kmJ per year, or nearly five times the recent inflow rates [62]. Obviously, restoration of the Aral Sea is not a linear process: its depletion happened over just three decades while restoration may take at least a century. And it is clear that even then the sea would hardly ever return to its original state before desiccation [7].

TABLE 4. Areas treated with insecticides (thousand ha) against the Migratory and Italian Locusts in the Amu Darya delta (compiled from different sources)

YEAR Locusta migratoria L. Calliptamus italicus (L.)

1946 500 35

1959 66 7

1960 60 40

1966 480 383

1967 384 198

1968 504 185

1972 178 151

1973 200 135

1974 380 140

1977 204 103

1978 203 103

1979 197 106

1980 166 115

1983 40 60

1990 55 21

1993 6 23

1994 18 49

1998 23 50

1999 20 40

51

Indeed, the Kazakhstan (the Syr Darya) portion of the Aral Sea, which formed the Small Aral in 1987, is considered much more salvageable than the still-dying Large Aral fed by the Amu Darya. In 1997, the local government built a 20-km long and 25-m wide dike between the Small and Large lakes. Protected from the larger, contaminated body, the smaller lake's shoreline began to stretch again towards the ships' cemetery on the old littoral area. The water rose from 34.5 m to 37.5 m and is still rising [62]. The fluvial forests along the Syr Darya showed signs of recovery: the saxaul and tamarisk shrubs started to regrow. Camels and wild horses are returning to their original habitats. Reed areas increased, even supporting some cattle breeding. Birds reappeared, including gulls, swans, and pheasants. The population of sole is also growing [14].

Although the Aral Sea itself is situated within the territories of two countries, Kazakhstan and Uzbekistan, its basin, including the Amu Darya and Syr Darya rivers, involves three more adjacent CIS countries: Turkmenistan, Tadjikistan and Kyrgyzstan. These five countries signed a water-sharing agreement in 1992 that included a provision to provide at least 14 km3 of water per year to the Aral Sea [14]. In practice, however, the

52

deliveries to the sea are considerably smaller. Agreements to improve the situation signed in 1993, 1994, and 1995 included the creation of a fund for the sea to which the nations agreed to donate 0.3% of annual national income [62].

The major focus of efforts in recent years has been concentrated on improving the living conditions of people around the sea through the implementation of improved medical services, wastewater treatment, and drinking water purification. Although some attention has been devoted to improving the ecological conditions ofthe Amu Darya delta, the preservation of the Large Aral has been abandoned. The amount of inflow from the Amu Darya necessary merely to stabilize the Aral Sea is viewed as unobtainable. It would require either drastic reductions in irrigation and subsequent economic calamity, or investments of tens of billions of dollars to rehabilitate the inefficient irrigation system [62].

There were several projects aimed at increasing the inflow to the Aral Sea. Some of these started during the Soviet period (before 1991) and involved the utilization of nonlocal water resources, even trans-basin water transfers. The "Sibaral", best known as "project of the century", was the idea of diverting several powerful Siberian rivers, the tributaries of the Arctic Ocean, to the semi-desert and desert south of the former Soviet Union. The water in the Sibaral project was to come from the confluence of the Ob and the Irtysh rivers and to be directed via a 2 500-km long canal through the Turan lowlands to the Amu Darya [14]. The plan was to use 27 km3 annually to provide more water for irrigation. The dominant factors for the cancellation of the technically feasible project were the excessive costs compared to the expected benefits and the growing concern for unpredictable environmental consequences [32]. After the collapse of the USSR in 1991, the attempts to scale down the Sibaral project included a proposal to divert water exclusively for the Aral Sea and not for irrigation expansion, transferring the water only to the northern part ofthe sea. This would substantially reduce the costs by shortening the route. However, it is not clear if the delivery of 10 to 15 km3 of water per year to the Aral would outweigh the environmental harms in Western Siberia. Another project of water transfer considers the possibility of constructing a 450-km long canal between the Caspian and the Aral Seas, transferring to the latter about 40 km3 of low saline water (4-6 gil) annually [14]. However, as the Caspian Sea lies below sea level and therefore below the level ofthe Aral Sea, such a delivery would require electric energy input, and the technical complications seem unsurmountable at present. By the year 2000, neither of the proposed projects had started.

A compromise solution which would alleviate the ecological conditions in the Amu Darya delta is the plan to build a dike in the southern part ofthe exposed sea bottom. This dike would withhold water in a reservoir, raising its level in the delta and allowing partial restoration of its former ecological conditions and preservation of low salinity in those areas which receive river inflows. It would create a network of artificial lakes, stabilizing the delta's water level. Periodic floodings would additionally inundate the delta because the Amu Darya is not regulated [32]. In fact, a small dam has already been constructed near Muynak (see Fig. 1) providing little ponds for fisheries and water supply. Nevertheless, the water of these artificial lakes comes from the Amu Darya and is of poor quality. Furthermore, although such an approach has yielded some successes at a local

53

spatio-temporal scale (partial fishery restoration), in the long-term perspective, this is Pyrrhic's victory: such artificial "irrigation" of the delta further prevents the Large Aral from stabilizing.

An economic solution to water crisis in the Aral Sea basin would be water pricing. Traditionally, water allocation for agricultural purposes has been provided "free", promoting inefficiency and waste. Although it is generally recognized that water pricing would be beneficial, its implementation faces many problems, such as lack of waterdelivery measurement facilities and outdated agricultural prices. Another obstacle is the traditional Islamic attitude towards water which is considered as a gift of God and can neither belong to anyone nor be sold. This corresponds to the Moslem agrarian law that does not recognize private property [14].

An agricultural solution calls for a structural reform in local land-use. It should be directed towards food-oriented and low water-consuming crops, instead of cotton and rice monocultures [48]. Among other restructuring measures are attempts to protect soil resources by intercropping of trees with seasonal crops, the utilizing natural fertilizer (dung) instead of mineral fertilizers, and diminishing use of pesticides and defoliants. However, cotton will still remain an important commodity for export, as long as the alternatives are not realized on a large scale. Furthermore, the transformation of cottonsown areas into food crops is very limited due to severe soil salinization. The salinized arable land could better be transformed into pastures with an intensification of cattle breeding, but instead, there is a tendency to plant opium poppy. This illegal but highly profitable crop increased substantially in recent years, and Central Asia, with an estimated 120000 ha is regarded as one of the largest producers in the world [28].

Some specialists proposed planting halophytes to stabilize the former sea bottom and prevent the salinized soil from aeolian erosion [47]. However, such measures are difficult to do and require a long time: these cultivations are only possible on the sea bed which has been exposed 15-20 years. The younger the exposed sea bottom is, the more saline and contaminated are the soils, making the restoration difficult or even impossible [44].

To summarize, in the year 2000, there are some signs of recovery in the Aral Sea basin, especially in the Syr Darya delta. Unfortunately, the ecological situation in the Amu Darya delta still remains critical.

4.2. LOCUST AND GRASSHOPPER MANAGEMENT

For decades, large-scale chemical locust control in the Aral Sea basin contributed to the deterioration of local ecological conditions. Thousands of tons of highly toxic and extremely persistent organochlorine pesticides (DDT and BCH) were carried into the Aral Sea between the 1950s and 1980s. During the 1980s and 1990s, the relay was taken by organophosphates and pyrethroids. Even now, extensive chemical treatments are still the panacea from acridid pest infestations in the Amu Darya delta: in 1999, more than 60000 ha were treated by synthetic pyrethroids (zetamethrin, esfenvalerate) and organophosphates (malathion). Being the epicentre of an ecological catastrophe, the ecosystems in the Amu Darya delta should be considered with the greatest diligence and care when planning

54

interventions. The most rational way to minimize the negative impact of the acridid pest management on the environment is to develop and implement a comprehensive, balanced concept of integrated pest management (IPM). Such a strategy should be based on the extensive surveys by both remotely sensed and ground methods and localized, preventive treatments of the key infested areas. Blanket applications of broad-spectrum pesticides should be replaced by barrier treatments with selective acridicides of moderate persistence. This would reduce the adverse impacts on non-target organisms and prevent repetitive treatments of the same areas, which are frequently practiced now, mostly due to a very long hatching period of the Italian Locust [13]. Grasshopper treatments with reduced doses of insecticide applied to the infestation in alternating parallel swaths (Reduced Agent! Area Treatments, or RAATs) showed promising results in the United States [25]. The strategy is, in fact, a combination of chemical and biological control, based on the preservation of the natural grasshopper enemies in the untreated swaths [26, 37]. We think that the RAA Ts approach could be successfully adapted to the conditions of the Aral Sea basin for use against grasshoppers (Oxya, Acrotylus) and the Italian Locust.

As long as the chemical method still prevails in the acridid pest management, the choice of insecticide for use in the conditions of the Amu Darya delta should be a question of particular scrutiny. For example, although such insecticides as deltamethrin, chlorpyrifos or diflubenzuron have good efficacy against locusts and grasshoppers, their use in a close proximity to water bodies (e.g., in the delta) is highly undesirable because of the negative impacts on non-target aquatic invertebrate fauna [2].

Since most formulations used in the acridid pest control are water-based emulsifiable concentrates (EC), the most commonly used sprayers are large, conventionaltype sprayers pulled by a tractor. They disperse from 150 to 300 I of insecticide solution per ha. Such a method of application requires large quantities of water and contributes to further contamination of the soil and ground water with insecticides and their metabolites. Recently, a new application technique was introduced. It consists of spraying non-diluted insecticide formulations from ultra-light aircraft (motorized deltaplans). In this case, an Ee product is actually used as an ultra-low volume (UL V) formulation, significantly reducing the amount of water involved. In 1999, more than 60% of anti-locust treatments in Uzbekistan were implemented using this technique. In our opinion, such a method is economical, efficacious and logistically simple, and it deserves further refinement.

Development of the alternative control strategies, for example, using biocontrol agents, could be very promising in the zone of ecological disaster. In this respect, it is necessary to recall that in the 1980s, extensive screening of possible candidates for microbiopesticides was undertaken by Nurzhanov [40, 41 , 42]. He dissected 1 500 nymphs and adults of the Italian Locust and 800 nymphs of the Migratory Locust from the Amu Darya delta, and also 445 nymphs of the Moroccan Locust from southern Uzbekistan. Over a dozen fungal, bacterial and protozoan pathogens were found and tested for antiacridid activity. The main result of this work was the isolation of a highly virulent strain of a fungal pathogen, Beauveria tenella (Siem.) Delacr., and the proposal for its use against different species of locusts and grasshoppers, primarily the Italian and Migratory Locusts [41]. A liquid suspension of spores of this fungus showed very good results in laboratory conditions (95% mortality of the 2nd instar Italian Locust hoppers in 7 d at a dose of

55

6.4xl08 spores/ml) [42]. In small-scale field trials in the Amu Darya delta, the same dose rate caused 68% mortality of the early instars of the Italian Locust [43]. Another fungal pathogen from southern Uzbekistan, Metarhizium anisopliae (Metch.) Sorok., caused 73% mortality of the Italian Locust hoppers in the field conditions at a dose rate of 1 x 108

spores/mIlO d after application [43]). The Metarhizium spp. fungi are considered among the most promising candidates for mycopesticide agents for locust and grasshopper control worldwide [8, 15,27,33,46].

Furthermore, a new species of a microsporidian pathogen Nosema sp. n. was discovered from the Moroccan Locust population from southern Uzbekistan [16]. Subsequently, the pathogen was tested in the laboratory and field conditions for control of the Italian Locust [9, 41]. In the laboratory, this microsporidian caused the wholesale mortality of the Italian Locust hoppers 2 wk after application [41], although attempts to use the pathogen in a bait mixture under field conditions failed. Despite a well-developed methodological basis for field screening and subsequent testing of microbiological pathogens for anti-locust activity [10], the pursuit of this very promising avenue of research was discontinued, due to financial constraints, soon after the collapse of the former Soviet Union. Still, the strains used in this work are maintained at the All-Russian Institute for Plant Protection (VIZR) in St. Petersburg. This collection could serve as a solid starting point for future projects on the development of alternative, fewer environmentally hazardous methods of controlling the acridid pests in the Aral Sea basin. The Nosematidae and some other Protozoa present interesting prospects as agents for control of grasshoppers and locusts [12] although in vivo production methods may present a logistical obstacle.

In our opinion, the current locust and grasshopper situation in the Amu Darya delta necessitates most careful attention from specialists in different fields: acridologists, ecologists, and environmental toxicologists. To resolve this problem with minimal harm to the ecosystem, it is necessary to develop a team project involving the best international experts. Provided the essential funding can be secured, this challenging task could be entrusted to the Association for Applied Acridology International (AAAI), a group of twenty-five specialists from six continents possessing key expertise in these domains.

5. Conclusions

The continuing desiccation of the Aral Sea remains one of the deepest economical, ecological and socio-political problems of Central Asia. Because of the poor water quality and deteriorating environmental conditions, the 5 million people of the basin experience severe health problems. The dwindling agricultural resources have eroded the living conditions of people below the poverty threshold. One of the consequences of the ecological catastrophe is the aggravating locust and grasshopper problem in the region. The Migratory Locust, which once required control at irregular intervals, has lost its economic importance to the Italian Locust and some grasshopper species. They inflict damage on a regular basis, requiring large-scale curative control to assure crop protection. Recurrent chemical interventions further deteriorate the fragile environmental conditions

56

in the zone of the ecological catastrophe. There is an urgent need for building an efficient and less harmful strategy of locust and grasshopper control in the Aral Sea basin integrating the best achievements of today , s applied acridology with the restoration oflocal ecological conditions.

Apparently, the structure of the present article is unbalanced: the "tools" and "solutions" occupy a modest space compared to immense "problems." We believe, however, that this disproportion adequately reflects the current state of the locust and grasshopper situation in the Aral Sea basin where the problems still outnumber the solutions by a large margin. And it is implausible to solve the problem of pest acridids out of the context of ameliorating the ecological situation in the Aral Sea basin in its entirety.

6. References

1. Agachanjanc, O.E. (1988) Wasserbilanz und wasserwirtschaftliche Probleme der mittleren Region der UdSSR (Mittelasien und Westsibiries), Petermanns Geographische Mitteilungen 2, 109-115.

2. Anonymous (1998) Evaluation offield trial data on the efficacy and selectivity of insecticides on locusts and grasshoppers. Report to FAD by the Pesticide Referee Group. Seventh meeting, Rome 2-6 March 1998, Food and Agriculture Organization of the United Nations, Plant Production and Protection Division, Rome.

3. Bei-Bienko, GSa. and Mishchenko, L.L. (1951) The Acridids of the Fauna of the USSR and Adjacent Countries. Vol. I, II, Academy of Sciences of the USSR, Moscow/Leningrad (in Russian). English translation 1963-1964: Locusts and grasshoppers of the USSR and adjacent countries, Parts I, II, Israel Program for Scientific Translations, Jerusalem.

4. Blackith, R.E. (1971) Morphometrics in acridology: a brief survey, Acrida I, 7-15.

5. Dech, S.W. and Ressl, R., (1993) Die Verlandung des Aralsees. Eine Bestandesaufahme durch Satellitenfernerkundung, Geographische Rundschau 6, 345-352.

6. Dukhovny, V.A. (1988) Aral: looking into the face of truth, Melioratsiya i Vodnoe Khozyaistvo 9,37-32 (in Russian).

7. Glazovsky, N.F. (1990) The Aral crisis, Priroda 10-11,27-41 (in Russian).

8. Goettel, M.S. (1992) Fungal agents for biocontrol, in CJ. Lomer and C. Prior (eds.), Biological control of locusts and grasshoppers, CAB International, Wallingford, Oxon, pp. 122-132.

9. Issi, LV. and Krylova, S.V. (1987) The Microsporidians of the acridids, in E.M. Shumakov (ed.), Locusts - Ecology and Control Methods, Proceedings of the All-Union Institute for Plant Protection (VIZR), Sl. Petersburg, pp. 58-62 (in Russian, with English summary).

10. Issi, LV., Latchininsky, A.V. and Gogolev, A.N. (1993) Recommendations on the screening of microbiological agents and their testing on anti-locust activity, Russian Academy of Agricultural Sciences, Moscow (in Russian).

II. Jago, N. (1963) A revision of the genus Calliptamus Serville (Orthoptera: Acrididae), Bulletin Br. Mus. Nat. Hist. (Entomology) 13, 289-350.

12. Johnson, D.L. (1997) Nosematidae and other Protozoa as agents for control of grasshoppers and locusts: Current status and prospects, Memoirs of the Entomological Society of Canada, 171,375-389.

13. Khudanov, S.K. (1998) lrifluence of the anthropogenicfactor on the acridids in the Aral Sea basin and the improvement of means of their chemical control, Ph.D. Dissertation (plant Protection), Uzbek Institute for Plant Protection, Tashkent (in Russian).

14. Killtzli, S. (1994) The Water and Soil Crisis in Central Asia - a Source for Future Conflicts? ENCOP Occasional paper No. 11, Center for Security Policy and Conflict Research Zurich/Swiss Peace Foundation Berne.

57

15. Kooyman, C. and Shah, P. (1992) Exploration for locust and grasshopper pathogens, in C.1. Lomer and C. Prior (eds.), Biological control of locusts and grasshoppers, CAB International, Wallingford, Oxon, pp. 208-213.

16. Krylova, S. V. and Nurzhanov A.A. (1987) A Microsporidian Nosema maroccanus sp.n. (Nosematidae) from the Moroccan Locust, Dociostaurua maroccanus (Thunb.) (Orthoptera), Bulletin VIZR 68, 10-15 (in Russian with English summary).

17. Latchininsky, A. V. (l990a) Phase characters in the population dynamics of Migratory Locust, Locusta migratoria (L.) from Karakalpakia, in G.S. Shamuratov and A.K. Matekeev (eds.), Problems of agricultural productivity in Karakalpakia, Karakalpakstan, Nukus, pp. 35-37 (in Russian).

18. Latchininsky, A. V. (I 990b ) Morphometrica1 characters of phase status oflocust populations, in Ecological Problems of Plant Protection, Proceedings of the All-Union Conference, Leningrad, November 21-24, 1990, VIZR, Leningrad, p. 44 (in Russian).

19. Latchininsky, A.V. (1991) On the number of generations in the Karakalpakian population of the Migratory Locust Locusta migratoria migratoria (L.) in G.S. Shamuratov and A.K. Matekeev (eds.), Problems of agricultural productivity in Karakalpakia, Karakalpakstan, Nukus, pp. 75-77 (in Russian).

20. Latchininsky, A.V. (1993) Characterization of the phase status of gregarious locust populations with morphometrical traits, in Kerzhner, I.M. and Pesenko Yu.A. (eds.), Advances in Entomology in the USSR: Ecology and Faunistics, Smaller Orders of Insects (Proceedings of the 1 (fl' Congress of the All-Union Entomological Society), Russian Academy of Sciences; Zoological Institute; Russian Entomological Society, St. Petersburg, pp.87-88 (in Russian).

21. Latchininsky, A. V. (1998) Moroccan Locust Dociostaurus maroccanus (Thunberg, 1815): a faunistic rarity or an important economic pest? Journal of Insect Conservation 2, 167-178.

22. Latchininsky, A.V. and Gapparov, F.A. (1996) Les consequences du dessechement de la mer d' Aral sur la situation acridienne dans la region, Secheresse 7, 109-113.

23. Latchininsky, A.V. and Launois-Luong, M.H. (1992) Le Criquet marocain, Dociostaurus maroccanus (Thunberg, 1815) dans la partie orientale de son aire de distribution. Etude monographique relative a l'exURSS et aux pays proches, CIRAD-GERDAT -PRIFAS, MontpellierNIZR, Saint-Petersbourg.

24. Lepeshkin, S.N. (1934) The system of control of the Oasis locust, in Lepeshkin, S.N., Zimin, L.S., Ivanov, E.N. and Zakhvatkin A.A. (eds.), The Acridids of Central Asia, Saogiz, Tashkent, pp. 229-236 (in Russian).

25. Lockwood, 1. A., Schell, S.P., Foster, R.N., Reuter, C. and Rachadi, T. (1999) Reduced agent-area treatments (RAA Ts) for management of rangeland grasshoppers: efficacy and economics under operational conditions. International Journal of Pest Management (in press).

26. Lockwood, J. A. and Schell, S.P. (1997) Decreasing economic and environmental costs through reduced area and agent insecticide treatments (RAA Ts) for the control of rangeland grasshoppers: Empirical results and their implications for pest management. Journal of Orthoptera Research 6, 19-32.

27. Lomer, C.J., Prior, C. and Kooyman, C. (1997) Development of Metarhizium spp. for control of grasshoppers and locusts, Memoirs of the Entomological Society of Canada, 171,265-286.

28. Lubin, N. (1993) Dangers and dilemmas. The need for Prudent Policy Toward Central Asia, Harvard International Review XV (3), 6-9.

29. Mainguet, M., Letolle, R. and Glazovsky, N. (1995) Aridite et secheresse dans la region aralo-caspienne, Secheresse 6, 135-143.

30. Micklin, P. (1987) Irrigation and its future in Soviet Central Asia: a preliminary analysis, in Holtzner, R. and Knapp, 1.M. (eds.), Soviet Geography Studies in Our Time, Festschrift for Paul E. Lyndolph, Milwaukee, pp. 229-261.

31. Micklin, P. (1991) The water crisis in Central Asia, in Pryde, P.R. (ed.) Environmental Management in the Soviet Union, Cambridge.

32. Micklin, P. (1992) Water management problems in Soviet Central Asia: problems and prospects, in Stewart, 1. M. (ed.), The Soviet Environmental Problems, Policies and PolitiCS, Cambridge, New York, pp. 88-114.

58

33. Milner, RJ. and Prior, C. (1994) Susceptibility of the Australian Plague Locust, Chorthoicetes termini/era, and the Wingless Grasshopper, Phaulacridium vil/atum, to the Fungi Metarhizium spp., Biological Control 4,132-137.

34. Mishchenko, L.L. (1952) The Acridids (Catantopinae), Fauna of the USSR, IV (2), New series No. 54, Zoological Institute of the Academy of Sciences of the USSR, Moscow/Leningrad (in Russian).

35. Nadvorny, V.G. (1985) The changes in the soil mesofauna following the drainage, in Fauna of Poles'e (Belorussia), Proceedings of the 4th regional scientific conference, Gomel, pp. 109-110 (in Russian).

36. Nikolsky, V.V. (1925) The Asiatic locust Locusta migratoria L .. A Monograph, Trudy Otdela Prikladnoi Entomologii 12 (2), pp. 1-332 (in Russian).

37. Norelius, E. E. and Lockwood, lA. (1999) The effects of standard and reduced agent-area insecticide treatments for rangeland grasshopper (Orthoptera: Acrididae) control on bird densities, Archives of Environmental TOXicology 37,519-528.

38. Novitzky, V.Y. (1953) The Asiatic Migratory Locust and methods of its control, Karakalpakskoe Gosudarstvennoe lzdatelstvo, Nukus (in Russian).

39. Novitzky, V.v. (1963) Ecological conditions of the permanent breeding area of the Asiatic Migratory Locust in the Amu Darya delta, Entomologicheskoe Obozrenie 42, 252-263 (in Russian).

40. Nurzhanov, A.A. (1988) Mycoses of the acridids in Karakalpakia in Shamuratov, G.S. (ed.) Crop protection from principal pests and weeds in Karakalpakia, Karakalpakstan, Nukus, p. 130-133 (in Russian).

41. Nurzhanov, A.A (1989) Entomopathogenic microorganisms of locusts in Uzbekistan and perspectives of their use in the biological plant protection, Ph.D. dissertation (Entomology), All-Union Institute for Plant Protection (VIZR), Leningrad (in Russian).

42. Nurzhanov, A.A. and Latchininsky, A.V. (1987) Pathogenic microorganisms of gregarious locusts in Uzbekistan, in E.M. Shumakov (ed.), Locusts - Ecology and Control Methods, Proceedings of the AIIUnion Institute for Plant Protection (VIZR), St. Petersburg, pp. 58-62 (in Russian with English summary).

43. Nurzhanov, A.A. and Pavliushin, V.A. (1990) Virulence of the entomopathogenic fungi for the Italian Locust nymphs, in Ecological Problems of Plant Protection, Proceedings of the All-Union Conference, Leningrad, November 21-24, 1990, VIZR, Leningrad, p. 246 (in Russian).

44. Perera, J. (1993) A sea turns into dust, New Scientist 140 (1896), October 23, pp. 24-27.

45. Precoda, N. (1991) Requiem for the Aral Sea, AMBIa 20 (3-4), 109-114.

46. Prior, C. (1992) Discovery and characterization of fungal pathogens forlocust and grasshopper control, C.]. Lomer and C. Prior (eds.), Biological control of locusts and grasshoppers, CAB International, Wallingford, Oxon, pp. 159-180.

47. Rafikov, AA (1988) The fight against desertification in the lower reaches of the Amu Darya, Melioratsiya i Vodnoe Khozyais/Vo 9, 32-35 (in Russian).

48. Rumer, B. Z. (1989) Soviet Central Asia. A Tragic Experiment, Unwin Hyman, Boston.

49. Samways, MJ. (1997) Conservation biology ofOrthoptera, in Gangwere, S.K., Muralirangan, M.C. and Muralirangan, Meera (eds.), The Bionomics of Grasshoppers, Katydids and their Kin, CAB International, Wallingford, New York, pp. 481-496.

50. Sergeev, M.G. (1986) Patterns ofOrthoptera Distribution in Northern Asia, NaukaPublishers, Novosibirsk (in Russian).

51. Shamuratov, G.S. (1975) The Asiatic Migratory Locust in Karakalpakia, Karakalpakstan, Nukus (in Russian).

52. Shamuratov, G.S. (1979) The Principles of Crop Protectionfrom Pests in Karakalpakia, Karakalpakstan, Nukus (in Russian).

53. Shamuratov, G.S. and Kopaneva, L.M. (1984) The Acridids in Karakalpakia, Karakalpakstan, Nukus (in Russian).

54. Shamuratov, G.S. and Latchininsky, AV. (1991a) Migratory Locust in the Amu Darya delta, in Shamuratov, G.S., Shamshetov S.N. and Matekeev AK. (eds.), Principal Agricultural Insect Pests in Karakalpakia, Karakalpakstan, Nukus, pp. 6-7 (in Russian).

59

55. Shamuratov, G.S. and Latchininsky, A.V. (1991b) Italian Locust in the Amu Darya delta, in Shamuratov, G.S., Shamshetov S.N. and Matekeev A.K. (eds.), Principal Agricultural Insect Pests in Karakalpakia, Karakalpakstan, Nukus, pp. 8-11 (in Russian).

56. Sinnot, P. (1993) The Physical Geography of Soviet Central Asia and the Aral Sea Problem, in Lewis, R.A. (ed.) Geographical Perspectives on Soviet Central Asia, London, New York, pp. 75-97.

57. Siyazov, M.M. (1912) The Control of Locusts in Turkestan, Departament Zemledeliya, Tashkent (in Russian).

58. Smith, D.R. (1992) Salinization in Uzbekistan, Post Soviet Geography 1, 21-32.

59. Stolyarov, M.V. (1966) Specific composition of the Orthopterans (Orthoptera) of Karakalpakia and some particularities of their ecological distribution, Zoologicheskii Zhurnal 45, 1017-1022 (in Russian with English summary).

60. Stolyarov, M.V. (1967a) Italian Locust, Calliptamus italicus (L.), in the Volga region and some data on its forecasting, Zoologicheskii Zhurnal46, 365-370 (in Russian with English summary).

61. Stolyarov, M. V. (1967b) Italian Locust, Calliptamus italicus (L.) (Orthoptera, Acrididae), in Karakalpakia, Entomologicheskoe Obozrenie 46, 615-628 (in Russian).

62. Stone, R. (1999) Coming to grips with the Aral Sea's grim legacy, Science 284, 30-33.

63. Tsyplenkov, E.P. (1970) Harmfol Acrididae of the USSR, Kolos Publishers, Leningrad (in Russian). English translation 1978: Amerind Publishing Co., New DelhilBombay.

64. Tsyplenkov, E.P. (1979) Methodological Instructions for Anti-locust Control, Kolos Publishers, Moscow/Leningrad (in Russian).

65. Uvarov, B.P. (1957) The aridity factor in the ecology oflocusts and grasshoppers of the Old World, in Arid Zone Research VIII. Human and animal ecology. Reviews of research, UNESCO, Paris, pp. 164-198.

66. Uvarov, B.P. (1966) Grasshoppers and Locusts. A Handbook of General Acridology. Vol. I. Anti-Locust Research Centre, University Press, Cambridge.

67. Uvarov, B.P. (1977) Grasshoppers and Locusts. A Handbook of General Acridology. Vol. II. Centre for Overseas Pest Research, London.

68. Vasiliev, K.A. (1962) Italian Locust, Calliptamus italicus (L.) in central Kazakhstan, Trudy n.-i. Instituta Zastchity Rastenii KazSSR 7, 124-190 (in Russian).

69. Verdier, M. (1972) The different life-cycles in Locusta in relation to climatic and genetic diversity, in Hemming, C.F. and Taylor, T.H.C. (eds.), Proceedings of the international study conference on the current andfoture Problems of Acridology, London, 1970, Centre for Overseas Pest Research, London, pp. 335-338.

70. Vilkamaa, P. (1987) The impact ofthe drainage on the numbers and biomass of soil invertebrates in pine swamps, in Soil Fauna of the Northern Europe, Academy of Sciences of the USSR, Moscow, pp. 103-109 (in Russian).

71. Vorontzovsky, P.A. (1932) Contribution to the study of the entomofauna of Karakalpakia, Trudy Kompleksnogo n.-i. Instituta Karakalpakskoi Avtonomnoi Oblasti (seriya bioI.) 2, 3-98 (in Russian).

72. Zakharov, L.Z. (1930) Draining the reedbeds near the Azov Sea and the locust problem in Kuban, Izvestiya Severo-Kavkazskoi Kraevoi Stantsii Zastchity Rastenii 5,97-104 (in Russian with German summary).

73. Zimin, L.S. (1931) On the biology and ecology ofCalliptamus italicus (L.), Izdanie Sredne-Aziatskogo Instituta Zastchity Rastenii 24, 94-251 (in Russian).

3. WHAT IS THE THE ROLE OF GRASSLAND VEGETATION IN GRASSHOPPER POPULATION DYNAMICS?

O.OLFERT Agriculture and Agri-Food Canada 107 Science Place Saskatoon, SK Canada S7N OX2

Abstract

Grasshoppers are the most important insect pests of small-grain crops in Canada. The concept of utilizing resistant host-plants to manage grasshopper populations and reduce crop damage has received increased attention in recent years. The dual role of resistant plants is in antibiosis and antixenosis. Effective suppression of grasshopper populations would reduce the overall need for insecticides, an approach which is supportive of the concept of managing grasshopper outbreaks without risking environmental disaster.

1. Problem

The bulk of arable land and range land in Western Canada occurs within the Prairie Ecozone, a northern extension of the Great Plains of North America (Fig. 1). The Prairie Ecozone has been subdivided into seven ecoregions as classified by the Canadian Soil Information System [1]. The grasslands, composed of the Mixed, Fescue and Moist Mixed Ecoregions, make up approximately 60% of the Prairie Ecozone. This large expanse of native grasslands has been irrevocably altered by human settlement over the past 150 years. Annual crops, predominantly small-grain crops, now occupy approximately 60% ofthe 50 million ha of the Prairie Ecozone [2]. The remaining 20 million ha are still in native grassland or in seeded perennial grass and forage crops. This alteration has resulted in not only a net loss of native habitat for grasshoppers but has also resulted in extensive fragmentation of remaining native habitat. The full impact of these dramatic changes on the diversity of grasshoppers is not known, however, recent studies have begun to address the issue in relation to other insect groups [38, 49].

Grasshoppers have been a feature of prairie agriculture since settlement of the Northern Great Plains [54]. Only a small proportion of the more than 90 species described by Brooks [11] in this region are of recurring economic importance. However, the need to control grasshopper populations has been a constant feature in the production of smallgrain crops. The availability of suitable food plants in the crop and suitable

61


62

Figure I. Ecozones of Western Canada as classified by the Centre for Land and Biological Resources (1995), Ottawa

oviposition sites in grassy areas has significantly contributed to a problem far greater than occurred in the grass prairie alone [52, 54]. As a result, the threat to production of small grains arises from migration of the hatchling populations into crop land from roadsides, headlands and field margins. Grasshopper damage to cereals and grasses consists of defoliation at all stages of plant growth [50], head clipping [37] and feeding on flower structures and seeds [12]. The relationships between grasshoppers and cereal crops [40] and between grasshoppers and rangeland plants [32] have received considerable attention.

The predominant method of grasshopper control during outbreaks is to apply insecticides. The economic and environmental concerns about pesticide use have stimulated a broad interest in alternative strategies to manage these pests. These include monitoring and forecasting population abundance, and implementation of cultural and chemical controls. Such strategies have been adopted under the banner of integrated pest management, where the overall aim is to manage economic levels of pest populations through a variety of measures in an effort to reduce the reliance on chemical insecticides. In recent years, management strategies for control of insect pests, such as grasshoppers, have broadened into the concept of ecological pest management. Management strategies are no longer just integrated at the level of the pest species complex but consideration is now being given to broader ecological issues at the landscape level, such as the role of alternative farming systems on insect population abundance and diversity [9].

Although the current proportion of annual crops to grass lands is not likely to change dramatically in the near future, crop production practices will continue to evolve

63

and will continue to impact grasshopper populations in different ways. New technologies to conserve agricultural soil and water are being designed to improve the sustainability of farming systems. To address the issues of profitability and soil degradation farmers are encouraged to diversify their production and to reduce farm inputs. However, climatic limitations favour some cropping options such as small-grain cereals, cool-season oil seeds and pulses, and perennial forages. In the context of these changes, the role of host-plants in managing grasshopper populations needs to be addressed. Effective suppression of grasshopper populations would reduce the overall need for insecticides, an approach which is supportive of the concept of managing grasshopper outbreaks without risking environmental disaster.

2. Tools

2.1 HOST-PLANT RESISTANCE AS A TOOL FOR MANAGING INSECT PESTS

Host-plant resistance refers to the genetic traits of a plant species that confer upon it some level of protection against herbivores. In the context of this paper, the term is broadly defined to include any plant attribute which contributes in any way to a reduction in insect damage potential. These would include plant compounds which are toxic to insects, chemical or morphological attributes of a plant that deter insect feeding, and plant growth characteristics which permit tolerance of insect damage. From an insect perspective, resistance mechanisms may deter insect feeding, oviposition, and etc. (antixenosis) or they may have deleterious physiological effects on the insect (antibiosis).

Host-plant resistance has been described as an effective and ideal tool in the management of insect pests [66]. One of the earliest records of differential susceptibility of cereals to insect damage is the wheat cultivar Underhill as being resistant to the Hessian fly, Mayetiola destructor [22]. Since that time, it has been estimated that there has been a $3.4 billion increase in farm income due to cultivars resistant to the Hessian fly [56]. Cereal cultivars with resistance characteristics to this and other insect pests are now relatively common. There are both wheat and barley cultivars with resistance to cereal leaf beetle, Oulema melanopus [64]; wheat cultivars with resistance to wheat stem sawfly, Cephus cinctus [30]; wheat, barely and oat cultivars with resistance to greenbug, Schizaphis graminum [63]; wheat cultivars with resistance to Russian wheat aphid, Diuraphis noxia [53]; and a wheat cultivar with resistance to orange wheat blossom midge, Sitodiplosis mosellana, [3]. The concept of utilizing resistant host-plants to manage grasshopper populations and reduce crop damage has also received increased attention in recent years [28].

The sources of resistance to the insect pests mentioned above have been primarily derived from conventional sources. The major cereal and grass crops of today are commonly grown around the world under a wide range of environmental and agronomic conditions. As a result, the thousands of extant cultivars of cereals, grasses and forbs provide a rich genetic resource from which to search for resistance to grasshoppers. Today, utilization of biotechnological procedures provides even further opportunities for facilitating the transfer of genetic material from germplasm sources between plants that are

64

too genetically divergent to permit conventional crossing.

2.2 PLANT ATTRIBUTES CONTRIBUTING TO GRASSHOPPER RESISTANCE

In general, the attributes which render plants less susceptible to insect attack can be chemical or morphological in nature. Morphological characteristics, such as leaf trichomes, pubescence and lignification do have negative effects on certain insects [5, 64]. However, the significance of these features in deterring grasshopper feeding will depend on the degree of feeding specialization of the species in question. Gangwere et at. [18] reported that there is evidence of specialization in mandibular morphology and digestive tracts of grasshoppers which allow them to cope with the physical characteristics of hostplants. The surface layer of protective waxes on the leaves of plants can also influence grasshopper feeding [67]. From a plant perspective, foliar waxes provide protection against dessication, plant pathogens and herbivores. From an insect perspective, they can influence whether an insect will begin feeding and how much it will continue to feed. The initial response of grasshoppers to cereals in binary preference studies suggests that leaf surface repellancy is a feature of oats and some cultivars of barley [26].

Secondary plant chemicals, such as alkaloids, tannins and protease inhibitors, are also implicated in host-plant resistance. The most common alkaloids found in the Poacea are gramine and hordenine. Relatively low levels of gramine in grasshopper diets can negatively influence feeding, growth and survival [20, 65]. Although gramine is expressed in some cultivars of barley [19] there is stronger evidence that these compounds playa more important role in non-cereal grasses. 8ernays and Chapman [7] found that grass species which were favored by grasshoppers had a lower occurrence ofthese compounds in the leaves. Tannins, which have been reported to occur in the heads of barley [8], are thought to disrupt the digestive processes of insects due to their protein binding properties [17]. The effects of condensed tannins in Lotus species on M. sanguinipes have been documented [39]. Protease inhibitors are widely reported in the seeds of legumes and cereals but they are also found in the leaves of barley [34] and non-cereal grasses [57]. Protease inhibitors can cause severe protein attrition by stimulating the over-production of proteolytic enzyme in insects [10]. The gut response of M. sanguinipes to oats, for example, is to produce a 2.8-fold increase in trypsin and chymotrypsin, and is soon accompanied by the symptoms of protein attrition [27]. In general, studies indicate that grasshopper responses to secondary plant chemicals vary considerably and can be correlated with dietary breadth and preferences [36]. In some species, feeding deterrence in the presence of secondary plant compounds is so pronounced that the plants do not require the production of toxic compounds for protection against herbivory [4]. At the other extreme, grasshoppers can be attracted to and readily feed on plants which produce deleterious compounds [26, 44].

2.3 HOST-PLANT RESISTANCE TO GRASSHOPPERS IN CEREALS AND GRASSES

In North America, early attempts to utilize host-plant resistance in a breeding program as a tool to manage grasshoppers were recorded by Painter [48]. The first sources of

65

grasshopper resistance in cereals and grasses were identified during the major grasshopper outbreak that devastated the Great Plains of North America in the 1930s. Clark [13] noted that a cultivar of spring wheat, Ceres, was more resistant to grasshopper damage than other cultivars in North Dakota. As a result, Ceres was used in breeding programs and in grasshopper damage studies in an attempt to reduce the devastating impact of grasshoppers on wheat production [62]. Subsequent grasshopper damage studies focused on the observation that grasshoppers differentially clipped the heads off of maturing bread wheat, durum wheat and barley [33, 37]. These studies found a significant correlation between degree of head clipping and days to maturity. However, this relationship did not extend to other symptoms of grasshopper damage such as plant defoliation. Although there were significant differences in the level of defoliation within 44 cultivars of wheat and barley, no significant correlation was observed between the extent of plant defoliation and days to maturity [42].

In grasses, Hermann and Eslick [23] quantified the extent of grasshopper feeding on 28 species of grasses at the seedling stage. The least damaged grasses were Elymus smithii, Elymus canadensis and Phalaris arundinacea. In a similar study, Hewitt [24] reported that grasses with high toughness ratings, such as Elymus dasystachyum and Stipa viridula were damaged the least by Melanoplus sanguinipes. Although Hewitt [24] found that there was a poor correlation between feeding preferences of nymphs and adults in his study, Chu and Knutson [12] reported that there was a consistent response by nymphs and adults to the species which were least preferred of the II species tested. In general, seedling grasses of commercial species appear to be more distasteful to grasshoppers than more mature plants, however, this may not be true with North American wild species of grasses [6].

Resistance to grasshoppers in cereals and grasses involves the effect of the food plant on grasshopper growth, survival and reproduction. The summed effects of these factors are measurable as the grasshopper's biotic potential; expressed as the net reproductive rate. The term biotic potential is used here to define a quantitative assessment of the net growth potential of popu lations where a value of less than, or more than, unity results in population decline and increase, respectively [43]. In univoltine species of grasshoppers where there does not appear to be a trade-off between life span and reproduction [16], any plant attribute which retards the physiological development of grasshoppers can be expected to impact on population growth negatively. Further reductions in population growth can be attained by attrition. Targeting the neonates is a reasonable objective because grasshopper mortality tends to be the greatest during the first two instars [47]. As a result, recent investigations of host-plant resistance to grasshoppers have focused on using neonates to assess the potential resistance of wheat, barley, oats and rye [25, 29] and grasses [45]. Olfert et at. [43] reported that popUlations of M sanguinipes, which were fed oats and a wheat cultivar HY320, had significantly higher mortality and lower rates of growth than those which were fed other crops or cultivars. When feeding on oats or HY320 was extended over their entire life cycle, these diets reduced the biotic potentials to such low levels that they suffered significant population decline [26,43]. In field studies using naturally-occurring populations of grasshoppers, the cultivars which suffered the least amount of defoliation in an extensive series of cereal cultivars correlated well with those showing the greatest negative effects on biotic potential

66

in laboratory studies [42,43]. Grasshoppers also have significant differential feeding preferences and growth

responses on different grass species. Hewitt [24] reported that, of the eight species evaluated, Elymus spicatum, E. desertorum and Stipa viridula were the least suitable food plants to support development of M. sanguinipes. Sanchez and On sager [58] also found that natural populations of M. sanguinipes were suppressed by diets of E. desertorum. In a 21-day assay using 12 grass species, Olfert et al. [45] reported that growth and development were significantly higher in M. sanguinipes nymphs feeding on E. smithii and the lowest when feeding on Bromus inermis. However, nymphal mortality among the groups fed the different grasses was not significantly different over the 21 days.

3. Solutions

It is evident from the literature that differential feeding by grasshoppers does occur in cereals and grasses, and that this can be correlated with their biotic potential. Host-plants which are responsible for reducing the biotic potential of grasshoppers below unity will have a suppressive effect on population growth. These same host-plants frequently are less preferred by grasshoppers, are consumed at a significantly lower rate and delay growth and development of nymphs. The net result is that damage to field crops is lessened and the area of infestation will be buffered which, in tum, contributes to a reduction in the overall amount of insecticides due to fewer instances where population density levels warrant applications. These findings offer insights for the potential utilization of host-plants in the development of solutions to manage grasshopper populations without risking environmental disaster. Host-plant resistance strategies can be employed in: 1) Selection of resistant grass species for roadsides and field margins; 2) Selection of crop cultivars that are resistant to grasshoppers to minimize yield loss; 3) Cultural control of grasshoppers by utilizing barrier strips (resistant plant species) or traps strips (attractive plant species); and 4) Evaluation offarming systems to ensure that new agronomic practices do not contribute further to grasshopper population outbreaks.

Rangeland, roadsides, headlands and field margins are favoured oviposition sites throughout the prairies in Western Canada. Because these sites contain the food plants critical for survival of the early-instars, widespread deployment of resistant host-plants in these areas would afford some level ofpopulation suppression. This approach is supported by the findings of several studies involving grass species [15, 24, 45, 52, 58, 60]. The major factor limiting successful implementation of such a strategy is the presence of weeds in these areas. The beneficial effects of having access to weed species in grasshopper diets has been documented [51]. In Saskatchewan, where there are approximately 26 000 km of roads, records of the annual grasshopper abundance survey indicated that 30% of the survey stops reported a 'weedy' roadside [45].

The prospect of managing grasshopper populations by growing resistant annual crops is encouraging [28]. Utilization of oats [43] or peas [45] in areas that are heavily infested with grasshoppers will reduce the need for chemical intervention and will not exacerbate the problem for future years by dampening the biotic potential of the population. This feature is particularly important for farmers whose land holdings are

67

adjacent to ecologically-sensitive areas and who are required by law to exercise a buffer strip along field margins when implementing chemical controls.

Both extremes of the host-plant suitability spectrum can be employed as cultural control strategies. Highly attractive host-plants can be grown as trap strips in fields which are adjacent to infested areas. Grasshoppers tend to migrate from infested areas when required to search for food. Traps strips of highly-preferred food plants will attract grasshoppers and concentrate the population within a relatively confined area in a short time span, where they can be controlled efficiently with insecticidal sprays [41]. Trap strips were recommended as early as 1919 [14] and are still a recommended control method [21]. Highly resistant plants, on the other hand, can be employed as a barrier strip next to infested areas to protect more susceptible crops. In addition to the beneficial attributes of resistant plants described above, a dense stand of resistant plants will retard the rate at which grasshoppers disperse through the crop [28]. This usually provides enough opportunity for spring-seeded annual crops to become well established and to lessen the damage potential of grasshoppers.

In conclusion, resistant host-plant species should be a component of future recommendations as farming systems continue to evolve. Alternative farming practices need to be carefully assessed to ensure that new crops or rotations do not exacerbate grasshopper problems. An example of such a program is the long-term alternative cropping systems study initiated in Saskatchewan in 1994 [9]. This study was designed to monitor and assess alternate input and cropping strategies with respect to (i) biodiversity, (ii) pest dynamics, (iii) farm profitability, (iv) soil quality, (v) food safety. The experimental framework is based on a matrix of three levels of input use, and three levels of cropping diversity. The design, data collection and evaluation are based on the collaborative efforts of crop, pest, economic and soil scientists. In relation to arthropods: (i) a baseline data set containing density and diversity of organisms found in the crop, in and on the soil has been compiled; and (ii) the effect of long-term rotations and farming systems on arthropod populations is being evaluated [9]. In a specific example, the resistance to grasshoppers of several legumes which were under consideration as green manure crops was assessed prior to implementation ofthe technology [46]. Green manures were being recommended on the prairies as an alternative to summer fallow. The advantages are that green manures provide a protective ground cover against wind erosion, they have the ability to fix nitrogen and they improve soil quality. The assessment concluded that, of the four legume species evaluated, Lathyrus tingitanus and L. sativus should be recommended for green manure crops because they significantly suppressed grasshopper populations when included in their diets. These findings suggested that green manure crops ofthese resistant legumes would not likely become epicentres of grasshopper outbreaks [46].

4. Acknowledgements

The manuscript is Contribution Number 1329 of the Saskatoon Research Centre.

68

s. References

1. Anonymous (1995) A National Ecological Frameworkfor Canada, Ecological Stratification Working Group, Agriculture and Agri-Food Canada Ottawa.

2. Anonymous (1997) Agriculture Profile of Canada, Statistics Canada, Ottawa. 3. Barker, B. (1998) Pursuing wheat midge resistance, Top Crop Manager 28,16-19. 4. Bernays, E.A. (1990) Plant secondary compounds deterrent but not toxic to the grass specialist Locusta

migratoria: Implications forthe evolution of graminivory, Entomologia Experimentalis et Applicata 54.53-56.

5. Bernays, E.A and Chapman, R.F. (1970) Experiments to determine the basis of food selection by Chorthippus parallelus (Orthoptera: Acrididae) in the field, Journal of Animal Ecology 39, 761-776.

6. Bernays, E.A. and Chapman, R.F. (1976) Antifeedant properties of seedling grasses, Symposia Biologica Hungarica 16, 41-46.

7. Bernays, E.A. and Chapman, R.F. (1977) Deterrent chemicals as a basis of oligophagy in Locusta migratoria. Ecological Entomology 2, 1-8.

8. Brandon, MJ., Foo, L.Y., Porter, L.J., and Meredith, P. (1982) Proanthocyanidins of barley and sorghum; composition as a function of maturity of barley ears, Phytochemistry 21,2953-2957.

9. Brandt, S.A.,Thomas, AG., Olfert, 0., Moulin, AP., Boyetchko, S.M., Selles, F., Zentner, R.P., Kutcher, R., Fernandez, M., Frick, B.L., Weiss, R., and Sonntag, G. (1998) Assessing alternative crop production systems for the Canadian Prairies. Proceedings of Agriculture in Harmony with the Environment, Canada Saskatchewan Agriculture Green Plan. Saskatoon, February 16-19, 1998. pp. 1-2.

10. Broadway, R.M. and Duffy, S.S. (1986) The etfect of plant protein quality on insect digestive physiology and the toxicity of plant proteinase inhibitors, Journal of Insect Physiology 34, 1111-1117.

11. Brooks, AR. (1958) Acridoidea of southern Alberta, Saskatchewan and Manitoba (Orthoptera: Acrididae), The Canadian Entomologist, Suppl. 9, 1-92 pp.

12. Chu, I.W. and Knutson, H. (1970) Preferences of eight grasshopper species among eleven species of cultivated grasses, Journal of Kansas Entomological Society 43, 20-31.

13. Clark, J.A (1936) Improvement in Wheat, United States Department of Agriculture Yearbook 1936. 14. Criddle, N. (1920) Locust control in the Prairie Provinces, Canadian Department of Agriculture

Entomology Division Proceedings, 1-20. 15. Davis, E.G. (1949) Reducing grasshopper damage by regrassing weedy roadsides and fence rows, U.S.D.A.

Circular No. 813,1-11. 16. Dean, J.M. (1981) The relationship between lifespan and reproduction in the grasshopper Melanoplus,

Oecologia 48, 385-388. 17. Feeny, P. (1969) Inhibitory effect of oak leaf tannins on the hydrolysis of proteins by trypsin,

Phytochemistry 8, 2119-2126. 18. Gangwere, S.K., Mulalirangan, M.e., and Mulalirangan, M. (1989) Food selection and feeding in acridoids:

a review, Contributions of the American Entomological Institute 25, 1-56. 19. Hanson, AD., Traynor, P.L., Ditz, K.M., and Reicosky, D.A. (1981) Gramine in barley forage - effects of

genotype and environment, Crop Science 21, 726-730. 20. Harley, K.L. and Thornsteinson, A.J. (1967) The influence of plant chemicals on the feeding behaviour,

development, and survival of the two-striped grasshopper Melanoplus bivillatus, Acrididae: Orthoptera, Canadian Journal of Zoology 45, 305-319.

21. Harris, J.L., Buchan, J.A., Ewen, A.B., Frowd, I.A., Minja, L.A., Mukerji, M.K., Riegert, P.W., and Weisbrot, D.R. (1984) Grasshopper Control. Saskatchewan Agric., Regina.

22. Havens, 1.N. (1792) Observations on the Hessian Fly, New York Society of Agricultural Transactions, Arts and Manufacturing \,89-107.

23. Hennann, W. and Eslick, R. (1939) Susceptibility of seedling grasses to damage by grasshoppers, Journal of the American Society of Agronomy 31.333-337.

24. Hewitt, G.B. (1969) Twenty six varieties offorage crops evaluated for resistance to feeding by Melanoplus sanguinipes, Annals of the Entomological Society of America 62. 737-741.

25. Hinks, e.F., Olfert, 0., and Westcott, N.D. (1987) Screening cereal cultivars for resistance to early-instar grasshoppers, Journal of Agricultural Entomology 4, 315-319.

26. Hinks, C.F., Olfert, 0., Westcott, N.D., Coxworth E.M., and Craig W. (1990) Preference and performance in the grasshopper Melanoplus sanguinipes (Fab.) (Orthoptera: Acrididae) feeding on kochia, oats and wheat: implications for population dynamics, Journal of Economic Entomology 83, 1338-1343.

69

27. Hinks, C.F., Cheeseman, M.T., Erlandson, M.A., Olfert, 0., and Craig, W. (1991) The effects ofkochia, wheat and oats on digestive proteinases and the protein economy of adult grasshoppers, Melanoplus sanguinipes, Journal of Insect Physiology 37,417-430.

28. Hinks, C. F. and Olfert, O. (1992) Cultivar resistance to grasshoppers in temperate cereal crops and grasses: a review, Journal ofOrthoptera Research 1,1-9.

29. Hinks, C.F. and Olfert O. (1993) Growth and survival to the second instar of neonate grasshopper nymphs, Melanoplus sanguinipes (F.) fed cultivars ancestral to hard red spring wheat, Journal of Agricultural Entomology 10,171-180.

30. Holmes, N.D. and Peterson, K.L. (1962) Resistance of spring wheats to the wheat stem sawfly, Cephus cinctus (Hymenoptera: Cephidae). 11. Resistance to the larvae, The Canadian Entomologist 94,348-365.

31. Jacobson, L.A and Farstad, C.W. (1941) Some observations on differential feeding on maturing wheat varieties by grasshoppers, The Canadian Entomologist 73, 158-159.

32. Joern, A (1987) Behavioural responses underlying ecological patterns of resource use in rangeland grasshoppers, in J.L. Capinera, (ed.), integrated Pest Management on Rangeland - A Shortgrass Prairie Perspective, Westview Press, Boulder, pp. 137-161.

33. Jones, E.T. (1943) Insect resistance in wheat, Journal of the American Society of Agronomy 35,695-703. 34. Kirsi, M. and Mikola,.l. (1977) Occurrence and heterogeneity of chymotrypsin inhibitors in vegetative tissue

of barley, Physiologia Plantarum 39, 110-114. 35. Kogan, M. and Ortman, E.F. (1978) Antixenosis-a term proposed to define Painter's 'Nonpreference'

modality of resistance, Bulletin of the Entomological Society of America 24, 175-176. 36. Kogan M. (1982) Plant resistance in pest management, in RL. Metcalf and W.H. Luckman (eds.)

introduction to Insect Pest Management, John Wiley and Sons, New York, pp. 93-134. 37. McBean, D.S. and Platt, AW. (1951) DitTerential damage to barley varieties by grasshoppers, SCientific

Agriculture 31,162-175. 38. Melnychuk, N. (1999) Predatory Insects in Saskatchewan Farming Systems, M.Sc. Thesis, University of

Saskatchewan. 39. Muir, A.D., Gruber, M.Y., Hinks, C.F., Lees, G.L., Onyilagha, J., Hallett, R., Xia, F., Soroka, J., and

Erlandson, M. (1999). The effect of condensed tannin in the diets of major crop insects, in G.G.Gross, R.Hemingway, R. and T.Yoshida (eds.), Plant polyphenols 2: Chemistry and Biology, Plenum Press, New York. (In press)

40. Olfert, O. (1970) Quantitative Evaluation of Grasshopper Defoliation of Wheat in Saskatchewan, PhD Thesis, University of Saskatchewan.

41. Olfert, O. (1986) Evaluation of trap strips for grasshopper (Orthoptera: Acrididae) control in Saskatchewan, Canadian Entomologist 118, 133-140.

42. Olfert, 0., Hinks, C.F.,Westcott, N.D., Crowle, W.L., Dziadyk, D.A., and Duczek, LJ. (1988) Differential feeding by grasshoppers and levels offoliar diseases in various cultivars of spring cereals, Crop Protection 7,338-343.

43. Olfert, 0., Hinks, C.F., and Westcott, N.D. (1990a) Analysis of the factors contributing to the biotic potential of grasshoppers (Orthoptera: Acrididae) reared on different cereal cultivars in the laboratory and in the field, Journal of Agricultural Entomology 7,275-282.

44. Olfert, 0., Hinks, C.F., Craig, W., and Westcott, N.D. (1990b) Resistance of Kochia scoparia to feeding damage by grasshoppers (Orthoptera: Acrididae), Journal of Economic Entomology 83, 2421-2426.

45. Olfert 0., Hinks C.F., Weiss RM., and Wright S.B.M. (1994) The effect of perennial grasses on growth, development and survival of grasshopper nymphs (Orthoptera: Acrididae): Implications for population management in roadsides, Journal of Orthoptera Research 2, 1-3.

46. Olfer! 0., Hinks C.F., Biederbeck, V.O., Slinkard, AE., and Weiss R.M. (1995) Annual legume green manures and their acceptability to grasshopper nymphs (Orthoptera: Acrididae), Crop Protection 14,349-354.

47. Onsager J.A and Hewitt, G.B. (1982) Rangeland grasshoppers: Average longevity and daily rates of mortality among species in nature, Environmental Entomology 11,127-133.

48. Painter, RH. (1951) Insect Resistance in Crop Plants, MacMillan, New York. 49. Pepper, J. (1999) Diversity and community assemblages of ground-dwelling beetles and spiders on

fragmented grasslands of southern Saskatchewan. M.Sc. Thesis, University of Regina. 50. Pfad!, RE. and Hardy, D.M. (1987) A historical look at rangeland grasshoppers and the value of grasshopper

control programs, in J.L. Capinera, (ed.) Integrated Pest Management on Rangeland - A Shortgrass Prairie Perspective, Westview Press, Boulder, pp. 183-195.

70

51. Pickford, R. (1962) Development, survival and reproduction of Melanoplus hilituratus (Orthoptera: Acrididae) reared on various food plants, The Canadian Entomologist 94,859-869.

52. Pickford, R. (1963) Wheat crops and native prairie in relation to the nutritional ecology of Cam nul a pellucida (Scudder) (Orthoptera:Acrididae) in Saskatchewan, The Canadian Entomologist 95,764-770.

53. Porter, D.R., Baker, C.A, and Webster, J.A (1998) Inheritance of Russian wheat aphid resistance in PI 140207 spring wheat, Plant Breeding 117, 293-294.

54. Riegert, P. W. {I 980) From Arsenic to DDT: a History of Entomology in Western Canada, University of Toronto Press, Toronto.

55. Riegert, P.W., Pickford, R., and Putnam, L.G. (I965) Outbreaks of Camnula pellucida (Orthoptera: Acrididae), in relation to native grasslands and cereal crops in Saskatchewan, The Canadian Entomologist 97,508-514.

56. Roberts,J.J., Foster, J.E., and Patterson, F.L. (1988) The Purdue-USDA small grain improvement programa model of research productivity, Journal of Production Agriculture 1, 239-241.

57. Ross, C.W. and Detling, J.K. (1983) Investigations oftrypsin inhibitors in leaves offour North American prairie grasses, Journal of Chemical Ecology 9,247-257.

58. Sanchez, N.E. and Onsager, J.E. (1988) Life history parameters in Melanoplus sanguinipes (F.) in two crested wheatgrass pastures, Canadian Entomologist 120, 39-44.

59. Smith, C.M. {I 989) Plant Resistance to Insects: a Fundamental Approach, John Wiley & Sons, New York. 60. Smith, D.S. (1959) Utilization offood plants by the migratory grasshopper Melanoplus hilituratus (Walker)

(Orthoptera: Acrididae), with some observations on the nutritional value of the plants, Annals of the Entomological Society of America 52, 674-680.

61. Smith, D.S., Handford, R.H., and Chefurka, W. (1952) Some effects of various food plants on Melanoplus mexicanus (Sauss.) (Orthoptera:Acrididae), Canadian Entomologist 84, 113-117.

62. Smith, R.W. (1939) Grasshopper injury in relation to stem rust in spring wheat varieties, Journal of the American Society of Agronomy 31, 818-821.

63. Tyler, J.M., Webster,1.A., and Merkle, O.G. (1987) Designations for genes in wheat germplasm conferring greenbug resistance, Crop Science 27, 526-527.

64. Webster, J.A., Smith, D.H., and Hoxie, R.P. (1982) Effects of cereal leaf beetle on the yields of resistant and susceptible winter wheat, Crop Science 22, 836-840.

65. Wescott, N.D., Hinks, C.F., and Olfert, O. (1991) Dietary effects of secondary plant compounds on nymphs of Melanoplus sanguinipes (Orthoptera: Acrididae), Annals of the Entomological Society of America 85, 304-309.

66. Wiseman, B.R. (I990) Plant resistance to insects in the southeastern United States - an overview, The Florida Entomologist 73,351-358.

67. Woodhead, S. and Chapman, R.F. (1986) Insect behaviour and the chemistry of plant surface waxes, in AS. Juniper and R. Southwood, (eds.), Insects and Plant Surfaces, Edward Arnold, London, pp. 123-\36.

Part 2. GRASSHOPPER POPULATION ECOLOGY AND MANAGEMENT

4. HOW DO SPATIAL POPULATION STRUCTURES AFFECT ACRIDID MANAGEMENT?

Abstract

M.G. SERGEEVI2, O.V. DENISOVA', and I.A. VANJKOV A' 'Department of General Biology, Novosibirsk State University 2 Pirogova St., Novosibirsk 630090 RUSSIA J Institute for Systematics and Ecology of Animals Siberian Branch of Russian Academy of Sciences I I Frunze St., Novosibirsk 630091 RUSSIA

The significance of spatial population structures is discussed relative to locust and grasshopper management. The general scheme of spatial population distribution is characterized. The spatial population structures ofthe model species (Calliptamus italicus, C. abbreviatus, Chorthippus parallelus, Ch. montanus, Ch.fallax, and Podisma pedestris) are described for temperate Eurasia. The temporal stability of many features of local colonies is shown. The boundaries between the different parts of a population system are evaluated as being very distinct. Movements of grasshoppers (including good flyers) are limited by local landscape barriers. Approaches and procedures of acridological population studies are discussed relative to management systems for pest and rare species.

1. Problems

Management oflocust and grasshopper populations is usually based on the knowledge of dynamic patterns. However, we should delineate observed differences of dynamics in different parts of a population. As we know, a population is organized in space, and the pattern of the spatial structure is usually rather complicated [1-5]. A few classic papers [6, 7] have been primarily concerned with the internal structures of local populations, their dynamics and inner organization.

Usually local populations (demes) are distributed over a species' range in concordance with natural conditions, especially the geographical landscape structure of the

71


72

Earth's surface. These populations may be connected with each other in one landscape unit or may be divided by various barriers which determine the level of gene flow. A barrier is considered to designate something which prevents the spread of animals or plants. We can describe this general pattern as the spatial population structure of a species [8]. In many theoretical studies, this population structure is termed a metapopulation [9, 10]. So the spatial population structure is a system of all local populations in the limits of a species' range (or a species' area). This approach is expected to be very valuable for understanding general ecology and biogeography of locusts and grasshoppers and for developing management strategies.

The environmental factors and the biological properties of a species are variable in time and space, so that their relations can be understood only by systematic studies of populations throughout a species' range [7]. We know of some fine examples of studies of spatial distribution of individuals, hatching sites, swarms etc. However, nothing is known of the spatial population structures for the majority of species, especially in connection with population dynamics. To solve this problem we must understand where populations are situated and how they change in space and time. When we eventually understand the exact patterns of spatial population structures, we will be able to forecast dynamic patterns in each local population, to manage each local population, and to manage migrations between populations. While this will be hard to do for tropical locusts, the approach will provide a real opportunity for managing temperate acridid species. In the future, the studies of spatial population structures will allow us to manage some local populations or parts of population systems without serious risk to the environment.

Spatial population structures can be described in a system of geographic gradients [II, 12]: (1) The latitudinal (or zonal) gradient is mainly determined by solar radiation distribution. (2) The longitudinal (or sectoral) gradient is chiefly associated with precipitation distribution, temperature amplitudes, and climatic severity which increases from coastal to central parts of continents. (3) The vertical gradient is determined by both altitudes (elevations) and gravitational flows of energy and matter. As a result, it should be related to not only mountain territories but also to each landscape system composed of connected landscapes (e.g., an individual catena sensu lato) or other such units.

Spatial population structures can be studied and estimated at different spatial scales [1, 13]. (1) At a range level, or at a broad scale, we investigate these structures as a system oflocal populations throughout a species' range. We can evaluate general spatial patterns of a species' distribution relative to geographic patterns (life zone and continental sector distributions, etc.) and identify barriers that divide regional parts of a species. (2) Regional (meta)populations are parts of a species and are limited by faunistic or geographic boundaries. At a regional level, or at a medium scale, different spatial structures can be described inside each region inhabited by a species that are more or less uniform from faunistic or geographic points of view. (3) Local populations are parts of regional metapopulations and are found within the basin limits of small rivers, the landscapes of which are uniform or closely connected by energy and matter flows. As we know, at the local level, identifying paths of wandering and potential contact between populations is possible [4, 13]. The knowledge of grasshopper distribution within such small regions allows for an understanding of the pattern of populations, community organization, and

73

dynamics. The main goals of this chapter are: (1) to characterize general patterns of spatial

population structures of locusts and grasshoppers, (2) to show some applications of this approach by exemplifying some model species distributions, and (3) to describe the relationships of this approach to management of locust and grasshopper populations.

2. Tools

2.1. METHODS AND MATERIALS

This study is based on quantitative and qualitative data collected on regularly distributed survey areas, as well as on the analysis of earlier published works (basically for the last two decades) and specially developed range maps.

Field sampling was organized along local transects [11]. Each local transect crossed a river valley or a lake basin from a local flood plain to a watershed plain or a montane slope, in cases where local watershed plains were almost absent or fully cultivated (see Fig. 1). Local transects were distributed according to their real positions inside three geographic gradients (zonal, sectoral, and vertical). The length of local transects varied from hundreds of meters to several kilometers. Essentially it consisted of local transect sets, extended to several hundreds of kilometers. Local transects were embedded in the long (up to several hundreds or thousands of kilometers) gradient transects (e.g., transzonal).

Grass landscapes and similar anthropogenic habitats along local transects were studied mainly in connection with specificity of temperate grasshoppers and locusts [2, 3, 11, 12]. Each part of local transects (low and upper flood plains, low and upper terraces, watershed plain) usually was investigated separately. Samples were also collected for different variants and units of the natural and anthropogenic landscapes. A sample area and time could vary in different landscape units and were limited by the landscape unit area.

Insects were caught by a standard net for a fixed period of time [II, 14], and the results were extrapolated to an hour. Orthopteran density was estimated from a set of plots [15, 16]; in our case usually from an area of25 square meters. Our data show that these methods allow us to obtain stable results suitable for long-term observations and geographical studies. These methods allowed us to obtain repeatable results over a number of years and permitted precise, fme-scale definition of distributions at the popUlation level.

The materials were collected from 1976 to 1999 in South Siberia, Kazakhstan, Kyrgyzstan, Uzbekistan, Turkmenistan, and Tajikistan. The data archived in the Department of General Biology, Novosibirsk State University, (1972-1981) were also analyzed for Siberia and Central Asia.

Three groups of relatively abundant species have been chosen to illustrate the utility of this approach. All these model groups include common species of the Palaearctic. The first group consists of the Italian locust, Calliptamus italic us (L.), and its nearest relative C. abbreviatus Ikonn. of the subfamily Calliptaminae (or the tribe Calliptamini). These species are mainly connected with the semi-desert habitats of the Mediterranean and

74

Central Asia. The second set of species includes the group Charthippus para//e/us of Gomphocerinae (Gomphocerini) - Ch. paral/elus (Zett.), Ch. mantanus (Charp.), and Ch. tal/ax (Zub.). These species are associated with the typical meadows, meadow-steppes and dry steppes of Eurasia. The widely distributed melanopline grasshopper Padisma pedestris (L.) belongs to the third group. In spite of its huge range this species is distributed in many insular populations where its abundance can be high.

C. italic us is usually a good flyer and can cover relatively long distances (up to few hundreds of kilometers). Almost all other species of Calliptamus are brachypterous, although macropterous forms are sometimes abundant. Ch. mantanus has both meso- and macropterous forms. However, migration by these species is strongly limited.

2.2. SPATIAL POPULATION STRUCTURES: GENERAL APPROACHES

The traditional approach - the analysis of spatial distributions, phylogenetic and genetic relationships between local populations inside a species range in two-dimensional geographical space [17] - is not sufficient to understand the ecology of many animals including locusts and grasshoppers. The inclusion of the vertical dimension is necessary [12, 18, 19]. The vertical axis is associated with not only general elevation but also vertical landscape structures determined by gravitational flows of energy and matter. These landscape structures can be observed as a system of slopes, terraces, depressions, etc. Such a three-dimensional approach is associated with geographic changes in habitats [12, 18] and can be expressed by vertical distributions of populations on maps of different scales.

Some earlier two-dimensional maps (especially for the breeding areas ofthe Desert locust (Schistacerca gregaria Forsk.) [20, 21] and the Moroccan locust [Daciastaurus maraccanus (Thnb.)] [22] are very interesting and important for understanding spatial population structures. They permit us to divide the species range into many zones with different types of population dynamics. However, in many cases, they can be interpreted as three-dimensional maps because the text description oflandscape features is added. For example, Adamovic [22] described the apparent restriction of the solitarious form ofthe Moroccan locust to small spots with optimal soils and vegetation.

Previously we described three main types of species colonies on plains [3, 12]: (1) The watershed populations are more or less diffusive, i. e., they occupy all

available habitats, and there are no evident barriers between local populations (Fig. 1, WP); (2) The valley populations occur on flood plains and low (moist) terraces (Fig. 1,

FL, FU, TL). As a rule, such populations may be characterized as insular or linear. The insular populations are distributed as islands with significant uninhabited areas between them. The linear pattern resembles a band associated with a linear landscape unit. Populations of this type are usually isolated from their neighbours.

(3) The slope (terrace) populations are distributed over upper and middle terraces of rivers and lakes (Fig. 1, TU).

Insular high-montane populations can be observed at high altitudes in montane regions.

As a rule, these types form a limited number of combinations in every region or locality. We described four combinations for plains [3, 11, 12]:

Ncn

h N

:;;rt

t:

-4.

C?t

i"pt>

3f)1

U,$

.1:

::ffc

u,)

e. f.:

ttfffp

frm1{

J,$

ab

hm

viM

us

L,~~

i<)\

~,<

!€!ve

f c:

4~tm

ndi.

mce,

1· i

s i

~~::i.

~(rj(B{

tW'I1

mediu

(}) h::O

'Y!>'l 0

-1 ab

:..in

ctao

c<';l,

2!)

~S9

il':<:

t p

N r·

Q;.;

::"

IIII

~~~~

. ;$

\.~l

Oi

«r;I.

::'1d&

O"1

l'':i3

. fn

oN

th.x

f: t~9

i{f{,.L

p6

r

c:J\":

(l''':

:!~i:'

ek''''

d !<

~bi~

t5

"'S:::

l:3<:,

t~,<,v

s::

:~~~

~dii

\! th

row

gh $

i.{:ih

(Qrx

?Q(::

1ic'

~~pi

tal$

"':J

$pt:

(~;:

e£ M

niM

l(;;

:

c. cn

odN

t}pu

S {J

3Fl?

JiieJ

us

CfC

1b

w fl

ood

pi~i

r:5

rl1Q

uppe

::' n

ood

rf»':!1

i!

[IT

] [ow

let(;

;}¢i:'

·~

L1

:Q]w

pp

er

lerr;

&r;(l

!<t

[~w~lershd p

Lain

t

0. C

t;ottl

!(&P

(JS

man

/am

.}::;

,0,

P<:,

-d;.

'W?)

~1 j'

.'f!:<

}f)::;

fr;.>

.;

L-::!:::

: .......

. i0p~

i:"t

)l;~

n pC

3":-;.i

lio:'1

Fig

ure

1. P

opul

atio

n di

stri

buti

on o

f m

odel

ort

hopt

eran

spe

cies

alo

ng n

orth

-sou

th t

rans

ects

fro

m W

est

Sib

eria

to

Taj

ikis

tan

(A,

L-E

) an

d tr

om m

e L

emra

l ,,!

Den

a lU

l

uva

(1:1

)

-.J

VI

76

1 - The optimal or main part is distributed over all available habitats at a high (but not extremely high) and comparatively stable level of abundance (the optimum of a range) (Fig. 1, black arrows).

2 - The transitional part is associated with the beginning of population dissociation (bifurcation) into the watershed and valley population subsystems. Populations are very rare or almost absent on the terraces.

3 - The basic part is found over watershed plains, flood plains and/or low terraces (sometimes over watersheds only); species abundance can be high locally. These areas often occupy almost all of the range.

4 - The marginal part is characterized by connection of populations with flood plains and low terraces; the populations are insular or linear.

The high-montane-valley part should be added to montane areas. There are systems of insular, linear and linear-insular local populations.

The analysis at the regional and local levels usually shows that every part of the range (especially the basic one) is a mosaic, where areas of different types of population distribution are interspersed throughout the dominating background. This is especially important for re-assessment of different ecogeographic barriers which constitute the limits for species spreading and migration. For instance, Schennum and Willey [23] have emphasized that the montane populations show a much higher degree of differentiation and discordance, and Lake Michigan acts as a natural barrier for gene flow [24]. In some cases, human activity sharply limits the spatial distribution of grasshopper populations. Conversely, new corridors for migration and colonization of previously uninhabited habitats can be established [1, 5, 25, 26 et al.].

2.3. TEMPORAL STABILITY OF SPATIAL STRUCTURES

Quantitative and qualitative samples of grasshoppers taken through a number of years in the comparatively uniform regions are highly correlated with each other and often have no meaningful differences [3,4]. This seems to be consistent for temperate grasshoppers and solitarious populations of locusts [7,22]. Our studies (unpublished data) in the Central Altai Mts. have shown that population distributions and frequencies of different variants of polymorphic traits exhibit some fluctuations but stay relatively stable in most cases.

Quantitative accounts on a constant set of habitats in the vicinity of Novosibirsk were conducted to study the long-term stability of abundance levels and popUlation distributions in 1981-1999 [27; unpublished data]. An analysis of the long-term data confirmed that the landscape distribution of Ch. parallelus did not change significantly [3]. In weakly disturbed habitats, its abundance varies, but maintains approximately the same level. As a whole, Ch. parallelus demonstrates constant preference for the steppe meadows of watershed plains. Therefore, in spite of certain variation, we contend that landscape distribution of populations and their parameters are comparatively stable for temperate grasshoppers.

77

2.4. SPECIES CO-EXISTENCE

Co-existence of species in every region is sustained by the differentiation of so-called "place niches" [28]; this co-existence is supposed to be a result of specific habitat elimination [29] and different species connections with succession processes [30]. The problem of competition at different levels is not solved to date, but in most cases the role of competition is not too significant [31]. Mulkern [32] analyzed multidimensional overlapping in resource utilization by prairie grasshoppers and found that there was little overlap among co-existing species.

The comparison oflocal population distributions of orthopteran species in the small river basins within the main biogeographical regions allows us to describe three main trends of grasshopper co-existence for temperate Eurasia [13, 33]:

(1) The majority of species inhabiting small river basins have strongly localized populations, and the density of each population is usually low. Such situations occur in the Far East and the deserts of Central Asia. There are a few species with widely distributed populations.

(2) Most species are distributed over all available habitats of the local river basin at high levels of abundance. Such forms spread through anthropogenic landscapes. As a rule, the maximum density of one species does not coincide in space and time with the maximum of others. This situation is observed in the forest-steppes and in some parts of the mountains of South Siberia.

(3) The intermediate type includes local populations distributed over a number of habitats, which are usually connected with each other. Densities can be high to very high. In the plain steppes, such species prefer local watersheds. Both in the mountains of South Siberia and in the northern part of the Central Asian mountains, these species are mainly associated with local southern slopes. On the contrary, in the southern part of these mountains, such orthopterans are linked with northern slopes.

2.5. SPATIAL POPULATION STRUCTURES OF THE MODEL SPECIES

To simplify discussion we compare spatial popUlation structures of the model species only along a main transzonal transect which crosses the internal part of Eurasia from the West Siberian forests to the southern desert of Tajikistan and Turkmenistan (Fig. 1, A, C-E) [3, 5]. This transect crosses almost all ranges of the model species from north to south. It consists oflocal transect sets representing vertical landscape organization. Their positions are determined by distributions of life zones and their subzonal parts. Therefore, we can characterize the spatial population distribution along this main transect as the zonallandscape form.

2.5.1. Calliptamus italic us The Italian locust is a common species and the most important pest in the Mediterranean and Saharan-Gobian regions. Many of its outbreaks were described for these areas and for the neighbouring steppes and mountains.

The Italian locust is an intermediate form between typical gregarious and solitarious

78

acridid species [7]. For its gregarious form distinctive migrations are described over comparatively small distances of about 100-200 kIn [34].

Ecological and geographical particularities of this species are described in many publications and for different parts of its range. Analysis of the main part of these studies is given by Uvarov [7], who emphasized the insufficiency of our knowledge of the Italian locust during periods between outbreaks. For the last two decades, more than 20 studies of this species were published but nearly all of these are limited to outbreak periods.

General particularities of spatial distribution. The main part of the species' range is located in the semi-desert zone where many outbreaks have been often observed and as a rule, abundance is high [5, 7]. This pattern coincides with the distribution of plants preferred by this species - xeromorphic dicotyledons like sagebrush [7, 37], as well as with the availability of the terrain used for egg laying (nearly bare, often sandy, soil) [7]. Based on available data, the Italian Locust appears to prefer landscapes with a mosaic of xeric plant cover. It is interesting to note that in spite of the overall abundance of the Italian Locust, its several restricted populations in Western Europe are critically endangered or even likely to go extinct [35, 36].

Zonal-landscape distribution over the studied transect. On the main transect, the Italian locust distinctly prefers the semi-desert of the Kazakh Uplands, or so-called Sary-Arka (Fig. 1, A). Nearly constant high abundance (observed in 1975, 1976, 1986, 1991) and distribution in all available habitats suggest that this region is optimal for the species. This area covers the semi-desert of East Kazakhstan and confirm earlier results [2,38].

In the northern part of the transect, the Italian locust is actually distributed more widely than was indicated by Stebaev [38]. Here, it inhabits the forest-steppes on overgrazed plain pastures, but usually occurs very locally and at low levels of abundance. In the steppes, its popUlations occupy not only local watershed plains, but the dry parts of upper flood plains and lower terraces, and the stony southern slopes of hills. In this area, another optimal habitat of this species occurs in the dry steppes of Kulunda [2]. This habitat partly covers a zone of the dry pine forests of the Irtysh region, and outbreaks often occur here [34, 39].

The most recent C. italicus outbreak started in this optimal habitat in the summer of 1999. (However, in neighbouring Pavlodar Region of Kazakhstan, the outbreak began in 1997). Densities of the Italian locust increased sharply after a warm and dry May, especially in fallow fields. Its populations occupied almost all suitable habitats including openings of pine forests, fallow agricultural fields, stony slopes and wet meadows at comparatively high levels of abundance (1-5 individuals per square meter). Some agricultural fields (mainly sunflowers) were damaged. However, our data show that all studied populations are intermediate between the typical solitarious and gregarious populations.

The presence of two optima near the east boundaries of Kazakhstan (main and additional) corresponds to the situation described by Stolyarov [40] for West Kazakhstan. Both optima seem to stretch east-west through the whole eastern part of the locust's range.

In the southern part of the transect, populations of the Italian locust once again

79

become more local. In the northern deserts, they are distributed on plain watersheds (Fig. I, A), and their local density can be rather high. The locust's abundance decreases to the south where its main populations are associated with river valleys and lake basins. Here the Italian locust often inhabits not only sagebrush vegetation in the steppe habitats but also meadows. In this part of its range, the Italian locust is widely distributed through irrigated fields, including alfalfa, and canal borders.

Spatial population distribution of the Italian locust at periods between outbreaks both within the whole area, and on along the model transect, generally corresponds to the rule of habitat change [18]. In the northern part, the species is distributed over very dry habitats; in the central zone it prefers medium to xeric and varied habitats of the steppe and semi-desert; in the southern part, its habitats are usually localized in meso-hygrophylous areas of river valleys or in the mountains.

If the distribution ofthe Italian locust along the model transect is compared with the distribution of relatively well-studied grasshoppers such as the white-striped grasshopper [Ch. albomarginatus (Deg.)] [18] and Ch. parallelus [3], it becomes absolutely clear that its distribution is very different from them [3, 12]. In the first place the Italian locust dominates in the additional (steppe) optimum of the population distribution. Besides that, the Italian locust's range includes a comparatively small mountain area, where conditions for its existence are nearly optimal [5].

It is important to note that popUlation dynamics within main and additional optima differ significantly, because they are situated in different climatic zones. In even greater degree, this is true for the marginal areas of high abundance and probably for the populations occupying different ecosystems within a small region (e.g., flood plain and watershed plain). For forecasting of the Italian locust (and other locusts) it is necessary to locate permanent survey sites in different parts of its range, including both natural and anthropogenic landscapes.

2.5.2. Calliptamus abbreviatus C. abbreviatus is another species of this genus in Central and East Asia [41]. Its abundance can be rather high sometimes even reaching pest levels [42]. C. abbreviatus belongs to the north temperate group of the genus Calliptamus, more adapted to humid and cool boreal and sub-boreal conditions, but its isolation from C. italic us has occurred due to isolation in the Far East and Central Asia [43].

General particularities of spatial distribution. As with the Italian locust, C. abbreviatus prefers the dry steppes and semi-deserts with a mosaic of xeromorphic vegetation and stony soils [11,41]. Both species are distinguished from the nearest relatives - WestMediterranean C. wattenwylianus (Pantel) and Irano-Turanian C. turanicus Sergo Tarb. The first prefers the typical, often disturbed, Mediterranean habitats [43], and the second is mainly associated with the deserts of piedmont plains. C. abbreviatus is obviously adapted to conditions of the Mongolian area with maximum precipitation and corresponding growth of plants occuring in late summer. This pattern corresponds to the distribution of of its preferred plants such as xeromorphic dicotyledons (sagebrush) [37].

80

Zonal-landscape distribution over the studied transect. Because this species is distributed in the eastern half of temperate Eurasia, it does not occur along the main transect As a result, we used another transect for the comparative analysis of its spatial distribution. This additional transect crossed the southern part of Krasnoyarsk Region, Khakasia and Tuva Republics (Fig. I, B).

Here C. abbreviatus distinctly prefers the semi-deserts ofthe Mongolian type (Fig. 1, B). In these semi-deserts, it is distributed through all available habitats, from flood plains to watershed plains and southern slopes. Comparatively high levels of abundance (in the second half of summer up to 300-500 individuals per hour of collecting) and distribution in all suitable habitats indicate that this region is optimal for the species.

In the northern part of the transect, in the dry steppes, C. abbreviatus is found in nearly all primary habitats, but its density decreases. However, in the montane steppes and forest-steppes, C. abbreviatus is generally absent even on the driest southern slopes. It has rare populations on the plains and arid slopes in the low part of the montane steppe belt (usually not above 1500 m), as well as in the semi-desert of the central part of Tuva. C.abbreviatus is not found northward.

In the southern part of the transect, its populations again become very local. In the most dry variants of the semi-desert and in the deserts, they are on the watershed plains (Fig. I, B) and southern slopes.

In the Far East, C. abbreviatus is usually found in dry regions [41]. In contrast with the South-Siberian transect, its abundance here is usually low; popUlations are. local and associated with xeric habitats, from the stony flood plains to the southern slopes. The broad amplitude of its distribution is observed in the hill area near Khanko Lake. In the northern forest-steppes of the Amur Region and in the Daurian steppes, C. abbreviatus occurs in xeric plain habitats and southern slopes.

In the forest regions of the Far East, in spite of wide distribution of some xeric habitats, particularly southern slopes, C. abbreviatus is a rare species. Only in the most southern part of the Russian Far East, is it occasionally found on dry south declivities and on the stony flood plains.

The distribution pattern of C. abbreviatus is similar to that of the Italian locust. However, careful analysis reveals some essential differences: (l) C. abbreviatus populations are distributed more locally and gravitate almost exclusively to xeric habitats with a mosaic vegetation cover and stones; (2) Unlike the other species of the genus, C. abbreviatus does not inhabit anthropogenic habitats. Moreover, its migratory potential is probably very limited even at the intralandscape level.

As a result, it appears that in spite of the extensive range and rather high densities in some habitats, the general pattern of its distribution is similar to many endemic species of grasshoppers with their highly localized and usually fragmented populations [I]. This makes C. abbreviatus a perfect species for testing the models developed in modem conservation biology relative to endemic species with highly localized populations. This type of distribution also helps to better understand the reasons for the absence of C. abbreviatus's outbreaks, since typically, outbreaks are characteristic of locusts and grasshoppers with well integrated population systems. However, in the long term, changes in the ecological and climatic setting (including the anthropogenic influence) will result in

81

creating the conditions favourable for increasing C. abbreviatus abundance.

2.5.3. Chorthippus parallelus The spatial population structure of Ch.parallelus was described by Kazakova and Sergeev [3]. The species has a classic pattern of a population distribution with an optimum in the northern steppes [12], making it suitable for spatial extrapolations.

General particularities of spatial distribution. The species is common and abundant in meadow and meadow-steppe habitats of the temperate Palaearctic. Rare cases of migrations ofCh. parallelus are mostly of an intrapopulation nature [44]; macropterous specimens ofthis species can not usually fly [45]. Our experimental data for the West Siberian forest-steppes also show that brachypterous (normal) specimens of Ch. parallelus do not cross forest-bush belts 10-15 m in width.

Zonal-landscape distribution over the studied transect. The northern part of the main transect is occupied mostly by scattered populations of this species associated with watershed meadows (Fig. 1, C) and flood plains. Its local abundance can be very high. Further south, in the south forest -steppe and particularly in the steppe life zone, abundance of this species is usually high in all meadow and meadow-steppe habitats. Therefore, this is an optimal area for this species.

In the southern part of the studied transect, the species is found almost exclusively in river valleys, near lakes, and in the mountains. Its populations are distributed sporadically and occupy grass meadows on flood plains and low terraces. The species' abundance is usually low but can be very high in some habitats. This means that contact between these local populations is limited.

For the plains area as a whole, a border between the steppe and the semi-desert parts of the range (i.e., the border between the natural life zones associated with significant decreases of precipitation) is the most significant barrier for dividing spatial population structures. As follows from landscape distribution of populations, this transition corresponds to the change from diffusive and nearly diffusive distribution of Ch. parallelus to the insular distribution.

The distribution pattern of this species is suitable for spatial extrapolations. However, genetic and ecological differences between its local populations can be very significant. This means that we should study every local population of the species for exact temporal forecasting, but in practice this is impossible. Therefore, the compromise would be to study a wide set of populations in different habitats and to evaluate main types of dynamics.

2.5.4. Chorthippus montanus This species inhabits the northern (boreal and subboreal) part of temperate Eurasia [11]. It prefers wet meadows with grasses.

Our data show that its optimal area lies near the northern part of the model transect (i.e., in the boreal forests). The southern portion (Fig. 1, D) lies in the south of the forest zone. Some insular populations of the species are situated in the forest-steppes and the

82

steppes. They are very local, and species abundance is usually very low. As a result, populations of the species are sporadic and migrations between them are limited.

This pattern coincides with the species' distribution in the Russian Far East [46] where Ch. montanus prefers meadows of the south forests. In the mountains of South Siberia this species is mesohygrophilous [33], and has an insular population structure.

The general distribution of Ch. montanus differs from the distribution of similar species (Fig. 1, C-D). This is a result of ecological and evolutionary differences, probably reflecting the species' origination in the forest life zone.

2.5.5. Chorthippusfal/ax Ch. fa/lax is a typical East -Palaearctic species [11]. Its range spreads westward over xeric, anthropogenic habitats, like overgrazed rangeland and road sides. It does not reach the main transect. However, in the near future, this species will probably be able to use disturbed habitats of the local watershed plains and upper terraces to spread westward in the steppe zone (Fig. 1, D, pointed arrows).

In the Kulunda steppe, Ch.fallax is distinctly associated with dry plots of watershed plains, terraces and flood plains. As a rule, its populations are highly localized and occur in disturbed habitats. Its abundance can fluctuate significantly.

The optimum zone of its range lies in the steppe regions of Mongolia and the South Siberian Mountains. Here the species prefers the steppe and dry meadow habitats [33]. In the dry steppes and semi-deserts, its populations are found in river valleys, from flood plains to middle terraces. However, in the Far East, its populations are distributed over mesic habitats (meadows, slopes with openings etc.) [46].

Therefore, three close species of the Ch. parallelus group differ significantly in their ecological preferences. Ch. parallelus is associated with the central part of Eurasia and prefers meadows and meadow steppes. Ch. montanus occupies the boreal area and prefers wet meadows. Ch. fal/ax inhabits the eastern part of temperate Eurasia and uses dry meadows and steppes. Their population distributions are usually limited by a consistent set of conditions. As a result, they occur very rarely in the same habitats at similar levels of abundance.

2.5.6. Podisma pedestris This species is widely distributed in the Palaearctic. It occurs from the mountains of South Europe to East Siberia. P. pedestris usually prefers meadows and openings of pine forests.

This species is found in the northern part of the model transect (Fig. I, E). Here its populations are extremely sporadic and occur in meadow openings and forest edges on watershed plains and upper terraces. In the typical steppes, P. pedestris is observed in meadow openings of the riverine forests. Its abundance is usually very low.

However, in some areas (e.g., in the Central Altai Mts., in some years, the abundance of this species can be extremely high [33]. As a result, local populations of P. pedestris can colonize new habitats and areas. Here P. pedestris uses anthropogenic habitats like road sides with ruderal vegetation, and cropland. In contrast, human activities in the vicinity of Novosibirsk have resulted in almost complete extirpation of local populations inhabiting the clearings of the pine forests.

83

2.6. PHENOTYPIC TRAITS IN SPATIAL POPULATION ANALYSIS

The analysis of species distribution allows us to distinguish parts of a species range and its population system. However, this approach is not sufficient, especially at the local level. Completely isolated populations are exceptional [7] and usually restricted to non-flying montane or island forms in specific natural conditions. Continuous populations with uniform structure are very rare too. Therefore, it may be difficult to differentiate uniform local populations solely on the basis of distribution patterns.

Investigations of frequency distribution of colour and other phenotypic traits can show more complicated relations between the different parts of population systems. The investigation of colour morph frequencies allows us to separate some parts of population systems, which can not described by conventional methods [3].

For instance, the spatial distribution of Ch. parallelus in the optimal area shows that there are no obstacles for migration and gene flow. However, phenotype frequencies among valley and watershed populations demonstrate their separate status [3]. Similar local studies of good flyers, [e.g., Bryodema tuberculatum (L.)], show that even a small montane river (10-15 m wide) can be a serious obstacle to migration [4].

3. Solutions

Local populations of each species can be significantly different in ecological, geographical and, probably, genetic characteristics. This is usually the case for neighbouring populations, which may require different management procedures. This means that we must evaluate these differences and these for population management.

Unfortunately, related species oflocusts and grasshoppers can have different spatial population organizations. Sometimes they are similar in general distribution pattern (e.g., C. italicus and C. abbreviatus)(Fig. I, A-B), butthey are significantly different at regional and local levels.

This is important both for potential pest species and for rare grasshoppers and other orthopterans. Consequently, species with spatially discrete population systems are not suitable subjects for forecasting and modelling. Probably this is also true for typical locusts, such as the Migratory (Locusta migratoria L.) and the Desert locusts.

Local populations of rare species may be very different. Some of them are under the threat of extinction or critically endangered [26, 47]. Few populations can be stable for a long time at low levels of abundance [47, 48]. However, some popUlations of rare grasshoppers can be abundant for a long time in avery limited space. It is interesting that some local popUlations of pest species (e.g., P. pedestris or especially C. italicus) may be almost extinct, while others are thriving.

All this means that if we want to create a productive management system we must study dynamic patterns of the spatial popUlation structures of each species across a range of scales from the general to the local (Fig. 2). These studies should include three main levels: general (the whole species' range), regional and local. At the general level of studies, a species range and its boundaries can be explored relative to the ecological limits

84

General spatial studies of a species

[Range and its boundaries

I Ecological limits

I General dynamics

... Regional studies

Spatial structures

I (landscape distribution, barriers, ways for migrations) i

[ Regional dynamics I I I I

Genetic and phenotypic

I ~. trait distribution

I \\ Local studies

Local spatial structures (patches, corridors, distances, habitat connectivity) and migrations

A (emigration/immigration)

Species \. I Dynamics (seasonal and long-co-existence "

\ I term)

'. I

\ \ Genetic structures \

\ \ .. .

Population management strategies

Rare species

.. ... Conservation strategies

Rare/pest species

• Pest species

Forecasting systems

~

I

I

Figure 2. Procedure of spatial studies of locust and grasshopper populations for management systems

85

and general dynamic patterns. These data may be used to establish the main area for regional studies with approximately similar conditions (usually geographic regions and life zones).

Regional studies should include research of spatial population structures, evaluating barriers, and paths for migration, long-term dynamics, and genetic and phenotypic trait distributions. Such studies should result in the choosing of model plots for local investigations.

At the local level of study, we should assess spatial structures (patches and gaps, distances, population connections, barriers and corridors) and local migration systems (parameters of emigration and immigration) associated with a landscape matrix. We should monitor seasonal and long-term dynamics oflocal populations and try to evaluate their temporal patterns relative to climatic and ecosystem changes and to peculiarities of each population. At this level, some special genetic investigations should be undertaken and characters of gene flow should be estimated. Generally, a sound basis for population management strategies can be created as a result of regional and local studies.

Population management strategies are manifest as three main types (Fig. 2). The fIrst strategy is for rare forms. As a result, conservation strategies can vary in different parts of a range. The second one is for typical pest species. The results of spatial and temporal research at different scales are important for forecasting, and spatial patterns are very important in optimizing the use of insecticides, especially fogger treatments. In this case, the optimal results of insecticide applications are only possible if control programs are strongly associated with spatial heterogeneity of populations [49, 50]. The third strategy is for forms with populations distinguished by their status in different areas. In some territories, such species are subjects for conservation measures and in others their dynamics and distribution must be managed to prevent damage.

4. Acknowledgements

We wish to express our sincere thanks to Professor A. Joern and Dr. I. Sokolov for valuable comments. We thank the Russian State Programme "Universities of Russia" (grant 1759), the Russian State Programme "Biological Diversity", the Russian Foundation for Basic Research (grant 97-49399), and the Federal Programme "Integration" (projects 274 and 275) for vital fmancial support.

5. References

l. Sergeev, M.G. (1997) Ecogeographical distribution ofOrthoptera, in S.K. Gangwere, M.C. Muralirangan and M. Muralirangan (eds.), The Bionomic of Grasshoppers, Katydids and their Kin, CAB International, Oxon and New York, pp. 129-146.

2. Sergeev, M.G. (1997) Metapopulations of locusts and grasshoppers: spatial structures, their dynamics and early warning systems, in S. Krall, R. Peveling and D. Ba Diallo (eds.), New Strategies in Locust Control, Birkhauser Verlag, Bazel et aI., pp. 75-80.

86

3. Kazakova, LG. and Sergeev, M.G. (1992) Spatial organization of species populations system of short-homed grasshopper Chorthippus paral/elus, Zhournal Obstshej Biologii 53, 373-383. (In Russian with English summary)

4. Kazalcova, LG. and Sergeev, M.G. (1997) About determination problem of grasshopper population boundaries, Sibirskii Ecologicheskii Zhournal4, 315-321. (In Russian)

5. Sergeev, M.G. and Vanjkova, LA. (1996) Zonal-landscape distribution of populations of the Italian Locust Calliptamus italicus L. (Insecta, Orthoptera, Acrididae), Sibirskii Ecologicheskii Zhournal3, 219-225. (In Russian)

6. Richards, O.W. and Waloff, N. (1954) Studies on the ecology and popUlation dynamics of British grasshoppers, Anti-Locust Bulletin 17, 1-182.

7. Uvarov, B.P. (1977) Grasshoppers and Locusts. A handbook of general acridology. Vol. 2, Centre for Overseas Pest Research. London.

8. Shilov, LA. (1977) Ecological and physiological base of population relations of animals, Moscow University Press, Moscow. (In Russian)

9. Levins, R. (1970) Extinction, in M. Gustenhaver (ed.), Some Mathematical Question in Biology, Vol. 2, Providence, Rhode Island, pp. 77-107.

10. Hanski, I. (1998) Metapopulation dynamics, Nature 396, 41-49. II. Sergeev, M.G. (1986) Patterns of Orthoptera distribution in North Asia, Nauka Publ., Novosibirsk. (In

Russian). 12. Stebaev, LV., Sergeev, M.G. (1982) The intemallandscape-population structure of area, as exemplified by

Acrididae, Zhournal Obstshej Biologii 43, 399-410. (In Russian with English summary) 13. Sergeev, M.G. Biological diversity ofOrthoptera in the North-East Palaearctic: Population distribution,

Sibirskii Ecologicheskii Zhournall, 547-554. (In RUSSian) 14. Gause, G.F. (1930) Studies on the ecology ofthe Orthoptera, Ecology 11, 307-325. 15. Riegert, P.M. (1968) A history of grasshoper abundance surveys and forecast of outbreaks in Saskatchewan,

Memoir of Entomological Society of Canada 52, 3-99. 16. Onsager, 1.A. (1977) Comparison of five methods for estimating density of rangeland grasshoppers, J. of

Economic Entomology 70, 187-190. 17. Tupikova, N.V. (1969) Zoological mapping, Moscow University Press, Moscow. (In Russian) 18. Bey-Bienko, GJ. (1930) The zonal and ecological distribution of Acrididae in West Siberian and Zaisan

Plains, Bulleten Zaststshity Rastenii, Entomologia 1(1), 51-90. (In Russian with English summary) 19. Zenkewitch, L. and Brotzky, V. (1939) Ecological depth-temperature areas of benthos mass-forms of the

Barents-sea, Ecology 20, 569-576. 20. Davies, D.E. (1952) Seasonal breeding and migrations of the Desert Locust (Schistocerca gregaria ForskM)

in North-Eastern Africa and the Middle East, Anti-Locust Memoir 4, I-57. 21. Fourtescue-Foulkes, J. (1953) Seasonal breeding and migrations ofthe Desert Locust (Schistocerca gregaria

ForskM) in South-Western Asia, Anti-Locust Memoir 5, 1-36. 22. Adamovic, Z.R. (1959) The Moroccan Locust (Dociostaurus maroccanus Thunberg) in North Banat, Serbia,

Bull. Mus. Hist. Nat. Belgrade, Serie B 13, 1-123. 23. Schennum, W.E., Willey, R.B. (1979) A geographical analysis of quantitative morphological variation in

the grasshopper Arphia conspersa, Evolution 33, 64-84. 24. Willey, R.B. (1987) Pattern of gene flow disruption in the evolution of the genus Arphia (Acrididae:

Oedipodinae), in B. Baccetti (ed.), Evolutionary Biology of Orthopteroid Insects, Ellis Horwood Ltd, Chichester, pp. 592-600.

25. Samways, M.J. and Sergeev, M.G. (1997) Orthoptera and landscape change, in S.K. Gangwere, M.C. Muralirangan and M. Muralirangan (eds.), The Bionomic of Grasshoppers, Katydids and their Kin, CAB International, Oxon and New York, pp. 147-162.

26. Sergeev, M.G. (1998) Conservation of orthopteran biological diversity relative to landscape change in temperate Eurasia, J. Insect Conservation 2, 247-252.

27. Stebaeva, S.K. and Sergeev, M.G. (1995) Structure of collembolan and chortobiont communities in grass urboecosystems, Polskie Pismo Entomologiczne 64, 199-206.

28. Clarke, G.L. (1954) Elements of ecology, Wiley & Sons, New York and Chapman & Hall, London. 29. Elton, C.S. (1930) Animal ecology and evolution, Clarendon Press, Oxford. 30. Buckley, R. (1983) A possible mechanism for maintaining diversity in species-rich communities: an

addendum to Connel's hypotheses, Oikos 40, 312.

87

31. Evans, E. W. (1992) Absence of interspecific competition among tall grass prairie grasshoppers during a drought, Ecology 73, 1038-1044.

32. Mulkern, G.B. (1982) Multidimensional analysis of overlap in resource utilization by grasshoppers, Transactions of the American Entomological Society 108, 1-9.

33. Kazakova, 1.0. and Sergeev, M.O. (1992) Regularities of distribution of the Orthopteroids populations in natural and anthropogenic landscapes of mountain depressions in Southern Siberia, Entomologicheskoje Obozrenije 71, 731-741. (In Russian with English summary) (English translation - Entomological Review 72,70-80)

34. Bunin, L.D. (1979). The Italian Locust (Calliptamus italicus L.) in the soil-saving agriculture zone in East Kazakhstan and development of treatments, Ph.D. Thesis, Leningrad. (In Russian)

35. Wallaschek, M. (1992) Stand der faunistischen Erfassung der Oeradfliigler (Orthoptera s.l.) in SachsenAnhalt Articulata 7,5-18.

36. Jiirgens, K. and Rehding, O. (1992) Xerothermophile Heuschrecken (Saltatoria) im Hegau -Bestandssituation von Oedipoda germanica und Calliptamus italicus, Articulata 7, 19-38.

37. Pshenitsyna, L.B. (1987). Food preference of grasshoppers relative their pressing on the steppe ecosystems, Ph.D. Thesis, Novosibirsk. (In Russian)

38. Stebaev, LV. and Kozlovskaya, E.B. (1979) Patterns of quantitative distribution of complexes of steppic and meadow pest grasshoppers in the Irtysh region and South-East Kazakhstan relative to regionalization of their potential pest activity, in LV. Stebaev (ed.), Voprosy ecologii, Novosibirsk State University, Novosibirsk, pp. 31-51. (In Russian)

39. Bugaev, O.S. (1977) Habitat distribution of grasshoppers in the zone of band pine forests in North-East Kazakhstan, Vestnik selskokhoziastvennoi nauki Kazakhstana, N6, 37-40. (In Russian)

40. Stolyarov, M.V. (1974) The Italian Locust (Calliptamus italicus L.) in West Kazakhstan, Trudy Vsesojuznogo entomologicheskogo obstshestva 57, 98-111. (In Russian)

41. Sergeev, M.O. and Vanjkova, LA. (1999) Zonal-landscape distribution of Calliptamus abbreviatus (Orthoptera, Acrididae), Zoologicheskii Zhournal78, 31-36. (In Russian with English summary)

42. Mistshenko, L.L. (1972) Order Orthoptera (Saltatoria), O.L. Kryzhanovskii and E.M. Dantsig (eds.) Nasekomye i klestshi - vrediteli selskohozjastvennyh kultur I, Nauka Publ., Petersburg, pp. 16-115. (In Russian)

43. Jago, N.D. (1963) A revision of the genus Calliptamus Serville (Orthoptera: Acrididae), Bull. Brit. Mus. (Natur. Hist.) Entomology 13, 289-350.

44. K5hler, G. and Brodhun, H. (1987) Investigation on population dynamics in Central European grasshoppers (Orthoptera: Acrididae), Zool. Jb. Syst. N114, 157-191.

45. Ritchie, M.O., Butiin, R.K., and Hewitt, O.M. (1987) Causation, fitness effects and morphology of macropterism in Chorthippus parallelus (Orthoptera: Acrididae), Ecological Entomology 2, 209-218.

46. Stebaev, LV., Murav'eva, V.M., and Sergeev, M.O. (1988) Specificity of ecological standards of the Orthoptera in landscapes with herbaceous vegetation in the Far East, Entomologicheskoe Obozrenie 67,241-250. (In Russian with English summary) (English translation - Entomological Review 68, 1-10)

47. Ingrisch, S. and K5hler, O. (1998) Die Heuschrecken Mitteleuropas, Westarp Wissenschaften, Magdeburg. 48. Chernyakhovskiy, M.E. (in press) Can micropopulation management protect rare grasshoppers?, in J.A.

Lockwood, A.V. Latchininsky, M.G. Sergeev (eds.) Grasshoppers and Orassland Health, Kluwer Academic Publishers, Dordrecht (this volume)

49. Sobolev, N.N. and Sergeev, M.O. (1985) Population dynamics of grasshoppers in the agrocoenoses of North Kazakhstan, O.S. Zolotarenko (ed.), Antropogennye vozdejstvija na soobstshestva nasekomyh, NaukaPubl., Novosibirsk, pp. 96-104. (In Russian)

50. Sergeev, M.O., Bugrov, A.O., Kazakova, I.G. and Sobolev, N.N. (1988) Regulation of population dynamics of grasshoppers in agrocoenoses with aid of the aerosol fogger, in O.S. Zolotarenko (ed.) Landshaftnaja ecologia nasekomyh, Nauka Publ., Novosibirsk, pp. 63-69. (In Russian)

5. CAN MICROPOPULATION MANAGEMENT PROTECT RARE GRASSHOPPERS?

Abstract

M.E. CHERNY AKHOVSKIY Department o/Zoology and Ecology Moscow Pedagogical State University 6/5 Kibal'chicha St., Moscow J 29278 Russia

Ecological and biogeographic peculiarities of grasshopper micropopulations are discussed relative to conservation biology. Examples of micropopulations are described for different natural regions of temperate Eurasia. Some peculiarities of micropopulation dynamics are discussed relative to egg pod structures.

1. Problems

A micropopulation is a colony of a species characterized by following features: (1) Limited numbers of specimens (usually hundreds) or fluctuating numbers (tens

to thousands). (2) Occupation of a small area, usually not more than hundreds of square meters. (3) Evident isolation from other parts of a species population system. Distance

between area occupied and the nearest population of the species is more than distance that an individual can move during its life.

(4) Comparative stability for several years.

Small local populations (micropopulations) of grasshoppers permanently occupying small habitats and sharply limited by geographical and ecological barriers, are common in nature. These micropopulations are very interesting for conservation biology of grasshoppers because they are relatively stable at a low level of abundance. Furthermore, as a rule, there is no evident gene flow between different micropopulations of a species.

2. Tools

2.1. MICROPOPULATIONS IN DIFFERENT REGIONS

Several acridid micropopulations were studied by the author in different parts oftemperate Eurasia.

89


90

2.1.1. Arctic Tundras A plot of tundra with lichens and Rhododendron sp. was studied at South Tchukotka (North-East Siberia), near Nagomyj Settlement (63° N, 179"20' E) in 1987. Its area was 40 m2• The plot was inhabited by a small population (~ 20 individuals) of Chorthippus montanus (Charp.).

2.1.2. Boreal Forests A grass-forb meadow on the coastal escarpment, occupying an area of nearly 300 m2, was studied in 1994, 1995, and 1997 in the vicinity of Pechora City, Uleshovo Village, on the left bank of the Listvianichnaya River, Komi Republic (European Russia) (65°30' N, 57° E). The lower part of the meadow was occupied by Omocestus viridulus (L.), but the upper, drier zone was inhabited by Chorthippus biguttulus (L.) at a density of about 0.1-0.2 perm2.

2.1.3. Sub-boreal Forests In 1995-1998, I investigated a small sedge opening with an admixture of reeds and forbs in Moscow, Park "Losiny Ostrov" ("Moose Island"), on the flood plain of the Jauza River (56° N, 38° E). This 20-m2 plot was separated from the lake, the road and the willow forest by water. A micropopulation of Stethophyma grossum (L.) occurred here. Usually, 2 or 3 specimens were in sight at a time.

2.1.4. The Caucasus A solid rock with fine grain soil was discovered in 1986 in an irrigated orchard with high vegetation providing 85-90% canopy cover. This orchard was located in Megri City, Armenia (38°50' N, 46°15' E). The area studied occupied 20x70 m, from 5 to 10 min height, with declivities up to 30 degrees. The vegetation cover included sagebrush and grasses with some forbs. Plant cover did not exceed 30%. The rock was separated from the nearest mountain slope by more than 200 m. The orchard was also separated from the mountain slope by a highway 6 m in width.

The discovered assemblage consisted of8 species of grasshoppers. Ch. biguttulus (0.3-0.4 per m2), Acrotylus insubricus insubricus (Scop.) (0.6 per m2), Oedipoda caerulescens (L.) (0.4-0.45 per m2), and Calliptamus tenuicercis Tarb. (0.3 per m2) inhabited both the rock and the orchard. On the contrary, other acridids such as Oedipoda miniata miniata (Pall.) (0.2 per m2), Notostaurus anatolicus (Krauss) (0.15 per m2),

Chorthippus macrocerus macrocerus (F.d.W.) (0.3-0.4 per m2), and the pyrgomorphid grasshopper Pyrgomorpha quentheri Burr. (0.07 per m2) were not observed in the orchard. The behaviour and abundance at this assemblage was not different from the slopes of the surrounding mountains.

2.1.5. Pamiro-Allay A plot with ephemeral forbs on fine grained and stoney soils was studied at an altitude of2 500 m in Romit Canyon, 60 km northeast from Dushanbe, Gissar Range, in Tajikistan (38°40' N, 69"20' E). Studies were conducted in 1978-1980 and 1988. At this location a micropopulation of the wingless montane grasshopper Conophyma olsufjevi Mistsh. was found.

91

At this location the micropopulation of the wingless montane grasshopper Conophyma oisufjevi Mishch. was found.

More examples of such micropopulations can be reported from other geographical areas of Russia and all temperate Eurasia. For instance, Ingrisch and K()hler studied a lot of micropopulations of grasshoppers and other orthopteran insects [I]. These data show that such micropopulations are more or less common both for non-flying and flying forms, especially in the forest regions of Western Europe. Often they occur in small plots and include few individuals. In all cases, such micropopulations can be stable for rather a long time.

2.2. MICROPOPULA TION DYNAMICS AND EGG POD STRUCTURES

Studies of dynamics are essential for conservation management. This is important because abundance of isolated micropopulations can increase rapidly. Sometimes this means that a species' status may be temporarily changed from rare to pest.

2.2.1. Micropopuiation Fluctuations I studied densities of the huge, brachypterous pamphagid grasshopper Saxetania cultricollis (Sauss.) on the cobble slope with forbs near Kara-Kala Village (Garrygala on the modem maps) (Kopetdag Mts.) in Turkmenistan (38"20' N, 46°15' E). Its general area was about 1000 m2•

In spring ofl966 the species' density was 0.3 perm2• In 1968,1970-1972,1982, and 1988 the densities of S. cultricollis did not exceed 0.01-0.02 per m2 on the same and similar plots. Similar changes in abundance here were observed in 1966 for the pamphagid grasshopper Melanotmethis fuscipennis (Redt).

The second part of these studies was conducted on the cobble-stone plot in the altitudinal belt of savanna-like vegetation with trees, bushes, forbs and grasses in Romit Canyon (Gissar Range) in Tajikistan. This plot was about 200 m2• In 1977 the density of the endemic apterous species C. olsufjeviMistsh. was 0.5 perm2• Later, in 1978-1981 and 1983 on the same area abundance of C. olsufjevi was very low and did not reach 0.05-0.06 perm2•

What are the possible reasons for such sharp fluctuations of micropopulations of rare grasshoppers?

I investigated the high montane endemic Conophyma prasinum Mistsh. near Anzob Pass (39°N, 68°50' E) and in Romit Canyon in Tajikistan [2]. This species had three hatching peaks at altitudes 2 500-2 700 m and two maxima at even higher altitudes (3 300 m). Guseva [3] described a similar pattern for hatching of Ch. biguttulus in the Naurzum State Reserve (Kazakhstan). She explained this phenomenon as afunction of temperature fluctuations.

2.2.2. Egg Pod Structures and Dynamics These patterns can be explained by egg pod structures. There are two different types of egg pod structures: (1) complex and (2) simple.

The complex egg pod structure is considered an adaptation to repeated hatching and is described for a number of species. The most obvious example of it is the pamphagid

92

thick (Fig. I). One can suggest that when only the upper layers of soil at a depth of 20 to 30 mm become wet, the upper eggs will hatch. In this case, the lower part consists of eggs which constitute a kind of reserve. Under wet conditions, when the soil is deeply drenched, all of the eggs hatch.

Figure 1. Egg pod of Atrichotmethis semenovi: complex, two-level (A) and simple (8) structures

Three main types of hatching patterns may be observed for this type of the egg pod depending on moisture:

(1) Hatching of the upper part of the egg pod. Eggs of the lower part are retained as a kind of reserve for subsequent years.

(2) Hatching from the upper part occurs first. Later in the year, with more rain, hatching of the lower portion occurs. In this case, two maxima of hatching can be observed.

(3) Under very wet conditions all eggs can hatch at once. As a result, abundance of the species may increase sharply.

93

The simple type of egg pod is typified by the species of the genus Saxetania Mistsh. (Fig. 2) [5]. The upper part of the egg pod is filled by a large-cell frothy glue. The egg capsule is irregular and surrounded by a wall of2-3 mm thick. The wall is very hard and resembles stone. The capsule includes soil with eggs. Th female lays 2-3 egg pods at a depth of 5-10 mm. In the pod, a female usually produces 90-92 eggs, and the second pod consists of 30-40 eggs. However, the general size of the egg capsule is not changed from the beginning to the end of egg laying. As a result, the less numerous eggs in later pods are covered by thicker soil layers. Eggs hatch at the end of October and the beginning of November. If the weather is wann and rainy, most eggs hatch. If there is a drought, "late" egg pods with thick walls are not wetted and eggs do not hatch.

Figure 2. Egg pod of Saxetania cultricollis

A curious phenomenon was observed for these species. During the first stage of egg laying, the female usually produced, as we have already indicated, 90-92 and sometimes up to 102 eggs. However, the egg pod consisted of 80 eggs. Where were the remaining 10-20 eggs? I observed that females often deposited these eggs on the soil surface. In some cases, females laid an additional egg pod with these remaining eggs. In such egg pods, the walls and soil layers were stronger and thicker than nonnal. One can suggest that these eggs can survive in the soil until the next spring or even the next autumn. If this is so, the adaptation can be

one of the main ways by which a micropopulation can be maintained under adverse climatic conditions. Therefore, additional egg pods are a kind of a reserve for significant climatic fluctuations. When conditions are favourable, almost all eggs hatch. As a result, a sharp increase in abundance occurs both for common and rare grasshoppers.

Simple structures were also observed in some endemic species. For example, C. olsujjevi and C. prasinum have cylindrical egg pods of the simple type (Fig. 3). Their length is 12-15 mm and the thickness of the soil walls of these egg pods is 0.5 to 2 mm. The egg pod is deposited at a depth of20 mm under the soil surface. One can suggest that in this case hatching be detennined by both environmental fluctuations and egg pod structures (wall thickness, depth of laying, pod size). However, there are probably

94

intrapopulation mechanisms of regulation, (e.g., long diapause). By means of these mechanisms, hatching can be extended or delayed for sufficiently long terms. This allows parts of a micropopulation to select the most favourable conditions for further development.

A B

Figure 3. Egg pods of Conophyma olsufjevi (A) and C. prasinum (8)

3. Solutions

These data show that during outbreaks requiring widespread control treatments, there is a real opportunity to preserve local micropopulations of rare grasshoppers if take into account the characteristics of such micropopulations. If there is a micropopulation inside of an outbreak area, we can use one of the three following types of treatments:

(1) Ifwe know where such micropopu!ations are located we should restrict the treatments and not use insecticides in this area. The downside would be that local populations of pest species may use this territory as a reserve.

(2) Ifwe know the timing ofrare species' hatching, treatments should be made before the second or third peak of hatching.

(3) If a micropopulation includes egg pods with long-term diapause there is a chance for good re-establishment of the colony after even a blanket treatment.

95

4. References

I. Ingrisch, S. and Kohler, G. (1998) Die Heuschrecken Mitteleuropas, Westarp Wissenschaften, Magdeburg. 2. Chernyakhovskiy, M.E. (1990) Ecological peculiarities of montane populations of Conophymaprasinum

Mistsh. (Orthoptera, Acrididae), in I.H. Sharova (ed.) Stuctura i Dinamica Populjacij Pochvennyh i HazemnyhBespozhvonochnyh Zhivotnyh, Vol. 2, Moscow Pedagogical State University, Moscow, pp. 27-40. (In Russian)

3. Guseva, V.S. (1979) Biotic potential and its realisation examplified by three grasshopper species (Orthoptera, Acrididae), Entomologicheskoe obozrenie 58, 522-537. (In Russian)

4. Chernyakhovskiy, M.E. (1987) About egg pods of grasshoppers (Acridoidea), Zoologicheskii Zhournal66, 832-839. (In Russian)

s. Chernyakhovskiy, M.E. (1982) Morpho-ecological characteristics of the grasshopper Saxetania cultricollis (Sauss.) (Orthoptera, Acridoidea, Pamphagidae), in F.N. Pravdin (ed.) Morpho-ecologicheskie adaptacii nasekomyh v nazemnyh soobstshestvah, Nauka Publ., Moscow, pp. 109-116. (In Russian)

6. WHAT FACTORS GOVERN ORTHOPTERAN COMMUNITY STRUCTURE AND SPECIES PREVALENCE?

Abstract

T. KISBENEDEK and A. BALDI Animal Ecology Research Group a/the Hungarian Academy a/Sciences, Hungarian Natural History Museum Ludovika sq. 2. Budapest H-J 083 Hungary

Twenty-seven steppe patches were chosen in the forested landscape in the Buda Hills, Hungary (1) to evaluate whether the composition of orthopteran communities is stochastically or deterministically organized, and (2) to estimate landscape factors influencing presence of individual species in small steppe patches. "Nested Temperature Calculator" was applied for calculating nestedness of the communities, and a logistic regression model was used to model the distribution of species over the patches. Significant nestedness of studied communities was found. The logistic regression models revealed that the organization of orthopteran communities seems to be rather deterministic in the dry mountain grasslands. The most important predictors are an area, number of corridors, and grass height, but consistent relationships were failed to be found for the studied years.

1. Problem

The expansion of habitat destruction by humans is considered as one of the main causes of species extinctions and depauperation of communities [1-3]. Grasslands are among the most threatened habitats, because they can be easily turned to agricultural lands. In densely populated Europe, only a few natural or semi-natural grasslands remained. In Hungary, only 12% of the country belongs to grasslands, mainly managed and used as pastures, contrasting to 29% in the 19th century [4]. Therefore, to study grassland fragmentation is an urgent goal of nature conservation. The orthopterans are a better model taxon than many other insect taxa, because of the small species numbers, relatively stable taxonomical status, comparatively easy sampling, and wide distribution [5-8]. They are, however, more or less insensitive to fine-scale heterogeneity of their environment [3, 9-10].

To understand orthopteran community structure is essential for grassland

97

J. A. Lockwood et al. (eds.), Grasshoppers and Grassland Health. 97 -107. © 2000 Kluwer Academic Publishers.

98

management. First, several species are threatened. The IUCN Red List includes 65 species threatened or extinct [11]. In spite of its small territory (93 000 km2), Hungary is extremely responsible for orthopteran conservation, because 11 from 65 threatened species of the IUCN Red List occur in the country [11]. Second, outbreaks of some species can seriously damage crops [e.g., 12-15], although this is very rare in Hungary [16, 17]. Thorough knowledge of community ecology may provide new insight in both grassland management and population level studies.

Earlier we showed that the orthopteran species number is not randomly distributed over the steppe patches. Species diversity on small patches is relatively richer than can be expected in case of random events [18]. The next step is to evaluate the species composition of the communities and to test for the null-hypothesis that species stochastically occur in communities, as predicted by the island biogeography theory [19]. In contrast, the deterministic approach suggests that species composition can be predicted for islands [20].

The nested pattern is a recently emerged theory, which belongs to the deterministic approach. A community is nested if its species are a subset of species in the larger patches, and they harbor all species of the smaller patches [e.g., 21]. Wright et al. [22] showed that nestedness seems to be a general character of communities. Thus, one can expect significant nestedness of orthopteran communities. Is there any patch or landscape factor, which may be responsible for the nested pattern? To answer the question and to evaluate different factors we modeled the occurrence of the common orthopteran species by multiple logistic regression.

The aims of our study are (1) to evaluate whether the composition of orthopteran communities is stochastically or deterministically organized, and (2) to determine factors influencing the occurrence of individual species in small steppe patches.

2. Studied Area

The study was conducted in the Buda Hills (47°35' N, 18°55' E), Hungary, on Kutya Hill and Nagy Szemis Hill (Fig. 1). Twenty-seven steppe patches were chosen in the forested landscape (Fig. 2). The average size of patches was 17200 m2, ranging from 100 m2 to 101 700 m2 [these values were calculated based on 26 steppe patches, because the boundaries of the largest habitat patch (>400 000 m2) were indefinable]. The present state of patches is determined by dolomite basic rocks and, to less extent, human activities, like forestry. The patches, however, were already present in the beginning of the 20th century, therefore, they can be characterized as permanent and pristine habitats, at least for the orthopterans [18]. The steppe patches consisted of two main vegetation association types - the open dolomitic grasslands (Sesseli leucospermo-Festucetum pal/entis) and the dolomitic slope steppes (Chrysopogono-Caricetum humilis). The habitat patches were in a matrix of two zonal forest types - the turkey-sessile oak forest (Quercetum petraeae-cerris) on the lower and warmer plots and the oak-hornbeam forest (Querco petreae-Carpinetum) on the higher, cooler and moister plots [23].

Figure 1. Outlined map of Europe with Hungary (A); outlined map of Hungary, the

o

bold square is the Buda Hills (B); sketchy map C of our study plot (C)

99

-=-o 13

o

o C

>

o o ~ o

o

o

00

t?c:. ~

CO

0 ~

c:::::

:. D

0

0 0

0

0 """" 1::\

~ <

= ~

o o o

o o

o ~O

\)

~

o 0

Fig

ure

2. S

ketc

h m

ap o

f the

ste

ppe

habi

tat p

atch

es i

n th

e B

uda

Hill

s, H

unga

ry.

The

pat

ches

wer

e di

stri

bute

d on

the

25 k

m2

plot

of t

he B

uda

Hill

s.

Tw

enty

-sev

en

patc

hes

wer

e ch

osen

for

stu

dy;

they

are

sig

ned

by b

oldl

ine

and

num

bere

d

- o o

101

3. Tools

The samples were collected by sweepnetting along I meter wide transects. The length of each transect in a patch was proportionate to the area of this patch. In large patches, the transects were placed to cover all vegetation types within the patch. Transects were not shorter than 30 m, but on the largest patches transects were longer than 200 m. The samples were collected during July of 1992 and 1993.

The 11 most abundant species (3 tettigoniid species, 7 acridid species and 1 mantid species) were chosen for further analyses from all collected orthopteran and mantid species. These species altogether represented 85% of the total abundance of Orthoptera. The following species were involved in the analyses, in decreasing order of rank:

Stenobothrus crassipes is very frequent in steppe grasslands; S. lineatus prefers dry habitats. It is a dominant grasshopper of heath and dry

grasslands, and occurs on roadsides and wastelands, and more rarely on fairly damp meadows, as well. In Hungary, it is rarer in the plain than in the mountain steppe grasslands;

S. nigromaculatus inhabits very dry places with sparse vegetation, occurring, for example, in dry stony grasslands, steppe-like gravely plains and sand dunes;

Euthystira brachyptera prefers both wet and dry habitats. It lives in marshy meadows as well as in dry fields with tall grasses, sometimes even in extremely dry, stony plots;

Euchorthippus declivus very often occurs in dry meadows, southern slopes, gravely plots with sparse vegetation;

Leptophyes albovittata is a thermophilic form and prefers sunny forest borders and dry, bushy grasslands;

Chorthippus (Glyptobothrus) biguttulus occurs in fairly dry places, such as meadows or roadsides;

Ch. (G.) brunneus is more dependent on dry habitats than Ch. biguttulus. This widely distributed species inhabits, for example, sand pits, dry grasslands and dry forest clearings;

Platycleis grisea is a thermophilic katydid. It lives on dry plots with sparse vegetation, especially on southern stony slopes;

Metrioptera bicolor is also a thermophilic species and occurs only in dry meadows. It slightly prefers the less dense grassy scrub habitats;

Mantis religiosa occurs locally on sunny grassy slopes. This species, which is of the Mantodea order, was involved in our analysis, because it occurred frequently in the studied patches. This species is a close relative of Orthoptera and shares habitats with the orthopteran species. In addition, it is the main predator of orthopterans.

Nestedness was calculated using "The Nestedness Temperature Calculator" of [24]. This metric measures the unexpected species absences and presences on individual islands, that is, the disorder of the system. In this context, heat and disorder are equivalent [25]. The temperature of a perfectly nested matrix is O°C, a random matrix is 100°C. This is true, however, only for 50% filled square matrices. In our case, due to the

102

low fill, and elongated shape ofthe matrix, the temperature (i.e., disorder) of the random system was calculated using Monte-Carlo randomization of the matrix by 500 times. Then, this value was used as the temperature of the totally random matrix.

Thirteen variables were chosen to characterize the structure of the habitats, the spatial distribution of the patches and the possible connectivity among the patches (Table 1). The variables can be classified into four groups:

(A) The variables which are connected with patch features, such as an area, perimeter, and shape.

(B) The variables which characterize the degree of isolation of a given habitat island, such as distances of a given habitat patch from the nearest small steppe patch, the nearest large steppe patch, and from the so-called continent, as well as the number of the potential corridors.

(C) The landscape structure variables, such as proportions of the similar habitats inside circles with 250 meter and 500 meter radii, and altitudes of the habitat patches.

(D) The variables which characterize the structure of habitat in an island, such as the average height of grasses.

TABLE 1. Variables characterizing landscape: its area (1-3), structure (4-7, 10-13), and habitat structure (8)

Variable Abbreviations Range of values

min. median max 1. Area [m2] Area 180 14501 101700 2. Perimetere [m] Perimeter 60 598.2 2730 3. Shape Shape 1.10 1.66 2.57 4. Distance to the nearest small patch [m] Dist. small 13 74.04 232 5. Distance to large steppe patch [m] Dist.large 4 100.4 250 6. Distance to nearest continent [m] Dist. cont. 17 574.6 1385 7. Number of corridors Corridor 0 1.96 5 8. Average height of grass [cm] Grass heigtht 15 33.1 70 9. Height bove sea level [m] HASL 410 466.6 525 10. Percentage proportion of the area of steppe % steppe 250 2 19.1 45

habitat within 250 meter radius circle I II. Percentage proportion of the area of steppe % steppe 500 5 16.0 35

habitat within 500 meter radius circle I 12. Number of patches of steppe habitats within no. steppe 250 4.3 8

250 meter radius circle 13. Number of patches of steppe habitat within no. steppe 500 4 9.7 16

500 meter radius

Area, perimeter, altitude and landscape structure values were obtained from 1 : 1 0 000 maps. A shape measures the rate of the actual perimeter to the perimeter of a circle that has the same area as the patch. Connections among the patches were estimated by the number of the potential corridors. Potential means that we did not test corridors for permeability, simply considered all grass strips between steppe patches as the corridors. The average heights of grasses were estimated by eye.

We applied the multiple logistic regression analysis to model the species

103

distribution over the patches. We used the forward stepwise method with the likelihoodratio statistic as the test for removal [26]. The logistic regression analysis is a frequently used method, partly because it does not require quantitative data, there is no restriction to linear relationships, and is insensitive to the violation of multivariate normality assumption [27,28].

4. Solutions

Altogether 692 specimens were caught in 1992, and 1201 in 1993. The communities were highly significantly nested in both years. The temperature of the orthopteran species/steppe patch matrix was 12.63°C in 1992, and 16.44°C in 1993. Both values are at several standard deviations apart from the temperatures expected under random distribution of species: 43.82°C ± 4.56 (1992) and 55.53°C ± 4.78 (1993).

Two steppe patches mainly contributed to the observed disorder of the system. These two patches were at the border of the forest and were connected with agricultural or abandoned fields on one or more sides. Thus, they are different from the other patches. These patches were excluded from further analysis.

Five species proved to be idiosyncratic; that is they had comparatively many unexpected presences in small patches. These species were E. braehyptera, M bieolor, and P. grisea in 1992, and S. lineatus and Ch. brunneus in 1993. Probably, some species specific habitat selection characters caused the idiosyncratic patterns. For example, E. braehyptera preferred the shadow at the forest/steppe patch edges, or the small shadowy patches mostly with tall grasses. The explanation for the other four species may include preference for microhabitats (e.g., bare ground), interspecific interactions (e.g., exclusion from optimal, large islands by a superior competitor or predator), or other factors. We measured 13 potential variables and applied the multivariate logistic regression analysis to select the most important variables for each species.

We found that 15 out of the possible 22 (11 species, 2 years) logistic regression models were significant (Table 2). Area, number of corridors, and grass height were the most important predictors of presence across all species. Area had positive, while a number of the corridors had negative effect on the presence of Ch. biguttulus, S. erassipes, S. lineatus, and S. nigromaeulatus at least in one year studied. Thus, these species prefer large, isolated patches. This may indicates that the emigration probability for these species from the small patches is higher than from the larger patches. Grass height and a number of the steppe habitats around the patch influence the presence of Ch. brunneus and P. grisea, which may reflect that the presence of these species depends, at least partly, on habitat quality, and is independent of the size or the isolation degree of a given patch. The probability of the presence of E. declivus and L. albovittata is opposite the presence of the above mentioned species: The former species prefers the forested landscape, while the latter prefers the steppe landscape with many small patches. E. braehyptera avoids the proximity of large steppe patches. M bieolor prefers elongated patches. The presence of these species can depend on the proportion

104

of the edge and inside size of the patches. The results for P. grisea are contradictory: in 1992 grass height was a negative predictor, while in 1993 it was a positive predictor. However, this species prefers elongated patches with many other patches in the proximity. Finally, Mantis religiosa prefers many large patches. The presence of these orthopteran species in some given patches can be connected with their dispersal ability, behaviour, and habitat requirements, too. The species of the genus Stenobothrus, and E. braehyptera, E. deelivus, M bie%r, P. grisea, M religiosa are vagrant, while Ch. biguttulus and Ch. brunneus are wandering species [29].

TABLE 2. Models based on the logistic regression analysis. The models tell us which variables of common occurrence are responsible for the presence of a given species

1992 1993 Variable Coefficient SE Signifi- Variable Coefficient SE Signifi-

Species of cance of cance of regression of the regression the

model model

Ch. biguttulus Area 6.9854 4.3301 0 Corridor -3.8181 2.5641

Ch. brunneus Grass 9.0845 3.9497 0.0332 height No. steppe 8.1948 4.6634 500

S. erassipes Area 12.9926 13.4073 0.0024 Area 30.9828 38.7614 0.001 ~orridor -1.9441 0.9487

S.lineatus Area 5.1524 2.4601 0.Q31

Dist. cont. 2.7794 1.6182

Corridor -2.2866 1.1339

S.nigrollUleu--~

Area 2.6099 1.2421 0.0306 latus

Corridor -1.1456 0.6933 -

E. declivus Vo steppe -7.6493 4.3715 0 500 Dist. small 5.8494 2.5506

L. albovittata Dist. smaIl -6.1746 3.1103 0.0038 % steppe 12.9936 6.0687 500

E. braehyptera Dist.large 3.6381 1.3727 0.0005 M. bieolor Shape 17.2792 9.9705 0.0156 P. grisea Grass -7.9884 4.1628 0.0262 prass 1 28.1846 0.0045

height height ~o. steppe 8.1643 4.3403 ~50 Shape 42.4101 23.3995

M. religiosa Area 2.1847 0.9006 0 ~o. steppe 5.6779 2.7479 0.0127 500

For one part of studied species the results of the logistic regression analysis identified mainly significant predictor variables for only one of the study years. For

105

another part different factors were identified in 1992 and 1993. In one case (P. grisea) there were contradictory results between 1992 and 1993. This suggests that although the presence of the orthopteran species was influenced by some characteristics of the landscape structure, presence was influenced by other factors, as well. Such influencing factor may be weather. In 1993 there was a heavy drought with extremely high air temperature in this region. As a result, population abundance of most species increased slightly in early July, but more than half of the species disappeared on the studied sites till the middle of August.

The results of the logistic regression analysis can help to identify variables that caused idiosyncratic distribution of five species in the nestedness analysis. The logistic regression analysis did not incorporate area into the model for four out of the 5 species (Ch. brunneus, E. braehyptera, M bie%r, P. grisea). Therefore, the occurrence of these species was influenced mainly by habitat characteristics, landscape structure, and the proportion of edge habitats, and was independent of area. The fifth species, S. lineatus, prefers large, disconnected patches. Its idiosyncratic distribution may be due to its occurrence in almost all patches, with a few unexpected absences in larger patches. A possible hypothesis to explain this pattern is its abundance-area relationship: if it is negative (more individuals in small, than in large patches), then we can predict unexpected absences in large, and unexpected presences in small patches.

Our two analyses reflected a limit of the nestedness analysis, namely it focuses on one simple measure: the size (usually an area) of islands, and excludes many other potential factors. Since ecological communities are spatial-scale dependent [30, 31], to understand the mechanisms of nested patterns requires investigations across several spatial scales, at the habitat, patch and landscape levels.

5. Conclusions

Based on the nestedness models the organization of the orthopteran communities seems to be rather deterministic in the dry mountain grasslands. The distribution of species was species-specific, and important influencing factors included habitat, patch, and landscape characteristics, as well. In addition, the different results of two study years suggest that the effects of stochastic factors, e. g., a drought, also may be important for the community organization.

6. Acknowledgements

We thank the nature conservation and forestry authorities for their permission to collect the insects for the study. The analysis was supported by the Hungarian Scientific Research Fund (OTKA 29242).

106

7. References

I. Ehrlich, P. R. (1990) Habitats in crisis: Why we should care about the loss of species, Forest Ecology and Management 35, 5-11.

2. Maynard Smith, J. (1989) The causes of extinction, Phil. Trans. R. Soc. Lond. B 325,241-252. 3. Swnyi, G. and Kincsek, I. (1986) Indication of spatial heteromorphy and community structure of

Acridoidea-communities in a sandy grassland, Acta BioI. Szeged 32, 141-156. 4. Kelemen, J. (ed.) (1997) Guidelines of grassland managementfor nature conservation, TermeszetbUvar

Alapitvany Kiad6, Budapest. 5. Kisbenedek, T. (1997) Egyenesszarnyuak - Orthoptera, in Form Laszl6 (ed.), Nemzeti Biodiverzitas

monitorozo Rendszer V Rakok szitkotOk es egyenesszarnyitak (Monitoring System of National Biodiversity V. Crustaceans. dragoriflies and orthopterans), Magyar Termeszettudomanyi Muzeum, Budapest, pp. 55-81. (In Hungarian)

6. Baldi, A.and Kisbenedek, T. (1997) Orthopteran assemblages as indicators of grassland naturalness in Hungary, Agriculture Ecosystems and Environment 66, 121-129.

7. Fischer, F. P., Schubert, H., Fenn, S., and Schulz, U. (1996) Diurnal song activity of grasslands Orthoptera, Acta Oecologia 17 (5), 345-364.

8. Fischer, F. P., Schulz, U., Schubert H., Knapp, P., and Schmiirger, M. (1997) Quantitative assessment of grassland biotope quality: Acoustic determination of population sizes of orthopteran indicator species, Ecological Applications 7 (3), 909-920.

9. Galle, L., Gy6rffy, Gy., Kiirmiiczi, L., Sz6nyi, D. G., and Harmat, B. (1985) KiiliinbiizeS kiizziissegtipusok eleShely heterogenitas indikacioja homokpusztai gyepen (Indication of space heterogeneity from various community types on sandy soil grassland), OKTH Evkonyv, 230-271. (In Hungarian)

10. Varga, Z. (1997) Trockenrasen im pannnonischen Raum: Zusarnmenhang der physiognomischen Struktur und der fIoristischen Komposition mit den Insektenziinosen, Phytocoenologia 27 (4), 509-571.

11. mCN (1996.)/996 JUCN Red List of Threatened Animals. mCN, Gland ,Switzerland. 12. Popov, G. B. (1988) Sahelian Grasshoppers, Overseas Development Natural Resources Bulletin, 5, 0-85. 13. Maiga, B. (1992) Mali: notes on the acridid problem in Mali, in C.J. Lomer and C. Prior (eds.), Biological

control of locusts and grasshoppers, CAB International, Wallingford, pp. 82-85. 14. Hewitt, G. B. (1977) Review offorage losses caused by rangeland grasshoppers. USDA-AARS Misc. Pub.,

1348,0-22. 15. Latchininsky, A. V. (1995) Grasshopper problems in Yacutia (Eastern Siberia, Russia) grasslands. J.

Orth. Res. 4, 29-34. 16. Nagy, B. (1993) Magyarorszagi saskagradaciok 1993-ban. (Hungarian grasshopper outbreaks in 1993),

Novenyvedelem 29 (9), 403-411. (In Hungarian) 17. Nagy, B. 1995. Are locust outbreaks a real danger in the Carpathian Basin in the near future? J. Orth.

Res. 4, 143-146. 18. BaIdi, A. and Kisbenedek, T. (1999) Orthopterans in small steppe patches: an investigation for the best-fit

model of the species-area curve and evidences for their non-random distribution in the patches, Acta Oecologica 20, 125-132.

19. MacArthur, R. H. and Wilson, E. O. (1967) The theory of island biogeography, Princeton University Press, Princeton.

20. Whittaker, R. J. 1992. Stochasticism and determinism in island ecology, Journal of Biogeography 19, 587-591.

21. Patterson, B. D. and Atrnar, W. (1986) Nested subsets and the structure of insular mammalian faunas and archipelagos, Bioi. J. of the Linnean Soc. 28,65-82.

22. Wright, D. H., Patterson, B. D., Mikkelson, G. M., Cutler, A., and Atmar, W. (1998) A comparative analysis of nested subset patterns of species composition, Oecologia 113, 1-20.

23. Pecsi, M. (ed.) (1959) Budapest termeszeti foldrajza (Natural Geography of Budapest), Akademiai Kiad6, Budapest. (In Hungarian)

24. Atrnar, W. and Patterson, B. D. (1995) The nestedness temperature calculator: a visual basic program. including 294 presence-absence matrices, AICS Research Inc., University Park, NM, and The Field Museum, Chicago IL.

107

25. Atmar, W. and Patterson, B. D. (1993) The measure of order and disorder in the distribution of species in fragmented habitat, Oecologia 96,373-382.

26. Nourusis, M. 1. (1990) SPCC/PC+ 4.0, Base Manualfor the IBM PClXT/AT, SPSS, Chicago. 27. Brito, 1. c., Crespo, E. G., and Paulo, O. S. (1999) Modelling wildlife distributions: logistic mUltiple

regression vs overlap analysis, Ecography 22, 251-260. 28. Kemenes, I. and Demeter, A. (1994) Uni- and multivariate analyses of effects of environmental factors on

occurrence of otters (Lutra lutra) in Hungary, Annis. hist. nat. Mus. natn. Hung. 86, 133-138. 29. Nagy, B. (1992) Role of activity pattern in colonization by Orthoptera, in Proceeding of the 4th

ECE/XIIl SIEEC (G6d61161991), pp. 351-363. 30. Wiens, 1. A. (1989) Spatial scaling in ecology, Functional Ecology 3, 385-397. 31. Fuisz, T. and Moskat, C. (1992) The importance of scale in studying beetle communities: hierarchical

sampling or sampling the hierarchy? Acta Zool. Hung. 38 (3-4),183-197.

Part 3. GRASSHOPPER AND LOCUST CONTROL STRATEGIES AND TOOLS

7. HOW CAN ACRIDID POPULATION ECOLOGY BE USED TO REFINE PEST MANAGEMENT STRATEGIES?

Abstract

M.LECOQ Centre de Cooperation Internationale en Recherche Agronomique pour Ie Developpement (ClRAD) P RIF AS - Acridologie operationnelle B.P. 5035 - 34032 Montpelier Cedex I, France

The problem posed in Brazil since 1984 by Rhammatocerus schistocercoides - an important pest locust in the Mato Grosso state - is exemplary. It shows how the occurrence of a new locust phenomenon has given rise to fruitless opposition between supporters of insecticide treatments, on one hand, and environmental conservationists on the other. Extensive control measures were implemented and large quantities of insecticides sprayed, with all of the drawbacks concerning risks to the environment and to the health of the indigenous people that eat locusts. Intensive ecological field work, carried out for 10 years, gradually has revised our understanding of the origin of this species' outbreaks and the way to control them. On the basis of solid scientific experience, solutions can now be posed that are realistic, efficient, respectful of the environment and the interests of farmers and indigenous populations.

1. Introduction

In Brazil, plague locusts and grasshoppers represent a serious concern in various regions. The species involved are diverse, and the areas where they swarm vary from one year to another. Nevertheless, in certain regions they are discussed more frequently [21]. The main areas affected during the last 20 years are the Rio Grande do SuI (Rhammatocerus conspersus Bruner 1911), and the Northeast (Schistocerca pallens Thunberg 1815, Stiphra robusta Melo-Leitao 1939 and Tropidacris collaris Stoll 1813),

109

J. A. Lockwood et al. (eds.), Grasshoppers and Grassland Health. 109-129. © 2000 Kluwer Academic Publishers.

110

and Mato Grosso (Rhammatocerus schistocercoides Rehn 1906). In southern Brazil, Schistocerca cancel/ata (Serville 1838) swarms coming from Argentina now seem to be eradicated, and for about 50 years Brazilian territory has not been invaded [13]. Some other locust species are the source of local and very sporadic outbreaks (Dichroplus bergii brasiliensis Bruner 1906, Staurorhectus longicornis Giglio-Tos 1876, Parascopas obesus Giglio-Tos 1894, Xyleus discoideus Serville 1831) with local aerial treatments sometimes being necessary [4].

One of the most interesting and best documented cases is without a doubt the locust of the Mato Grosso, R. schistocercoides. The problem posed by this species, since 1984, is remarkable. It shows how the occurrence of an apparently new locust along with the local absence of experience and accurate knowledge have given rise to severe and fruitless opposition between supporters of massive insecticide treatments, on the one hand, and environmental conservationists on the other.

At the beginning, the outbreaks were explained based on a priori hypotheses, without significant scientific foundation. Locust invasions threatened numerous Brazilian states. The origin was unknown, but workers suspected that it was tied to recent and intensive cultivation in the southern limits of the Amazonian forest. Extensive measures were implemented and large quantities of insecticides were sprayed. Of course, associated drawbacks included contamination of the environment as well as considerable risks to the health of the indigenous people that eat locusts.

The control strategy that was developed during the emergency undoubtedly limited the outbreaks. With a broader perspective, however, this strategy seems poorly adapted, being completely out of step with the reality of the phenomenon. Ecological work, carried out for 10 years, gradually revised our understanding of the origin of this species' outbreaks and the way to control them. We now have a completely different vision of the factors that prevailed at the beginning. On the basis of solid scientific experience, realistic, efficient, and environmentally sound solutions can now be proposed.

2. Problem

2.1. NATURE OF THE PROBLEM OF LOCUST OUTBREAKS IN MA TO GROSSO

In Brazil, starting in 1983, outbreaks of a locust species until then unknown started to pose serious problems in areas of the states of Mato Grosso and Rondonia, in a large region that extends in an outline between the southern parallels of 12° and 15°, and between 52° and 61° west longitudes [18]. The affected areas are predominantly "cerrado" areas (wooded savannahs) located immediately to the south of Amazonian forests, from the Rondonia border to the west and from the Araguaia river valley to the east (the Chapada dos Parecis region sensu lato) (Figure 1). These areas had been recently cultivated and developed. Large developments of thousands or tens of thousands of hectares were implemented in almost virgin territories formerly occupied only by Indians. The main crops that were developed included soybean, sugar cane,

111

rice, and com. The new agriculturists coming primarily from the south of Brazil were

confronted very quickly with what was considered immediately as a true locust invasion. The first major swarms were noted in September of 1984 [7]. In 1985, the infested area was estimated at 15 million ha and then at 20 million ha in 1986 [7, 9]. In addition, this locust invasion seemed to proceed to the east and threatened the rich neighbouring states of Goias and Sao Paolo.

Figure 1. Outbreak area of Rhammatocerus schistocercoides in Mato Grosso according to Miranda et al. [37]

In white: forest areas. In dark grey: savannah areas (campo and campo-cerrado) where most outbreaks occured. In light grey; savannah areas which are less favourable for the locust. GO, Goias; MT, Mato Grosso; RO, RondOnia; TO, Tocantins. Rivers: 1 - Juina; 2 - Juruena, 3 - Papagaio; 4 - Sacre; 5 - do Sangre; 6 - Sacariuina; 7 - Arinos; 8 - Xingu; 9 - dos Mortes; 10 - Araguaia; 11 - Paraguai; 12 - Guapore.

112

The species responsible for the outbreaks was identified with some difficulty. First named Rhammatocerus pictus, it was later considered by some as a new species, and the name R. parecis was suggested for a time. It was finally demonstrated that it was Rhammatocerus schistocercoides, a locust of the Gomphocerinae subfamily, described by Rehn in 1906 [2] (Figure 2) and known only from Mato Grosso and neighbouring regions [2, 43].

Invading the newly cultivated areas in Mato Grosso, this locust created such a serious problem that it cast doubt on the viability of recent and important agricultural investments. In order to carry out and coordinate control operations, a national programme was initiated in 1986 by the Brazilian government, and the "Locust of Mato Grosso" constituted one of its main concerns [I, 8, 9]. Large-scale treatments were implemented, first under the auspices of the Department of Agriculture and the national control programme, then by the farmers themselves [I, 5, 6, 7, 34, 35]. These treatments were especially important in certain years and had both economic and environmental consequences [48]. During the last 15 years (since the tremendous agricultural expansion in the region) major outbreaks were recorded in 1984 and 1988, as well as, to a lesser degree, in 1992 and 1993. The problem currently appears to be chronic.

Figure 2. Rhammatocerus schistocercoides (Rehn 1906).

113

The economic importance of this species fluctuates from year to year and is fairly difficult to estimate. Some figures can give an idea of the magnitude of losses during years of heavy outbreaks. For 1984, the losses were estimated at 84 million litres of alcohol from crops, 600 000 tons of soybeans, 75000 tons of rice, and 22500 tons of corn [9]. In 1986, 2.5 million US dollars were spent on control operations, preventing losses estimated at 500 million US dollars in agricultural production in Mato Grosso. From 1985 to 1989, a total of 5 million US dollars were spent on the anti-locust campaign in Mato Grosso of which 41% was for aerial spraying and 43% for insecticides and labor (F olha do estado, Cuiaba, 13/3/1995). Tens of thousands of liters of insecticides were used, predominantly fenitrothion and malathion (Table 1).

TABLE 1. Results of the aerial anti-locust campaigns in Mato Grosso from 1984 to 1988, after Curti and Britto [9] and Kuffner [14).

Years 1984 1986 1987 1988

Areas controlled (ha) 100000 245000 99387 71303

Pesticides (litters) :

Fenitrothion UL V 50000 67310 26591 21887

Malathion UL V 20400 10750

Aircraft hours:

Plane 500 835 264 200

Helicopter 1031 581

Cost of control operations (US$) 1 710462 1 189487 1 146262

(-, no data; data from 1985 not available)

Insecticide treatments of such magnitude could not lead to indifference. Many questioned their true efficiency and impact on the environment. Furthermore, the problem was complex due to the co-existence of two fundamentally different modes of living and land occupation in the locust outbreak areas: recent immigrants, farmers or ranchers coming mainly from states in the south of Brazil, and the traditional population of local Indian tribes, Parecis, Nambiquaras and Baiquiris. For the former, the locust was a plague to be fought by every means; for the latter, the locust was a natural element of the ecosystem, at times a mythological creature, and frequently an important element of daily food. Two "parties," or perspectives, therefore confronted each other regarding the nature of the plague.

114

2.2. INITIAL HYPOTHESES

What were the initial hypotheses to explain this invasion of locusts and to justify the control operations?

Until the beginning of the 1980s, R. schistocercoides had been a virtually unknown locust species that never caused any economic problems. The first outbreaks and damages were observed shortly after agricultural development of the affected areas. The explanation for these outbreaks was naturally sought in the extensive modifications brought on by agriculture to the local environment over the course of a few years.

The popular hypothesis from the beginning of the locust problems in Mato Grosso assumed that new biotopes favourable to locust reproduction had been created through intensive deforestation (agricultural groundbreaking) and transformation of the land into crops or pastures. An ecological disequilibrium also developed that would have caused a serious reduction of the locusts' natural enemies thus favouring the pests' outbreaks. In particular, it was explained that with the land clearing the avian predators of the locusts would have become much easier prey for hunters [1, 48]. The new immigrants appeared to be responsible themselves for the plague that affected them. This thesis was defended by the Indians and a large number of ecologists.

For farmers, the locust plagues had to be controlled by all means. They believed that the locusts mainly reproduced in naturally vegetated areas and particularly in the Indian reserves where it was unfortunately impossible to carry out insecticide treatments. The Parecis Indian reserve was primarily accused of being the origin of the invasion that spread very quickly towards the east of the state. In this hypothesis, the first occupants of the land appeared responsible for the problem. In addition, converting natural areas to agricultural purposes should have limited the reproduction areas of this locust and gradually reversed the problem. At first, treatment of the Indian areas as potential reserves for this locust seemed essential. It was also assumed that R. schistocercoides had a great capacity for migration, and the apparent spread of the invasion towards the east made this locust a threat to the neighbouring states of Goilis, Tocantins and Sao Paulo. From there, a control strategy was implemented that aimed at cutting off the invasion in its supposed eastern route [7, 9].

These diverse hypotheses about the origin of the outbreaks and the dynamics of the invasion fed a great controversy about the respective roles of the farmers and the Indians, as weII as conflicts regarding how to carry out treatments [40, 46].

Nevertheless, until the end of the 1980s, none of the hypotheses described above were scientifically examined. At this time, we found ourselves facing an apparently new problem: • that seemed tied to the development of new territory and to the manner of habitat conversion; • that needed large-scale insecticide treatments, without evidence that this strategy was valid; • that exposed the environment to risk from large quantities of insecticides (possibly aggravating the supposed natural disequilibrium induced by the introduction of agriculture and exacerbating the outbreaks);

115

• about which there were neither experience nor accurate scientific data.

The plague was managed, since it could not be ignored under pressure from farmers and politicians. But, for many, the situation was not satisfactory, and the problem warranted greater understanding.

3. Tools: research and its results

3.1. METHODS

The controversy over the ongm of the invasion, the need for large-scale treatments, and the risks to the environment via insecticides, all showed that it was necessary to try to better understand the phenomenon and to find new strategies and methods for control.

A research project in collaboration between EMBRAPA-NMA (Brazilian agricultural research corporation - Satellite Monitoring Centre) and CIRAD-Prifas (French research organization on agriculture in the tropics and subtropics - Operational acridology unit) began in 1991. The main objective of this project was to analyze the causes of R. schistocercoides' outbreaks. In particular, it dealt with analyzing the impact of the recent agricultural occupation of land and of anthropogenic landscape modifications on the locust outbreaks in those territories located in the confines of the Brazilian Amazon. EMBRAPA and CIRAD combined their respective competencies in remote sensing and acridology in order to assemble the necessary ecological bases for formulating more efficient, economical, and acceptable management strategies.

This research was based, in principal, on the hypothesis (which seemed realistic) that the problem was related to accelerated land development. Very quickly, however, it turned out that this hypothesis "did not fit" with the facts and observations, and it was necessary to reconsider the basic aspects of the locust problem in Mato Grosso [25].

A series of ecological field studies was then undertaken. This work posed the problem in a very general context. Its objectives were to clarify the biological cycle of this locust in its natural environment, to understand the exact nature of its biotopes, its ecology and the causes of its outbreaks, to determine the real influence of recent agricultural development, and to propose appropriate control strategies [37].

Research domains were biological, ethological, ecological, bibliographical, historical, economic, cartographic, meteorological, and pest management. The research methods and techniques integrated quite varied resources. Among them were interviews with the inhabitants of Mato Grosso (Indians, old farmers, pioneers who colonized Mato Grosso, anthropologists, Jesuit and Salesian missionaries, herdsmen, rubber collectors, etc.) as well as the use of satellite images, and the creation of geographical information systems (GIS).

Research on the biology, ecology and behaviour of this locust under natural conditions played a truly fundamental role. Intensive land studies using sampling

116

methods designed to comprehend the situation within the natural biotopes have quickly provided clarification about aspects of the phenomena that were poorly understood to date.

The use of satellite images and GIS equally warrants special mention. At the time in Brazil, it was the first use of these new tools for solving on an entomological problem of economic importance.

3.2. RESULTS

The numerous results accumulated since 1991 repudiate previous hypotheses pertaining to the character of locust outbreaks in Mato Grosso, to the influence of agricultural development, and particularly to the threat that this locust represents for neighbouring states. The following summary allows us to highlight some of the significant results as well as to review the practical lessons regarding the recommended control strategy.

3.2.1. Distribution area

A few years ago, it was thought that R. schistocercoides was an endemic species from Brazil, mainly from the Mato Grosso area. Research carried out on the collections from diverse museums and information from the field, showed that R. schistocercoides is not limited to Mato Grosso and neighbouring areas. This species is now known from other countries of South America, particularly Columbia, where severe outbreaks occurred in 1994-1996 [3, 10,20,30,33]. Recently, some local outbreaks were reported from the Brazilian states of Slio Paolo and Minas Gerais (Cosenza, personal communication).

3.2.2. Biology and ecology

The general facts of the insect's life-cycle were known at the beginning of the project.[7]. However, it was necessary to study other, specific parameters on populations in their natural environment rather than in cages. The accumulated data required a critical re-evaluation of numerous aspects of the biology and ecology of R. schistoceroides, including: termination of adult diapause, sexual maturation, oviposition, length of nymphal development, number of instars, adult moulting, etc.. For example, the number of hopper instars was revealed to be much higher than previously assumed: 8 or 9 instars instead of 5 or 6. Development periods were also established under real conditions in the field [22, 37].

R. schistocercoides has only one generation per year. The hoppers develop slowly over the 5 or 6 months of the rainy season (Figure 3). Adults live through the dry season like nomads, and they reproduce at the beginning of the following rainy season. Eggs are deposited from the end of September until the end of October, at the beginning of the rainy season. Embryonic development lasts about 15 days. The hoppers develop during the rainy season, from the end of October until the middle of April [22, 37]. The adults appear around the middle of April and spend the majority of the dry season, from May to September, apparently in diapause. During this time, the ovaries do not function

117

and their volume remains reduced. Reproductive activity begins at the end of August, regardless of the rains, initiating the first egg-laying at the end of September [16]. The length of developmental delay in the dry season is about 4 months, from the middle of April to the middle of August. Sexual maturation lasts approximately I month (end of August to end of September); the cycle of egg-laying is about 15 days, and apparently there are no more than 3 pods per female (average of 1.5).

Figure 4 is an example of the annual dynamics of R. schistocercoides in the Chapada dos Parecis. It represents an average cycle that takes into account the characteristics (size and density) typically noted in hopper bands and swarms, as well as the average values observed or assumed in birth and death rates at the different developmental stages.

The role of certain natural enemies has been studied, especially that of Prionyx thomae (Fabricius) (Hymenoptera, Sphecidae). The attack and nest-building behaviour of this wasp was described, as well as the defensive reactions of the locust [24].

Studies of the pigmentation of R. schistocercoides revealed an important pigment variability. This work clearly showed green-brown chromatic polymorphism, tied to sexual maturation, that had not been reported previously [26].

Finally, the first case of geophagy (consumption of soil) in locusts was noted. Adults and hoppers sometimes consumed considerable quantities of soil, even when there was green vegetation nearby [26].

3.2.3. Absence o/phase polymorphism

In spite of a very obvious gregarious behaviour both in the nymphal and adult stage, it has not been possible to show morphological differences between individuals from high vs low density natural populations. Therefore, it does not seem that there are solitary and gregarious phases as with true locusts. The species thus resembles a grasshopper [42], which does not affect the conceptual approach to developing a control strategy.

3.2.4. Migration behaviour and capacity

R. schistocercoides exhibits a notable gregarious behaviour, and it forms hopper bands and very dense swarms that can cover dozens of hectares. The behaviour of the hopper bands and swarms has been studied in detail [27,28]. In particular, the swarms have a much more limited migratory capacity than previously believed, moving only a few hundred meters to a few kilometers per day.

It was shown that the outbreaks in Mato Grosso were basically local, and they did not threaten the other states in Brazil. The hypothesis of locust invasions progressing towards the east was refuted. Conversely, the structure of the habitats, meteorological conditions (the regular alternation of north or northeast and south winds and the absence of west or east winds during the dry season), and the existence of "barriers" formed by gallery forests, create small ecological units that are essentially isolated from one another. As a result, swarms can only wander locally on a mainly north-south/south-north directi<m (not west-east). These displacements show

118

characteristic migratory behaviour (Le., obligatory flight regardless of habitat quality) but only on a restricted scale.

eggs L1 L2 L3 L4 L5 L6

( 0 N D J F M

00000 11111 1

22222 2 333333 3

44444444 4 55555555 5

66666666 6

A M J J A s J

L7 7777777 7 La 1M IV

PPPPP ,_I_P~

(M )(211111000

PP 300

o N D J F M

BBBBBBB B

A

111111111111111111111111111111111111111111 VVWV

P

0111222223333333444444442)

• ..... • M J J A s

Figure 3. Life cycle of Rhammatocerus schistocercoides, according to Miranda et al. [37].

0, eggs; L1 to L8, hoppers; 1M, immature adults; IV, maturing adults; IP, laying females; M, migratory capacity of swarms (from 0, none, to 4, high); PP, rains (mm).

1 ha 4 ha 250 adults/m2 •

2.5 million adults? ~M. ~

A A

250 adults/m2 10.0 million adults

0.5 ha 500 L8/m2

25 million hoppers

0.1 ha 10000 L 11m2

10.0 million hoppers

Figure 4. Example of R. schistocercoides annual dynamics in Chapada dos Parecis, according to Miranda et al. [37].

The following assumptions were used (the values vary as a function of ecological conditions):

• Each major swarm lays on different sites and gives birth to a dozen of hopper bands.

119

• Ten young hopper bands, moving during the rainy season end up grouping together and giving birth to 4 major hopper bands.

• Four young swarms, moving during the entire dry season, can give rise to only one major swarm.

• The sex-ratio is I: I, the number of egg-pods per female is 1.5, the number of eggs per pod is 30, the average embryonic mortality is 10%, hopper mortality is 90%, and adult mortality during the dry season before laying is 10%.

120

10 m

5

0

Wind >/ Flying locusts (max. 15%)

~ --+ --" ~ -.

// .-. ~

--+

Settled locusts {min. 85%)

~

'-'- " Figure 5. Behaviour of locusts in the interior of a "rolling" swann of R. schistocercoides,

according to Lecoq and Pierozzi [27].

The bold arrows indicate the locusts' direction of movement.

1000m

3.2.5. Antiquity of outbreaks

One of the most important project accomplishments was the demonstration that Mato Grosso has always been a theatre of important and chronic outbreaks of R. schistocercoides [23]. Bibliographical research and field data agree on the fact that the outbreak phenomenon of locusts in Mato Grosso is very old and that the current outbreaks do not seem to be either more severe or more frequent than in the past. The acquired data revealed the occurrence of outbreaks since as early as the end of the 19th

century [23, 44, 45]. At the beginning of the 20th century, some specialists even considered them to be one of the factors that prevented the development of agriculture in Chapada dos Parecis, which gives an idea of the magnitude of those outbreaks [12]. So, the phenomenon of locust outbreaks in Mato Grosso is not new. However, what is "new" is the economic problem that the locusts pose in Mato Grosso, and it naturally goes back to the intensive development of agriculture in the 1980s. Crops were planted in areas of habitual locust oubreaks. New farmers, immigrants coming mainly from southern Brazil, did not know about this hazard. Ignorant of old observations, they saw their first crops threatened or destroyed by an insect that they had not ever seen, although in reality it was an important component of the local entomofauna. The "illusion" of a "new" problem was thus created. It gained acceptance so much more easily, since it emerged in a particularly favourable cultural context, aware of the

121

problems connected to the deforestation of the Amazon, agrarian conflicts between the Indians and the new farmers, and criticisms of the new immigrants' cultivation methods [23].

It is now clear that recent modifications to the environment of Mato Grosso (transformation of landscapes via human activities, deforestation, development of "cerrado" areas, expanding areas for crops and pastures, reducing wild fauna etc.) did not influence locust outbreaks [23, 37]. This documentation of historical R. schistocercoides outbreaks constitutes a capital fact that has radically changed our understanding of the phenomenon and its causes, and it has led to the rejection of previous hypotheses. It can no longer be assumed that the locust outbreaks are a result ofthe introducing intensive agriculture to the frontiers of Brazil [23,37].

3.2.6. Characterisation and cartography of the habitats

The nature of the R. schistocercoides habitats has been clearly established. It was possible to map them using satellite remote sensing (LANDSAT -TM data) and a GIS [37,38,39,41]. The structure of these habitats is particularly clear in the main outbreak area ofthis locust, in the western part of Chapada dos Parecis, from the Arinos river to the Guapore river. Globally, these habitats correspond to a group of areas in which plant formations of the "cerrado" (wooded savannahs) or "campo cerrado" (open wooded savannahs) have established. These savannah formations with shrubs and wooded areas are generally very open and maintained by regular fires. They differ from the interior of the savannah regions by the type of soil [37] (Figure 6):

• The breeding habitats where females will lay eggs, where the hoppers will carry out the majority of their development during the rainy season, and where young, immature swarms will appear at the very beginning of the dry season correspond to sandy soil areas. These areas are generally related to the hydrographic network (a few kilometres from both sides of the river). The areas between rivers are flatter, with heavier, sandyclayey soils that are less favourable for egg-laying.

• The wandering habitats where the adults, generally grouped in swarms, move according to the wind during the entire dry season, from May to September are areas of heavy, sandy-clayey soils. These areas are between rivers and are currently occupied to a large extent by crops. The breeding habitats are too sandy and are generally not used by the farmers.

3.2.7. Relationship between method of land use and locust outbreaks

Having been able to define the R. schistocercoides habitat characteristics, it was possible to elucidate the relationships between the method of land use and locust outbreaks.

122

These human/locust relationships are complex depending on the season, method of crop cultivation, and pasture practices [37].

A B

Figure 6. Schematic structure of the R. schistocercoides habitats of the western part of Chap ada dos Parecis, according to Miranda et af. [371.

Light grey areas, breeding habitats (campo and campo-cerrado on sandy soils); white areas, wandering habitats (campo and campo-cerrado on clay-sandy soils) ; dark grey areas, gallery forest; square pattern, crops. Upper frame, plane view; lowerframe, section A-B.

Up until recently, Chapada dos Parecis was occupied almost exclusively by Indians. This wide plateau remained for a long time almost mythical, an area frequented only by "garimpeiros" (gold searchers), "seringueiros" (rubber collectors), "boiadeiros" (herdsmen), and mIssIOnaries, without long-lasting agricultural occupation. Anthropologist Claude Levi-Strauss frequented it in the 1940s in order to study the Nambiquaras Indians (who were major consumers of locusts), a stay that was the prelude to the writing of his already classic work, "Tristes Tropiques" [31]. It was only at the end of the 1970s and at the beginning of the 1980s that agriculture was implemented massively and quickly in Chapada with the arrival of new immigrants from the south of Brazil. It was these colonists who, at the beginning of the 1980s, "discovered" the locust swarms at the same time that they started to exploit the cerrado and develop their crops.

123

The crops in Chapada dos Parecis are grown specifically in the high elevation areas between rivers, in sandy-clayey soil. The sandiest parts, generally linked to the hydrographic network, are more commonly preserved as natural vegetation due to lack of agricultural interest. Therefore, crop areas were juxtaposed with traditional locust reproduction areas, creating a situation which favoured crop damage. Locust development was not enhanced; rather crops invaded traditional wandering zones.

Ground and current occupation maps of the land show very clearly the juxtaposition of the locust habitats and of the cultivated areas, the latter being truly "surrounded", in numerous places, by the locust breeding biotopes. They can reproduce with impunity in their traditional environments and later slowly invade the crops, either as hopper bands or as swarms.

The finding that reproduction areas were juxtaposed with the crop areas, along with the discovery of the decreased capacity of swarm movement, shows that the locusts that threaten crops develop in their proximity, and they do not come from distant areas. The accurate cartography of locust habitats that was made within the project framework allows us to estimate the potential risks as a function of the proximity and importance of the locust breeding areas for each "fazenda" (farm).

3.2.8. Sexual maturation and outbreaks causes

The previous results show that the introduction of agriculture has been either neutral or unfavourable for the locust. The cause of the outbreaks should be viewed from another point, the annual variability of ecological conditions.

The study of R. Schistocercoides' sexual maturation has taken on very specific importance since it has permitted understanding certain key aspects associated with the cause of the outbreaks. Precise studies of the ovarian function in nature have shown that sexual maturation of populations appears to be induced independently of rainfall conditions. There apparently exists an imaginal diapause, with a fixed starting date. Sexual maturation appears to be regularly initiated in the second half of August, even before the first rains [16]. Such photoperiodic determination of sexual maturation has important implications for the of outbreaks. If maturation must begin at a fixed date, locust populations can find themselves in ecological conditions radically different from one year to the next. The dry-season/rainy-season transition period, in August and September, is effectively the season in which the ecological conditions are the most variable. Specifically, this is the case for rainfall and the magnitude of bush fires. Their annual variability should have considerable consequences for survival in the dry season, fecundity of the females, and completion of embryonic development. It must explain to a large extent the level of locust populations and the outbreak cycles.

These considerations have direct, practical implications on the organisation of control strategies. In view of prevention, evaluating the ecological conditions in the pivotal period during transition of the dry season to the beginning of the rainy season seems to be of utmost importance for determining a level of locust populations before they start invading the crops a few months later.

124

4. Solution: a new control strategy

The previous summary of the principle research results shows that it has been possible, within a few years of field study, to obtain a set of scientific data to better understand the phenomenon of locust outbreaks in Mato Grosso. The concepts from the beginning to the end of the research project have completely changed the view of the phenomenon (Table 2). Thus, a review of the strategy and of the means of management of the R. schistocercoides populations can be proposed.

The new control strategy is based on the set of research results described above and mainly on the fact that the problem seems to be primarily local and not tied to the recent development of agriculture. The locusts invading the fields do not arrive from hundreds of kilometres away, as it was believed before. They come from nearby areas of natural vegetation which generally are not frequented by farmers.

This new strategy intends to be fundamentally preventative [29,37]. It is based on precise cartographic knowledge of the locust habitats and on localized control actions. These actions must be agreed upon by producers in a region and performed as prevention in natural vegetation areas against hopper bands, before the locusts enter the crops. Through these timely and regular actions, the strategy focuses on locally reducing the level of these populations.

Therefore, the strategy to be implemented does not consist in trying to globally reduce the level of populations in the Mato Grosso area. There is no intention to "finish" with the locusts and exterminate them, a goal that was declared only a few years ago. On the contrary, it deals with implementing an integrated control strategy that allows farmers to co-exist with a species that inhabited Mato Grosso long before this region had its name.

Current treatment techniques (products, materials, application conditions) were also reviewed. Specifically, in order to carry out these preventative control operations, the use of bio-pesticides seems to be a good substitute for traditional insecticides due to various reasons:

• the preventative strategy implies actions outside of crop areas, therefore knockdown treatment is not needed;

• rainfall and humidity are elevated, thus facilitating germination of fungal spores (2000 mm of rain per year; relative humidity from 80 to 100%);

• duration of hopper development (more than 5 months) allows monitoring and control operations to precede any emergency situation.

Experiments were carried out, and the first results have shown good efficiency of a local strain of Metarhizium anisopliae var. acridum [11, 32]. The use of such a myco-pesticide should offer an elegant solution and is tailor-made for the Mato Grosso locust problem. It integrates well into the prescribed strategy, it respects the environment, and it addresses the interests of farmers and indigenous popUlation.

125

TABLE 2. Old concept of the locust problem in Mato Grosso and facts established by recent studies, according to Miranda et al. [37].

Old concept

New phenomenon

Outbreaks are a result of an ecological disequilibrium (deforestation, decrease of natural enemies, creation of new suitable biotopes for the locust).

Swarms have a high migration capability.

New concept

Old phenomenon

Outbreaks are the result of a sequence of favourable ecological conditions (good rains by the end of the dry season I beginning of the rainy season) inducing good breeding and low mortality for eggs and hoppers. Favourable impact of bushfires.

Swarms have a low migration capability.

Locust species with a phase transformation Grasshopper species without any phase phenomenon. phenomenon.

Invasion, beginning in 1983, apparently progressing eastward. Origin of the invasion in the Indian Parecis reservation.

Suitable biotopes for the locust are not known. Some people suppose this species is from humid areas, others that dry areas are more suitable.

No invasion; the locust has been distributed from old time on the whole Chapada dos Parecis campo-cerrado areas. Local outbreaks may occur everywhere in suitable areas.

Locust biotopes are well known. Savannah areas of Chapada dos Parecis, known as "campo" and "campo-cerrado," are the most suitable areas. Two main categories are recognized:

breeding habitats on sandy soils;

wandering habitats on heavier soils.

Indian reservations are source areas for outbreaks Indian reservations are just one of the several and invasions. natural areas for the locust to breed.

Regular eastward migration. Swarms wandering with the wind, mostly in a

N-S I SoN orientation.

Swarms are a threat for other Brazilian States Outbreaks are only a local phenomenon. Swarms (Goias, Tocantins, Sao Paolo ... ). from Mato Grosso are not a threat for other

The control strategy is curative and directed to stop the eastward progression of the invasion.

Brazilian states. However, as the species has a large geographic distribution, small local outbreaks may occur in many places according to local ecological conditions (eventually in Goias ... ).

The control strategy must be preventative and directed to the elimination of the hopper bands at an early stage within the outbreak areas, not far from the crop areas.

126

5. Conclusions

In Brazil, facing an apparently new locust problem in the recently developed areas of Mato Grosso, the development of field ecological research has allowed better and quicker understanding of the problem and creation of a control strategy for locust populations that combines efficiency and maximum respect for the environment. The latter is derived from more localized actions, prevention rather than cure, and the possible use of non-chemical products that are expected to be available soon.

In this respect, the general objectives of the research undertaken within the framework of the control of R. schistocercoides are in line with the current general objectives of operational acridological research in the world, i.e., limit treatments, develop preventative strategies, and find credible alternatives to traditional chemical insecticides.

This work shows once again that a sound control strategy must be based on accurate knowledge of the insect and good understanding of the causes of the outbreaks. Simple field ecology studies can quickly provide the elements of an answer to the most important questions. Based on these field studies, various new tools will be revealed as being important, as was the case for satellite remote sensing, which provided an overall spatial view of the problem. It is likewise the case with geographic information systems that allowed us to study the relationships between diverse, key ecological factors and establish a map oflocust habitats and potential risk zones. These measures provide a better understanding ofthe ecological complexity within which the locust has evolved, and allow study of different scales ranging from the habitat to the entire distribution range.

The acquisition of scientific knowledge is therefore fundamental. Nevertheless, one of the main problems which we face in the battle against the locust plague in Brazil, as in any other place, is not so much scientific as it is political and institutional.

The greatest risk is certainly that of to continuing to respond to the locust problem only in crisis periods and not doing anything beyond these emergency operations. This has very often been the case up to now. Administrative and fmancial authorities seem to be interested, unfortunately, in the locust problem only under emergency situations, when the swarms are highly visible (Figure 7). On other continents, the last large outbreak of the Desert locust [15, 17, 47] and the last invasion of the Migratory locust [19] show us that acting only during an emergency leads to situations which are very difficult to manage and to catastrophes regarding organizational logistics and the environmental impact of control operations. Monitoring and preventative control actions must be maintained permanently, outside of crisis periods. Likewise, long-term funding must be maintained in order to carry out research designed to improve preventative strategies and to favour updating new, less harmful methods. Only at this price can we better control locust and grasshopper outbreaks without simultaneously risking environmental disaster.

6. References

Plague/recession vs

motivation/disinterest

Motivation : lot of money,

equipements, emergency operations,

new researches

Disinterest: Locust organisation non

operational Few research es

Figure 7. The vicious cycle of the locust problem.

Problem forgotten or considered as

resolved

127

I. Barrientos, L.L. (1995) The present state of the locust and grasshopper problem in Brazil, Journal of Orthoptera Research 4,61-64.

2. Carbonell, C.S. (1988) Rhammatocerus schistocercoides (Rehn, 1906), especie per judicial para la agricultura en la region centro oeste de Brasil (Orthoptera, Acrididae, Gomphocerinae), Boletim do Museo Nacional de Rio de Janeiro, Zoologia 318, 1-17.

3. Cardenas, G.D., and Devia, H. (1995) EI problema del saltamontes Rhammatocerus schistocercoides plaga en pasturas nativas, introducidas, cultivadas e importancia de con troles biol6gicos naturales existentes en el altillanura colombiana, in XUe Congresso nat. de ciencias biol6gicad (oct. 8 alii, 1995), Association Colombiana de Ciencias Biol6gicas, Bogota, pp. 47-49.

4. Cosenza, G.W. (1977) Uso da apJicayao aerea e terrestre de inseticidas para 0 controle do gafanhoto em Minas Gerais. An. Soc. Entomol. Brasil, ltabuna, Bahia 6, 295-300.

5. Cosenza, G.W. (1985) 0 gafanhoto ataca novamente. Revista brasileira de Extansiio Rural, Brasilia 6,28.

6. Cosenza, G.W. (1987) Biologia e controle do gafanhoto Rhammatocerus sp., EMBRAPA-CPAC (Documentos 25), Planaltina, DF, Brazil.

7. Cosenza, G.W., Curti, 1. 8., and Paro, H. (1990) Comportamento e controle do gafanhoto Rhammatocerus schistocercoides (Rehn, 1906) no Mato Grosso. Pesquisa agropecuaria brasileira, Brasilia 25, 173-180.

8. Cosenza, G.W., Ribeiro, 1.G.B., and Carvalho, 1.S. (1994) Programa Nacional de Controle ao Gafanhoto. Manual Tecnico. EMBRAPA-SPI, Planaltina, DF, Brasil.

9. Curti, J.B" and Britto, 1.S. (1987) National Program of locust control. Ministerio da AgriculturalSDSV, Brasilia.

128

10. Duque, J.E. (1996) Reunion tecnica regional sobre biologia y control de la langosta (Rhammatocerus schistocercoides). Cuiabti. Brazil. 6 a 10 de mayo de 1996. IICA, Lima, Peru.

II. Faria, M.R., Almeida, D.O., and Magalhat:s, B.P. (1999) Food consumption of Rhammatocerus schistocercoides Rehn (Orthoptera: Acrididae) infected by the fungus Metarhizium jlavoviride Gams & Rozsypal. Anais do Sociedade Entomolagica do Brasil 28,91-99.

12. Hoechne, F.C. (1951) Mato Grosso, contraste do seu nome. Relatario anual do Instituto de Bottinica referente ao exercicio de J 949, Silo Paulo, Brazil, pp. 45-51.

13. Hunter, D.M., and Cosenza, E.L. (1990) The origin of plagues and recent outbreaks of the South American locust, Schistocerca cancel/ata (Orthoptera : Acrididae) in Argentina. Bulletin of Entomological Research 80, 295-300.

14. Kuffner, J.R. (1988) A campanha de combate ao gafanhoto. Delegacia federal de agricultura em Mato Grosso, Cuiaba, Brazil.

15. Launois-Luong, M.H., and Lecoq, M. (1988) Une nouvelle invasion du Criquet pelerin, Agritrop 112, 83-95.

16. Launois-Luong, MH, and Lecoq, M. (1996) Sexual Maturation and Ovarian Activity in Rhammatocerus schistocercoides (Orthoptera : Acrididae) a Pest Grasshopper in the State of Mato Grosso in Brazil, Environmental Entomology 25, 1045-1051.

17. Lecoq, M. (1991) Le Criquet pelerin. Enseignements de la derniere invasion et perspectives offertes par la biomodelisation, in A. Essaid (eds.). La lutle anti-acridienne. AUPELF-UREF, John Libbey Eurotext, Paris, pp. 71-98.

18. Lecoq, M. (1991) Gafanhotos do Brasil. Natureza do problema e bibliografia. EMBRAPAlNMA, Campinas, Brazil and CIRAD/PRIFAS, Montpellier, France.

19. Lecoq, M. (1998) Lutte contre les invasions acridiennes : les le~ons du (mauvais) exemple malgache, Marches Tropicaux 2741, 1094-1095.

20. Lecoq, M., and Assis-Pujol, C.V. (1998) Identity of Rhammatocerus schistocercoides (Rehn, 1906) Forms South and North of the Amazonian Rain Forest and New Hypotheses on the Outbreaks Determinism and Dynamics [Acrididae, Gomphocerinae, Scyllininij, Transactions of the American Entomological SOCiety 124, 13-23.

21. Lecoq, M. and Pierozzi Jr, I. (1994) Rhammatocerus schistocercoides (Rehn. 1906). criquet ravageur des Etats du Mato Grosso et du Rondonia au Bresil. Essai de synthese bibliographique. Cirad-GerdatPrifas, Montpellier, France.

22. Lecoq, M. and Pierozzi Jr, I. (1994) Les stades 1arvaires de Rhammatocerus schistocercoides (Rehn 1906) [Orthop. Acrididae Gomphocerinaej, criquet ravageur de l'Etat du Mato Grosso, au Bresil, Bulletin de la Societe entomologique de France 99,447-558.

23. Lecoq, M. and Pierozzi Jr, I. (1995) Rhammatocerus schistocercoides locust outbreaks in Mato Grosso (Brazil) : a long-standing phenomenon, The International Journal of Sustainable Development and World Ecology 2, 45-53.

24. Lecoq, M. and Pierozzi Jr, I. (1995) Attaques de Prionyx thomae (Fabricius, 1775) (Hymenoptera, Sphecidae) sur un criquet ravageur, Rhammatocerus schistocercoides (Rehn, 1906) au Bresil (Orthoptera, Acrididae), Bulletin de la Societe entomologique de France 100, 515-520.

25. Lecoq, M. and Pierozzi Jr, I. (1995) Le criquet du Mato Grosso: I'agriculture est-elle responsable ? Tropicultura 13, 32-33.

26. Lecoq, M. and Pierozzi Jr, I. (1996) Chromatic Polymorphism and Geophagy: Two Outstanding Characteristics of Rhammatocerus schistocercoides (Rehn 1906) Grasshoppers in Brazil [Orthoptera, Acrididae, Gomphocerinae j, Journal of Orthoptera Research 5, 13-17.

27. Lecoq, M. and Pierozzi Jr, I. (1996) Comportement de vol des essaims de Rhammatocerus schistocercoides (Rehn, 1906) au Mato Grosso, Bresil (Orthoptera, Acrididae, Gomphocerinae), Annales de la Societe entomologique de France 32,265-283.

28. Lecoq, M., Foucart, A. and Balanca, G. (1999) Behaviour of Rhammatocerus schistocercoides (Rehn, 1906) hopper bands in Mato Grosso, Brazil (Orthoptera: Acrididae, Gomphocerinae), Annales de la Societe entomologique de France 35,217-228.

29. Lecoq, M., Miranda E. E. de, and Pierozzi Jr., I. (1997) A new approach to the control of Rhammatocerus schistocercoides (Rehn, 1906) in Brazil, in : S. Krall, R. Peveling and Ba Daoule Diallo (eds.), New strategies in locust control, Birkhlluser Verlag, Basel, Suisse, pp. 505-506.

129

30. Le6n, M.G. (1996) EI grillo de los llanos orientales. Biologia, habitos y recommandaciones para su manejo. Boletin Tecnico abril de 1996, Corporaci6n Colombiana de Investigaci6n Agropecuaria. CORPOICA REGIONAL 8, Villavicencio, Colombia.

31. Levi-Strauss, C. (1955) Tristes Tropiques. Pion Ed., Paris. 32. Magalhlles, B.P., Faria, M.R., Tigano, M.S., and Sabral, B.W.S. (1997) Characterization and virulence of a

Brazilian isolate of Metarhizium flavoviride Gams & Rozsypal (Hyphomycetes), Memoirs of the Entomological Society of Canada 171,313-321.

33. Martinez, O. J., and Gomez, J.J. (1996) EI grillo. Deteccion del grillo de los pastos en los llanos y pautas para su manejo integrado. Instituto Colombiano Agropecuario and Fondo Nacional del Ganado, Bogota.

34. McCulloch, L. (1986) Control ofRhammatocerus swarms in Mato Grosso State, Brasil. Interim report and recommendations. FAOrrCP/4508/BRA report. Food and agriculture organization of the United Nations, Rome.

35. McCulloch, L. (1986) Locust control in Mato Grosso state, Brasil. Final Consultancy Report to FAO. FAOrrCP/4508IBRA report. Food and agriculture organization of the United Nations, Rome.

36. Miranda, E.E. de, Pierozzi Jr, I., and Lecoq, M. (1996) Bi6topos de Rhammatocerus schistocercoides (Rehn, 1906) Gafanhoto Praga do Mato Grosso. EMBRAPA-NMA I ECOFOR<;A I ClRAD-GERDATPRIFAS ICE, Campinas, Brazil, 2 polychrome maps, scale: 1:250.000.

37. Miranda, E. E. de, Lecoq, M., Pierozzi Jr, I., Duranton, J.-F., and Batistella, M. (1996) Le Criquet du Mato Grosso. Bi/an et perspectives de 4 anm!es de recherches: 1992-1996. Ropport final du projet « Environnement et criquets ravageurs au Bresi/ ». EMBRAPA-NMA, Campinas, Bresil I CIRADGERDAT-PRIFAS, Montpellier, France.

38. Miranda, E. E. de, Pierozzi Jr, I., Batistella, M., Duranton, J.-F. and Lecoq, M. (1994) Static and dynamic cartographies of the biotopes of the grasshopper Rhammatocerus schistocercoides (Rehn, 1906) in the state of Mato Grosso, Brazil. Revista SELPER, Technical review for ibero-american and worldwide integration 10,67-71.

39. Miranda, E.E. de, Pierozzi Jr, I., Lecoq, M., Duranton, J.F., and Batistella, M. (1996) Unidades de vegetacilo. EMBRAPA-NMA I ECOFOR<;A IClRAD-GERDAT-PRIFAS ICE, Campinas, Brazil, 8 polychrome maps, scale: 1:250.000.

40. Oliveira, S. (1987) Operaciio gafanhoto, Globo rural 3,82-93. 41. Pierozzi Jr, I., and Batistella, M. (\995) Gafanhotos via satelite, Fator GIS II, 36-37. 42. Pierozzi Jr, I. and Lecoq, M. (1998) Morphometric Studies on Rhammatocerus schistocercoides (Rehn,

1906) [Orthoptera, Acrididae, Gomphocerinaej in Brazilian and Colombian populations, Transactions of the American Entomological Society 124,25-34.

43. Rehn, J.A.G. (1906) Notes on South American grasshoppers of the subfamily Acridinae (Acrididae), with descriptions of new genera and species, Proc. U. S. Nat. Mus. 30, 371-391.

44. Rondon, C.M. da S. (1919) Relatorio apresentado a Diretoria Geral dos Telegrafos e a Divisiio de Engenharia do Departamento da Guerrapelo Tte. Coronel Candido Mariano da Silva Rondon, Chefe da Commissiio. (2° volume) Construylio (1907-1910), Publ. n° 39, Pap. Macedo, Rio de Janeiro.

45. Roquette-Pinto, E. (\ 975) Rondonia, Brasiliana vol. 39, Compo Edit. Nac., Silo Paulo, Brazil. 46. Santos, J. (1987) Gafanhotos devastam a lavoura do Mato Grosso. 0 Globo / Economia (17 August

1987), pp. 13. 47. Showier, A.T. and Potter, C.S. (1991) Synopsis of the 1986-1989 desert locust (Orthoptera: Acrididae)

plague and the concept of strategic control, Amer. Entomol. 37, I 06-110. 48. Zahler, P.M. (1987) Agricultura vs Ecologia no combate ao gafanhoto Rhammatocerus, Ciencia e

Cultura, Silo Paulo 39,703-706.

8. WHAT ARE THE CONSEQUENCES OF NON-LINEAR ECOLOGICAL INTERACTIONS FOR GRASSHOPPER CONTROL STRATEGIES?

Abstract

A.JOERN School of Biological Sciences University of Nebraska-Lincoln Lincoln, NE 69599-0118, USA

Grasshopper populations from grasslands have been notoriously difficult to predict, a prerequisite for developing forecasting methods. In part, lack of good studies that uncover mechanisms driving population dynamics has contributed to this state of affairs. More importantly, grasshopper population dynamics may not follow the simple climate-driven, niche-based models that underlie most current explanations and control strategies. More likely, non-linear responses resulting from combined effects of abiotic forces (climate) and biotic interactions (e.g., competition, predation, parasitism) cause this unpredictability. A variety of examples is provided and a synthesis of new possible directions is described.

1. Problem

Complexes of grasshopper populations wax and wane in abundance, sometimes reaching sufficiently high densities to impact human resources. In parts of the world, legendary swarms of phase-polymorphic locusts engulflarge areas and cause great economic damage before retreating to a fraction of the outbreak area and population size. In North America, fortunately, no phase polymorphic grasshopper species are known that exhibit such dramatic shifts in life cycle, population surges, or corresponding economic damage. However, grasshopper populations periodically reach high densities with resulting loss of forage causing economic impact [12]. In each of these situations, multiple factors collaborate to limit and regulate grasshopper populations [31]. In many cases, significant differences exist concerning the underlying ecological processes that actually apply, often making it difficult to assign causal relationships. Whatever the details, these multiple, natural processes critically affect the successful development of control policies and their success, but in reality are typically ignored as a serious backdrop for decision making.

Federal government sanctioned grasshopper control programs in North America emphasize western grassland regions, programs that focus on forage for cattle in rangeland. Agricultural crops in this region can also be at risk, but control of grasshoppers on crops

131


132

is typically left to individual farmers. Over the years, grasshopper control efforts in North America readily embraced the development of new technologies. Even so, this history is relatively short (since ca. 1880). Until the development of broad-spectrum chemical control capabilities, most technologies were environmentally benign, labor intensive and local in their impact. However, the problem captured national attention from the beginning. Grasshoppers and the need for their control were instrumental in the establishment of professional entomology in the United States in the late 1800s in response to widespread outbreaks of the Rocky Mountain Locust (Melanoplus spretus). However, even after this species went extinct early in the 1900s, other periods of significant, widespread grasshopper damage periodically occurred. Early control efforts relied on a variety of mechanical contraptions, simple chemical control including natural poisons and moats filled with fuel oil, or natural control agents. For example, an entire agency within the USDA was initially established to study and evaluate the importance of birds to insect pest control (including grasshoppers), although this group disbanded in the 1930s.

With the development of synthetic insecticides in mid-century, grasshopper control in North America shifted from local application to the wide-ranging, broad spectrum chemical control of grasshopper populations, with insecticides aerially sprayed over large areas (minimum 4 000 ha blocks). Recent chemical control guidelines were developed with application efficiency in mind, given the large expense and effort needed to apply insecticides. Costs were typically shared equally by rancher, state and federal governments on private land and by federal or state governments alone on public areas, a large fraction of the total land area in the western USA. Control action was considered when populations reached > ~9 individuals/m2 although higher densities typically were needed to actually initiate action. All grasshoppers were included in the counts until recently despite the fact that only a tiny fraction of these species can be considered economically important; no effort was made to adjust estimates to include only those species that were likely to cause economic damage or to conditionally adjust risk offorage loss according to species composition. Moreover, no real effort was made to either assess the actual success of chemical methods on long-term grasshopper control, or to evaluate the impact of grasshopper chemical control technologies on non-target, often beneficial arthropod species, including the natural enemy complex that could be important in limiting grasshopper numbers in most years. After the demise ofthe "bird unit", little thought was given to environmental risks in government circles except by some thoughtful USDAI ARS research scientists stationed at the Bozeman, MT lab (sometimes at career risk). What critical environmental risks to large-scale chemical control of grasshoppers exist in western North American range? What more needs to be done to pursue these issues?

Most current grasshopper control efforts in North America are based on a very simple underlying ecological model of grasshopper population dynamics. Unspecified environmental conditions (usually directly reflecting climate) lead to the rapid build-up of grasshopper densities using an exponential popUlation growth model, requiring a 2-3 year period of a density buildup. At this point, grasshopper populations may cause sufficient damage to be economically important. The contribution of density-dependent or frequency-dependent factors, including more complex responses reflecting non-linear interactions among species within and among trophic levels, are seldom acknowledged.

133

Focused only on grasshopper population complexes, current strategy relies on repeated knock-down of population numbers, a successful approach that awaits the build-up of populations during periods of suitable environmental conditions until they once again require control. If the exponential population growth model applies to natural populations, this response is legitimate with regard to the specific goal: controlling grasshopper populations. The model simplifies ecological reality in a way that accommodates simple approaches to grasshopper control.

But, what ifthis simple underlying model is not an appropriate framework within which to design control? Will control efforts contribute to the problem? What ifnatural enemies are critically important to natural limitation and control? If so, not only does the current chemical-control strategy not work, it has negative environmental impact because broad-spectrum insecticides kill many native arthropods. Future grasshopper control efforts in North America must be sensitive to the possible degradation of biological diversity at local and regional scales, both because of current political and public pressure and because natural checks and balances can only be maintained by keeping these players in the system.

Recent historical analyses of grasshopper control suggest that the current model underlying control is incorrect, and that continued use of this model to justify large-scale grasshopper control may contribute to grasshopper problems. Long-term grasshopper population responses were compared between Wyoming with little chemical control and Montana with significant chemical control on government lands [34]. In Wyoming, where large-scale chemical control was not regularly applied, acute problems rapidly disappeared and long intervening periods of low grasshopper density persisted. Montana populations, on the other hand, exhibited chronic, long-term increases in grasshopper population densities after a history of control with associated economic significance. Despite that fact that these two regions are adjacent, contain essentially the same habitat, and experienced the same general climatic conditions during this period, very different population trajectories were observed after the period with extensive chemical control in Montana. More importantly, population responses from the two sites were the opposite of what one expects if the basic population model underlying control efforts is true.

Interactions among grasshopper species with other species in other trophic levels, and their non-linear consequences are largely ignored in the current ecological model underlying North American grasshopper control programs. In part, this reflects a lack of commitment by federal agencies responsible for grasshopper control programs to consciously incorporate basic ecological understanding of grasshopper population dynamics. Most funding has been directed at either chemical methods of control or bioinsecticides, and North American agencies have been quite successful in developing these methodologies. Less effort is directed at their long-term effectiveness and essentially no effort is allocated to the environmental impact of these policies. The irony here is that of these three important research priorities, two issues require basic understanding of grasshopper population processes in North American range systems. Clearly more research is required to work out the critical details, and future environmental risk from grasshopper control depends on these considerations.

134

I would like to emphasize the following take-home points:

(1) Successful control strategies must act in concert with natural processes or they wiII either fail at best and possibly exacerbate the problem at worst. In this context, better information is needed at all stages of the process, ranging from the basic understanding of what is going on in terms of species-specific population dynamics and multiple species assemblies, to having access to accurate data on which species are contributing to current conditions and why.

(2) New conceptual models of grasshopper population and community responses are needed that incorporate important non-linear interactions among species, with an emphasis on developing critical understanding at different temporal and spatial scales, including their linkage. Intense research should be directed at understanding how multiple sources of mortality contribute to basic population processes, including those resulting from mUltiple species interactions, within and among trophic levels.

(3) Natural enemies play an important role at all levels. While we are not always certain exactly what dynamic role these species play under specific but variable environmental conditions, it is certain that they are important. Moreover, it is certain that natural enemies play different dynamic roles under different environmental conditions and that different sites will exhibit different responses.

(4) Models that exhibit multiple equilibria with thresholds that define domains of attraction at low and high densities are currently attractive from my perspective. If such models result from complex density- and frequency-dependent, non-linear interactions among species and multiple domains of attraction, standard time-series analysis to search for density-dependent relationships will often fail to either project future change or uncover underlying ecological mechanisms.

(5) Protecting the biodiversity of native species must be a main consideration in developing grasshopper control programs in North American western rangeland. Not only will incorporation of this goal typically contribute to long-term, effective grasshopper management, but basic public concerns for protecting biodiversity can be met simultaneously.

2. Tools: A Conceptual Framework of the Population Problem

Basic assumptions underlying insect popu lation dynamics, including that for grasshoppers, often reflect basic propositions of Uvarov [55], ideas generalized by Andrewartha and Birch [I]. In this "climatic release hypothesis" [36, 56], good and bad environmental periods occur because of changes in climate, affecting population growth parameters for a period of time until conditions switch. The length of this period determines if insect populations increase or decrease, and to what extent. Climate is a conspicuous possible determinant for such population responses, and it certainly plays a role no matter what the eventual mechanism underlying insect population dynamics proves to be. This view has utility in that it provides insect ecologists and economic entomologists with a simple, easily-recorded set of environmental variables that can be used to predict grasshopper

135

population changes. In addition, an unstated assumption that typically accompanies this view is that the process responsible for grasshopper population outbreaks is physiologically driven. Since grasshoppers are ectotherms, their key basic physiological processes are temperature driven and changes in climate can alter the ability to extract nutrients, and then allocate these to basic demographic attributes that affect fitness and population growth, among other possibilities. How far does this logic take us for understanding grasshopper population dynamics?

2.1 DIRECT EFFECTS OF CLIMATE

Repeated efforts have correlated grasshopper population changes with weather variables, including lagged effects [14, 21, 25, 35]. Significant correlations between weather variables and grasshopper population responses are often uncovered. As a rule of thumb, population densities tend to increase with hot, dry weather in northern US grasslands and decrease with these same conditions in southern plains [12, 14]. However, despite the significant relationships, correlations between climate variables and populations typically explain only a small fraction ofthe variance associated with the relationship, usually less than about 30%; the 70% or so ofthe variance in the relationship unexplained. It is almost always acknowledged that weather data from standard recording stations are used for convenience. Moreover, the data themselves are only correlated with the actual environment occupied by individual insects, especially given the regional rather than local nature ofthe data. Thus, significant correlations probably do reflect real relationships. At issue, however, is the causal nature ofthe interaction underlying the relationship. Unless underlying ecological mechanisms can be described, the relationship between grasshopper population change and weather provides only a tenuous link with which to direct control programs, even if the underlying dynamics eventually prove to be logically simple and straightforward. Effective control programs will work only when they operate in concert with natural mechanisms and processes!

Several examples document the important influence of climate on mechanistic population processes, including the clear fact that temperature does playa role! For example, food processing strongly interacts with temperature [60]. Coxwell & Bock [1995] documented a positive relationship between grasshopper abundance and distribution and diurnal surface temperatures in alpine situations. Ritchie [46] finds a significant correlation between grasshopper density and food quality, but only in warm years. In a food web context, Chase [15] documented how thermal conditions profoundly alter trophic cascades. And finally, Carruthers et al. [14a] document the adaptive consequences of increasing grasshopper body temperature to thwart disease and increase survival. There is no question that abiotic factors emanating from a variable climate are important. At issue is the need to understand exactly how they influence dynamic responses by grasshoppers at the population level in a varying environment.

2.2 CLIMATE AND FOOD PLANT QUALITY

Climate also directly affects food plant quality, thus indirectly influencing grasshopper

136

population dynamics. Stressed food plants often contain higher concentrations of key nutrients, especially N-containing compounds such as protein, that are often limiting to insect herbivores [37,38,57,58]. There may be a direct link between climatic variation, food quality and grasshopper population dynamics. If so, increased food quality under extreme environmental conditions may lead to increased survival and reproduction, with consequent increases in grasshopper population size.

Physiological, behavioural and demographic responses in grasshoppers are sensitive to changes in food quality, especially N [3,4,27,29,30,60-62]. These various responses are consistent with predictions from a nutrient quality hypothesis of population dynamics. Densities of grass-feeding grasshoppers often increase with increased fertilization levels under otherwise natural conditions [27]. At the behavioural level, grasshoppers select food based on the presence of limiting amino acids in a remarkably specific manner [3, 4], among many other factors that affect diet selection [9]. Grasshopper demographic responses often respond positively to total-N content of food, but not always [29, 30]. For all grasshopper species studied, maximal egg production rate occurred at ca. 4% total-N for the diets used, mostly reflecting the rate at which egg pods were produced and laid rather than the number of eggs per pod. Survival, development and weight gain often respond positively to increased total-N in the food, although this can be highly species-specific [29, 30]. Food processing capabilities strongly vary with food quality [60-62], compensating for suboptimal diets. However, the main point of these examples is that food quality has repeatedly been shown to influence physiological and demographic responses at the individual level. The next critical question is whether these individual-level responses translate into population-level responses.

Have we identified the key mechanistic link between grasshopper population dynamics and climatic variation based on food quality as just described? If true, two key predictions must be satisfied. The first is that there is a link between physiological stress and plant nutritional quality. This appears to be satisfied, even for the system in which the above species occur (Joem & Mole, unpublished data). Second, there should be strong positive relationships between increased N-concentrations in host plants and population size. Table I provides correlation coefficients for the most common species for 10 years of available plant quality data from a sandhills grassland site (Arthur Co., Nebraska, USA). Significant correlations exist but most explain relatively small proportions of the observed variance in grasshopper population size. Moreover, many of the correlations between population size and average total-N are negative (although not necessarily significantly different from zero), counter to predictions of the plant stress-nutrient quality hypothesis. Like investigations of the direct relationships between climate and grasshopper population dynamics, critical mechanisms are ignored in this type of analysis.

3. Tools: A Conceptual Framework of Dynamic Interactions in Food Webs

Grasshopper populations do not occur in an ecological vacuum, and species from many trophic levels routinely interact in a variety of ways: plant-plant, plant-grasshopper, grasshopper-grasshopper, and grasshopper-natural enemy. Basic density- and frequency-

137

dependent processes guide most of these species interactions within and between trophic levels. Moreover, a wide variety of direct and indirect responses that affect grasshopper

TABLE I. Pearson product moment correlations between adult grasshopper densities and plant nutritional quality. In these analyses, plant quality was assessed as Total-N and Total Nonstructural Carbohydrates (TNC) based on average values tor 7 grass species for all seasonal measures. Data were obtained from a

sandhills grassland in Nebraska (Arapaho Prairie, USA) over a ten year period

Species Total-N TNC

All grasshoppers -0.30 0.21 Grass-Feeders -0.10 0.08 Ageneoteltix deorum 0.05 0.10 Amphitornus coloradus -0.13 -0.08 Melanoplus angustipennis -0.31 0.39 Melanoplusfoedus -0.51 0.33 Melanoplus sanguinipes -0.29 0.67 Opeia obscura -0.51 0.33 Phoetaliotes nebrascensis -0.23 0.Q3

populations cannot be detected and certainly not predicted except within this complete set of interactions. These relationships in all of their complexity are seldom recognized much less accommodated by the prevailing grasshopper control policies in North America. This may prove to be a big mistake if grasshopper management unknowingly disrupts critical links.

3.1 COMPETITION

Naturally occurring species are embedded in food webs in all natural communities [45]. Many grasshopper species (20-50 species) regularly coexist at a single site, providing significant opportunities for interactions, including competition [7, 16, 17, 20, 43] for limited food. Contrary to previous views [54], it is now recognized that insect herbivores regularly compete [20] and many opportunities exist for competition among grasshoppers in some years when quality food is limited. Ironically, the complex of relatively uncommon, non-economically important grasshoppers may keep damaging species in check by the diffuse effects of competitive interactions from many species [7]. However, competition among grasshoppers is variable and may not occur at all times and places [22, 23].

3.2 PREDATION

Natural enemies have widespread impact in most natural communities [18]. Grasshoppers are routinely preyed upon by a long list of arthropod and vertebrate predators [6, 8, 11, 28, 31, 32, 41, 42]. Many fungal, bacterial and viral pathogens attack grasshoppers as well

l38

[53]. This combined natural enemy complex is well recognized but largely underappreciated in terms of its importance for the limitation and regulation of North American grasshopper populations. Recent results from a variety of studies indicate the importance of understanding grasshopper population dynamics in the context of food web dynamics if we are to understand primary sources of mortality and how these sources of mortality interact [2, 15,47,48,49,50]. Many of these responses will be strongly convoluted and difficult to predict based on correlation analysis and simple linear responses.

Many studies have documented the role of natural enemies in limiting grasshopper populations [6, 11,24,28,48,50], although this role is not always easy to demonstrate [7]. In Nebraska sandhills grassland (USA), arthropods and vertebrates provide significant impact [28, 32, 41, 42]. Specifically, nymphs are most susceptible to predation from wandering spiders. Moreover, while nymphs experience significant predation pressure, they also experience many other sources of mortality, with younger nymphs most affected. Adults are primarily taken by large predators such as birds and occasionally robber flies and large orb web spiders [6, 41, 42]. Avian predation on adults is variable in time and space [28]. Direct depressing effects of bird predation on adult grasshoppers are routinely observed in Nebraska grasslands [28, Joern unpublished] but seldom observed in Montana Palousse grassland [6]. However, even when direct effects of bird predation are not observed, the impact on grasshopper dynamics can be quite pronounced, including responses in the Montana system [see multiple stable states below, 5].

3.3 COMPENSATORY VS. ADDITIVE MORTALITY

Multiple sources of mortality undoubtedly combine to limit grasshopper populations. How do these factors combine to determine final densities? When all mortality sources contribute to mortality in a largely independent fashion, mortality is considered additive. However, sources of mortality often interact in a nonlinear fashion where the final population density may reach the same level independent of the number of factors involved, a compensatory response. Understanding how mortality factors interact is critical in that increasing mortality pressure at one life stage may have the effect of decreasing it at a later stage with the consequence that there is no real impact for overall population level dynamics, just where and how mortality occurs. These types of interactions routinely occur in grasshopper popu lations [42]. Food quality ( starvation) and spider predation interact as multiple mortality factors to affect final population sizes. Under ambient food quality conditions, predation and starvation act in a compensatory fashion. However, when plots are fertilized and food quality increases, the impact on mortality is non-compensatory and the effects of both are evident in the final population numbers. In a separate experiment, Oedekoven & Joern [42] showed that survivors in treatments that included spider predation were better able to withstand starvation, suggesting that particular qualities of individuals within a population shift in response to interacting sources of mortality.

139

3.4 TROPHIC CASCADES

Interacting sources of mortality operating within food webs exhibit a series of indirect interactions among trophic levels referred to as trophic cascades [48-50]. Presence or absence ofthe top predator in a food web [chain] results in different impacts on the basal food level [45,48,50]. The key point is that final densities in natural food webs reflectthe combined set of dynamic interactions among populations at different levels of a food web, the combined impact of within trophic level effects of competition, and the between trophic level effects of consumption (both herbivory and predation). Both direct and indirect interactions combine to determine final densities, often in complex ways.

For grasshoppers, the control target should be focused on the plant forage base of the food web, not the absolute number of grasshoppers. If true, and if the presence of predators can increase forage biomass over the long term, all efforts should be focused on maintaining the complete system. Moreover, the nature of interactions among trophic levels in food webs is wildly difficult to predict. Indirect interactions can often be disentangled only through careful experimentation [2,47,48,49,50], and the sources of indirect interactions that contribute to final popUlation densities are highly nonlinear.

3.5 MULTIPLE EQUILIBRIA

The standard explanation for grasshopper outbreaks follows from Andrewartha & Birch [2]: unpredictable but regular changes in climate make environmental conditions more or less suitable for grasshopper popUlation growth. Are there alternate explanations for the existence of sometimes low vs. high grasshopper densities other than merely changing climate? Are population responses more regulated than they appear to be and not merely results of unpredictable environmental conditions?

There is good evidence that complex, nonlinear dynamics account for endemic vs. epidemic population levels in insect herbivores, including grasshoppers [5, 10,44,51, 52]. Basic biotic interactions (e.g., competition, predation, herbivory) in association with climatic state variables can determine response thresholds between grasshopper population responses over short temporal and spatial scales [5]. Threshold boundaries prescribe multiple equilibria, determining which domain the system will lie in. Identifying which ecological factors drive the dynamics of the system and which parameters (and their values) determine the location of thresholds defining domains of attraction is important. Deriving these responses in natural systems becomes an important empirical challenge, especially when trying to manage a grasshopper population. As described in Belovsky & Joern [5], there are good reasons to believe that multiple equilibria occur in North American grasshopper populations, and any feasible control strategy must take this new view into account. Much research remains to be performed on this important topic, however.

140

4. Solutions

North American grasshopper control programs in western rangeland have reached a critical juncture: programs must preserve and conserve natural occurring biodiversity, including the majority of grasshopper species which should not be targeted for control under any situation. Current grasshopper control approaches present significant risk to many species, particularly other arthropods. These risks must be explicitly addressed and increasing public pressure will demand that this be done. Moreover, the primary goal of grasshopper control policy must shift from one of control to one of management, a mindset very similar to that required for managing renewable resources.

4.1 COMPETING ECOLOGICAL HYPOTHESES UNDERLIE CURRENT GRASSHOPPER CONTROL POLICY

Recent research focused on dynamic food web interactions illustrates the complex set of direct and indirect interactions that underlie grasshopper population dynamics. Densityand frequency-dependent interactions among species coupled to unpredictable climatic forcing functions lead to a wide range of largely unpredictable responses. However, the causal mechanisms can be uncovered and their consequences evaluated in the context of intervention policies. Unfortunately, the current population dynamics model underlying past and most current grasshopper control philosophies in North America will not accommodate these new ecological findings. The current model harkens back to Andrewartha and Birch [1], requiring nothing more than exponential population growth and unpredictable weather. This theoretical underpinning is clearly insufficient, even inappropriate, to develop grasshopper control programs and new approaches should be explored now.

Results of these recent studies suggest additional, important outcomes. While predictive models are desired to provide necessary lead time to organize grasshopper control programs, it seems unlikely that truly predictive models of future grasshopper population trends based on the current model can be developed. In particular, the existence ofnon-Iinearities strongly affects population dynamics. These relationships challenge the utility of time series analyses based on sequences of annual population data for either predictive purposes or for purposes of uncovering critical ecological mechanisms responsible for grasshopper population dynamics. Finally, we should shift our focus regarding grasshopper control in North America. Currently we emphasize grasshopper control, presuming that our efforts can control grasshopper outbreaks according to human desires and needs. More considerations must be included in the decision.

Scale-dependent responses for understanding grasshopper population dynamics have not been adequately evaluated and these really need attention. Small and large-scale visions are needed. However, it is clear that the emphasis on the 4 000 ha control block is arbitrary and population dynamics at this spatial scale should be ignored until it can be shown to be meaningful for ecological reasons rather than used in response to the economic needs of control agencies. Large-scale dynamics will probably be meaningful

141

and a scale-dependent spatial hierarchy should be developed that reflects appropriate ecological dynamics.

4.2 AMELIORATION OF RISK TO RANGELAND BIODIVERSITY FROM GRASSHOPPER CONTROL IS GOOD POLICY

Heightened public interest and pressure in North America to preserve biodiversity now requires that grasshopper control strategies on western rangeland protect native species even during periods of grasshopper outbreaks. Public concern for preserving biodiversity will most likely continue to grow. From a political view, this is reality. However, protecting biodiversity will be good management as well, given the important contributions of biotic interactions to the grasshopper population dynamics.

Natural controls that limit and regulate grasshopper populations in rangeland ecosystems must be accommodated. These natural limits are part of any rational grasshopper management program and most chemical control programs disrupt these processes, sometimes permanently. Interspecific competition and effects of natural enemies each have the potential to contribute significantly to grasshopper management programs. It is not yet clear how these processes can be brought to bear on the problem despite the fact that the basic interactions have been studied for decades. Ironically, these critical interactions are only now being described in sufficient detail, in just a few situations. Do otherwise non-economically important grasshopper species compete in such a way that economic species are limited? In what way are natural enemies a critical feature of these interactions? Managing grasshopper populations will require that these important limiting factors be maintained while managing for biodiversity at the same time. This is a win-win situation as both contribute to the same goal. Developing policy that builds on this assumption will provide the necessary balance for preserving biodiversity while appropriately addressing grasshopper issues, with both needs benefiting.

A look to the future is also worthwhile for assessing the relationship between grasshopper management and biodiversity. How will situations change in response to changing land use, orto global climate change? Each of these issues potentially influences grasshopper population dynamics: e.g., fragmentation of habitats that disrupt naturally limiting agents or altered climate that results in changed species distributions or altered temperature-dependent population-level mechanisms. I sense that these issues cannot be addressed based on current knowledge and there is no indication that appropriate actions are under consideration. Significant research on these types is warranted.

In sum, a significant paradigm shift is underway regarding the way in which grasshopper populations in North America will be managed. Past and current control policy now presents a significant risk to non-target native species on native rangeland if it continues as always. Preserving biodiversity in rangeland systems must be included as a major goal in the final package, both because it will retain natural controls on grasshopper populations and because the public demands thatthis be done. Unfortunately, much critical information regarding the basic ecological interactions are unknown. A high priority should be placed on developing the basic ecological framework that underlies grasshopper population dynamics.

142

5. References

I. Andrewartha, H.G. and Birch, L.C. (1954) The Distribution and Abundance of Animals, University of Chicago Press, Chicago.

2. Beckerman, A., Uriarte, M., and Schmitz, OJ. (1997) Experimental evidence for a behaviorally mediated trophic cascade in a terrestrial food chain, Proceedings of the National Academy of Sciences (USA) 94, 10735-10738.

3. Behmer, S.T. and Joem, A. (1993) Dietary selection by the generalist grasshopper, Phoetaliotes nebrascensis (Orthoptera: Acrididae) based on the need for phenylalanine, Functional Ecology 7, 522-527.

4. Behmer, S.T. and Joem, A. (1994) The influence of proline on diet selection: sex-specific preferences by the grasshoppers Ageneotettix deorum and Phoetaliotes nebrascensis (Orthoptera: Acrididae), Oecologia 98, 76-82.

5. Belovsky, G.A. and Joem, A. (1995) The dominance of different regulating factors for rangeland grasshoppers, in: N. Cappuccino and P.W. Price (eds.), Population Dynamics: New Approaches and Syntheses, Academic Press, San Diego, CA, pp. 359-386.

6. Belovsky, G.A. and Slade, J.B. '(1993) The role of invertebrate predators in a grasshopper community, Oikos 68, 193-20 I.

7. Belovsky, G.A. and Slade, J.B. (1995) Dynamics of two Montana grasshopper popUlations: relationships among weather, food and intraspecific competition, Oecologia 101, 383-396.

8. Belovsky, G.A., Slade, J.B., and Stockoff, B.A. (1990) Susceptibility to predation for different grasshoppers: an experimental study, Ecology 71, 624-634.

9. Bemays, E.A. and Chapman, R.F. (1994) Host Plant Selection by Phytophagous insects, Chapman & Hall, NY.

10. Berryman, A.A. (1987) The theory and classification of outbreaks, in: P. Barbosa and J.e. Schultz (eds.), Insect Outbreaks, Academic Press, San Diego, CA, pp. 3-30.

II. Bock, C.E, Bock, J.H., and Grant, M.e. (1992) Effects of bird predation on grasshopper densities in an Arizona grassland, Ecology 73, 1706-1717.

12. Capinera, J.L. (1987) Population ecology of rangeland grasshoppers, in: J.L. Capinera (ed.), integrated Pest Management on Rangeland: A Shortgrass Perspective, Westview Press, Boulder, CO, pp. 162-182.

13. Capinera, J.L. and Thompson, D.e. (1987) Dynamics and structure of grasshopper assemblages in shortgrass prairie, Canadian Entomologist 119, 567-575.

14. Capinera, J.L. and Horton, D.R. (1989) Geographic variation in effects of weather on grasshopper infestation, Environmental Entomology 18, 8-14.

14a. Carruthers, R.I., Larkin, T.S., Firstencel, H., and Feng, Z. (1992) Influence of thermal ecology on the mycosis of a rangeland grasshopper, Ecology 73, 190-204.

15. Chase, J.M. (I996a) Abiotic controls of trophic cascades in a grassland foodchain, Oikos 77, 495-506. 16. Chase, J.M. (1996b) Differential competitive interactions and the included niche: an experimental analysis

with grasshoppers, Oikos 76, 103-112. 17. Chase, J.M. and Belovsky, G.E. (1994) Experimental evidence for the included niche, American

Naturalist 143, 514-527. 18. Crawley, MJ. (ed.). (1992) Natural Enemies, Blackwell Scientific Publications, Oxford, UK. 19. Dempster, J.P. (1963) The population dynamics of grasshoppers and locusts, Biological Reviews 38, 490-

529. 20. Denno, R.F., McClure, M.S., and Ott, J.R. (1995) Interspecific interactions in phytophagous insects:

competition reexamined and resurrected, Annual Review of Entomology 40, 297-331. 21. Edwards, R.L. (1960) Relationship between grasshopper abundance and weather conditions in

Saskatchewan, 1930-1958, Canadian Entomologist 92, 619-624. 22. Evans, E.W. (1989) Interspecific interactions among phytophagous insects in tall grass prairie: an

experimental test, Ecology 70, 435-444. 23. Evans, E.W. (\992) Absence of interspecific competition among tallgrass prairie grasshoppers during a

drought, Ecology 73, 1035-1044. 24. Fowler, A.e. , Knight, R.L., George, T.L., and McEwan, L.e. (1991) Effects of avian predation on

grasshopper populations in North Dakota grasslands, Ecology 72, 1775-1781.

143

25. Gage, S.H. and Muke~ii, M.K. (1977) A perspective of grasshopper population distribution in Saskatchewan and interrelationship with weather, Environmental Entomologist 6, 469-479.

26. Greathead, OJ. (1962) A review ofthe insect enemies of Acridoidea (Orthoptera), Transactions of the Royal Entomological Society of London 114, 437-517.

27. Heidorn, TJ. and Joern, A. (1987) Feeding preference and spatial distribution of grasshoppers (Acrididae) in response to nitrogen fertilization ofCalamovi/fa longifolia, Functional Ecology I, 369-375.

28. Joern, A (1992) Variable impact of avian predation on grasshopper assemblies in sandhills grasslands, Oikos 64, 458-463.

29. Joern, A. and Behmer, S.T. (1998a) Impact of diet quality on demographic attributes in adult grasshoppers and the nitrogen limitation hypothesis, Ecological Entomology 23, 174-184.

30. Joern, A. and Behmer, S.T. (l998b) Importance of dietary nitrogen and carbohydrates to growth, survival and reproduction in adults of the grasshopper Ageneotettix deorum, Oecologia 112, 201-208.

31. Joern, A. and Gaines, S.B. (1990) Population dynamics and regulation in grasshoppers, in: R. F. Chapman and A Joem (eds), Biology of Grasshoppers, John Wiley & Sons, NY, pp. 415-482.

32. Joem, A. and Rudd, N.T. (1982) Impact of predation by the robber fly Proctacanthus milbertii (Diptera: Asilidae) on grasshopper (Orthoptera: Acrididae) populations, Oecologia 55, 42-46.

33. Leonardsson, K. (1994) Multiple density dependence in two sub-popUlations of the amphipod Monoporeia affinis: a potential for alternate equilibria. Oecologia 97. 26-34.

34. Lockwood, J. A., Kemp, W.P., and Onsager, J.A. (1988) Long-term, large-scale effects of insecticidal control on rangeland grasshopper popUlations, Journal of Economic Entomology 81, 1258-1264.

35. Lockwood. J.A. and Lockwood. D.R. (1991) Rangeland grasshopper (Orthoptera: Acrididae) population dynamics: insights from catastrophe theory, Environmental Entomology 20. 970-980.

36. Martinat, P.1. (1987) The role of climatic variation and weather in forest insect outbreaks, in P. Barbosa and J.e. Schultz (eds.). Insect Outbreaks, Academic Press, San Diego, CA, pp. 241-268.

37. Mattson, W.1. (1980) Herbivory in relation to plant nitrogen content, Annual Review of Ecology and SystematiCS 1 I, 119-161.

38. Mattson, W.J. and Haack, R.A. (1987) The role of drought in outbreaks of plant-eating insects, Bioscience 37, 110-118.

39. Messier, F. (1994) Ungulate population models with predation: a case study with the North American moose, Ecology 75, 478-488.

40. Nothnagle, PJ. and Schultz, J.e. (1987) What is a forest pest?, in P. Barbosa and J.C. Schultz (eds.), Insect Outbreaks, Academic Press, San Diego, CA, pp. 59-80.

41. Oedekoven, M.A. and Joern, A. (1999) Stage based mortality of grassland grasshoppers (Acrididae) from wandering spider (Lycosidae) predation, Acta Oecologia 19, 507-515.

42. Oedekoven, M.A. and Joem, A. (2000) Effects ofhost plant quality and spider predation on grasshopper populations (Acrididac): food-quality dependent compensatory mortality. Ecology 81,66-77.

43. Onsager, J.A. (1985) Relationships between survival rate, density, population trends and forage destuction by instars of grasshoppers (Orthoptera: Acrididae), Environmental Entomology 12, 1099-1102.

44. Peterman, R., Clark, W.e. and Holling, e.S. (1979) The dynamics of resilience: shifting stability domains in tish and insect systems, in R.M. Anderson. B.D. Turner and L.R. Taylor (eds), Population Dynamics, Blackwell Scientitic, Oxford, UK, pp. 321-342.

45. Polis, G.A. and Winemiller. K.O. (eds.). (1996) Food Webs, Chapman & Hall, NY. 46. Ritchie, M. E. (2000) Nitrogen limitation and trophic vs. abiotic influences on insect herbivores in a

temperate grassland, Ecology (in press). 47. Rothley, K.D., Schmitz, 0.1., and Cohen, J.L. (1997) Foraging to balance contlicting demands: novel

insights from grasshoppers under predation risk, Behavioral Ecology 8, 551-559. 48. Schmitz, O. J. (1994) Resource edibility and trophic exploitation in an old-tield tood web, Proceedings

of the National Academy of Sciences (USA) 91, 5364-5367. 49. Schmitz, 0.1., Beckerman, A.P., and O'Brien, K.M. (1997) Behaviorally mediated trophic cascades:

etlects of predation risk on food web interactions, Ecology 78, 1388-1399. 50. Schmitz, 0.1. (1998) Direct and indirect effects of predation and predation risk in old field interaction

webs American Naturalist 151, 327-342. 51. Sinclair, A.R.E. (1989) Population regulation in animals, in J.M. Cherritt (ed), Ecological Concepts,

Blackwell Scientific, Oxford, UK, pp. 197-241. 52. Sinclair, AR.E., Olsen, P.O., and Redhead, T.D. (1990) Can predators regulate small mammal

populations? Evidence from house mouse outbreaks in Australia, Oikos 59. 382-392.

144

53. Streett, D.A. and Maguire, M.A. (I990) Pathogenic diseases of grasshoppers, in R. Chapman and A. Joem (eds), Biology o/Grasshoppers, John Wiley & Sons, NY, pp. 483-516

54. Strong, D.R., Lawton, J.H., and Southwood, T.R.E. (1984) Insects on Plants: Community Patterns and Mechanisms, Harvard University Press, Cambridge, MA.

55. Uvarov, B.P. (1931) Insects and climate, Transactions o/the Royal Entomological Society, London 79, 1-247.

56. Wellington, W.O. (1957) The synoptic approach to studies of insects and climate, Annual Review 0/ Entomology 2, 143-162.

57. White, T.C.R. (1976) Weather, food and plagues oflocusts, Oecologia 22,119-134. 58. White, T.C.R. (1993) The Inadequate Environment, Springer Verlag, Berlin. 59. Willott, SJ. and Hassall, M. (1998) Life history responses ofBritish grasshoppers (Orthoptera: Acrididae)

to temperature change, Functional Ecology 12, 232-241. 60. Yang, Y. and Joern, A. (1994a) Influence of diet quality, developmental stage and temperature on food

residence time in the grasshopper Melanoplus differentialis, Physiological Zoology 67, 598-616. 61. Yang. Y. and Joern, A. (I 994b) Out size in relation to variable food quality and body size in grasshoppers,

Functional Ecology 8, 36-45. 62. Yang, Y. and Joern, A. (1994c) Compensatory feeding in response to variable food quality by

Melanoplus differentialis, Physiological Entomology 19, 75-82.

9. WHAT TOOLS HAVE POTENTIAL FOR GRASSHOPPER PEST MANAGEMENT?

A North American Perspective

Abstract

lA. ONSAGER i and O. OLFERT2

J USDA / ARS / Northern Plains Agricultural Research Laboratory 1500 North Central Avenue Sidney, MT 59270, USA 2 Agriculture and Agri-Food Canada 107 Science Place Saskatoon, SK Canada S7N OX2

Grasshoppers are the most important insect pests of grasslands within the Great Plains of North America. The extent to which this large expanse has been disturbed by human intervention varies considerably. At one extreme, there are some conservation grasslands where no human enterprise is pennitted; but at the other extreme, most of the grasslands in the Great Plains have been brought under cultivation. Current management tools include monitoring and surveying grasshopper abundance, chemical control, biological control and cultural control.

1. Problem

Grasshoppers are a chronic problem within the Northern Great Plains, varying from many years where the threat is relatively isolated to large-scale outbreaks that have a major economic impact on the agriculture sector [13, 26]. Of about 140 common species that inhabit grasslands in western North America, only 17 occur in high numbers on rangeland and five are considered economically important in cropland [36]. T~e threat to grasslands most often arises from grasshoppers hatching from oviposition sites within the grasslands and feeding on grasses and forbs. The threat to production of small grains arises from migration of the hatchling populations into cropland from grasslands and grassy roadsides, headlands and field margins. On average, grasshopper feeding results in an estimated annual loss of$400 million in grasslands [13] and $6 million annually in small-grain crops [8].

In the recent past, control of grasshoppers on grasslands and the associated

145

J. A. Lockwood etal. (eds.). Grasshoppers and Grassland Health. 145-156. © 2000 Kluwer Academic Publishers.

146

cropland in North America was not a particularly perplexing problem due to wide acceptance of a number of highly effective chemicals registered for use against grasshoppers. This was especially true in instances where the relatively high value of cultivated crops provided incentive to prevent excessive damage by insects. However, many of the grasshoppers that invade cropland develop on adjacent rangeland. In contrast to cropland, the value offorage produced on rangeland can often be significantly less than the cost of traditional chemical control treatments. Grasshoppers can then rigorously interfere with efforts to achieve economic gain, and the challenges of how to deal with them become a major issue.

As indicated, the primary method of grasshopper control during outbreaks is to apply insecticides. The economic and environmental concerns about pesticide use have stimulated an interest in alternative strategies to manage grasshoppers. These include monitoring and forecasting population abundance, and implementation of cultural and biological controls. Such strategies have been adopted under the banner of integrated pest management, where the overall aim is to manage economic levels of pest populations through a variety of measures in an effort to reduce the reliance on chemical insecticides.

2. Tools

2.1 FORECASTING POPULATION ABUNDANCE

As mentioned above, the Great Plains contains a large expanse of intermingled cropland and grass prairie. Monitoring and forecasting grasshopper abundance over this extensive area require a coordinated effort of several levels of government agencies and industry groups [31].

Grasshopper outbreaks tend to begin with localized infestations that appear to expand and change in a geographical area over successive years as rates of reproduction and survival change [19]. Factors that influence fluctuations in grasshopper abundance usually involve variables that have both spatial and temporal qualities including precipitation, heat, soil quality, vegetation type and other ecosystem descriptors. As a result these variables lend themselves well to analysis using Geographic Information Systems (GIS). These

Early warning of major changes in grasshopper abundance and geographic distribution is required to assess the threat to rangeland and to crops. In Canada, annual surveys of grasshopper abundance and distribution have been conducted since 1919, however, the first practical application of GIS in grasshopper surveys was implemented in 1989 [18]. On the ground, surveys of adult grasshoppers are conducted in late summer (August). At each survey stop, trained surveyors record grasshopper numbers along 100-m transects in the crop and in the field margin. Approximately 2 000-4 000 survey points are compiled each year and summarized to produce a forecast of the grasshopper damage potential for the subsequent growing season. Using GIS, the grasshopper density data is contoured to produce density/distribution maps that are suitable for overlays with weather, soil and ecosystem variables. Recent advances in models of grasshopper growth and

147

activity now incorporate spatial analytical techniques to forecast the expected threat of grasshopper damage [22]. Spatial analysis (map overlays of rainfall, temperature and hours of sunshine) and modeling are used to adjust popUlation density maps to produce updates of damage potential in the current growing season [19].

In the western United States, annual surveys of nymphal and adult populations were conducted by the United States Department of Agriculture's Animal and Plant Health Inspection Service (USDA/APHIS) [2]. The data that were generated supported useful models that predicted forage destruction [I] and probabilities of extended outbreaks [20, 23], while they simultaneously supported delimitation of serious infestations and planning for suppression activities. Unfortunately, these surveys were recently discontinued due to lack of support funding.

2.2 CHEMICAL INSECTICIDES

2.2.1 Rangeland In the United States, chemical insecticides approved for broadcast applications to rangeland are restricted to those that may be applied in the presence of livestock. These include carbaryl, malathion, and orthene. In Canada, two additional products, dimethoate and deltamethrin, are approved for rangeland application under the proviso that producers remove cattle prior to application and follow re-entry guidelines.

In the past, most treatments on rangeland in the United States were applied through cooperative programs administered by USDA/APHIS. Federal funds paid one third of the cost, some states contributed up to one third of the cost, and ranchers paid any remaining costs. GIS techniques were developed to identify environmentally sensitive areas within infested blocks [21] and electronic guidance systems on spray planes were used to assure that such areas were not treated. [39].

In the early 1990s, a Grasshopper Integrated Pest Management (GIPM) project learned that the economic threshold for grasshoppers on rangeland was much higher than had previously been assumed [4]. Consequently, higher grasshopper populations were required to justify treatments, and fewer acres of rangeland qualified for treatment. Since then, costs of treatment have been rising faster than ranch revenue, so even fewer range acres now produce forage that is worth as much as the cost of treatment. Since the mid 1990s, USDA/APHIS has not participated in large-scale cooperative programs for protection of range forage. Therefore, it is not clear where leadership will come from when we have the next major outbreak of grasshoppers in the United States.

Most of the chemical treatments applied by USDA/APHIS were broadcast spray applications. Carbaryl also is registered as a 2%-5% bait formulated on flaky wheat bran. This bait treatment has certain advantages over spray treatments in that it uses less than 10% as much toxicant per area treated, and is toxic only against organisms that consume wheat bran [7]. However, it has not gained wide acceptance, perhaps for three major reasons. First, logistical problems of treating, storing, shipping, handling, and applying bulky material are greater than for liquid treatments. Second, equipment to meter accurately and dispense dry material is not as available as for liquid treatments. Finally, although GIPM experiments showed that bait treatments eventually provided up to 75%-

148

85% control of treated infestations, the response to bait is not as rapid or as certain as response to most chemical sprays [33].

2.2.2 Cropland In cropland, insecticides are recommended to reduce population levels to economically acceptable levels when grasshoppers are not being satisfactorily suppressed by natural factors. There are a number of chemicals approved for control of grasshoppers in field crops; the list of approved insecticides will vary in relation to crop type and stage of crop development [10]. However, prior to utilizing chemicals, producers are asked to consider three important questions. They are:

1. What is the density of grasshoppers in or near the crop? 2. Is the density of grasshoppers high enough to cause damage? 3. Will the costs of the product plus application costs be greater than the value

of the crop? In order to answer these questions, producers must be able to locate the area of grasshopper infestation, estimate grasshopper density, and calculate an expected value of the crop being protected in relation to the costs of the control treatment.

In order to locate and assess the damage potential of grasshoppers, producers are encouraged to scout their fields on several occasions. The primary areas to be inspected should be those where grasshoppers were the most abundant the previous fall. These areas include grassy areas along roadsides and fence lines, fall-seeded crops, stubble fields, forage crops and rangeland. To obtain a preliminary estimate of grasshopper numbers, soil samples can be taken in spring to assess the number of eggs in the soil. This is rather timeconsuming and is not practised by most producers. Instead, producers will begin their monitoring program after grasshopper hatch has begun. In the northern Great Plains, grasshopper eggs usually begin hatching in May, although this can be delayed by cool wet weather. Once the hatch has begun, scouting becomes increasingly important. Insecticide applications should be timed to control third-instar nymphs unless the number of newlyemerged grasshoppers is extremely high. If the timing can be accommodated, most of the eggs should have hatched but the nymphs will not have yet migrated a great distance from the egg beds. The added benefit of treating these areas where the small grasshoppers are concentrated is that the costs of application are minimized because lower rates are required to kill young grasshoppers and the extent of the area to control is relatively small. If this strategy fails or is not possible, grasshoppers will need to be controlled in field crops when their density exceeds economic levels. Table I is presented as a guide to when appropriate action may be required [10].

2.3 BIOLOGICAL TOOLS

The list of natural enemies of grasshoppers in North America is quite extensive. For example, there are 24 species of paras ito ids (19 Diptera, 3 Hymenoptera and 2 Nematoda) known to parasitize Melanoplus sanguinipes [9]. In addition, there are a number of

TABLE 1. Chart of grasshopper infestation as indicated by density of nymphs and adults per square meter, and the appropriate action required

Density in indicated habitat Control measures

Field Roadside

0-3 0-6 not required

4-6 7-12 usually not required

7-12 13-24 may be required

13-24 25-28 usually required

over 25 over 50 required

149

known predators and pathogens. The role of natural enemies in suppressing grasshopper populations has not been well quantified. It has been suggested that they do playa significant role in most years but that these can be negated by hot, dry weather systems that are more favorable to grasshopper growth and survival [10]. Since natural enemies are already widely distributed geographically, it is unlikely that they can be exploited to prevent grasshopper outbreaks over extensive areas. As a result, the development of these biological tools for grasshopper management has not been as successful as we might wish.

Microbial pathogens have received the most attention in the quest for biological control tools to control grasshoppers. Pathogens can act on the host insect in a number of ways, including increased mortality, alteration of behavior, reduced fecundity or reduced feeding, to suppress grasshopper population growth [12]. However, the extent to which a pathogen is able to reduce food consumption and cause increased mortality will determine its potential as a crop protection tool. The most dramatic examples of grasshopper population control by a pathogen have been recorded in naturally-occurring epizootics of the fungus Entomophaga grylli. In 1963, E. grylli (pathotype I) was responsible for a drastic reduction in populations of Camnula pellucida [38]. In 1985 and 1986, E. grylli (pathotype II) was responsible for major infections (>40%) in populations of Melanoplus bivittatus [5].

Pathogens have not yet justified early optimism concerning their potential for large-scale utility. A bait formulation of Nosema locustae (a protozoan parasite) is registered and available in the United States, but has not enjoyed wide commercial acceptance. Sales are primarily to homeowners for lawn or garden protection, or to organic farmers or ranchers. Perhaps four factors discourage more extensive use on commercial rangeland. First, periodical grasshopper outbreaks cannot be predicted with certainty, so preventive tactics are not always justified. Second, the course of an epizootic of Nosema locustae cannot be predicted precisely and therefore may not be recognized and credited as such [11]. Third, by the time an outbreak is recognized, it may be too severe to reduce

150

with a slow-acting biological agent. Finally, relatively high costs dictated by a limited market plus the same logistic problems associated with carbaryl bait combine to inhibit use of Nosema locustae. We are aware of one organic rancher that utilizes Nosema locustae in a unique manner. Each year after hay from irrigated meadows is removed, the meadows are irrigated to stimulate regrowth and to attract adult grasshoppers. Then the fields are treated with Nosema locustae to debilitate females and to produce infected egg-pods.

Research established that two entomopox viruses were highly virulent, and gave promising results in lab and field trials [29, 41]. However, the virulence prohibited economical in vivo production, and in vitro methods or economical alternative host production methods have not been perfected.

Beauveria bassiana (a pathogenic fungus) is economical to produce in large fermentation vats, and it has appeared very promising in laboratory tests. However, it has not always performed as expected in field tests. Grasshoppers captured at intervals as long as three weeks after treatment suffered high levels of mortality when placed in cages for observation [6], but field populations do not always decline in density at rates that are commensurate with infection levels. Infected grasshoppers apparently are able to survive in the field by basking to raise internal temperatures to levels that inhibit fungal development. We are not aware ofresearch to determine the effects of such infections on forage consumption rates or reproductive potentials.

In general, there appears to be little potential for augmentation of natural populations of parasites or predators. Natural parasitism often occurs too late in the season to significantly reduce forage destruction or to prevent grasshopper egg production [40]. Parasites tend to overwhelm grasshopper infestations only after consecutive seasons of gradually increasing parasite incidence among what might be considered untenable grasshopper densities. Predators may be too opportunistic to be relied upon for suppression of grasshoppers. Predators consume many grasshoppers when grasshoppers are more abundant than alternate prey species or when depletion offorage and concomitant starvation increases vulnerability of grasshoppers. However, such predation provides mostly compensatory mortality and may do little to prevent economic loses.

In contrast to scepticism about augmentation of natural enemies, there appears to be great potential for effective conservation of parasites and predators. In experiments with carbaryl bait, lower survival rates were frequently observed among grasshoppers that did not succumb to the bait than among grasshoppers in adjacent untreated check plots. That phenomenon was attributed to enhancement of the parasite and predator to prey ratio through selective removal of grasshoppers by the bait [35]. A similar phenomenon probably contributes to the success of the reduced area and agent insecticide treatments (RAATs) approach to grasshopper control [24]. (One could here argue that discussion of RAA Ts should occur in the section on chemical control, but we chose to mention it here to emphasize the conservation of natural enemies.) RAATs has at least four major advantages working in its favor, especially when it is used with carbaryl as the toxicant. First, it reduces cost by treating only a portion (perhaps only about 50%) of the infested area in alternating strips. Second, it uses lower than normal rates of carbaryl to achieve further reductions in cost. The approach requires both courage and self-confidence on the part of the user, because it does not provide the immediate and total annihilation of

151

grasshoppers that some segments of the general public expect. On the other hand, it agrees with a 1978 study in which maximum economic return was achieved with a dosage of only 0.28 kglha, and economic benefits of higher dosages did not justify their additional costs [32]. The third advantage, which is caused by the residual toxicity of carbaryl, is that a portion of the grasshoppers from untreated strips will wander into treated strips and be killed within the first week to 10 days after treatment. Finally, and perhaps most importantly, the untreated strips serve as a refuge for parasites and predators. When the grasshopper population becomes depleted through movement into treated strips, the predator and parasite to prey ratio may become enhanced as in the case of the carbaryl bait treatments.

2.4 CULTURAL TOOLS

2.4.1 Rangeland In some Great Plains ecosystems, grasshopper populations can be suppressed and grasshopper outbreaks can be prevented through use of a rotational system of grazing management. The twice-over grazing system has been described elsewhere in detail [3, 25]. From the perspective of hypothetical beneficial attributes for grasshopper management, it can be described briefly as follows. It requires two cycles of sequential movement of a herd of livestock through a series of three to six pastures within a 4.5-month grazing season. The first grazing cycle occurs I June to 15 July. This cycle tends to distribute defoliation over a higher proportion of plants and tends to leave a more uniform canopy height than typically occurs when the herd has continuous access to the same total area for the same period of time. During the interval between grazing periods, new tillers tend to establish in bare spots. The canopy produces few reproductive stems, so it tends to remain relatively level while it grows in height. The second grazing cycle occurs 16 July to 15 October, when each pasture is grazed in the same sequence for twice as many days as it was grazed in the first cycle. The second cycle also appears to promote more uniform harvest than seen under season-long grazing. That may have an important impact on soil temperatures, grasshopper basking efficiency, and grasshopper egg development rates, because less soil is exposed to direct sunlight for a shorter time than under season-long grazing. The exit pasture becomes the entry pasture for the next season, but the same sequence of pasture use is retained. This assures that pasture use is systematically changed each year, which theoretically reduces the number of grasshopper species that would be favoured for successive generations in any given pasture.

Grasshopper population trends were monitored by the senior author during 1993-1995 and 1997-1998 on commercial native rangeland under twice-over rotational grazing vs traditional season-long grazing. The study was conducted during an interval of progressively warmer and dryer seasons, so grasshopper populations would be expected to increase from 1993 to 1998. A ubiquitous pest grasshopper, Melanoplus sanguinipes (Fabricius), occurred at every sample site during each year in numbers sufficient to provide life history parameters for comparisons between treatments. Stage-specific survival rates for nymphal instars 3 through 5 were significantly lower and less variable under rotational grazing (means were 0.51 versus 0.66 and variances were 0.03 versus 0.07). Nymphs

152

developed significantly slower under rotational grazing (means were 11.5 days versus 8.9 days), which significantly reduced the proportion that became adults. Average adult density was significantly lower under rotational grazing (0.6 versus 2.05 per m2).

Seasonal presence of all grasshopper species combined under rotational grazing averaged 30% as high as under season-long grazing. Local outbreaks that generated 18 and 27 adult grasshoppers per m2 under season-long grazing in 1997 and 1998, respectively, did not occur under rotational grazing. The outbreaks consumed 91% and 168%, respectively, as much forage as had been allocated for livestock, as opposed to 10% and 23%, respectively, under rotational grazing. Three grasshopper species that were primary contributors to outbreaks under seasonal grazing remained innocuous under rotational grazing. All three were late-season species that rely heavily on blue grama grass, a short-grass (under-story) species that becomes exposed under traditional seasonal grazing.

It should be emphasized that soil and air temperatures were monitored within grazing exclosures. All sampling sites had similar thermal potential to support grasshopper development, and there was no evidence of bias in site selection, at least with regard to factors like slope or aspect, which are known to influence soil and air temperatures. There should therefore have been no major differences between grasshopper developmental parameters either in the absence of grazing or under identical grazing. Because differences were observed among grasshopper life processes that are driven primarily by temperature, it was assumed that the grazing regimes differentially affected grasshopper microhabitat. It therefore appeared that outbreak suppression through grazing management is feasible in the northern Great Plains.

2.4.2 Cropland Similar to the rangeland situation, cultural control of grasshoppers in cropland is a preventative approach. The selection of cultural control tools includes early seeding of crops, rotation of crops, weed control, host plant resistance, trap strips and barrier strips.

Early spring seeding of field crops will reduce damage potential because older plants that are growing more vigorously at the time when grasshoppers hatch will be able to withstand more defoliation than younger plants. At the other end of the season, the early-seeded crops mature earlier and are less attractive to migrating grasshoppers than immature crops.

Crop rotation is recommended for cereal stubble fields that are infested with grasshoppers. Growing alternative crops to cereals, such as pulse or oilseed crops, may reduce the need for chemical intervention ifthe grasshopper infestations are not too severe.

Weed control is a very important grasshopper management tool because weeds are generally regarded as suitable food plants that support grasshopper population growth [37]. Cultivation of the soil is a very effective method of controlling weeds and suppressing grasshoppers in regions of the prairies that receive adequate rainfall. However, soil conservation strategies such as zero-till or minimum-till are now strongly being recommended in the more arid regions of the Northern Great Plains. In these regions, broad-spectrum herbicides have replaced tillage operations. The long-term impact of this weed management approach on grasshopper population growth is not well

153

documented. However, it has been observed that poor weed management can encourage high grasshopper infestations.

A comprehensive overview of host-plant resistance is presented elsewhere in this book and so it will not be given the same diligence here. However, a brief summary of this management tool is warranted here in the context of grasshopper control in cropland. Host-plant resistance to grasshoppers in field crops involves the effect of the food plant on grasshopper growth, survival and reproduction. The overall effect ofthese factors has been quantified in terms of grasshopper biotic potential [28]. Targeting young grasshoppers with resistant host-plants is probably the best use of this strategy because grasshopper mortality tends to be the greatest during the first two instars [34]. Recent investigations of host-plant resistance to grasshoppers have focused on using neonates to assess the potential resistance of wheat, barley, oats and rye [14,17] and grasses [16]. Populations of Melanoplus sanguinipes, which were fed oats and a wheat cultivar HY320, had significantly higher mortality and lower rates of growth than those which were fed other crops or cultivars [28]. When feeding on oats or HY320 was extended over their entire life cycle, these diets reduced the biotic potentials to such low levels that they suffered significant population decline. In field studies using naturally-occurring popUlations of grasshoppers, the cultivars which suffered the least amount of defoliation in an extensive series of cereal cultivars correlated well with those showing the greatest negative effects on biotic potential in laboratory studies.

The two remaining tools, trap strips and barrier strips, exploit the concepts of host-plant susceptibility and host-plant resistance. Trap strips of highly attractive food plants, such as weeds, can be used effectively to concentrate grasshopper populations in a relatively small area. Once the migration is complete, and before the vegetation is consumed, it is possible to control the grasshoppers quickly and effectively using a minimum amount of insecticide. Vegetative strips of approximately 10m wide were used to reduce grasshopper population levels by 60 -70%, as measured in reduced egg numbers, in infested field margins [26].

Barrier strips of highly resistant food plants, such as peas or oats [15] next to infested areas can be used to protect more valuable, more susceptible crops. Dense stands of resistant plants will retard the rate at which grasshoppers disperse through the crop [16]. This usually provides enough opportunity for spring-seeded annual crops to become well established and to lessen the damage potential of grasshoppers.

3. Solutions

It is evident from the literature that a number of tools, in addition to chemical insecticides, are available for managing grasshoppers in North America. The GIPM User Handbook, which has been cited numerous times in this paper, contains information on efficacy studies, effects of treatments on non-target organisms, sampling, survey, and calibration techniques, a computer-based decision support system, and a multitude of other useful topics within its nearly 100 chapters. (Although a limited number of hard copies were produced, electronic copies of the handbook will become available at

154

http://www.sidney.ars.usda.gov during early summer of 2000). While the GIPM handbook and other scientific literature offer insights for the potential utilization of alternate strategies to manage grasshopper populations without risking environmental disaster, successful implementation remains a challenge in many cases. Of all the management tools discussed above, cultural control has the potential to be the least costly grasshopper management strategy, as well as one of the least disruptive to other organisms in the environment.

4. Acknowledgements

The authors of this review would like to acknowledge earlier researchers whose many contributions made a significant impact on management of grasshoppers on the northern Great Plains. Due to a shift in agricultural priorities and government budgetary constraints, the number of scientists actively involved in grasshopper research has dwindled dramatically. The full impact of these reductions may not be felt until the next major outbreak of grasshoppers on the northern Great Plains.

5. References

1. Berry, 1.S., Kemp, W.P., and Onsager, 1.A. (1995) Within-year population dynamics and forage destruction model for rangeland grasshoppers (Orthoptera: Acrididae), Environmental Entomology 24, 212-225.

2. Berry, J.S., Onsager, l.A., Kemp, W.P.,McNary, T., Larden, D., Legg, D., Lockwood,l.A., and Foster, R.N. (1996) Assessing rangeland grasshopper populations, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, pp. VI.IO-I to VI.I0-12.

3. Biondini, M.E. and Manske, L.L. (1996) Grazing frequency and ecosystem processes in a northern mixed prairie, USA, Ecological Applications 6, 239-256.

4. Davis, R.M. and Skold, M.D. (J 996) Regional economic thresholds in grasshopper management, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, pp. VI.4-1 to VI.4-4.

5. Erlandson, M.A., 10hnson, D.L., and Oltert, O. (1988) Entomophaga grylli infections in grasshopper (Orthoptera: Acrididae) populations in Saskatchewan and Alberta, 1985-1986, The Canadian Entomologist 120, 205-209.

6. Foster, R.N., Reuter, K.C., Britton, J., and Bradley,C. (1996) Laboratory studies and field trials with the fungus Beauveria bassiana against grasshoppers, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, chapter VII. In Press.

7. Foster, R.N. and Onsager, lA .. (1996) Sprays versus bait, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, pp. 11.3-1 to 11.3-3.

8. Gage, S.H. and Mukerji, M.K. (1978) Crop losses associated with grasshoppers in relation to economics of crop production, Journal of Economic Entomology 71, 487-498.

9. Graham, A.R. (1965) A preliminary list of the natural enemies of Canadian agricultural pests. I'!formation Bulletin No.4, Canada Department of Agriculture, Belleville, ON, pp. 1-179.

10. Harris, J.L., Buchan, JA, Ewen, A.B., Frowd, lA, Minja, L.A., Mukerji, M.K., Riegert, P.W., and Weisbrot, D.R. (1984) Grasshopper control. Saskatchewan Agriculture, Regina.

II. Henry, I.E. (1972) Epizootiology of infections by Nosema locustae Canning (Microsporida: Nosematidae) in grasshoppers, Acrida 1, 111-120.

12. Henry, J .E. (1977) Development of microbial agents for the control of Acrididae, Revista de la Sociedad Entomologica Argentina 36,125-134.

155

13. Hewitt, G.B. and Dnsager, lA. (1983) Control of grasshoppers on rangeland in the United States - a perspective, Journal of Range Management 36, 202-207.

14. Hinks, C.F., Dlfert, D., and Westcott, N.D. (1987) Screening cereal cultivars for resistance to early-instar grasshoppers, Journal of Agricultural Entomology 4, 315-319.

IS. Hinks, C.F., Dlfert, D., Westcott, N.D., Coxworth E.M., and Craig, W. (1990) Preference and performance in the grasshopper Melanoplus sanguinipes (Fab.) (Drthoptera: Acrididae) feeding on kochia, oats and wheat: implications for popUlation dynamics, Journal of Economic Entomology S3, 1338-1343.

16. Hinks, C.F. and Dlfert, D. (1992) Cultivar resistance to grasshoppers in temperate cereal crops and grasses: a review, Journal ofOrthoptera Research I, 1-9.

17. Hinks, C.F. and Dlfert D. (1993) Growth and survival to the second instar of neonate grasshopper nymphs, Melanoplus sanguinipes (F.) fed cultivars ancestral to hard red spring wheat, Journal of Agricultural Entomology 10, 171-180.

18. Johnson, D.L. (1989) Spatial autocorrelation, spatial modelling, and improvements in grasshopper survey methodology, The Canadian Entomologist 121,579-588.

19. Johnson, D.L., Dlfert, D., Dolinksi, M. and Harris, L. (1995) GIS-based forecasts for management of grasshopper populations in Western Canada, Proceedings of International Symposium on Agricultural Pest Forecasting and Monitoring, pp. 109-112.

20. Kemp, W.P. (1987) Probability of outbreak for rangeland grasshoppers in Montana: application of Markovian principles, Journal of Economic Entomology SO, 1100-1105.

21. Kemp, W.P., McNeal, D., and Cigliano, M.M. (1996) Geographic information systems (GIS) and integrated management of insects, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, pp. VI.9-1 to VI.9-7.

22. Lactin, DJ., Holliday, N.D., Johnson, D.L., and Craigen, R. (1995) Improved rate model oftemperature development by arthropods, Environmental Entomology 24,68-75.

23. Lockwood, J.A. and Kemp, W.P. (1988) Probabilities of rangeland grasshopper outbreaks in Wyoming counties, Wyoming Agricultural Experiment Station Bulletin B-869.

24. Lockwood, J.A. and Schell, S.P. (1997) Decreasing economic and environmental costs through reduced area and agent insecticide treatments (RAA Ts) for the control of rangeland grasshoppers: empirical results and their implications for pest management, Journal of Orthoptera Research 6, 19-32.

25. Manske, L.L. (1994) Ecological management of grasslands defoliation, in F.K. Taha, Z. Abouguendia, and P.R. Horton (eds.), Managing Canadian rangelandsfor sustainability and profitability, Grazing and Pasture Technical Program. Regina, Sask., pp. 130-160.

26. Dlfert, D. (1986) Evaluation of trap-strips for grasshopper (Drthoptera: Acrididae) control in Saskatchewan, The Canadian Entomologist lIS, 133-140.

27. Dlfert, D. (1987) Insect pests of wheat in western Canada and their control, in A.E. Slinkard and D.B.Fowler, (eds.), Wheat Production in Canada - A Review, University of Saskatchewan, Saskatoon, Canada, pp.410-430.

28. Dlfert, D., Hinks, C.F., and Westcott, N.D. (1990) Analysis of the factors contributing to the biotic potential of grasshoppers (Drthoptera: Acrididae) reared on different cereal cultivars in the laboratory and in the field, Journal of Agricultural Entomology 7,275-282.

29. Dlfert, D. and Erlandson, M.A. (1991) Wheatfoliage consumption by grasshoppers (Drthoptera: Acrididae) infected with Melanoplus sanguinipes entomopoxvirus, Environmental Entomology 20, 1720-1724.

30. Dlfert D., Hinks, C.F., Weiss, R.M., and Wright, S.B.M. (1994). The etIect of perennial grasses on growth, development and survival of grasshopper nymphs (Drthoptera: Acrididae): Implications for population management in roadsides, Journal ofOrthoptera Research 2, 1-3.

31. Dlfert, D. and Kaminski, L.A. (1998) Keeping tabs on insect pests in prairie crops, Pest Management News 10,3-4.

32. Onsager, J.A. (1978) Efficacy of carbaryl applied to different life stages of rangeland grasshoppers, Journal of Economic Entomology 71, 269-273.

33. Dnsager, lA., Foster, R.N., and Jech, L. (1996) Bait acceptance by different grasshopper species and instars, in Grasshopper Integrated Pest Management User Handbook, USDA/APHIS Technical Bulletin 1809, pp. 11.12-1 to 11.12-5.

34. Dnsager lA. and Hewitt, G.B. (1982) Rangeland grasshoppers: Average longevity and daily rates of mortality among species in nature, Environmental Entomology II, 127-133.

35. Dnsager, lA., Rees, N.E., Henry, lE., and Foster, R.N. (1981) Integration of bait formulations of Nosema locustae and carbaryl for control of rangeland grasshoppers, Journal of Economic Entomology 74, 183-187.

156

36. Parker, J.R. (1952) Grasshoppers, in Insects, The Yearbook of Agriculture. United States Government Printing Office, Washington, D. C., pp. 595-605.

37. Pickford, R. (1962) Development, survival and reproduction of Melanoplus bilituratus (Orthoptera: Acrididae) reared on various food plants, The Canadian Entomologist 94,859-869.

38. Pickford, R. and Riegert, P.W. (1964) The fungus disease Entomophthora grylli and its effect on grasshopper populations in Saskatchewan in 1963, The Canadian Entomologist 96, 1158-1166.

39. Rodriguez, G. and Roland, T.R. (1996) Aircraft guidance for grasshopper control on rangelands, in Grasshopper Integrated Pest Management User Handbook, USDNAPHIS Technical Bulletin 1809, pp. 11.22-1 to 11.22-2

40. Sanchez, N.E. and Onsager, J.A. (1994) Effects of dipterous parasitoids on reproduction of Melanoplus sanguinipes (Orthoptera: Acrididae), Journal ofOrthoptera Research 3, 65-68.

41. Streett, D.A., Woods, S.A., and Erlandson, M.A. (1997) Entomopoxviruses of grasshoppers and locusts: biology and biological control potential, Memoirs of the Entomological Society of Canada 171, 115-130.

PART 4. GRASSHOPPER CONTROL AND GRASSLAND HEALTH

10. HOW CAN ACRIDID OUTBREAKS BE MANAGED IN DRY LAND AGRICULTURAL LANDSCAPES?

V.E. KAMBULIN Kazakhstan Quarantine Inspection (formerly) Astana, Kazakhstan

Abstract

This study considers changes occurring in the composition of dominant species and populations of locusts and grasshoppers in fallow lands in various stages of succession in northern Kazakhstan. In response to these changes, approaches to optimize the system of locust pest management are discussed. In particular, the application of barrier treatments using an insect growth regulator - diflubenzuron (Dimilin) - is considered.

1. Problem

Locusts and grasshoppers belong to the group of insects capable of reproducing over vast territories and causing catastrophic damage. Therefore, the problem of reducing their abundance and harmfulness is critical. Protection of agricultural lands against locusts is especially difficult in regions with changing agricultural landscape structures and in territories undergoing complicated and unpredictable anthropogenic transformations. Kazakhstan is exactly such a region. Currently the agricultural economy is facing serious difficulties, with chemical treatments against locusts and grasshoppers being undertaken in the area of >2.5 million ha annually [I]. To protect crops and rangelands against these pests it is necessary to use chemical products purchased from abroad, while for the control of other agricultural pests, domestically produced pesticides may suffice.

2. Status of populations

Locusts and grasshoppers are widely distributed in Kazakhstan. However, their dependence on soils and climatic conditions creates centers of high abundance and potential damage for different species in particular environments. For example, the Moroccan locust

157

J. A. Lockwood et al. (eds.). Grasshoppers and Grassland Health. 157-172. © 2000 Kluwer Academic Publishers.

158

Dociostaurus maroccanus (Thunb.), inhabits the clay deserts with ephemeral vegetation and reproduces in great numbers, mainly in Zhambyl and South Kazakhstan regions. The Asiatic Migratory locust, Locusta migratoria migratoria L., is distributed over the western, southern and southeastern parts of Kazakhstan. Its permanent outbreak centers are located among reed beds of large lakes and rivers. The Italian locust, Calliptamus italicus L., is one of the most common and harmful species of this group. Its natural habitats occupy practically the whole territory of Kazakhstan, but it is most dangerous in the western, northern, northeastern and central regions. In years of increasing reproduction, up to 80% of all treatments against locusts are applied to control this species. Grasshoppers also represent a serious threat, especially in the northern regions of the Republic, where they are the primary pests of agricultural crops.

In total, Kazakhstan is inhabited by 230 species and sub-species of locusts and grasshoppers belonging to 65 genera. Among them, 112 species and sub-species of 54 genera cause noticeable damage for crops and rangelands. Fifteen to twenty species comprise the most harmful pests, depending on soils and climatic zones.

Most species of locusts and grasshoppers inhabit natural grasslands from which they can spread to crops. Ploughing of vast areas of virgin and fallow lands in the 1950s resulted not only in a considerable reduction of areas favorable for locusts and grasshoppers, but it also led to the disappearance of some species of Orthoptera associated with steppe habitats. However, having good migration abilities, locusts and many grasshoppers moved to the "edge zones", where crops were adjacent to grasslands. After ploughing of virgin lands, periods of mass reproduction and increased damage by locusts and grasshoppers were comparatively short (2 to 3 years) followed by 5 to 6 years of relatively low abundance. Anti-locust chemical treatments were concentrated in limited centers representing locust breeding areas.

Agricultural development of lands in Kazakhstan started about 190 years ago. Since that time, four stages of arable farming have developed [2]. Local ploughing started in the 1860s, followed in 1954-1964 by an active ploughing of virgin and fallow lands. The next stage was characterized by the adoption of soil protection systems in agriculture (1965-1975). After that (1976-1992), some intensive technologies were used, when about 25 million ha of arable lands were planted to cereals.

The current, fourth stage started after 1992 and led to a considerable reduction of arable land areas devoted to cereal production. As a result, huge areas of land have turned into multi-stage fallow fields with diversified vegetation. The process of restoration of fallow lands usually includes three main stages: a weed grass and forb fallow, a couch-grass fallow, and a fallow with spongy bush and cespitose grasses.

The virgin type of vegetation in the northern regions of the Republic will be fully restored from fallow lands in 10 to 20 years. Further to the south, the restoration process slows down and will require 25 to 30 years. The speed of restoration depends on the use of uncultivated plots (pasture, hay harvesting, etc.), soils differences, meteorological peculiarities, and other factors.

The distribution of such fallows expanded and re-distributed the contact zones between cultivated and uncultivated lands. Therefore, outbreaks and migrations of locusts and grasshoppers on landscapes with crops adjacent to uncultivated lands became regular.

159

The most vulnerable crops are the cereals, especially spring wheat. In this connection, we have begun to assess the changes occurring in the composition of dominant species and of population groups of locusts and grasshoppers in the vast territory of the northern part of Kazakhstan.

The preliminary analysis of the collected data shows that C italieus started to dominate in the forest-steppe and steppe life zones. The second most important species is Doeiostaurus brevieollis Ev., followed by StenobothrusfisheriEv., Paraeyptera mieroptera F.d .W., and Aeropus sibirieus L. (the Siberian grasshopper). Abundance of Doeiostaurus kraussi Ingen. has also grown considerably.

In the Stipa-Festuea steppe, Citalieus, Stenobothrus eurasius Zub., Omoeeetus haemorrhoidalis Charp., and Euehorthippus pulvinatus F .d. W. can be found most frequently and in high numbers. The importance of P. mieroptera, S.fisheri, S. nigromaeulatus H.-S., Chorthippus albomarginatus DeG., Glyptobothrus biguttulus L. is expected to grow. C italieus can also become the dominant species here.

In the Stipa-Poa grass steppe, dominant species include C italieus, D. brevieollis, A. sibrieus, and P. mieroptera. Subdominants are S. eurasius hyalosuperficies Vor. and S. fisheri. In the sandy Stipa grass steppes, the importance of C italieus, A. sibirieus, S. fisheri, P. mieroptera, D. brevieollis, and some other species has increased.

In the northern Festuea-Artemisia semi-deserts, D. brevieollis, Calliptamus eoelesyriensis earbonarius F.d.W., and D. kraussi have become dominant, while in the Artemisia-Festuea and halophylous forb semi-deserts, D. brevieollis, Ramburiella tureomana F.d.W., and C barbarus Costa now prevail.

In the areas where the Festuea fallow patches are prevalent, abundance of P. mieroptera has increased, and S. eurasius and 0. haemorrhoidalis are able to become dominant. E. pulvinatus and represenatives ofthe genus Chorthippus (s. str.) follow them in prevalence. Abundance of mesophilic species increases in the lands restoring into couchgrass fallows. Most probably, Ch. albomarginatus and P. mieroptera will become dominant, and the abundance of Stauroderus sealaris F.d. W. will increase. At this stage of fallow succession, assemblages of the locusts and grasshoppers will mainly include grass-feeders, because mesic habitats always have tall grasses.

These changes influence the long-term population dynamics of both individual species and acridid assemblages. In recent years, the period of quantitative increase and mass reproduction has lasted for 4 to 5 years, rather than 2 to 3 years as it did earlier. In the future, until the lands are stabilized and reach the status of resilient sustainability, the interphase period of outbreaks will be even shorter, and the danger to agricultural crops will be chronically elevated. In connection with this, forecasting seasonal, annual and long-term changes in abundance and importance of unstable assemblages oflocusts will be extremely difficult.

3. Interim solutions

Since the beginning of agriculture development in Kazakhstan, there have been considerable changes in the methods and management approaches to reducing damage

160

caused by acridids. These changes were reflected the development of farming based on scientific discoveries and technical progress.

In the beginning, various mechanical control methods and approaches were used. Later on, chemical treatments were applied in combination with poisoned baits, although some mechanical methods were still used. The next stage involved the use of dusting and spraying in combination with baits and mechanical methods. In the 1970s application of organophosphates was common, followed by the use of synthetic pyrethroids and other chemical compounds. From that time until the I 990s, the use of conventional chemicals prevailed without economical limitations. Territories of high densities and damage underwent blanket treatments by air and ground application. However, this did not prevent increases in gregarious or non-gregarious acridid populations.

At present, for reasons not only of economic origin, but also due to the ecologically unstable agricultural environment created in Kazakhstan, the traditional approaches and methods or, more precisely, strategies and tactics of treatments, require considerable adaptation. To achieve prolonged reductions in the density of acridid populations while protecting agricultural landscapes, it is necessary to use formulations with selective action. Their selectivity should be based not only on the mechanism of action being different from conventional neurotoxic chemicals but also on peculiarities of acridid behaviour and feeding.

The dynamics which allow natural or semi-natural functioning of the trophic system originally showed periodic intervention from various biotic factors, including the interactions among locusts, plants, entomophages, and pathogens. This self-regulation was based on the synchronized trends of species and populations in a region that mediated the abundance and damage of dominant species. Therefore, chemical regulation of acridid populations also has to be periodical rather than chronic.

During declines in the population cycle of acrid ids, protective measures using pesticides are ill-advised. The treatments are optimal during periods of upsurge, mass reproduction, and population peaks. During periods of upsurge and mass reproduction, local or barrier treatments in nymphal habitats are sufficient. During population peaks, blanket treatments on precisely selected areas may be needed along with barrier treatments.

Our studies have shown that in the life zones where non-gregarious grasshoppers and assemblages with the Italian locust prevail (i.e., the northernmost regions of Kazakhstan), there are two to three waves of hatching that take 25 to 30 days in total. Conventional insecticides (mainly pyrethroids) with short residual action inevitability require repeated treatments of the same plots. In these conditions, higher and longer efficacy is ensured by using chemicals with a residual of about 7 to 10 days (e.g., fenitrothion 50%, 100% UL V; fenitrothion+esfenvalerat 50% e.c.; chlorpyrifos 48% e.c.). However, these compounds do not have high selectivity, and they are dangerous for the environment.

Pyrethroids are effective against the Moroccan, Asiatic and Italian locusts in areas inhabited by their gregarious forms, provided that the entire treatment protocol is strictly implemented. As a rule, these chemicals provide sufficient, short-term biological efficacy, and repeated treatments are not needed. But it is necessary to remember that pyrethroids affect non-target species as well. It is known that, to a considerable extent, efficacy and

161

safety of insecticides depend on application techniques and technology. A low-volume application is acceptable for most formulations of insecticides, but it is not very practical because ofthe huge territory infested byacridids. Also, it is often not acceptable in deserts due to the difficulties with water supply. An ultra-low volume application (UL V) is more practical because there are fewer problems with water supplies. However, not all chemicals can be applied with this method; the wettable powder formulations are excluded.

In recent years in Kazakhstan, ordinary aerosol foggers and aerosol foggers with a regulated droplet dispersion have been widely used. The practice of using this special machinery showed that chemicals formulated as emulsifiable concentrates or suspensions are highly effective at normal or reduced rates. At the same time, aerosol treatments are suitable for the blanket coverage of areas during the peak outbreaks of locusts. During the periods of upsurge, mass reproduction, and decline, it is most feasible to use ground ULV sprayers, deltaplanes and microplanes. Unfortunately, these machines are in critically short supply.

This once again proves that the techniques and technology used in Kazakhstan to control locusts and grasshoppers have considerable shortcomings. Expansion of territories suitable for locust reproduction forces us to reject traditional technologies of blanket treatments for both economic and environmental reasons. Therefore it is necessary to develop more efficient, economic and safe technologies. In 1996-1998, efficacy trials of new techniques and technologies to protect agricultural crops and rangeland from locusts were carried out in north Kazakhstan [3-5].

4. Solution

One of the promising technologies under investigation was barrier treatments, an approach that was widely used in earlier years. This technology has recently acquired particular importance due to the appearance of insect growth regulators (IGR) that have the potential for long residual action against locusts and nominal toxicity to non-target arthropods and terrestrial vertebrates. Dimilin (diflubenzuron) trials were carried out using ground sprayers of several modifications supplied to Kazakhstan by Uniroyal Chemical and Micronair Sprayers Ltd. Two formulations of Dimilin are accepted in the Republic for barrier and blanket methods.

The barrier treatment technology involves application of the IGR to vegetation in strips of not less than 40 to 100-m wide. In this case, droplets of a standard size can be sprayed without diluting with water. Uniroyal Chemical's Dimilin OF-6 (6% oil suspension of diflubenzuron) was applied with Micronair' s AU-811 0 sprayers with rotating atomizers. Trials of this application technology, new for Kazakhstan, were carried out in 1996-1997 in the Kachiry District of the Pavlodar Region.

In 1996, the registration trials ofDimilin OF -6 were carried out with the dose rate of 1.0 lfha, both for the barrier and block applications. The results proved that this formulation applied in barrier treatments against the gregarious locusts provided acceptable efficacy. The total cost per hectare decreased to a level comparable with that for pyrethroid

162

treatments. The efficacy was prolonged, as the residual action lasted up to the end of the locusts' activity, and there was no need for repeated treatments.

More detailed investigations were carried out in 1997 at "Beregovoe" and "Peschanoe" farms (Kashiry District, Pavlodar Region). The treatment areas consisted of wheat grass strips alternating with fallow strips. The grass fields were 6 to 8 years old, and the age of the fallows was 3 to 4 years. Plant cover was estimated at 35 to 55%, which is quite typical for fallow lands in the study area.

Locusts and grasshoppers were represented by 22 species. A complex of dominants consisted of species with different phenologies: C. italic us, D. brevicollis, D. kraussi kraussi, P. microptera, Ch. albomarginatus, Oedaleus decorus (Germ.), and Oediopoda miniata Pall.

The motorized AU-8000 knapsack sprayer was used to treat small plots, and the motorized AU-811 0 stand sprayer installed on a GAZ-52 truck was used for plots> 30 ha. The sprayers were calibrated for 200 !-Lm droplets. The average ground speed was 5 km/h for the AU-8000 sprayer and 20 km/h for AU-811 O.

In the barrier trials with Dimilin OF-6, the application rate of formulated material was 0.1 to I IIha (the chemical was diluted with diesel fuel). In the standard, blanket treatments, Decis (2.5% e.c. of deltamethrin at 0.4 IIha) was applied at a total volume of20 IIha (the chemical was diluted with water).

The area of each block was 3 200 m2, and the treatments replicated twice. On the day oftreatment only non-gregarious grasshoppers (3fd to 4th instars) were present in the experimental plots. Mass hatching of the Italian locust started later, 2 to 3 days after the treatment.

Decis attained its maximum efficacy (94.9%) 24 h after the treatment (Table 1), but residual action lasted for only 4 to 5 days. Further hatching of the Italian locust and migration of acridids from adjacent untreated plots considerably decreased the efficacy of Decis. Ten days after treatment, the abundance of locusts and grasshoppers in the Decis and untreated, control plots were actually equal.

A different pattern was observed in the plots treated with the Dimilin OF-6. The abundance of locusts and grasshoppers noticeably decreased 5 to 6 days after the application. After 10 days, the efficacy of the chemical was >90%, even with the dosage of 0.25 I/ha. The acridids that remained in the plots treated with the Dimilin OF-6 were mainly adults and older instars of non-gregarious grasshoppers. They moved slowly, lost hind legs, and their colouration was changed.

The observations of feeding by locusts in plots treated with Dimilin OF-6 revealed that the chemical interfered with food assimilation by these pests. The chemical probably impedes plant consumption in addition to providing direct mortality. As a result, it significantly decreases damage caused by locusts, although this aspect needs more research.

Two trials were carried out using lattice or grid (intersecting barriers) treatments of relatively large areas (100 ha each). The treatment was applied with the AU-8110 sprayer in the 4-year fallow field. The chemical was not diluted with diesel fuel. Nonswarming grasshoppers comprised 46% of the acridid community. The weighted average developmental stage was instar 3.6. The Italian locust was represented by I st and 2nd instar nymphs.

163

TABLE I. Biological efficacy ofDimilin OF-6 on non-gregarious grasshoppers and the Italian locust

under the blanket treatment (1997)

Trial variant Parameter Day after treatment

0 3 5 8 to 12 15

Dimilin OF-6, density, 29.0 30.0 24.0 21.0 18.0 8.1 2.1 1.5 O.lllha insects/m2

biological 23.1 33.3 44.7 56.1 77.5 94.7 96.5 efficacy, %

Dimilin OF-6, density, 28.0 28.0 29.0 20.0 14.0 3.5 1.2 1.0 0.2511ha insects/m2


Dimilin OF-6, density, 36.0 32.0 30.0 24.0 8.0 1.9 1.0 1.0 0.511ha insects/m2


DimilinOF-6 density, 40.0 31.0 27.0 17.0 7.0 1.8 l.l 1.0 0.7511ha insects/m2


Dimilin OF-6, density, 34.0 32.0 26.0 18.0 6.0 1.5 1.0 1.0 1.0l IIha insects/m2


Decis 2.5% e.c., density, 38.0 2.0 3.0 5.0 17.0 40.0 44.0 42.0 0.411ha insects/m2

(standard) biological 94.9 91.7 86.9 58.5 11.0 10.0 2.3 efficacy, %

Untreated density, 31.0 39.0 36.0 38.0 41.0 .36.0 40.0 43.0 control insects/m2

An area of 100 ha (400x2 500 m) was treated on 12 June. The width of the longitudinal barriers was 50 m; the width of the cross-barriers was 100 m, and the inter-barrier space was 300 by 380 m. As such, the actual treated area was 43 ha. The application rate ofDimilin OF-6 was 1.0 I/ha, taking into account double application in the places of barrier crossings. Mortality counts were conducted both within barriers and 150 m within the inter-barrier space.

164

The data given in Table 2 show that the abundance of locusts and grasshoppers within and between barriers actually did not differ on the 8th day after treatment. The efficacy of Dimilin OF-6 on the loth day reached 95.2 to 96.5%.

It was found that the 3,d to 4th instars of non-swarming grasshoppers were quite mobile. At population densities of 30 to 40 insects/m2 with plant cover up to 40 to 50%, nymphs migrated for 150 to 200 m in 2 to 3 days. This movement allowed them to reach the treated barriers and accumulate a lethal dose of Dimilin.

TABLE 2. Biological efficacy ofDimilin OF-6 on harmful locusts and grasshoppers under the block treatment

(1997)


0 4 8 10

Treated strip (barrier) density, insects/m2 36.0 11.0 1.5 1.1

biological efficacy, % 63.3 94.2 96.5

Protected area density, insects/m2 38.0 21.0 1.8 1.5

biological efficacy, % 30.0 93.1 95.2

Untreated control density, insects/m2 38.0 30.0 26.0 31.0

In the barriers and inter-barrier spaces, about 80% of insects had typical features of being affected by this chemical (missing hind legs, slow movements, etc.). An obvious decrease of abundance (up to 65%) was observed up to 50 m beyond the boundaries ofthe treated plot. As a result, there were 130 ha actually protected. In this case, the actual Dimilin OF-6 dose rate was 0.35 l/ha, as compared to an average of 0.45 I/ha within the area delimited by the barriers (100 ha).

Thus, the trial of the lattice scheme of application with Dimilin OF-6 proved the feasibility of using this method against a complex of harmful locusts and grasshoppers.

To improve tactics of chemical control (decreasing dependence on wind direction and speed, increasing productivity of sprayers, and diminishing chemical dose rates), the barrier treatment method was also tried. Treatments were carried out in two neighbouring fields on 16 June. The treatments were implemented in strips of perennial grasses which were 6 to 7 years old, alternating with the fallow strips which were 3 to 4 years old. The non-swarming grasshoppers (12%) were represented by 4th and 5th instar nymphs, and the Italian locust (88%) was primarily in the 2nd to 3,d instar (Table 3).

165

TABLE 3. Conditions for Dimilin OF-6 treatments under the scheme of parallel barriers (1997)

Parameter Field "A" Field "B"

Protected area 480 ha (2 400x2 000 m) 275 ha (I 250x2 200 m)

Barrier width 80-120m 80-120 m

Distance between barriers 100m 200m

Treated area 287 ha 121 ha

Dose rate of the chemical per 0.241/ha 0.661/ha barrier

Dose rate the chemical per 0.151/ha 0.291/ha protected area

Time spent for treatment 2.5 h I.5h

As shown in Table 4, a marked decrease in abundance oflocusts and grasshoppers in both fields was seen the 5th day. The efficacy on the 8th day (89.4 to 88.1 %) remained stable during the entire period of the study. Mortality counts after 40 days showed that locusts did not re-colonize the treated fields. The density of the pests ranged from 1.1 to 1.5 per m2, while there were 10 to 12 insects/m2 in the untreated field.

TABLE 4. Changes of acridid density and treatment efficacy under parallel barrier trials (1997)


0 5 8 10 12

Field "A" density, insects/m2 32.0 15.6 2.5 1.7 1.3

biological efficacy, % 44.3 89.4 90.8 91.1

Field "8" density, insects/m2 33.0 18.5 2.8 2.1 1.8


Untreated control density, insects/m2 30.0 28.0 23.5 18.5 14.5

Results of these trials showed that the barrier treatment method provides acceptable control of both Italian locust swarms and dispersed populations. Application of the Dimilin OF-6 at 0.15-0.30 l/ha with a barrier width of 100 m and an inter-barrier space of 100 to 200 m provides stable reduction of acridid abundance for 40 to 45 days.

166

In 1997 (on 25 June and 2 July) personnel ofthe Kachiry District Plant Protection Station assessed the efficacy of the barrier treatments against the Italian locust (78 to 90% adults), over an area of 1 800 ha. The treatment scheme was as follows: barrier width of 100m; inter-barrier space of 200 m; Dimilin OF-6 applied to barriers at 0.5 I/ha. The density of locusts at the time of treatment was 28 to 36 insects/m2. Mortality counts after 22 and 30 days showed pest densities in the treated fields were 2.1 and 1.6 insects/m2, respectively, compared to 15 and 14 insects/m2, respectively, in the untreated, control fields.

Thus, 2 years of monitoring control programmes for various species oflocusts and grasshoppers have shown that the action ofDimilin does not lead to rapid mortality. On the 3rd to 4th day after application mortality becomes evident and the following changes can be observed: slowed and uncoordinated movements, lack of jumping, reduced feeding, loss of hind legs, dispersion of dense swarms, limited migrations, and mass mortality of insects during molting. In case of survival after molting, there are abnormalities of legs, antennae and wings. The cuticle is discoloured, thin and soft, resulting in less resistance to unfavourable thermal and moisture conditions. The adult insects are unable to fly, and females have decreased reproduction ability.

Due to lower mobility, the locusts become easy targets for predators (insects, lizards etc.). The plots treated with Dimilin OF-6 attracted flocks of crows, jackdaws, starlings, and other birds that further reduced locust and grasshopper abundance. Nontarget arthropods were not killed here in contrast to plots treated with pyrethroids.

Comparative assessments of economic efficacy of conventional, blanket applications of insecticides for locust and grasshopper control and the new approach using barrier/lattice methods are shown in Table 5. The new technology has the following advantages compared to the conventional one:

• fewer personnel engaged in treatments; • much higher productivity (especially for barrier treatments), allowing treatments to be

applied to larger areas in shorter periods of time; • gusty winds >5 m/s are not a restricting factor for the use of the AU-8110 atomizing

sprayer (compared to aerial applications); • less harm to the environment; • long residual action (up to 45 days) with low dose rates of Dimilin (this is the most

important advantage especially in those areas where hatching of locusts is not synchronized, and repeated treatments with short-lived chemicals can lead to additional economic costs and to greater insecticide loads in the environment);

• low costs for treatment (for the barrier method, expenses are 1.5- to 3-times less compared to a single application ofpyrethroids).

Thus, the ground-based barrier/lattice application of Dimilin OF-6 against harmful locusts and grasshoppers was a considerable improvement in the technology, strategy and tactics of pest management. These methods allowed preventive treatments of the most important crops before locust hatching, during the first wave of hatching, and during autumn migrations of adults.

167

TABLE 5. Economic efficacy of different technologies of locust and grasshopper control

Pyrethroids Pyrethroids Dimilin OF-6 Dimilin OF-6 (ground (air spraying) (blanket (barrier spraying) treatment) treatment)

Required OP 2000 sprayer, AN-2 aircraft, AU 8110 sprayer, AU 8110 sprayer, machinery MTZ-80 tractor, water tank truck (GAZ-52) truck (GAZ-52)

Water tank vehicle (GAZ-vehicle (GAZ- 53), truck (GAZ-53), truck (GAZ- 52), fuel filler 52)

Attending 4 14 2 2 personnel

Width of coverage, 20 40 100-120 100-120 m

Average treatment 54 800-1 000 1200 3600 rate, ha/day

Water 10 800 20 000-25 000 0 0 consumption, I/day

Fuel consumption, 0.5 l.l 0.05 0.02 kg/ha

Insecticide Decis - 0.4 Decis - 0.4 0.25 0.125 consumption, l/ha Sumi-alpha - 0.3 Sumi-alpha - 0.3

Fastac - 0.15 Fastac - 0.15

Insecticide cost, 3-4 3-4 5.0 2.5 $/ha

Treatment cost, 2.5 4.5 0.16 0.04 $/ha

Total expenses, 5.5-6.5 7.5-8.5 5.16 2.54 $/ha

Duration of 4-6 (new 4-6 (new minimun 45, really up to the end of protection, days treatment may be treatment may be adualt activity, reduced reproduction

required) required)

On the basis of our trials in 1996-1997, the following objectives were set for the year 1998:

I. To introduce the developed technology of Dimilin OF-6 application for the control of harmful locusts.

168

2. To carry out registration trials of a new formulation, Dimilin SC-48, for control of harmful locusts. 3. To specify physiological, morphological and behavioural changes oflocusts induced by diflubenzuron. 4. To estimate the impact of diflubenzuron on beneficial and non-target organisms. 5. To train the experts of plant protection stations in the new technology oflocust control using Dimilin. 6. To study changes in locust populations in the areas treated by Dimilin OF-6 in 1997.

The operational application of Dimilin OF-6 was carried out on crops and rangeland in the Kachiry, Jelezinski, Irtyshski, and Pavlodarski districts. More than 20 species of acridids were found on the plots. The complex of dominant species included a range of phenologies. The earliest hatching species were P. rnicroptera and Bryoderna tuberculaturn (F.), which hatch at the end of April or the beginning of May. Their abundance is usually low, and they are insignificant as pests. In the middle of May comes the hatching the most abundant and harmful grasshoppers: Ch. albornarginatus, D. brevicollis, Ae. sibiricus, D. kraussi kraussi, Euthystira brachyptera (Ocsk.), andS.fisheri. In the beginning and middle of June, C. italicus, S. scalaris, and some other species hatch. Thus, the period of locust and grasshopper hatching lasts more than 1.5 months, which creates additional problems in the organization of control campaigns.

As a result of a late spring frost in the absence of snow cover, locust abundance decreased markedly on some hillocks due to destruction of eggs. However, infestation by the locusts, and especially by C. italicus, was higher than in 1997. Hatching of the early spring grasshoppers was observed at the end of May, with hatching of Dociostaurus and Charthippus in the middle of June.

The first adults of C. italic us were observed 28 June, and mass emergence took place in the beginning of July. In comparison with 1997, the density of the Italian locust increased, and very dense migrating bands of hoppers were observed.

The company "KazTechInvest" was involved in the locust control programme in the Pavlodar Region, where local acridid communities consisted of about 80% Italian locusts. This company provided an aerosol fogger with regulated droplet dispersion. The applications were made using the barrier method. The rate of Dimilin OF-6 in the barriers was 0.3 l/ha ( i.e., 0.15 IIha on the protected area). The treatments were done on 14 and 15 June at night, with an average air temperature 20 to 23°C. In total, 20 000 ha were protected with the barrier method (Dimilin was applied to about 10 000 ha). The pest density counts, which were carried out on the 8th day after application, showed a decrease in the locust density that ranged from 180 to 250 insects per m2 to a density of 1 to 3 insects per m2 • Thus, the biological efficacy of Dimilin OF-6 in this treatment regime was nearly 100%. The insects active after the treatment were mainly grasshopper adults with obvious symptoms of diflubenzuron intoxication (slow movements and missing hind legs). The survey of this area carried out in August (at the end of a summer generation of the Italian locust), revealed only 3 to 5 locusts/m2, whereas the neighbouring fields had an average density of 30 locusts/m2 .

About 15 000 ha of the crops and rangeland in the Kachiry, Irtyshski, and Jelezenski districts were treated with Dimilin OF-6 barriers applied with an AU 8115

169

sprayer mounted on a UAZ-469 truck. This sprayer was calibrated to deliver 200 Ilm droplets, and the average speed ofthe truck was 20 km/h. The swath width was determined by arranging water-sensitive cards; the average swath width was about 100 m. The density of the Italian locust and grasshoppers was reduced (after 7 to 10 days) from a range of35 to 50 locusts/m2 to 1.0 to 1.5 locusts/m2 on crops and rangeland. In these areas, the locusts were virtually nonexistent a month after application. Thus, in 1998, the high biological efficacy and long-term protective action of Dimilin OF-6 were confirmed. Next year, we plan to monitor the dynamics of the pests in the treated areas.

In 1998, the registration trials of a new formulation of Dimilin (48-SC) were carried out for control of locusts and grasshoppers. The first trial took place on 10 June on 10-year old wheat grass fields in the former state farms "Beregovoe" and "Mynjasarov". The soil was dry and compacted to a depth of up to 20 cm. On the west and the south, the site was bordered by four windbreaks of aspens. The applications were carried out using a "Micronair" AU 8000 sprayer mounted on a IJ-2715 vehicle which travelled at an average speed of20 kmlhour. The sprayer was calibrated for a droplet size of200 Ilm. The total rate of application was 2 I/ha (0.02 to 0.1 I/ha of Dimilin). The acridid assemblage included seven species, and P. microptera, C. italicus, D. brevicollis, and S. kraussi were dominant (79%). The density of locusts before spraying ranged from 18.6 to 33.2 insects/m2 • The average weighted ages of non-swarming grasshoppers and Italian locusts were instar 3.0 and 1.2, respectively. On the plots where Dimilin was applied, in the first 3 days after application, sharp fluctuations of locust abundance were observed. These dynamics were caused by mass migration of hoppers from the adjacent untreated sites to the treated areas. Consistent and high efficacies (95 to 96%) across all doses of Dimilin were observed on the 7th day after application, and these mortality rates were maintained for 14 days after insecticide application. The visible traits of Dimilin intoxication on the acrid ids were evident within 3 to 5 days of application (missing hind legs and antennae, deformation of ventral segments, decreased activity, etc.). Such insects were quickly taken away by ants, predatory ground beetles, or consumed by birds and lizards. In the standard application (Decis 2.5% e.c. at 0.35 I/ha), on the first day after the application, the efficacy was 56.2%; after 3 days it rose to 78.8%; dropped to 59.5% after 7 days; and fell to just 14.0% after 14 days. By that time the locust density was practically equal to the density in the control (untreated) plots.

Similar results were obtained in another trial, on 11 June, in the territory of the former state farms of "Bobrovskoie" and "Beisembaeva", where the Italian locust was dominant (78.3%). The average weighted ages of the Italian locusts and non-swarming grasshoppers were instars 1.2 and 3.2, respectively. The study plot consisted of wheat grass field strips regularly alternating with fallow fields (strip width 100 m). The soil was dry and compacted to a depth of up to 20 cm. The experimental plot was bordered on the west side by bush windbreaks. The applications were carried out using a "Micronair" AU 8000 sprayer mounted on IJ-2715 vehicle traveling at an average speed of20 km/h. The sprayer was calibrated for droplet size of 200 J,lm, and the total rate of application was 1.5 I/ha (0.02 to 0.1 Ilha of Dimilin).

Data on the biological efficacy of Dimilin 48-SC were compared with the control of the mixed acridid populations at the "Beisembaeva" farm. Acridid density before the application varied from 33.2 to 61.0 insects/m2 • As in the previous trial, the action of

170

Dimilin was not pronounced on the first day after treatment. After 3 days, the efficacy increased, ranging from 38.1 to 44.2%. After 7 days the efficacy was 82.8 to 89.5%, and after 13 days it reached 96.1 %. Sublethal intoxication by Dimilin was observed in the adults of the non-swarming species P. microptera and Angaracris spp.; they lost legs and could not fly. Thus, the results of the previous experiments with Dimilin 48-SC were confirmed: at 0.02 to 0.1 I/ha Dimilin controlled both the Italian locust and non-swarming grasshoppers. In the standard treatment (Decis 2.5% e.c. at 0.35 l/ha), the efficacy was comparatively higher immediately after application (up to 84.3% on the 3rd day), but mortality decreased to 56.9% by the 13 th day.

At all rates, Dimilin did not cause negative impacts to beneficial arthropods. In the plots treated with Dimilin, the abundance of entomophages did not differ significantly from that of the untreated, control plots. In plots treated with Decis, a decrease in the abundance of some entomophagous species was observed. Two days after application, ladybird beetles, meIyrid beetles, green lacewings, ants, and hymenopteran parasites (ichneumonid and braconid wasps) were completely eliminated, and the spider abundance was significantly reduced. After 10 to 14 days, the abundance ofthese arthropods (with the exception the melyrid beetles) was restored, probably due to immigration from the untreated sites.

On 17 June, we studied the efficacy of UL V and LV Dimilin formulations. The Italian locust was the prevailing species (72%) on the treated plots. Dimilin OF-6 was applied by the atomizer sprayer AU 8000, without being diluted by water, at a rate of 0.15 IIha. Dimilin SC-48 (0.02 I/ha) was applied at a total rate of 1.5 IIha. Dimilin SC-48 ( 0.04 and 0.06 I/ha) and the standard treatment (Kinmix 5% e.c.) were applied at a rate of 24 I/ha. The size ofthe trial plots was I OOx 100 m (I ha).

Table 6 shows that Dimilin SC-48 yielded a rather high efficacy (94.0 to 95.2% in 14 days) at all tested rates, regardless of the total volume applied per hectare. Similar results were obtained using Dimilin OF -6. These results demonstrate that it is not necessary to increase the volume ofthe carrier in order to control locusts with blanket applications of 0.02 to 0.06 l/ha. The efficacy of a minimal rate ofDimilin SC-48 did not differ from that of Dimilin OF-6. This probably could be explained by the identical rates of active ingredient per hectare (9 and 9.6 g a. i., respectively). The biological efficacy of Kinmix (standard) peaked on the first day after application, then declined sharply reaching just 14.1 % after 14 days.

Based on the results of testing the new formulation of Dimilin (SC-48), we recommended adding this product (0.02-0.03 l/ha in blanket application) to the approved list of materials for control of the Italian locust and non-swarming grasshoppers. The following rates of Dimilin SC-48 were recommended for barrier applications: treat mixed populations oflocusts and grasshoppers with 0.03 to 0.06 l/ha (in the barrier); treat locusts with 0.06-0.09 I/ha. It should be emphasized that the formulation of Dimilin as a suspension concentrate (SC) allows application by atomizer UL V equipment, as well as by aerosol foggers and conventional sprayers mounted on AN-2 aircraft, deltaplanes, and tractors.

171

TABLE 6. Biological efficacy of different formulations of Dimilin for acridid control (Kachiry District, farm "Beisembaeva") (1998)


0 3 7 9 14

Dimilin OF-6, 0.15 I/ha density, insects/m2 60.8 48.2 6.2 3.6 6.6


Dimilin SC-48, 0.02 density, insects/m2 57.8 49.2 10.2 3.8 3.8 I/ha


Dimilin SC-48, 0.04 I/ha density, insects/m2 59.2 48.4 9.8 2.6 3.0


Dimilin SC-48, 0.06 l/ha density, insects/m2 63.8 48.8 12.4 3.6 3.6


Kinmix 5% EC, 0.35 density, insectslm2 55.0 7.4 25.4 47.2 52.4 l/ha (standard)


Untreated control density, insects/m2 53.6 73.2 46.6 55.2 59.2

On 13 June, the study farm "Peschanoe" conducted a seminar with participation by the Head of Administration of the Pavlodar Region, during which the results of the 2 years of tests with Dimilin OF-6 were discussed, and the work of the atomizer sprayer AU 8000 was shown. The experts of the Kashiry District Plant Protection Station acknowledged the efficacy ofDimilin OF-6 for acridid control and recognized its ecological safety.

In summary there are several advantages of this new technology oflocust control by using the insect growth regulator Dimilin compared with traditional methods. Due to an extended residual activity, only one application is sufficient for effective locust control. Moreover, this application can be carried out early in the pest's phenology -- even before the Italian locust hatches, and the protective effect may be observed in the treated fields during the next year. Use of barrier applications reduces costs of control 1.5- to 3-times compared with applications ofpyrethroids. The low toxicity ofDimilin allows applications near villages, environmentally sensitive habitats, and water reservoirs. The opportunity of using different formulations ofDimilin allows pest managers great flexibility in deploying available application equipment (AU-8000, 8110 and 8115, aerosol foggers, aircraft AN-2, deltaplanes, and tractor sprayers).

172

5. Acknowledgements

I thank the Head of Kazakhstan Representative office ofthe Uniroyal Chemical Company, Mr. A.E. Slivkin, and leaders of Micron air Sprayers Ltd. for providing the opportunity to start trials ofDimilin with sprayers of different modifications in Kazakhstan. I express my gratitude to the companies' employees for their assistance in this work and active participation in large-scale experiments. I received ongoing support from leaders and experts of plant protection stations, representatives offarmers' groups in Akmola, Pavlodar and North Kazakhstan regions, and I thank them for their interest in developing this technology.

6. References

I. Kambulin, V.E. (1991) About the locust problem in Kazakhstan, Plantprotection in Kazakhstan 2,9-10. (In Russian)

2. Hossen, E.F. (1998) Basic groups of weeds in the north of Kazakhstan and ways to reduce their harmfulness, Plant protection in Kazakhstan 1,6-11. (In Russian)

3. Evdokimov, N.Ya., Kambulin, V.E., and Korchagin, A.A. (1998) About technologies oflocust control in Kazakhstan, Plant protection in Kazakhstan I, 12-11. (In Russian)

4. Suleimenova, Z.Sh., Zhukashev, A., and Baginsky, P. (1998) Experience of the use of a new technology of locust control, Plant protection in Kazakhstan 4, 23-26. (In Russian)

5. Evdokimov, N. Ya .. Temirgaliev, Zh. K .. and Dublyazhova. M. K. (1999) Problems of protection of agricultural crops and lands against harmful locusts, Plant protection in Kazakhstan 1,22-25. (In Russian).

11. CAN LOCUST CONTROL BE COMPATIBLE WITH CONSERVING BIODIVERSITY?

Abstract

M.J. SAMWAYS Invertebrate Conservation Research Centre School of Botany and Zoology University of Natal Private Bag XOJ Scottsville 3209 South Africa

There are four pest locust species in southern Africa, but only Locustana pardalina, the Brown locust, outbreaks frequently. It swarms most years in a 250 000 km2 area. Control of the Brown locust is principally by deltamethrin, which has low bird and mammal toxicity, albeit is harmful to many non-target invertebrates. The problem is whether the Brown locust can be controlled without having an adverse effect on indigenous and endemic biodiversity. Besides diazinon, fenitrothion and deltamethrin, the mycoinsecticide Metarhizium anisopliae var. acridum is now registered for use against the Brown locust. Most of the outbreaks are in small areas (0.1 - 0.5 ha) and are rarely in the same place in consecutive years. Although deltamethrin is highly effective against the target Brown locust, it impacts also many non-target, particularly flightless orthopterans, whose recovery takes a year or so. However, as the outbreak spots are small, this appears not to be a problem in terms of indigenous biodiversity. Although M anisopliae is effective, it does have, at least in the laboratory, adverse impacts on non-target hymenopterans. Its use, especially relative to persistence, needs to be monitored very carefully, particularly with regard to impacts on the large number of endemic, non-target invertebrates in the outbreak area.

1. Problem

There are four species of swarming and migratory locusts in southern Africa. Schistocerca gregaria jlaviventris (Burmeister), Locusta migratoria migratorioides (Reiche and Fairmaire) and Nomadacris septemfasciata Serville rarely form large swarms [5, 25]. There was however an outbreak of N. septemfasciata in 1996, the first since the 1940s [19]. These species contrast with the indigenous Locustana pardalina (Walker) (Fig. 1) which swarms in a 250 000 km2 area of the semi-arid Karoo of South Africa and southern Namibia (Fig. 2). It is known to have impacted on grazing land at least since 1795 [15].

173

J. A. Lockwood et al. (eds.), Grasshoppers and Grassland Health. 173-179. © 2000 Kluwer Academic Publishers.

174

This is the only orthopteran species in South Africa that is controlled over wide areas [I, 14, II]. It is the species where control is potentially the most hazardous to biodiversity, simply because of its abundance and the perception by some that there is an urgent need to control it (see the debates by Hockey [10] and Siegfried [24]).

Figure 1. The Brown locust Locus/ana pardalina (reproduced with permission from Annecke and Moran [I))

N

t

IH"tt.'rPo,1l frequency

,~High _Low

Figure 2. The outbreak area of the Brown locust in southern Africa

One of the problems is that outbreaks are so frequent that there have been only five years out of the past 47 years (prior to 1998) when there have been no control programmes. In seven outbreak seasons (1988/89 - 1994/95) alone, about 402 000 litres of ultra-low volume formulated insecticides were used to control 219 000 hopper bands

175

and 25 600 adult swanns, using either knapsack or vehicle-mounted mistblowers [26]. Control is mostly of hopper bands in spot locations of 0.1 to 0.5 ha. These spot locations rarely receive an insecticide application more than once per five years as they are so small in the vastness of the outbreak area [6].

In recent years, the synthetic pyrethroid deltamethrin has replaced the more persistent organophosphates in Brown locust contro!' This was driven partly by the fact that deltamethrin has low toxicity to birds and mammals [9] and the Karoo is an important 'Endemic Bird Area' [3]. However, deltamethrin is potentially hazardous to non-target invertebrates [22]. Indeed, Stewart [26] has shown that non-target grasshopper abundance is significantly reduced one day after treatment (in 0.25 ha plots), although the longer-term impact was relatively insignificant.

In short, the problem lies not with our ability to control the Brown locust, but whether it actually needs to be controlled (see below) and whether the control methods currently available are having an adverse effect on indigenous, and more importantly, endemic biodiversity in the outbreak areas.

2. Tools

As the Brown locust swanns on agricultural land, it is automatically monitored by the fanning community, which is obliged to report outbreaks to the government authorities. These, in turn, mobilize chemical and mycoinsecticide control programmes. Chemical pesticides currently available for use against the Brown locust in South Africa (Act No. 3611947) are diazinon, fenitrothion and deltamethrin [12]. Recently it was found that the exotic biopesticideMetarhizium anisopliaevar. acridum (formerly Mflavoviridae) isolate IMI 330189, when formulated as dry conidial powder in a paraffinic oil mixture and applied with a Micronair AU7000 fitted to a micro-light aircraft, was effective in causing up to 98% Brown locust mortality in three weeks [23].

The effectiveness of this mycoinsecticide resulted in its registration in 1999 in terms of Act 36 of 1947 under the trade name 'Green Muscle', and is a world first registration under this name [20].

The fungal spores have to impact on the insect cuticle, either directly from the spray or from secondary pick-up from treated vegetation. The spore residue can remain highly effective for up to three weeks [13]. On germination, the spores penetrate the host and develop in the haemolymph. Efficacy and speed of kill achieved by the fungus in the field are a complex interaction between pathogen and host which, in turn, is influenced by a range of environmental factors [20]. There is no minimum dose threshold and, in theory, even low dosages of the fungus can kill locusts under optimum conditions. Insects dying of the fungal infection can also release fresh spores which may then infect further hosts, especially under humid conditions.

Acute dermal and oral toxicity tests have indicated no ill effect on birds or mammals [20]. However, laboratory tests resulted in 100% mortality of the hymenopteran parasitoids Bracon hebetor (Braconidae) and Apoanagyrus lopezi (Encyrtidae), with average survival times of 1.6 to 4.1 days [7]. In the same study, however, and for reasons unknown, no infection resulted when two Sahelian tenebrionid beetles were topically treated at high conidial dose rates.

176

3. Solutions

Organophosphate insecticides such as fenitrothion are not a satisfactory solution, especially as locusts are an important food item for some vertebrates [21]. The synthetic pyrethroid deltamethrin, with its low toxicity to birds and mammals [9] is a more suitable alternative, but still with the disadvantage of being potentially harmful to non-target invertebrates [22]. However, when applied in relatively small-sized plots (0.25 ha, which is typical size of areas sprayed during Brown locust control) the recovery of the more mobile non-target grasshoppers from surrounding untreated areas occurs within about 64 days [26]. The flightless, low-vagility bushhoppers [Nepiothericles sp. (Eumastacidae), Phymella capensis (Pyrogomorphidae) Karruacris browni, Karruia paradoxa (Lentulidae)] are probably the most sensitive of the non-target orthopterans, and are important representatives of biodiversity as they are endemic to the Karoo. In Stewart's [26] study, their recovery took about one year, and was from eggs in the treated area rather than from immigration. However, the situation is complex as the time of recovery depends on when spraying is done relative to rainfall. Recovery takes longer at the dry end of the summer when populations are in natural decline.

In summary, the use of deltamethrin to control the Brown locust in the small-sized outbreak areas is a viable option for rapid knockdown on the one hand and environmental sensitivity on the other. However, it is essential not to treat the same area on a regular basis. But as the Brown locust outbreaks in different spots in different years, this problem is largely overcome.

The use of the mycoinsecticide Metarhizium anisopliae, although being hailed as a breakthrough, should nevertheless be considered with caution. Firstly, M anisopliae takes two to three weeks to kill the Brown locust, which is currently a disadvantage where locust control officers and farmers are accustomed to and require fast-acting insecticides. Also, the hopper band continues to wander and it is not clear after a while whether it was treated or not. Furthermore, spores are being widely dispersed, which is a high risk to nontarget grasshoppers. It has been stated that one plan is first to apply it in National Parks and other conservation areas where rapid control is less vital [20]. From a biodiversity point of view, and particularly for indigenous and migratory birds (e.g., abdim's stork, white stork, ludwig's bustard), such an approach is unacceptable.

From an environmental point of view by far the issue of most concern at present is the impact, not on non-target vertebrates, including humans, but on non-target invertebrates. Hymenoptera, whether parasitoids [7] or honeybees [8] decline when exposed to preparations of M anisopliae in the laboratory. Whether this is a serious problem under field conditions is clearly a pressing research issue.

A further consideration, but not an environmental one, is that the production process has to be carefully controlled, with spore batches needing to be >99.9% pure and with <5% moisture content and a spore viability of>85% [20]. This rules out the original ideal of establishing small-scale, rural production systems [20]. But then this also applies to organic chemical pesticides. An additional problem is whether it will be economically feasible unless subsidized.

177

Currently baseline data on grasshopper population levels are being gathered both in the Karoo (S. Gebeyehu, in prep.) and in Namibia (R. Kinvig, in prep.). These data and those of Stewart [26] will be useful in establishing the long-term impacts of introduced M anisopliae. This may not be easy to determine as the variability of environmental conditions may be either the prime impact or even synergistic with M anisopliae impacts. Adverse effects on insect diversity by M anisopliae nevertheless must be watched diligently on a regular basis to see whether non-target populations are declining from the introduced mycosis. While laboratory screening highlights which non-target taxa should be given specific attention in the field, only long-term monitoring will determine whether this mycoinsecticide is persistent or not and whether it poses a local or widespread threat to indigenous and endemic biodiversity.

4. But How Serious a Pest is the Brown locust. .. ?

The Karoo is an arid ecosystem, poor in organics and nitrogen [16, 18]. Large quantities of nutrients can become available only after rains and principally in the upper 30 cm of the soil.

The key question is how important is the Brown locust in the cycling of nutrients [4]? Locusts, in common with termites (but more diffusely) convert plants to nutrient-rich frass, as do sheep. But the insect frass being more finely divided, presumably becomes avaliable to plants faster than sheep dung which takes years to break down [17]. It has been well-established that grasshopper frass is a vehicle for rapidly recycling nutrients [2]. Boshoff [4] calculated that during the 1985/86 outbreak of the Brown locust that about 2.26 million tonnes of frass would have been produced by the locusts that were killed by spraying. This represents 14700 tonnes of nitrogen that would have been recycled. To purchase this amount of nitrogen would have cost between 66% and 107% of what was spent on spraying [10]. Boshoff [4] tentatively proposes that the Brown locust accelerates and escalates nutrient turnover during outbreaks, the function of which may be essential to the long-term stability and optimal productivity of the Karoo. Clearly, there are both financial and ecological nuances of Brown locust control that need addressing before we can be certain that control is necessary and worthwhile in all circumstances, at all places and at all times. The underlying pivotal issue may be one of spatial scale. While the findings of Boshoff [4] are compelling for abandonment of Brown locust control at the scale of the Karoo, this may not be the case at the scale of particular farms with a large outbreak. Not only is it necessary to reconcile these spatial scalar issues but it also appears that we need more data on the economic threshold, not just at the scale of the Karoo, but particularly at the scale of individual farms. This would not necessarily involve calculations on ail farms but rather, development of a set of guidelines from which the situation on individual farms can be deducted. Best then that control is by choice oflocustaffected individual or groups of individual farmers, rather than a generalized strategy for the whole Karoo.

178

5. Acknowledgements

Special thanks to the organizers and sponsors of the NA TO Advanced Research Workshops on Acridogenic and Anthropogenic Hazards to the Grassland Biome: Managing Grasshopper Outbreaks without Risking Environmental Disaster for the invitation to present this paper, and to Alexandre Latchininsky and Sue Milton for critically reading it, and Erika B6rner for preparation assistance.

6. References

I. Annecke, D.P. and Moran, V.c. (1982) Insects and Miles of Cultivated Plants in South Africa, Butterworth, South Africa.

2. Belovsky, G.E. (in press) Do grasshoppers diminish productivity? A new perspective for control based on conservation, in J.A. Lockwood, A. V. Latchininsky and M.G. Sergeev (eds), Grasshoppers and Grassland Health, Kluwer Academic Publishers, Dordrecht, (this volume)

3. Bibby, C.J., Collar, N.J., Crosby, MJ., Heath, M.F., Imboden, Ch., Johnson, T.H., Long, AJ., Stattersfield, A.J., and Thirgood, SJ. (1992) Putting Biodiversity on the Map: Priority AreasforGlobal Conservation, International Council for Bird Preservation, Cambridge.

4. Boshoff, C. (1988) Does the Brown locust play an important role in the nutrient cycling?, in B. McKenzie and M.Longridge (eds), Proceedings of the Locust Symposium in Kimberley, South African Institute of Ecologists Bulletin, Johannesburg, pp. 85-96.

5. Brown, H.D. (1986) New locust problem, Antenna 10, 11-13. 6. Brown, H.D. (1988) Current pesticide application: effectiveness and persistence, in B. McKenzie and M.

Longridge (eds), Proceedings of the Locust Symposium in Kimberley, South African Institute of Ecologists Bulletin, Johannesburg, pp. 101-117.

7. Danfa, A. and Van der Valk, H.C.H.G. (1999) Laboratory testing of Metarhizium spp. and Beauveria bassiana on Sahelian non-target arthropods, Biocontr. Sci. Tech. 9, 187-198.

8. Goettel, M.S., Poprawski, TJ., Vandenberg, J.D., Li, Z., and Roberts, D.W. (1990) Safety to nontarget invertebrates offungal biocontrol agents, in M. Laird, L.L. Lacey and E.W. Davidson (eds.), Saftty of Microbial Insecticides CRC Press, Boca Raton, pp. 209-231.

9. Greig-Smith, P J. (1993) Killing with care - can pesticides be environmentally friendly? Biologist 3, 132-136.

10. Hockey, P.A.R. (1988) The Brown locust - war or peace?, in B. McKenzie and M. Longridge (eds), Proceedings of the Locust Symposium in Kimberley, South African Institute of Ecologists Bulletin, Johannesburg,pp.145-159.

11. Kieser, M.E., Thackrah, A., and Twyman, L. (1999) A Brown locust early warning system (EWS). Proceedings of the Twelfth Entomological Congress, PotchefStroom. Entomological Society of southern Africa, Pretoria, p. 66.

12. Krause, M., Nel, A. and Van Zyl, K. (1996) A Guide to the Use of Pesticides and Fungicides in the Republic of South Africa, National Department of Agriculture, Pretoria.

13. Langewald, J., Ovambama, Z., Mamadou, A., Peveling, R., Stolz, I., Bateman, R., Attignon, S., Blanford, S., Arthurs, S., and Lomer, C. (1999) Comparison of an organophosphate insecticide with a mycoinsecticide for the control of Oedaleus senegalensis (Orthoptera: Acrididae) and other Sahelian grasshoppers at an operational scale, Biocontr. Sci. Tech. 9,199-214.

14. Lea, A. (1973) Locust control and research in southern Africa, Enl. Mem. Dept. Agric. Tech. Serv .. South Afr. 31, 1-14.

15. Lounsbury, C.P. (1915) Some phases of the locust problem, S.Afr. J. Sci. 11,33-45. 16. Milton, S.J. (1995) Effects of rain, sheep and tephritid flies on seed production of two arid Karoo shrubs

in South Africa, J. Appl. Ecol. 32,137-144. 17. Milton, S.l. and Dean W.R.J. (1996) Rates of wood and dung disintegration in arid South African

rangelands, Afr. J. Range Forage Sci. 13,89-93. 18. Milton, S.J., Dean,W.R.J., and Kerley, GJ.H. (1992) Teirberg Karoo Research Centre: history, physical

environment, flora and fauna, Tran. R. Soc. Sth. Afr. 48, 15-46.

179

19. MOller, E. and Price, R. (1997) Biological warfare against the red locust, Plant Prot. News Summer 1997, No. 50,13-15.

20. Moller, E. and Price, R. (1999) A century to develop a biopesticide for locusts, Plant Prot. News Autumn 1999, No. 54, 5-7.

21. Mullie, W.C. and Keith, J.O. (1993) The effects of aerially applied fenitrothion and chlorpyrifos on birds in the savanna of northern Senegal, J. Appl. £Col 30, 536-550.

22. Murphy, C.F., Jepson, P.C., and Craft, B.A. (1994) Database analysis of the toxicity of antilocust pesticides to non-target, beneficial invertebrates, Crop Prot. 13, 413-420.

23. Price, R.E., Bateman, R.P., Brown, H.D., Buller, E.T., and MOiler, E.1. (1997) Aerial spray trials against Brown locust (Locustana parda/ina, Walker) nymphs in South Africa using oil-based formulations of Metarhiziumjlavoviride. Crop Prot. 16,345-351.

24. Siegfried, W.R. (1988) Towards high-benefit and low-costs: a call for change in the Praetorian guard, in B. McKenzie and M. Longridge (eds.), Proceedings of the Locust Symposium in Kimberly, South African Institute of Ecologists Bulletin, Johannesburg, pp. 2-11.

25. Stewart, D.A.B. (1997) Economic losses in several crops following damage by the African Migratory locust, Locusta migratoria migratorioides (Reiche & Fairmaire)(Orthoptera: Acrididae), in the Northern Province of South Africa, Afr. Ent. 5,167-170.

26. Stewart, D.A.B. (1998) Non-target grasshoppers as indicators of the side-effects ofchemica1locust control in the Karoo, South Africa, J. Insect Cons. 2, 263-276.

12. HOW DOES INSECTICIDAL CONTROL OF GRASSHOPPERS AFFECT NON-TARGET ARTHROPODS?

Abstract

I.M. SOKOLOV Laboratory of Agrobiocoenology All-Russian Institute for Plant Protection (VIZR) Podbelsky Street, 3 St.-Petersburg - Pushkin 189620 Russia

Outbreaks of Calliptamus italicus L. and grasshoppers cause significant yield losses due to the recent extension of vast unmanaged agricultural lands in Russia. The problem ofa choice of adequate insecticides to control grasshopper outbreaks as well as the lack of knowledge on non-target effect of anti-grasshopper insecticides in Russia inspired the present research. To evaluate the side-effect of treatments against grasshoppers the new insecticide, fipronil, was chosen. Chlorpyrifos served as a standard control. Field trials were done near Ust-Ordynsky, 70 kIn to the north-east ofIrkutsk in Baikal region, Siberia. Four groups of plots were set up. The first (blanket treatment) was sprayed with chlorpyrifos (205 g a.i./ha), the second and the third (one blanket and one barrier treatment) were sprayed with fipronil (4 g a.i./ha), and the fourth plot was control (without treatments). Each plot occupied ~ 30 ha. Trials were conducted in pastures and fallows with a perennial grass Bromopsis inermis, to control grasshopper adults, particularly Aeropus sibiricus L., Stauroderus scalaris F.d.W., Chorthippus albomarginatus DeG., Pararcyptera microptera F.d.W., and Arcypterafusca Pall. Arthropods were collected with sweep-nets or with pitfall traps on the following sampling dates: one day before treatment and on the 1 st, 3d, 5th, 7th, 14th, and 21 st day after treatment. The analysis of the response of non-target arthropods to the fipronil application showed the following results. (1) Taxa living in the grass layer or regularly visiting it are most at risk (Chrysomelidae, Curculionidae, Miridae, Lygaeidae, Cicadellidae, Aphididae, Psyllidae, Agromyzidae, Pyralidae, Chalcidoidea, Ichneumonoidea, Thomisidae). (2) Within these groups the maximum hazards occur in herbivorous insects possessing mandibulate mouthparts (Chrysomelidae, Curculionidae); in herbivorous insects possessing sucking mouthparts and high grass-dwelling activity (Miridae, Lygaeidae); in predators (Thomisidae) and parasitoids (Chalcidoidea, Ichneumonoidea) inhabiting the upper part of the grass stratum. (3) Insects possessing sucking mouthparts and low grass-dwelling activity, or with grass-flying activity, are at less risk (Cicadellidae, Aphididae, Psyllidae). (4) The risk to ground-dwelling arthropods is less, and correlated with their ability to hide in litter, soil etc. (Cydnidae, Carabidae, Lycosidae). (5) Treated sites attract species which

181

J. A. Lockwood et al. (eds.). Grasshoppers and Grassland Health. 181-192. © 2000 Kluwer Academic Publishers.

182

use dead organic material for feeding and development (Silphidae). A comparison of the side-effects of fipronil and chlorpyrifos proved that fipronil was much more hazardous than chlorpyrifos. At the same time fipronil was more selective due to the reduced mortality of non-target arthropods with "cryptic" modes of life (Cydnidae, Cicadellidae, Lycosidae). An effective way to diminish the negative impact of grasshopper control treatments on non-target fauna appeared to be the method of barrier spray applications, which allows a quick recovery of paras ito ids, predators and herbivores with sucking mouthparts (Chalcidoidea, Ichneumonoidea, Thomisidae, Miridae, Lygaeidae, Cicadellidae). On the contrary, even in barrier treatments we did not observe a recovery of chrysomelid beetles (with mandibulate mouthparts), as in the target Acrididae. As a result of conducted research, the use of fipronil for grasshopper control in Russia is proposed.

1. Problem

Within the Orthoptera Calliptarnus italicus L. and grasshopper species (Aeropus sibiricus L., Stauroderus scalaris F.d.W., Chorthippus albornarginatus DeG. etc.) cause the most damage to agricultural crops in Siberia [7]. Grasshopper populations have increased significantly in recent years due to the appearance of vast fallow, abandoned and other untreated agricultural lands, and have caused serious economic problems [8, 11]. Hence the selection of appropriate insecticides to control grasshopper outbreaks has become an issue. Within the last 30 years, Russia has seen several generations of insecticides: organochlorines were replaced by organophosphates, the latter - by pyrethroids (at least partially). Currently a few new groups of promising insecticides are being tested such as the phenyl-pyrasol group. Fipronil, for example, is considered to be more effective and persistent, relative to previously used insecticides. However, fipronil is reported [2] to reduce biodiversity and to impair ecological processes within the treated ecosystems. Therefore, the potential dangers of fipronil to the environment, as well as the vacuum of knowledge about the side-effects of insecticides used to control grasshoppers in Russia [10], have provided a motivation for this research. The objective of the research was to understand how the treatments influence non-target invertebrates and to explore methods of insecticide application that will minimize the negative effect to the environment.

2. Tools

Two insecticides, fipronil and chlorpyrifos, were selected to evaluate the side-effects of grasshopper control treatments on non-target insects. Field trials were done near UstOrdynsky, 70 km to the north-east of Irkutsk in Baikal region, Siberia. The experimental site was located near Ust-Ordynsky on terrace slopes of the Kuda River in fields of perennial grasses (cropped with Brornopsis inerrnis Leys.) and in natural pastures. The experiments were conducted from 28 June till 25 July. Four experimental square plots were established; the first (blanket treatment) was sprayed with chlorpyrifos (205 g a.i./ha), the second and the third (one blanket and one barrier treatment) were sprayed with fipronil

183

(4 g a.i./ha), and the fourth plot served as a control (without treatments). Plots with chlorpyrifos (2 replications, each 15 ha) and one (blanket treatment) plot with fipronil (2 replications, each 16,2 ha) were located 200 m from each other in the fields of perennials. One more (barrier treatment) plot with fipronil was located 2 kIn from the blanket-treated plots and was represented by two replications (15 ha each): the first one was placed on natural pastures, the second on the field with Bromopsis inermis. Width of the treated and the untreated strips was the same - 30 m. A control plot (15 ha) was located between the blanket-treated and barrier-treated fields in the ecotone zone, which included both mesophytic and xerophytic patches with vegetation of various grasses and forbs. Spraying of all plots was done by a tractor ventilator sprayer OVT-2000 with a spray coverage of about 20 m and was carried out on 1 July (blanket treatment with chlorpyrifos), 3 July (blanket treatment with fipronil), and 4 July 1997 (barrier treatment with fipronil) to control high densities of non-swarming grasshopper adults (9.6-11.0 ind.lm2). The species complex consisted of Aeropus sibirieus L., Stauroderus sealaris F.d.W., Chorthippus albomarginatus DeG., Pararcyptera mieroptera F.d.W., and Arcypterafusea Pall.

Arthropods were collected with sweep-nets or pitfall traps at the following sampling dates: one day before treatment and the 1st, 3d, 5th, 7th, 14th and 21st day after treatment. On each sampling date, 40 pitfall traps and 20 sweep-net samples were taken from a control plot, 20 sweep-net samples were taken from the plot that received a barrier treatment, and 30 pitfall traps and 20 sweep-net samples were taken from each plot that received a blanket treatment. In the latter case, 12 pitfall traps and 10 sweep-net samples were located outside of the blanket treated area and used as a local control plot. The scheme of sampling is shown in Fig. 1.

The structure of the arthropod community collected during the investigation is shown in Tables 1-2. Arthropods collected in pitfall traps were represented by a small number of dominant (>5% of abundance) and subdominant (1-5% of abundance) taxa (=families). In contrast, arthropods collected by sweep-nets were represented by a small number of dominant and a large number of subdominant taxa.

2.1. EFFECT OF FIPRONIL BLANKET TREATMENT ON NON-TARGET PHYTOPHAGOUS INSECTS

Herbivorous insects appeared to represent the group most at risk by application offipronil. This is not surprising, since the insecticide is aimed against grasshoppers, which exhibit a similar type offeeding behaviour. The effect of fipronil on the relative abundance of arthropods is shown for the following herbivorous taxa: non-swarming grasshoppers (Acrididae, Orthoptera), chrysomelid beetles (Chrysomelidae, Coleoptera), plant bugs (Miridae, Heteroptera), and leafhoppers (Cicadellidae, Homoptera). The population densities of all herbivores were significantly reduced by the treatment (Fig. 2). The rate of population decrease and the time of recolonization varied among the different taxa. A minimal effect was registered for leafhoppers. Within 7 days they lost 60% of their popUlation; afterwards their abundance began to grow. The negative effect was observed also for Curculionidae (Coleoptera), Lygaeidae (Heteroptera), Aphididae and Psyllidae (Homoptera), Agromyzidae and ef Muscidae (Diptera), Pyralidae (Lepidoptera).

184

A 0 " B " ()

" <> () ()

()

• • •

• • • •

• • "

-!-.~. I ~ () ( ) ( ) ( ) ( )

'-:-Tj"

" ()

()

()

/ ()

()

• • •

• • • •

• •

(-, (I () (; \) ( ) (l () ()( l

()

" ()

\ ()

()

\ " () \ t

\ ()

D

')

• • •

• • • •

• • I :: 1.. __ ------'

• • •

• • • •

• • <)

Figure 1. Scheme of pitfall trap distribution ( circles) and routes of sweep net sampling (arrows) while collecting arthropods on experimental plots.

A - plot with blanket treatment with fipronil; B - plot with blanket treatment with chlorpyrifos; C - control plot, without treatment; D - plot with barrier treatment with fipronil. Filled circles - pitfall traps on treated

area; empty circles - pitfall traps on untreated area (on A and B = local control plots)

2.2. EFFECT OF FIPRONIL BLANKET TREATMENT ON PREDATORS, PARASITOIDS AND NECROPHAGOUS INSECTS

The impact offipronil on predators and parasitoids was less than on herbivores (Fig. 3). The changes in relative abundance are demonstrated for the following taxa: parasitoids (Chalcidoidea, Hymenoptera), ground-beetles (Carabidae, Coleoptera), ants (Formicidae, Hymenoptera), and necrophagous beetles (Silphidae, Coleoptera). These groups exhibited a wide spectrum of responses - from the absence of changes in relative abundance

I. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

I. 2. 3. 4. 5. 6. 7. 8. 9. 10. II.

TABLE 1. The structure of arthropod community on treated and control plots (target and 10 top-ranked non-target families captured with pitfall traps)

185

Fipronil Chlorpyrifos Control (without treatments)

Family, % Family, % Family, % Acrididae 33.8 Acrididae 38.3 Formicidae 30.6 Carabidae 22.4 Carabidae 22.7 Acrididae 18.2 Cicadellidae 12.6 Silphidae 17.1 Lycosidae 17.0 Silphidae 12.3 Cicadellidae 6.2 Carabidae lOA Formicidae 7.3 Lygaeidae 3.0 Cicadellidae 5.9 Lygaeidae 3.0 Chrysomelidae 2.5 Lygaeidae 4.6 Lycosidae 3.0 Lycosidae 1.8 Silphidae 2.3 Cydnidae 0.9 Cydnidae 1.7 Curculionidae 2.3 Dermestidae 0.8 Formicidae 1.2 Tenebrionidae 2.0 Reduviidae 0.7 Phalangiidae 1.2 Tettiogonidae 1.3 Phalangiidae 0,5 Dermestidae 1,1 Apidae 0,7

TABLE 2. The structure of arthropod community on treated and control plots (target and 10 top-ranked non-target families captured with sweep-nets)

Fipronil Chlorpyrifos Control (without treatments) Family, % Family, % Family, % Thomisidae 14.5 Thomisidae 24.3 Lygaeidae 22.1 Acrididae 14.0 Cicadellidae 17.7 Miridae 10.2 Cicadellidae 13.3 Chrysomelidae 15.6 Formicidae 6.9 Lygaeidae 9.3 Acrididae ll.l Thomisidae 6.9 Formicidae 5.2 Psillidae 5.8 Cicadellidae 5.3 cfMuscidae 4.6 Lygaeidae 5.0 Curculionidae 5.1 Ichneumonidae 4.4 cfMuscidae 3.2 Aphididae 4.4 Miridae 4.1 Tephritidae 2.2 cfMuscidae 4.1 Curculionidae 3.4 Ichneumonidae 2.1 Agromyzidae 3.7 Chrysomelidae 3.3 Agromyzidae 1.4 Chrysomelidae 3.7 Chalcididae 3.3 Pyralidae 1.4 Acrididae 3.5

(Formicidae and Silphidae) to various degrees of population decline after the treatment (Chalcidoidea and Carabidae). No effect was observed also in Lycosidae (Araneae), however, Thomisidae (Araneae) and Ichneumonoidea (Hymenoptera) populations declined after treatment. Chalcidoidea and Thomisidae demonstrated the strongest decrease in abundance. The arthropods belonging to these families started recolonization on the 7th day after treatment, noticeably later than representatives of the other taxa. In contrast to these groups, ants exhibited a slight increase in abundance after the treatments. This fact, as well as a comparatively fast recolonization of the treated areas by ground beetles, might indicate that migration plays a significant role in relative abundance of certain arthropods. The migration into the treated territories was also demonstrated by the changes in relative abundance of necrophagous beetles. The application of an insecticide leads to an abrupt increase in the amount of dead organic material, providing food to organisms specialized in utilization of cadavers. The general feature of such organisms is a high level of searching and migration activities, as well as a strong resistence to aggressive environments, due to perfect detoxification systems. The analysis of the abundance of SiJphidae showed that their increase in number takes place immediately after the insecticide

186

i j

Acrtdidae, Orthoptera

: : : .. --..... -~.' ....... -.--.. --~ .. --... -..... -.~ ····r··_·_······_··r·····_· __ ···t· i i !

..... -J .. -........... ~ ...... -..... ..l ............... 1 ... - .... __ .-.1 ............. ..[ .............. .1. .. .

l ······-1---·

tsttl·~·l~ :1:::1 .. ]. f 'r ~:~~:r::r~~: .:_:' .' f··· + .: ... : ..... :~.: '~). ~. 1., ........ :- ...... J .. ~..:::::.i . :Ii •.•.••.•..••... :-

1 '".;::. .. ::= •. "' ..• I"' .. = .. ~ .. = ... ~ .. f"i --(J' : ! c: 2 (, -·i····---·····i-········ ... ,:-, .. : ~

'.' ,,--,.---,---,---,.----;---,-----,-,

f 1 3,0

iii

!Chrysomelldae, :Coleopte~ . -1" .. -+. • J :

.. ·i .......... I ... ·----L~~·~:l:·~·-+ .. ···-:·~·:·· .. .I........... ... 1 ! 15 J ... j "~~==+= ... +j ... = .......... -t=. =;t-.. i=-=p

j i i Cicadellidae, Homoptera

......• ~.-•• -.-.•... I ... ---

'. i. • i --1 .-... -.-.. --~ ..... -. ····· .. t

! .... ! .... __ .. _ ... -I ...

._ ......... _._.~. __ . . .. L. ..... _

Daysartertreat .. ent

Figure 2. Effects of fipronil spray treatments on phytophagous insects in treated plots. Continuous line + circles = treatment; dotted line + squares = local control plot; dotted line + crosses =

control plot; arrows = points of treatment

'.' r-;----,----,.-----,----,.--~c---_,_--.,---,

.. 30 .. _ Chalcidoidea, Hymenoptera .. ; .. ,L.A",. .. ...... +.... ;-1

12,5 .--.~ -_ ....... + "'/ ................ +""-...... ,,, .... , ................ ! .... 1

~ ",;'::':;:_ .. ;'7 ~ "" 1 ::Tl·j:·,·-! :'~~"'="'~'(~" t::==;::;:;::{i;:::;:::=::J=~~l i ",.~;N·rT ............. , ; ................ ,; ................ , ................ + .............. .;. ... 1

~ . " . Formicidae, Hymenoptera -_·---·'1·········· __ ··-;-·

....... j ..... __ ....... + ............. _+ .. ! ! !

i .... +·t· ... · ... ~-... ·~··i·~-:··:-·~·:·-~t-·· . . .

········--····f··· j---- --- --- --t ----- -+--l '+':-" + .... ·······1······· __ ······, ··············t···

.....•........ __ .j ···············f ...... __ ...... + --_ ........... \ ... .

is. y ............... j ................ + ................ + ................ ; ................... , .. · ............ · ... _ ...... · .... + .. -1

~'I .... ·!\ .......... · .... i· ...... ··CY~ .. · .... ~,,; ...... · ...... · .. +· ...... ·· .. · .. ·+ ............... ; ................ + ... , ~

j31 .... ·; .. · ... · ...... ··' .. 1 !"··+ .. l .. ·· .... ····~, ..... ·+· .. ·-:~~--~ ~ ·'I .... ·!·)"~~,I .... · .. :·4 .... · ........... t····:: ........ ·?~· .... ··+· .... =~---='Y : ~ , 1 .. 1'.,. .. ·· .. ·:··· .. ·, .. : .. ·· ...... ·, .... · ...... · .... , ...... ·· .... ·+····,

iK V\ Silphidae, CoI~ptera

";J

'" •

~ ..n. ~ • .i f"n \-, .. ..

" Daysaftertrubnent

Figure 3. Effects offipronil spray treatments on parasitoids, predaceous and necrophagous insects in treated plots. Continuous line + circles = treatment; dotted line + squares = local control plot; dotted line + crosses =

control plot; arrows = points of treatment

187

application (Fig. 3). Probably, the same reasons (increase of appropriate food resources) lead to an increase in ants and to the quick recolonization of the ground beetle population.

2.3. EFFECT OF FIPRONIL BLANKET TREATMENT ON ARTHROPODS IN RELA nON TO THEIR HABITAT PREFERENCES

A comparison of groups of arthropods occupying different niches in the ecosystem (i.e., the grass layer and the ground surface) exhibits reduced treatment effect proportionate to the increase in arthropod's ability to hide. Variations in relative abundance of such groups are illustrated for bugs (Lygaeidae and Cydnidae) and spiders (Thomisidae and Lycosidae) (Fig. 4). The chinch bugs prefer moving in open spaces; their habitat embraces both the grass and the ground surface. The burrower bugs demonstrate an obvious tendency to a cryptic mode of life, inhabiting soil and the ground surface. Fig. 4 shows that the insecticide application reduced the Lygaeidae population but did not influence the burrower bugs (Cydnidae); their number actually grew after the treatment. The collected spiders belong to two major ecological and systematical groups: crab spiders (Thomisidae) and wolf spiders (Lycosidae). Crab spiders occupy the top part ofthe herb layer (spikes, flowers etc.), while wolf spiders spend most of the time on the ground surface. Fig. 4 shows that the treatments had a greater influence on the crab spiders than the wolf spiders. The latter are more protected either by the vegetation cover or by their ability to hide in the litter. These results are confirmed when comparing the effect of fipronil on the taxa collected by sweep-net and by pitfall traps. The results showed that 70% of the 10 most abundant groups collected by sweep-net decreased after the treatments, whereas only 30% of the 10 most abundant ground-dwelling taxa were reduced.

- .. ~ j 0,8 ..

~ I: 0,4

: ..

- i ! :6f;·· +········+··c~····+·~+··········i~'··+j

o N

J ~3f~'~~~~········+,~-··+·········+·-·+--r'~r~

i

Heteroptera

i 0' f···-j/·_··_·+-\_···+_····j··_····_·+····- .+_ ....... + ....... +1

l

3.' ,.,-_--,-_-,-_--,--_.,--_,--_,----,-,

i t IlYCOSid~e, Aralei i

r~ 0.0 '-"''---'-_-'-_...l.-_-'::---,':----':----:'c....J

Days after tre.tment

Figure 4. Effects of fipronil spray treatments on related arthropods with different habitat preferences in treated plots. Continuous line + circles = treatment; dotted line + squares = local control plot; dotted line +

crosses = control plot; arrows = points of treatment

188

2.4. COMPARATIVE EFFECT OF FIPRONIL AND CHLORPYRIFOS BLANKET TREATMENTS ON NON-TARGET ARTHROPODS

The relative effects offipronil and chlorpyrifos on the non-target arthropods are shown in Table 3. The most abundant non-target taxa can be divided into four categories according to the relative effect of fipronil and/or chlorpyrifos on the relative abundance: (1) groups in which the effect offipronil is greater than that of chlorpyrifos; (2) groups affected by both insecticides nearly equally; (3) groups in which the effect offipronil is less than that of chlorpyrifos; (4) groups in which the effect of both insecticides is obscure. On the basis of the number of taxa found in the Category 1 we may conclude that in the ecosystem studied fipronil had a more severe effect on the arthropod community than did chlorpyrifos. The cases where the effect offipronil was less than chlorpyrifos (Category 3) warrants further discussion (Fig. 5). The representatives of the families belonging to this category (Lycosidae, Cicadellidae, Cydnidae) are less likely to come into direct contact with the insecticides due to the peculiarities of their mode of life. The response of the populations ofLycosidae and Cicadellidae to the chlorpyrifos treatment is quick population decline (with minimal numbers on the 3rd day after treatment) and a short (with increasing numbers on the 5th day) period before the population begins to recover. We suppose this feature of population changes could be explained by the effect of chlorpyrifos vapours, since it belongs to the group of organophosphate insecticides with highly volatile active ingredients (volatility of chlorpyrifos is 104 times higher than of fipronil [12]).

TABLE 3. Comparative effect offipronil and chlorpyrifos on non-target arthropods on treated plots (dominant and subdominant families captured by different methods)

Type of effects I. The effect of fipronil is greater than that of chlorpyrifos

2. The effect is nearly equal 3. The effect of fipronil is less than that of chlorpyrifos 4. The effect is obscure

Pilfalliraps Acrididae, Lygaeidae

Carabidae, Cicadellidae Lycosidae, Cydnidae

Silphidae, Dermestidae, Formicidae, Reduviidae

Sweep-nets Acrididae, Lygaeidae, cf Muscidae, Jchneumonoidea, Miridae, Chrysomelidae, Agromyzidae, Pyralidae Thomisidae, Chalcidoidea Cicadellidae

Formicidae, Aphididae, Psyllidae, Curculionidae

2.5. EFFECT OF FIPRONIL BARRIER TREATMENT ON NON-TARGET ARTHROPODS

One of the effective ways to diminish adverse effects of grasshopper control treatments on the environment is the utilization of the barrier treatment technique. The data from this study show that the barrier treatment significantly reduced the negative effect of fipronil on arthropods. For instance, the barrier treatments lead to the quick recovery of many herbivores, predators and parasitoids: Miridae, Cicadellidae, Thomisidae, cf Muscidae and

Flpronll 1D~~ __ ~ ____ ~ ____ ~ __ ~ ____ ~ __ -,,-__ -,,

t, i I 8 .... .j..... . ...• + ............ 1.. .... -.... + Cicadellidae, Homoptera .[---

~ 6 .... 1.~ .. : ...... 1. .......... .1. .. j E • ~ . : 2 :E

L'F"';! ·';;r::+::::· .1 .. ___ .......................... .:. ........ .

Fipronil

Days after treatment

189

Chlorpyrlfos

.......... +

i 2S

1 20 I···'!·· 1···\········,···· : fa 15

............ , , i : ........ ; .............. +. Clcadellidae, Homoptera

,

·····--1---i

·· __ ······_-_·t· i . ... + .... __ ...... + .. , Jl E 10 . ·······t· ············-f .......... _-_ . .j.

L 0···············,··,.,.,,··0··· .J.., ·············)····1··· :E~~~~~

] ...... + ............ + ..

.j ...

Chlorpyrifos

, Aranei

........

i\~ .r ~ I '-

A ~ ~ J I' i/ \r-,.---->' ~ / ~ .. , 'J.

Days after treatment

Figure 5. Comparative effect of fipronil and chlorpyrifos spray treatments on non-target arthropods with "cryptic" mode of life. Continuous line + circles = treatment; dotted line + squares = local control plot; dotted

line + crosses = control plot; arrows = points of treatment

Natural pasture

! lJ. i, ... /" !;; It 16 •. -........ . Chrysomelidae, Coleoptera

i iI i"" i,,' ~ "" ... , ....... "/ ..... ,,, ................ 1. ........ '" ~r .... ": ::I::: ~ . , ... , .. -j .. l ....... ' ................. i .......... · .... , ...... ,,'\[···· .. r···· .. ·

!. 7 ~ .+. !.-+L .... ij ~ , .. , r+" j \r.' .. ·:.·t .. ~+.

t = 12

!il

l i!

!

Natural pasture

j' ... Lygaeidae, Heteroptera i ··+.i.. ' i

·1· .. ·· .. + ........... + ... :.: :.(., .. : ... :± ....... --.1.---._._--j .....

Field with Bromus sp. 3.' r-lr----,---,----,----.,-----,-----,----,--,

t 2.sl .. ii .. rtl;1..,-... i ..... ····)·· .. ·· .. · ...... ·j C:hI1,.ornelidae.

! I """-" i 2,0 1··..)··1 .. · ...... · .. ·,'· .... ".::n ... ,-............. i ............ ~ ............... + ............. -+ ....... -... + .. J

~ " ~+II ........... ii ........... +~ ..... ~.+ ......... + ...... ·; .. · ............ ·i .. · .. ··· .. '1··-1

I " "·il .. · ... + ............. ', ."~"", .... ). +· ...... · ...... ·' .... 1

!::JI ~D-~ Field with Bromus sp.

;.t 5 " .... , ............... + ....... -[:11 _ ... __ Lygaeida~ Heter~pter:¥l ____ .. _. . ~' " r f .. + ..... ,/I .... ij ......... + ........... "-... +-K ..... -.. ~ .. ::L+.·~·:·_:·~¥rL/· .. :: .. ~ .. · .. ~ ... +.~ ....................... ; .. ' .. j

])! 21 1 'rf 1

Figure 6. Effect of barrier fipronil spray treatments on herbivorous insects with different types offeeding behavior in treated plots. Continuous line + circles = treatment; dotted line + squares = local control plot;

dotted line + crosses = control plot; arrows = points of treatment

190

Agromyzidae, Ichneumonoidea and Chalcidoidea. Here this tendency is illustrated for herbivorous insects with chewing and sucking mouthparts (Fig. 6). Among the herbivores with sucking mouthparts, the chinch bugs (Lygaeidae) are the most susceptible to fipronil. The recovery in the case of barrier treatments began on the 21 st day, while in the case of blanket treatments no recovery occurred during the time of observation. Conversely, we did not observe a recovery of the population of chrysomelid beetles (with chewing mouthparts) even in the barrier treatment. This can be explained by the mode of action of fipronil via ingestion, the same mode of action that is used to control the target Acrididae.

In summary, the analysis of the response ofthe non-target fauna to the application of fipronil allows us to hypothesize that: (1) the taxa living in the grass layer or regularly visiting it are most at risk; (2) within these groups the maximum hazard can be expected in herbivorous insects with chewing mouthparts and/or in insects with sucking mouthparts and high grass-dwelling activity; (3) insects with sucking mouthparts and low grassdwelling activities, or ones with predominantly grass-flying activity, are at less risk; (4) the risk to ground-dwelling arthropods is less and correlated with their ability to hide in litter, soil etc.; (5) the treatment sites attract insect species which can use dead organic material for feeding and development.

The comparative analysis offipronil and chlorpyrifos brings us to the following two conclusions. (1) Generally, fipronil is a more hazardous insecticide than chlorpyrifos. (2) At the same time fipronil acts more selectively than chlorpyrifos causing a lower mortality ofthose non-target arthropods which had reduced contact with this less volatile insecticide.

3. Solutions

Without any doubt, broad-scale and uncontrolled application of highly toxic and persistent insecticides in grasshopper control represents a serious menace to biodiversity and to the functioning of natural ecosystems. It is evident that an improvement of insecticide application technologies should be oriented to the reduction of adverse effects on the nontarget fauna. The information about the principles of the response of non-target species to the insecticide treatments provides a basis for solving this problem. Results obtained in Russia, with fipronil applications, using the barrier treatment method, confirmed a high efficacy of this technique against grasshoppers. On the other hand, this method obviously minimizes the hazard offipronil to many non-target taxa, namely to phytophagous insects with sucking mouthparts, predators and parasitoids.

Atthe same time, some taxa of phytophagous insects with mandibulate mouthparts (i.e., Chrysomelidae) remain highly susceptible to fipronil even in this case. We suggest that all arthropods with mandibulate mouthparts - leaf-eating and detritivorous arthropods, and the pollinators, which eat pollen during its collection, should be placed among the groups which are most at risk. The approach to minimize the danger to these groups might be different. For actively moving groups -detritivores (e.g., Tenebrionidae) and pollinators (Apoidea), an optimization of the barrier width and barrier spacing might reduce the risk considerably. For less mobile groups (e.g., Chrysomelidae) which commonly live only on their host-plants, this approach might not give positive results. In the case of natural

191

patchiness of the resources (= host plants) and low dispersal capacity barrier treatment might lead to the fragmentation of the population and, depending on the degree of this fragmentation, to its local extinction. The level of risk for a species to become extinct will correlate with the ratio treated area / area occupied by the population. If endemic species dominate in the treated area, the use of insecticides like fipronil appears to be unacceptable. In these situations, the application of less hazardous preparations (e.g., bioinsecticides) seems to be more appropriate. We would recommend the following strategy for fipronil applications: (l) The blanket spray applications should be limited to territories where there is relatively low biodiversity (i. e., monoculture cereal vegetation; fields with perennial grasses). (2) The barrier method of application is recommended in most of the other cases. (3) In the case of ecosystems, which host endemic species, it is more preferable to use a less hazardous insecticide than fipronil. In the region under consideration, the latter situation seems to be more hypothetical than real, since the area that is typically damaged by non-swarming grasshoppers in Russia is within the steppe life zone, which generally lacks endemic species (for Carabidae see [6]; for Tenebrionidae - [5]; for Chrysomelidae - [9]).

It can be concluded that the level of risk to the ecosystem resulting from insecticide applications is determined by the composition of the indigenous fauna. This composition we interpret not only a set of different taxonomic groups sensitive to insecticides, but also as a combination of ecological types within and between taxa and the degree of their endemism. The routine "taxonomical" approach to assess the effects of insecticides on separate orders or families [e.g., 1, 3,4] does not allow us to compare results obtained in different regions. For example, only the Carabidae family contains a large number of ecological types of taxa: specialized predators (Carabini) alongside with specialized phytophages (most ofZabrini); grass-dwelling (Odacanthini and some Lebiini) and ground-dwelling (most ofPterostichini andAgonini) species; open-living (Bembidiini) and cryptozoic (Scaritini and Broscini) taxa and all transitional forms. Hence results obtained for Carabidae without taking into account the indigenous structure of the ecological types in the treated areas may lead to an assessment of insecticide effects which cannot be compared to the results from other studies.

4. Acknowledgements

The author is indebted to J.-F. Duranton (CIRAD-PRIFAS) and A.V. Latchininsky (University of Wyoming) for their kind help and advice in the realization of the research program during field experiments. The author is also thankful to anonymous reviewers for their invaluable comments and suggestions on the paper.

5. References

I. Balan~a, G. and de Vissher, M.-N. (1995) Effets des traitements chimiques antacridiens sur des Colc!opteres terrestres au Nord du Burkina Faso, Ecologie 26, 115-126.

2. BaJan~a, G. and de Vissher, M.-N. (I 997a) Effects of very low doses offipronil on grasshoppers and nontarget insects following field trials for grasshopper control, Crop Protection 16, 553-564.

192

3. Balan~a, G. and de Vissher, M.-N. (1997b) Impacts on nontarget insects of a new insecticide compound used against the desert locust [Schistocerca gregaria (ForskM 1775)], Arch. ETWiron. Contam. T oxicol. 32, 58-62

4. Everts, J. and Lamine, B. (1997) Environmental effects of chemical locust control, in J.W. Everts, D. Mbaye, D. Barry (eds.) Environmental side-effects of locust and grasshopper control, FAD, Locustox Project, Dakar, Senegal, 1, pp.l-14

5. Knor, l.B. (1973) Ecological and faunistical review of black beetles (Coleoptera, Tenebrionidae) of steppe and forest-steppe landscapes of West Siberia, in Fauna of Siberia, Nauka Pub!', Novosibirsk, pp.93-106 (In Russian)

6. Kryzhanovskij, D.L. (1983) The beetles of the suborder Adephaga: families Rhysodidae. Trachypachidae; family Carabidae (introduction and review of the fauna of USSR), Nauka Pub!., Leningrad .. (Fauna of the USSR. Coleoptera, Part I) (In Russian)

7. Latchininsky, A.V. (1997) Grasshopper control in Siberia: strategies and perspectives, in S. Kral, R. Peveling, and D. Ba Diallo (eds.) New Strategies in Grasshopper Control, Birkhauser Verlag, Basel/Switzerland, pp.493-502.

8. Latchininsky, A.V. and Sergeev, M.G. (1999) Acridological forum of CIS countries, Plant Protection and Quarantine 7, 43-44 (In Russian)

9. Lopatin, l.K. (1977) The leaf beetles (Chrysomelidae) of Middle ASia and Kazakhstan, Nauka Pub!', Leningrad. (Identification Keys of Coleoptera. published by Zoological Institute of Academy of Sciences of USSR. 113) (In RUSSian)

10. Murphy, C.F., Jepson, PJ., and Croft, B.A. (1994) Database analysis of the toxicity of antigrasshopper pesticides to non-target, beneficial invertebrates, Crop Protection 13, 413-420

II. Nurmuratov T.N. (1999) The ways of diminishing of grasshopper abundance, Plant Protection and Quarantine 4,15-17 (In Russian)

12. Tomlison C.D.S. (ed.) (1997) The pesticide manual, II th edition, British Crop Protection Council, Farnham, Surrey, UK

SUMMARY

The Risks of Grasshoppers and Pest Management to Grassland Agroecosystems: An International Perspective on Human Well-Being and Environmental Health

I.A. LOCKWOOD andA.V. LATCHININSKY Entomology Section Department of Renewable Resources University of Wyoming Laramie, WY 82071-3354 USA

1. Introduction

The papers presented at the workshop, which form the core of this book, were essential ingredients in the process and served as the common intellectual currency for subsequent deliberations. However, these presentations were a "means to an ends", rather than being the purpose of the workshop. These provocative and informative thoughts provided the raw intellectual material for the primary focus of the workshop -- a series of directed discussions, specifically designed to identify problems and solutions pertaining to the management of grasshoppers and locusts on grassland ecosystems. The goal of this chapter is to summarize the salient points, novel perspectives, and areas of dissension that emerged in the course of the workshop's discussion sessions.

2. Methods: The Strategies Used to Find Common Ground

To generate the most productive discourse possible, four methods were employed. With each method, the communications were translated to assure full participation in the project. First, a survey was conducted prior to the workshop in order to assess the participants' views on two essential issues: the risks associated with acridid outbreaks and the management strategies used in context of these risks.

Second, we designed a set of four "problems" to structure an international dialogue along a line of inquiry intended to generate a final set of recommendations. Each ofthe problems was addressed by a small (4-6 members) team with an assigned leader who kept the discussion focused and a reporter who summarized the 2 hr sessions. All of the teams included at least three nationalities and two of the teams operated in Russian and English, with the assistance of translators, to assure an international synthesis of ideas.

193


194

The problem sets consisted of the following:

1) What are the theoretical and empirical risks of taking no action during a grasshopper outbreak?

2) What are the theoretical and empirical risks of pest management actions during a grasshopper outbreak?

3) What are viable tools/solutions and what prevents their implementation (with particular emphasis on assessing currently viable tactics and then evaluating the roles of policy, law, education, and funding as obstacles to the employment of these tactics)?

4) What are the research priorities in optimizing environmentally sound management of acridid outbreaks? Each group was given 100 units of (hypothetical) funding for each of the next 5 years, and the task of the group was to allocate these funds among the possible research avenues that would most effectively address how to manage grasshopper outbreaks without risking environmental disaster.

Third, before each of the aforementioned "problems", there were three to four, 30 min. presentations intended to stimulate thought and discussion pertaining to the issue. These presentations were the substance for the chapters that form the "core" of this book.

Fourth, we concluded with each participant submitting list ofthe most important and specific risks (or class of risks) and the most promising and constructive solutions (including the tools or methods that are part of the solution). The workshop organizers synthesized these ideas into a set of proposed recommendations. These recommendations were then discussed by all participants, and each recommendation was assigned to a writing group, charged with expressing the issue and its resolution in concise terms. These texts were presented to the entire group and revisions were made until a complete consensus was achieved, a process that required several hours.

3. Results of the Workshop: Synthesis and Consensus

3.1. THE SURVEY

The survey conducted prior to the workshop addressed two essential issues: the nature of risks associated with acridid outbreaks on grasslands and the current/future tools for managing these outbreaks while minimizing environmental harm (Tables 1 and 2). The survey generated 19 responses, with 7 from Russia/CIS, 6 from Europe/Africa, and 6 from North America. For most of the questions, the responses ranged from 0 to 10, and in every case the range spanned at least 5 units. Despite this extent of dissension, some areas of agreement emerged.

195

3.1.1. Analysis of Risks The topic addressed in this element of the survey was, "Risk is a function of both likelihood of occurrence and severity of harm. Based on your knowledge and experience, rate the following in terms of the severity of harm and likelihood of occurrence with regard to risk to the environment."

Ifwe first consider the element of "likelihood", the participants rated the most likely sources of risk as: 1) ecological ignorance, 2) harm to non-target acridid species, and 3) economic ignorance. The least likely factors were considered to be: 1) harm to the environment via inaction, 2) harm to humans via action, and 3) harm to humans via inaction. Thus, it appears that ignorance is the most likely source of risk in acridid outbreaks. Our own uncertainty about the ecology and economics of the grassland system would seem to be a serious matter for concern. Consistent with the notion that the grassland biome itself is an entity of considerable complexity that is yet to be resolved, the least likely sources of risk were associated with inaction or impacts to humans.

Turning attention to the second element of risk, the most severe damage associated with acridid outbreak management was associated with: 1) harm to non-target acridid species, 2) harm to non-target other species, and 3) harm to non-target acrididlother processes. The factors associated with the lowest severity of harm were: 1) taxonomic ignorance, 2) harm to the environment via inaction, and 3) harm to humans via inaction. As such, it is apparent that the most severe harm is likely to be manifested in damage to non-target organisms. Conversely and logically, the consequences of inaction were deemed to have the least severe harm.

Given that risk is a function of both the likelihood and severity of harm, we generated the product of these two elements in an effort to derive a synthetic measure of risk. With this approach, the greatest risk was associated with: 1 )harm to non-target acridid species, 2) harm to non-target other species, and 3) ecological ignorance. The lowest risk factors included: 1) harm to the environment via inaction, 2) taxonomic ignorance, and 3) harm to humans via action or inaction. Based on this analysis, it is evident that the workshop participants believed that the highest risk of acridid pest management was to non-target species, although it seems that we are largely ignorant of their ecological roles. The least environmental risk was associated with taking no pest management action and with the apparently nominal harm to humans when action is taken.

Several other important conclusions can be extracted from this portion of the survey. It is worth noting that the greatest dissension among the participants was associated with their assessment of the risk to the environment via inaction and the harm to target acridids. Those surveyed generally agreed on the high likelihood of our uncertainty, but not the severity of its consequences, concerning the "three great unknowns": ignorance of taxonomy, ecology and economics. Interestingly, neither the likelihood nor the severity of risk dominated the synthetic value derived for risk. As such, it appears that we should not focus on only one of these variables in understanding or managing environmental risk. Indeed, it was entirely possible to have factors with: high likelihood and high severity (harm to non-target acridids), high likelihood and low severity (ecological ignorance), low likelihood and high severity (harm to humans), and low likelihood and low severity (harm to environment via inaction). Finally, the respondents

196

identified a number of "other" sources of risk, including: agricultural (crop) damage, diversion of resources from other problems, social instability, insecticide resistance, and non-target pest outbreaks.

3.1.2. Analysis of Tools In this element of the survey, respondents were instructed that, "Tools that allow us to 'solve' (manage) acridid outbreaks without risking environmental catastrophe can take many forms. The solutions may be extant or potential. Rate the following in terms of their capacity to manage acridid outbreaks without risking environmental harm."

With regard to the current status of control methods, those approaches judged as having the greatest capacity to manage outbreaks while minimizing harm were: 1) direct survey, 2) ultra-low volume (UL V) applications of insecticides, and 3) blanket applications of neurotoxins. Conversely, the methods with the least capacity included: 1) neoclassical biological control (the use of exotic organisms to control native species), 2) classical biological control (the use of exotic organisms to control exotic species), and 3) augmentative/conservation biological control (increasing the abundance of a native organism or protecting such an organism from practices that would diminish its efficacy). In summary, it appears that the participants found the traditional methods to combat outbreaks (direct survey and chemical control) to be reasonably effective, while all forms of biological control are currently of limited effectiveness.

In terms of the future capacity of control methods to manage acridid outbreaks while minimizing environmental harm, the approaches viewed as having the greatest potential were: 1) direct survey, 2) integrated pest management (lPM), and 3) preventative tactics. Interestingly, three forms of biological control were rated as having the lowest potential: 1) neoclassical biological control, 2) classical biological control, and 3) augmentative biological control. Thus, traditional uses of insecticides were not perceived as having a significant role in the future of acridid pest management. While chemicals may have a place, acceptable control strategies will require developing methods to use them effectively, rather than simply finding new products or using more of existing products. Despite two decades of optimism for biological control, it appears that the "standard" forms ofthis technology which involve adding organisms to the ecosystem (versus conserving those that naturally occur) are not likely to solve acridid management problems. Rather there appears to be a general recognition that acridid outbreaks on grasslands are ecologically and economically complex events, and the most appropriate solutions will reflect this complexity. Simple, unimodal and universal approaches do not seem likely to provide the ultimate solutions for pest management of acridids.

Other methods that respondents added to the list of tools are worth mentioning, although none were rated as having great extant or potential capacity. These included: traditional methods of native people, changing cultivars, applications of insecticides via fogging devices, and use of UL V applications in reduced agent-area treatments (RAA T; the application of low rates of insecticides to discrete swaths alternating with untreated swaths to control acridids directly in treated swaths, indirectly via movement of acridids into treated swaths with intoxication by feeding on treated vegetation, and indirectly via conservation of native, naturally occurring predators and parasites in the untreated swaths).

197

In an effort to reveal the management strategies that ought to attract the greatest attention of researchers and managers, we improvised an index of future importance:

Future value = (potential capacity - extant capacity) x potential capacity

This index included both the element of the gap between extant and potential value and the absolute measure of potential value. The former element was intended to reflect the notion that a method warranting particular attention should have a large (positive) difference between potential and extant capacity (i.e., a large capacity for "growth and development"). The latter element was intended to reflect the sense that a method should have not only a large capacity for growth but a high absolute potential (i. e., a large, maximum growth potential).

Using the index of future value, the methods deserving the greatest investment in the foreseeable future were: 1) remote survey, 2) IPM, 3) decision support, 4) prevention, and 5) conservation biocontrol. Those methods requiring the least attention in terms of research and development were: 1) (neo )classical biological control, 2) UL V applications of insecticides, 3) solid bait formulations of insecticides, and 4) blanket applications of neurotoxins and insect growth regulators. Thus, it is apparent that there is imminent need for developing tactics that integrate information (e.g., remote sensing, IPM, and decision support). These approaches are information based and would generally rely on knowledge to replace costly, external inputs. Biological control is clearly not an approach that can be dismissed. While the relatively intrusive, high-input methods in which organisms are added to the environment were not viewed as being viable, conservation biological control did have considerable future value. This finding is consistent with the notion that subtle, complex, ecologically integrated, knowledge-based strategies largely define the future of acridid pest management. Continued work on forceful or extensive restructuring of grassland ecosystems by adding exotic organisms or chemicals in large-scale interventions was rated ofvery low importance. It appears that the traditional, simplistic methods have largely "run out", having failed to solve complex ecological problems. In this context, it is worth noting that the only chemical application strategy that was rated as having above average future value was the reduced agent-area treatments (RAA T), which relies on understanding the ecology and behavior of acridids and conserving their natural enemies.

3.2. RISK ASSESSMENT OF ACRIDID PEST OUTBREAKS AND MANAGEMENT ON GRASSLANDS

The first task of the workshop was to discover and explore the nature of the risks associated with acridid outbreaks and the pest management practices used to suppress them. This assessment was divided into two phases: defining the risks of taking no action during an outbreak and then defining the risks of pest management interventions during an outbreak.

198

3.2.1. The theoretical and empirical risks of taking no action during a grasshopper outbreak The central dilemma of pest management is that we cannot avoid risks by failing to take action. Allowing a pest outbreak to proceed without intervention constitutes a decision on our part - one with potential harm to humans and the environment. For the purposes of this analysis, an outbreak was defined as the condition in which acridids destroyed sufficient forage to cause human suffering, economic loss, or environmental damage. In this context there were three types of risk associated with a failure to take action during a grasshopper or locust outbreak on grasslands.

First, ECONOMIC HARM may result from inaction. In the grassland ecosystem, losses to the livestock industry or subsistence graziers and other pastoral people may be severe. The extent and consequent harm of economic loss varies widely with cultural, economic, and biological geography, but all grasslands are susceptible to significant losses under some conditions. The severity of loss is also heterogeneous within a region, as acridid outbreaks are not uniform, and site-specific economic and management conditions also create a diversity ofimpacts. A particularly severe and perhaps increasingly common, indirect consequence of an untreated acridid outbreak is the loss of land ownership by the agriculturalist. In addition to the economic consequences of displacing and employing these landless people, the subsequent owner of the land may use the resource in a manner that generates poorer economic returns. Finally, the possibility that the acridids infesting grasslands will consume this resource and then move into cropping systems cannot be overlooked. Indeed, the economic risk of inaction is arguably a function of the value of the affected resource, so migration of acridids from grasslands to croplands deserves serious consideration in a pest management system operating in a heterogeneous landscape of agroecosystems.

Second, there are several ENVIRONMENTAL CONSEQUENCES associated with doing nothing and allowing an acridid outbreak to proceed. The greatest risks are to overgrazed, damaged, or otherwise stressed grasslands. In these areas, acridid feeding can lead to soil erosion and contribute to desertification. Severe damage to the native plant community by grasshoppers and locusts can create opportunities for invasion by weeds, including many aggressive, exotic plants. Of course, this deterioration of biological diversity is not exclusively a matter of concern from an environmental perspective, but expansion of noxious weed populations can have significant economic consequences. Finally, there is a risk that parallels the economic concern that arises with the transfer of land ownership that may follow an untreated acridid outbreak. The new use of the land (e.g., tilling the land for crop production, developing the land for human housing, etc.) is often much more harmful to the environment than responsible livestock production.

Third, there are SOCIO-POLITICAL RISKS that arise from failing to intervene in an acridid outbreak. If citizens suffer in terms of hunger, poverty, or displacement as a consequence of an untreated outbreak, the government and society will need to respond to these conditions. In cases of severe loss to the grassland biome (and movement of acridids to crops), the affected country may have broad-scale socio-economic consequences with respect to the balance of trade, food security, etc. The possibilities of famine and destabilization of a federal government must be considered in some national settings.

199

In summary, it is essential to understand that livestock production on grasslands can be a sustainable, low-input, economically viable form offood and fibre production that is compatible with the conservation of wildlife and biodiversity over immense areas ofland that are not suitable for intensive, crop-based agriculture. As such, risking this form of resource management by allowing acridid outbreaks to progress unabated may not represent a sound or defensible course of (in)action.

It is also important to realize that the risks of inaction can be mitigated using approaches other than pest control. The harm of inaction is reduced as the extent of resource (forage) extraction is lessened, which in many cases is consistent with responsible stewardship and sustainable levels of livestock grazing. In a holistic sense, the risk of inaction may also be relatively lessened if the resources that might have been used to suppress the acridid outbreak are used to generate greater benefits by application to more serious problems (e.g., using funds to build roads or other elements of an infrastructure may have greater benefits to the region than an acridid control programme). Also, with effective planning of the grassland-cropland landscape, understanding of the ecological roles of acridids and other species, and monitoring of the resource status and pest populations on grasslands, it is possible to assess and minimize the economic, environmental, and sociopolitical effects of a grasshopper or locust outbreak and thereby responsibly include non-intervention as a management strategy.

Finally, we must recognize that the risks of inaction are scale-dependent. As the spatiotemporal scale increases, the severity of harm generally diminishes. This scaling phenomenon can be used to conduct an analysis that supports a preconceived management decision (e.g., focusing on large-scale, long-term effects to marginalize local impacts and thereby justify inaction), so it is important to assess the effects of an outbreak across a range of scales to assure a complete understanding of the consequences of failing to treat an outbreak.

3.2.2. The theoretical and empirical risks o/taking pest management actions during a grasshopper outbreak The most familiar formulation of risk analysis in terms of pest management is to assess the likelihood and severity of harm arising from taking action. There is no doubt that altering the course of acridid outbreaks has incumbent risks, and these deserve our full consideration. Perhaps the most effective means of classifying the risks of intervening in an acridid outbreak is to cast these in terms of the non-target entities that are harmed: organisms, ecosystems, and humans.

The risk to non-target organisms is most easily and effectively expressed in terms of the species and roles of these entities. The impacts ofacridid control can be a manifest via:

1) elimination of a species (local or regional extirpation or possibly true extinction of highly endemic species),

200

2) changes in population dynamics (i.e., suppression of organisms susceptible to the intervention [e.g., pollinators] or enhancement of organisms released from control by natural enemies suppressed by the intervention [e.g., secondary outbreaks of acridids)),

3) changes in the ecological functions performed by the non-target species (e.g., reductions in the reproductive rate of insect-pollinated plants, or diminishment of nutrient cycling processes), and

4) loss of biodiversity (including genetic variation which may be unique to localized populations).

In general, the greatest impacts to non-target organisms are associated with chemical insecticides, although the severity of harm is highly variable and depends on the nature of the product and the mode of application. The organisms at greatest risk would include non-target acridids, insectivorous birds, arthropod natural enemies, pollinators and other organisms with high rates of exposure and limited potential for metabolic detoxification. Biological control would risk both harm to native, natural enemies via competition and harm to non-target species within the host range of the agent. Finally, even cultural controls may adversely effect non-target species and processes.

From a more holistic perspective, acridid control programmes on grasslands represent risks at the ecosystem level. Chemical insecticides may directly or indirectly degrade soil, air, and water quality and may bioaccumulate in the food web. Introduced biological control agents may undermine the integrity of co-evolved relationships within the grassland communities and thereby adversely effect ecosystem functioning at many levels. Even though the insecticide loads on grasslands are markedly less than those on croplands, some natural habitats are extremely vulnerable (e.g., locust outbreaks may originate in wetlands). Furthermore, the spatial scale of treatments on grasslands can be enormous, making recovery of ecosystem functioning via recolonization very slow.

With respect to non-target effects on the human population, chemical and, to a lesser extent, biological insecticides represent health risks. Unfortunately, the long-term effects of exposure and the impacts of exposure to mUltiple agents are poorly understood. Direct intoxication during treatment programmes represents the most acute risk, but bioaccumulation in human diets is also a matter of serious concern. In a social context, the effects of using toxic chemicals and introducing exotic organisms represent significant political risks. Finally, the costs of misguided control programmes can represent enormous economic losses, particularly with respect to chemical and biological interventions, although even cultural control practices may have significant costs.

The risks to non-target entities and processes can be markedly enhanced or reduced by the nature ofthe pest management strategy. For example, environmental and human risks are significantly aggravated by a number of factors, including needless treatments due to lack of a surveyor inappropriate economic analysis, excessively large treatments due to a failure to have intervened early in the outbreak process, overapplication and misuse of materials due to inadequate training, use of outdated materials,

201

reliance on inappropriate but overstocked agents, reliance on incorrect efficacy data and invalid environmental assessments, and increased application rates due to real or perceived resistance in the target acridid populations. Essentially all of these aggravating factors can be eliminated with knowledge in the form of training or surveys. Risks can be further mitigated by understanding the ecological functions of grassland systems and their constituent communities and species. Finally, the risks of our interventions in natural processes can be minimized with an approach that integrates pest management, production practices, and conservation strategies at a landscape scale. This strategy would avoid the unilateral, myopic, and disjointed programmes that too often characterize acridid control programmes.

3.3. ANALYSIS OF MANAGEMENT TOOLS AND THE OBSTACLES TO THEIR EFFECTIVE IMPLEMENTATION

3.3.1. Solutions and the obstacle o/pragmatics Virtually all plausible solutions to acridid pest management are limited by five factors. First, communication has become limiting, as the network of applied acridologists has dwindled. The need for effective communication across language barriers is extremely keen, as many outstanding works and viable methods are effectively lost across national or linguistic borders. Second, the paucity of training has become an extremely serious issue as the complexity of emerging management tools and strategies has increased. Technology transfer has not kept pace with technological sophistication and innovation. Third, international and national pest management infrastructures are being rapidly dismantled. The decline of research laboratories and control networks in applied acridology has dramatically reduced our capacity to develop and implement safer and cheaper pest management practices. Fourth, the costs of pest management, including labor, materials, technology, and information continue to increase at a rate far greater than that of agricultural commodities. Hence, it has become critical to bring the cost of management into line with the value of agricultural production. Fifth, there has been little concentrated effort on developing truly integrated forms of acridid pest management. Although some strategies make use of multiple systems (e.g., RAAT integrates conservation biological control and insecticidal control), no approaches incorporate the full context of grassland health, including the processes of nutrient cycling, herbivore ecology, landscape patterns, and socioeconomic dynamics.

The current "fragmentary" solutions to acridid pest outbreaks fall into three general classes: decision support, direct interventions, and indirect interventions. Each of these types of approaches has unique advantages, and each is limited by a set of economic, educational, informational, political, and legal obstacles.

DECISION SUPPORT. The single most important element of addressing an imminent or ongoing acridid outbreak is the matter of whether an intervention is environmentally and economically justified. In effect, the "decision to treat" is the most fundamental and least understood component of pest management. Although progress has been made with some computerized decision-support systems, for the majority of grassland ecosystems in the world, there is no valid, rigorous, objective economic threshold established, and the

202

integration of environmental concerns with extant thresholds and systems is primitive, at best. The greatest obstacle to developing useful tools for enabling pest managers to make sound decisions with respect to acridid outbreaks is knowledge. We simply lack the information, and in some cases the analytical methods, to provide this fundamental component of a responsible pest management system. Even where such knowledge is extant, the decision support tools still remain largely confined to the computer laboratories of scientists and a few technologically sophisticated practitioners.

DIRECT INTERVENTIONS. Assuming that a form of direct intervention or treatment is justified, there are several obstacles that hinder the optimal use of available tools. While all of the approaches could be improved with better information on timing, efficacy, environmental effects, and economic costs, each of the methods also has its own impediments.

The neurotoxic chemical insecticides are largely limited by the following five factors. First, with respect to both product and application costs, insecticides may be prohibitively expensive to use in context of the low-input economics of grassland agroecosystems. Second, application technologies for insecticides may be unavailable or inadequate in regions where these agents are needed. Third, insecticides are often harmful to environmental health, and all of these products involve some damage to non-target organisms. Fourth, these products often have negative effects on human health; many are neurotoxins and some have carcinogenic properties. Fifth, rates of degradation may be too slow to be environmentally sound and/or too rapid to be useful in barrier and RAA T programmes.

The insect growth regulators face five obstacles. First, the rate at which they induce mortality is slower than the neurotoxins, and this delay may not be acceptable in some control programmes. Second, the conditions under which they are effective are limited to the nymphal stages, so these products may not be viable in later phases of a campaign. Third, these insecticides also may be prohibitively expensive for many agriculturalists. Fourth, as with other insecticides, appropriate application technologies may be unavailable. Fifth, the risks to environmental and, particular, human health are reduced relative to neurotoxins, but non-target organisms may be negatively impacted (e.g., aquatic arthropods).

The bioinsecticides and augmentative biological control products are constrained by seven factors. First, the complexity of producing organisms with in vivo methods limits the availability of both the productions systems and the fmal products. Second, the nature of the production systems often makes the costs of these products prohibitively high and narrows the potential market. Third, the efficacy of some products may be too low to justify the associated costs. Fourth, the biological agents are slow, relative to chemical insecticides, in inducing mortality, and the damage done before the agent becomes lethal may be unacceptable. Fifth, the environmental conditions under which these products are efficacious may be narrow, such that the already limited market is further reduced. Sixth, there is concern about the effects of bioinsecticides and biological control on human and, particularly, environmental health; the host ranges and impacts to non-target organisms have not been fully investigated. Seventh, some of the biological products have limited

203

shelflives and rigorous storage parameters, thereby limiting their viability in some settings and programmes.

The botanical insecticides are rather severely limited in four ways. First, the cost of commercial production is often very high. Second, the breadth of activity may be limited to early developmental stages (nymphs). Third, there is a lack of formulation standards, such that the user cannot be assured of efficacy. Fourth, the production system is such that the amount and availability of material may be limited, so pest managers cannot depend on these products during a control programme.

Attractants and repellents are limited by our knowledge ofbio-active compounds and acridid behaviour. In addition to a paucity of information on potential attractants and repellents, existing agents appear to have generally poor residual activity and delivery methods are not well developed.

Application strategies such as barrier applications, RAA T programmes, and preventative or "hot spot" treatments are limited by five factors. First, preventative methods are constrained by our poor understanding of population dynamics, source areas, and other ecological factors. Second, these methods may involve high costs for survey, monitoring, or application technologies. Third, these approaches are relatively sophisticated, and their implementation demands a training or education programme. Fourth, as promising as these approaches appear to be, they have yet to be fully optimized, so further research and development are needed. Fifth, these methods do not generate the extremely high levels of mortality associated with traditional, blanket applications of insecticides, and this level of efficacy may be required to manage some severe acridid (especially locust) outbreaks.

Genetic modifications of organisms have not been introduced for acridid pest management, but it appears that such approaches are constrained by a complex off actors including social resistance, legal obstacles, environmental concerns, and extremely high research and development costs.

INDIRECT INTERVENTIONS. There are several approaches that can be used to manage acridid outbreaks without directly reducing population densities. This appears to be an exciting area for further research, as these methods are generally in their early stages of development and it would seem that other, alternative tactics have yet to be discovered.

Perhaps the most intriguing of the indirect methods is the use of an insurance system to compensate agriculturalists for losses due to acridids. Insurance programmes have been developed for other natural disasters (e.g., hail and drought), and the extension of these systems to acridid outbreaks appears plausible. The obstacles to implementing such a system are considerable, however. An insurance programme could undermine autonomy and food security at a local, regional, or national scale. In addition to the loss of self-reliance, such a system would be highly prone to abuse and fraud. Finally, the complexities of implementing an insurance programme (e.g., calculating premiums based on historical risk) would be significant.

The most promising form of indirect control of acridid outbreaks is probably cultural control, which includes practices such as twice-over grazing, reduced grazing intensities, timing of harvest, trap cropping, and land use planning. The barriers to developing and implementing such systems are economic, educational, legal, and political.

204

In economic terms, the actual rate of return from these programmes has not been clearly established, and implementing such strategies can include significant labor and opportunity costs. With regard to educational obstacles, these approaches are relatively subtle and complex in terms of affecting acridid population dynamics, so training would be essential. These methods are generally preventative, rather than curative, and even their capacity to prevent an outbreak is dependent on their being employed on a chronic basis. As such, they are "slow acting" and may be difficult to sustain. Legal and political issues emerge as the freedom of agriculturalists is restricted by setting grazing limits or agricultural "zoning" practices, and these matters may become overwhelming obstructions to change.

The indirect intervention that has the most proven record of contributing to effective acridid pest management is survey and monitoring. No responsible pest management system can function in the absence of information on the density and distribution of the target organism(s). Any hope of a preventative intervention programme succeeding is firmly connected to a sound survey system. The value of such knowledge is tremendous in many contexts, but it appears that the obstacles to acquiring this fundamental information have become significant. The most severe impediment is political. As the time from the last outbreak increases, the political interest in maintaining a survey programme wanes, with the result that we often enter into an upsurge with little or no warning. The dismantling of monitoring systems just prior to outbreaks is a pattern that is repeated in time and space, being seen over and over again within nations and currently being witnessed on three continents. Acridid surveys are not simple matters, and other difficulties must be recognized. Such programmes can be fairly costly, as the scale of monitoring becomes immense. Some of these logistical difficulties might be overcome with remote sensing technologies, but these tactics are currently limited by training, experience, and costs.

3.3.2. Solutions and the obstacle of ignorance The allocation ofa fixed pool of research funding provided an effective means offorcing the participants to prioritize the needs for work in areas that addressed methods for managing acridid outbreaks without risking environmental disaster. Based on the values generated by three of the four groups, it was apparent that some fields of inquiry were extremely important while others were essentially dismissed (Table 3). Of course, the parameters of the exercise (i.e., acridid pest management on grasslands in the next 5 years) may have generated a different set of priorities than if the participants had evaluated the needs pertaining to locust management in tropical cropping systems, for example.

The most important topic for research was clearly identified as acridid ecology and biology, with 44% of the funding being directed into this area. Fully one-quarter of the funding was earmarked for studies of acridid ecology in recognition of our need to better understand the roles of these insects in the grassland ecosystem. It should be noted that such studies would necessarily include a great deal of work on topics peripherally related to acridids per se, as the participants were clear that an understanding of herbivore ecology (e.g., the roles of native ungulates and livestock on grassland production, nutrient cycling, soil quality, etc.) and belowground communities (e.g., nematodes, arthropods, and other organisms in the rhizosphere) would be essential to putting acridids in an ecological

205

context. The next most important research topic in this category was the study of environmental impacts of our pest management practices. Other areas of interest included acridids as livestock fodder and human food (a somewhat surprising allocation based on the almost total lack of funding for this topic at the present time) and the effects of global climate change on acridids. Basic studies of physiology, taxonomy, and systematics were given low priority.

Research on management and control was the next most important field of investigation with 32% of the total funding. The development of novel approaches (e.g., RAA T, barriers, and "hot spot" or incipient infestation treatments for preventative management) were provided with the most support within this category. Cultural control methods also fared quite well in terms of recommended funding. The balance of the allocation was distributed across a range of other tactics. One might have expected conservation biological control and IPM to have garnered significantly more support, but perhaps these approaches were partially subsumed within other lines of funding.

The importance of social and economic research was evident, with 13% of the funding being directed to such work. The importance of finding effective means of training and technology transfer led the funding priorities in this category. Rather surprisingly, economic analyses were poorly funded. Perhaps if the exercise had focused on locusts the allocations would have changed, or perhaps this area of research was conceptually integrated into other lines of funding (e.g., IPM).

Monitoring and survey (II % of the funding) were clearly important, and their priority may have increased if the exercise had included the allocation of management funds, rather than research funds. Even so, the study of remote sensing and geographic information systems received the fourth greatest funding of any particular research field and more support than any of the studies within sociology and economics.

The allocation summary included three of the four participant groups. The fourth group took a radically different approach, and so their results were not included in the averaging process. This group allocated the entirety of its funding to a single initiation -the development of an international center for acridid pest management. This proposal reflects the critical need for a cohesive, collaborative, and cooperative group of acridologists and supporting scientists. Such institutes have arisen and collapsed at the national level (e.g., the Anti-Locust Research Centre in the UK), but no such center has taken form internationally. Perhaps the nearest approximation to date is the Association for Applied Acridology International (AAAI).

The AAAI is composed of25 Associates representing some of the finest applied acridologists and supporting scientists from around the world and 12 of the best organizations dedicated to the study and management of acridids. The Association is funded through a coalition of private industries. In principle, international markets for pest management products provide the incentive for industrial and international support of AAAI, and the world-wide nature of outbreaks provides continual opportunities for consulting, training, and research. Because the quality of international experience and knowledge in acridology is extremely high but highly dispersed, AAAI aims to consolidate the existing resources in acridology. The concept is one of bringing together the world's best minds in the field of applied acridology to collaborate in an intellectually rigorous,

206

professionally unbiased context to sustain natural and agricultural systems during grasshopper and locust outbreaks. The Assocation appears to have good promise, but time will tell if this innovation will provide the necessary elements of critical mass, infrastructure, and strategy.

3.4. CONSENSUS RECOMMENDA nONS

A true consensus emerged from the arduous but ultimately productive effort to reach consensus on a set of recommendations pertaining to the creation of an acridid pest management system that would simultaneously recognize the socioeconomic needs of humans who use the world's grasslands and the environmental necessity of sustaining this complex agroecosystem. What emerged was a three-tiered approach to balancing the anthropogenic and acridogenic risks to the grassland biome (Fig. 1). The recommendations are cast in terms of risk, where risk is defined as the likelihood and severity of harm, and harm is an undesirable outcome of an action or inaction. So, for example, ecological ignorance is not a harm per se, but harm could arise from our ignorance. The challenge then was to phrase the analysis in terms of problem-solving - to identify a risk and then recommend an approach to minimizing the associated harm.

Tier I: Grassland Health. The comprehensive formulation of risk associated with acridid pest management was expressed in terms of grassland health. Grasslands are renewable resources, and the management of these biomes must maintain ecosystem integrity (including sustainability, resilience and biodiversity) within limits imposed by local environmental conditions. To do otherwise is to risk losing these immensely important landscapes and their environmental and economic benefits. In light of this fundamental position, it is apparent that we need more and better knowledge of grassland ecology, including the role of grasshoppers, to support sound ecosystem management. Only multidisciplinary research will provide the necessary insights to implement sound ecosystem management practices. In terms of acridids, such management requires the integration of all tactics within the bounds of ecological processes so that chemical suppression is the tactic of last resort.

Tier II: Human and Environmental Values. The anthropocentric values of economic wealth, social well-being, and political stability were perceived to be on level with environmental values. The notion here is that these three sources of value (economic, socio-political, and environmental) function as the legs of a potentially stable stool; these elements are the fundamental sources of inputs of our primary concern - grassland health. Ignoring anyone of these three inputs would seriously impinge on the sustain ability of grassland ecosystems.

The risk of economic loss and the strategies to minimize such harm are most effectively captured in a flowchart (Fig. 2). In this formulation of risk, it is evident that losses can accrue both by taking unjustified action and by failing to take defensible action. As such, intervention is not presumed always to lead to lower economic losses. The model also emphasized the necessity of continually assessing the outcomes of our actions to

207

update the decision-making process with empirical feedback. Effective assessments will require that appropriate tools be developed in this context and that personnel be trained in their use. Finally, this approach explicitly includes not only the options of "take existing action" and "avoid action" but "change action". The principle here is that there are alternatives and the range of these options is expanding; we are not forced into a simplistic dichotomy, but we have the capacity to modify and revise the nature of our management programmes.

Significant socio-political risks may result from grasshopper outbreak mismanagement, which could take the form of unjustified or inappropriate action or inaction. The harms may be primarily manifest in three ways. First, mismanagement may jeopardize local and national food security which can lead to serious human conflicts. Second, failed management programmes can result in human displacement, leading to reduced capacity for food production and increased pressure on urban services. Third, failure to properly manage an acridid outbreak results in the loss of good stewardship of the land, further aggravating the potential for both economic and environmental losses. To minimize these socio-political risks, we must develop research and management programmes along three lines. First, our efforts need to take into account the dynamic changes in agricultural policy priorities. It makes no sense to develop strategies that are inimical to the prevailing and foreseeable social and political goals pertaining to the grassland biome. Second, it is crucial that we establish active communications with sociopolitical decision-makers. Only through a system of structured feedback and ongoing dialogue with those who shape public policy can scientists and managers develop and implement viable acridid pest management systems. Third, we must learn lessons from the mistakes and successes of others. There are sufficient similarities in grassland ecosystems throughout the world to justify an active network of communication among scientists and managers to avoid repeating errors already committed and to adapt successful approaches to local and regional conditions.

The risks to environmental quality were summarized using a matrix ofthe primary ecological impacts and the relative harm expected by employing various types of intervention (Table 4). In general, the greatest risks are associated with standard, blanket applications of neurotoxic chemicals. The risks can be reduced substantially by employing cultural controls, alternative agents (insect growth regulators or bioinsecticides), or low dose methods (RAAT and barrier applications). Interestingly, these three approaches all resulted in approximately the same level of risk to the grassland environment. Taking no action also reduces environmental risk, relative to the traditional control tactics. However, inaction did not generally reduce environmental risk relative to the alternative pest management approaches. The two critical factors in assessing the consequences of failing to control an acridid outbreak were the health of the grassland (uncontrolled infestations could severely damage lands in poor condition) and the condition of the human population (locales or regions dependent on the food extracted from and near grasslands could be severely harmed by uncontrolled infestations).

Tier III: Imminent Risks. With respect to minimizing the economic, sociopolitical, and environmental risks of mismanaging acridid outbreaks on grasslands, there

208

emerged three contemporary issues. These issues are relatively specific and serve as vivid examples of how current conditions are increasing the risks of acridid pest management.

The risk of failing to prevent an outbreak (i. e., the increase of an acridid population to damaging levels) can be manifest in two ways. We might fail to prevent an outbreak because we chose not to intervene under conditions that would have allowed us to stop an outbreak (failure by having avoided appropriate action), or we might fail to prevent an outbreak because our attempted intervention was ineffective or misguided (failure by having taken inappropriate action). To minimize these harms, we must maintain a continuous, standardized survey. Such a fundamental programme should include monitoring of acridids, weather and vegetation at the landscape level in order to understand the ecosystem and acridid population dynamics. Efficient and effective surveys will employ remote sensing and geographic information systems. Incorporating these technologies and other methods to optimize survey systems requires the establishment of effective communications among scientists and managers at scales ranging from the local to the international.

The risk of losing research and management capacities through the continual erosion of funding is becoming increasingly serious. At national, regional, and local scales, the cyclic or erratic character of acridid outbreaks leads to discontinuities in funding of research and management activities. This boom-and-bust pattern in which funding tracks acridid outbreaks with a time-lag of months or years means that research and management efforts are poorly timed with the pest population dynamics. Moreover, it is difficult to confidently assess the role that most acridid management programmes play in the decline of an outbreak, particularly when the control efforts coincide with, and are thereby confounded by, a natural decline in the population. For example, the contribution of the pest management interventions in the collapse of the desert locust plague in 1986-1988 has yet to be resolved. Further complicating matters is the paradoxical situation in which a truly successful acridid management programme that employs preventative control tactics may sow the seeds of its own demise, as a lack of severe outbreaks can lead to a socio-political complacency and an eventual shortage of funding.

Fortunately, there are viable tactics to address these risks to the integrity and progress of management systems. There can be no doubt that we need to better justify the necessity for continuous programmes to government agencies and fmancial institutions and to better communicate what we do know about acridid problems (e.g. surveys have established that strategic, spatiotemporal management interventions can significantly reduce forage losses). We must also generate relevant research and effective management programmes to ensure the continuity and sustainability of funding, recognizing that these efforts need to track dynamic social, economic, and political goals. Finally, the participants of the workshop recommended that we should take advantage ofthe current, critical situation in the CIS countries of Central Asia to demonstrate our capacity to take relevant and effective actions that provide long-term solutions.

The risk of losing the human resources (i.e., the experience, knowledge, and wisdom of innovative managers, creative scientists, and effective infrastructures) that are fundamental to sustaining grassland health must be recognized. Not only are funds for management and research in applied acridology declining, but the human capital necessary

209

to develop new methods and respond to emerging pest outbreaks is rapidly eroding. The complexity and scale of the grassland ecosystems (which may be nearly pristine preserves, economically productive tracts, or seriously degraded habitats) and the stakeholders in these lands (commercial producers, subsistence agriculturalists, environmentalists, private industries, government agencies, consumers, etc.) mean that the risks of pest management are enormously diverse. To stem the loss of human resources needed to address these risks, scientists and managers must communicate the nature of these risks to grassland ecosystems with a unified voice. While we may not agree on the value of particular elements of pest management programmes, there is a consensus that we have not achieved the optimal balance between human and environmental values in our approach to grassland stewardship. While we continue to press for sustaining or enhancing investment in the human capital needed to responsibly manage grassland ecosystems, we can also stem the loss of resources by establishing avenues of international communication and programmes for education and training of current and future managers and scientists in this field. We must carefully, consciously, constantly, and creatively transfer the experience, knowledge, and wisdom across borders of space and time. And in the end, facilitating this transfer was perhaps the single most important accomplishment of the NATO workshop.

210

Table 1. Results of the survey of NATO workshop participants in response to these instructions. "Risk is a function of both likelihood of occurrence and severity of harm. Based on your knowledge and experience, rate the following in terms of the severity of harm and likelihood of occurrence with regard to risk to the environment. A score of 10 indicates cause for extreme concern and a score of 0 indicates cause for no concern."

Risk Factor Likelihood of Severity Risk Occurence of Harm

Taxonomic ignorance (not knowing what species are present 6.1 ±0.39 4.0±0.67 24.4 in the environment)

Ecological ignorance (not knowing what role a species plays 7.5 ±0.29 5.5 ±0.54 41.2 in the environment)

Economic ignorance (not knowing the types or quantities of 6.8±0.42 5.6 ±0.55 38.1 plants that are damaged)

Harm to non-target, non-acridid species (risk suppressing or 6.3 ±0.43 6.6±0.34 41.6 eliminating species by management action)

Harm to non-target, non-acridid processes (risk of 5.5 ±0.45 6.4 ±0.39 35.2 suppressing or eliminating the functions of species by management action)

Harm to non-target, acridid species (risk suppressing or 6.9±0.42 6.7 ±0.37 46.2 eliminating species by management action)

Harm to non-target, acridid processes (risk of suppressing or 6.0±O.42 6.4 ±0.34 38.4 eliminating the functions of species by management action)

Harm to target acridid species (risk of suppressing or 6.0±O.68 5.6±O.62 33.6 eliminating "pest" acridids by management action)

Harm to target acridid processes (risk of suppressing or 6.5 ±O.80 5.2 ±O.56 33.8 eliminating the functions of "pest" acridids by management action)

Harm to humans (risk damaging our health, social structure, 4.6 ±0.54 5.6 ±O.46 25.8 political stability, or economic vitality by management action)

InactionlEnvironment (risk to the environment of an 3.8 ±O.74 4.0±O.87 15.2 outbreak for which we take no management action)

InactionlHumans (risk to health, social structure, economic 5.2 ±O.50 5.0±O.60 26.0 vitality of an outbreak for which we take no management action)

211

Table 2. Results of the survey of NATO workshop participants in response to these instructions. "Tools that allow us to 'solve' (manage) acridid outbreaks without risking environmental catastrophe can take many forms. The solutions may be extant or potential. Rate the following in terms of their capacity to manage acridid outbreaks without risking environmental harm. A score of I 0 indicates an extremely sound approach and a score of 0 indicates an extremely poor approach."

Management Tool/Solution Extant Potential P-E Future: Capacity: Capacity: (P-E) x (E) (P) P

Blanket, liquid applications of neurotoxins 4.8±0.68 4.2±0.82 0.6 2.5

Barrier, RAAT, or "interval" applications of 3.9±0.80 6.2±0.65 2.3 14.3 neurotoxins

Blanket, liquid applications of growth regulators 3.3±0.65 4.0±0.30 0.7 2.8

Barrier, RAAT, or "interval" applications of IGRs 2.8±0.80 5.1±0.62 2.3 I I.7

Ultra-low volume (ULV) applications ofliquids 6.2±0.49 6.5±0.49 0.3 2.0

Solid bait formulations (e.g., wheat bran) 3.6±0.92 4.1±0.80 0.5 2.0

Conservation biological control (preserving natives) 2.4±1.l2 5.5±0.50 3.1 17.0

Preventative strategies (control high risk 4.7±0.63 7.2±0.37 2.5 18.0 infestations)

Augmentative biological control (increasing natives) 2.0±1.26 4.0±0.68 2.0 8.0

(Neo)-Classical (exotic controls native) 1.l±1.50 1.9±0.94 0.8 1.5

Classical biological control (exotic controls exotic) 1.5±1.54 1.9±1.24 0.4 0.8

Cultural control (actively managing the habitat) 2.8±0.99 5.2±0.62 2.4 12.4

Conservation/Restoration (landscape changes) 3.4±0.95 5.9±0.56 2.5 14.8

Direct survey (ground observation) 6.4±0.36 7.8±0.28 1.4 10.9

Remote survey (remote sensing) 3.3±0.68 6.9±0.39 3.6 24.8

Forecasting (models to anticipate changes) 3.7±0.46 6.4±0.35 2.7 27.3

Decision support (to make pest management 3.7±0.49 6.8±0.31 3. I 21.1 decisions)

rPM (combination of methods in a unified approach) 4.2±0.54 7.4±0.37 3.2 23.7

212

Table 3. Proposed allocation of research funds for acridid pest management on grasslands over the next 5 years by NATO workshop participants.

Research Category Allocation (% oUotal)

Ecology and Biology

Ecology (nutrient cycling, population biology, trophic interactions, feeding, etc.) 25

Environmental impacts of management (chemical, biological. and cultural) 9

Acridids as a resource (human or animal food) 6

Effects of climate change on acridid ecology and population dynamics 3

Acridid physiology <I

Acridid taxonomy and systematics <I

Management and Control

New methods and tactics (RAAT, barriers, and "hot spot" treatments) 10

Cultural controls (grazing systems, harvesting, and landscape planning) 7

Conservation biological control 4

Integrated Pest Management systems 4

Bioinsecticides 3

Refining existing classes of control agents 2

Botanical insecticides I

Developing new classes of agents and novel products <I

Classical biological control <I

Sociology and Economics

Methods to assure effective technology transfer and training 5

Studies of rural sociology and public policy 4

Economic analyses of pest outbreaks and pest management practices 2

Investigations of insurance programmes against acridid damage 2

Monitoring and Survey

Remote sensing and geographic information systems 8

Forecasting models 3

Tab

le 4

. E

nvir

onm

enta

l ri

sk m

atri

x fo

r ac

ridi

d pe

st m

anag

emen

t pra

ctic

es o

n gr

assl

ands

; a

risk

inde

x o

f 0 r

epre

sent

s lo

w p

oten

tial

for

har

m,

and

a va

lue

of

S re

pres

ents

hig

h po

tent

ial

for

harm

..

A c

En

vir

on

men

tal

risk

A

ctio

n N

o ac

tion

Sta

nd

ard

, A

lter

nat

ive

inte

rven

tion

s ch

emic

al

inse

ctic

ides

C

ult

ura

l co

ntro

l A

lter

nati

ve a

gent

s L

ow d

ose

met

hods

(b

ioin

sect

icid

es, I

GR

s)

(RA

AT

, b

arri

ers)

Bio

dive

rsit

y S

3

2 2

2

Ext

inct

ion

S

2 I

I I

Pol

luti

on

4 0

3 I

0

Out

brea

ks

I I

I I

SA

Hum

an h

ealt

h S

0

08

2 OC

Tot

al

20

6 7

7 8

--

-- The

env

iron

men

tal

harm

of

inac

tion

may

be

very

hig

h if

gra

ssla

nds

are

in a

deg

rade

d st

ate;

thi

s ri

sk i

s re

duce

d as

a f

unct

ion

of g

rass

land

hea

lth,

whi

ch is

of

ten

rela

tive

ly p

oor.

The

act

ual

risk

to h

uman

hea

lth

by a

lter

nati

ve a

gent

s is

pre

sum

ed to

be

low

at t

his

tim

e, a

ltho

ugh

the

empi

rica

l ev

iden

ce t

o su

bsta

ntia

te t

his

judg

men

t is

ge

nera

lly

lack

ing.

The

ris

k to

hum

an h

ealth

may

be

incr

ease

d su

bsta

ntia

lly

iffo

od s

hort

ages

ari

se a

t lo

cal

or r

egio

nal

scal

es d

ue t

o th

e un

trea

ted

acri

did

infe

stat

ions

.

N .... \.;.>

IV -~

Los

s o

f Gra

ssla

nd H

ealt

h

Env

iron

men

tal

Eco

nom

ic

Soci

o-po

liti

cal

Los

s L

oss

Los

s

Fai

lure

to

Pre

vent

L

oss

of R

esea

rch

and

Los

s o

f Hum

an

an O

utbr

eak

Man

agem

ent

Cap

abil

itie

s R

esou

rces

Fig

ure

1.

A c

once

ptua

l m

odel

for

the

ris

k as

soci

ated

wit

h ac

ridi

d ou

tbre

aks

on g

rass

land

s an

d ou

r de

cisi

on to

ign

ore

or r

espo

nd t

o th

ese

even

ts.

215

Acridids remove vegetation which mayor may not lead to economic loss Forage Loss Assessment (Does the damage represent an economic loss?)

Yes No

\V \V

Cost-benefit Analysis and I Avoid Action I Environmental Impact Assessment of ,\

1 existing actions (Is intervention justified?) I Assess outcome

Yes No No \ \

Take existing action I I Develop new integrated approach (Is I a novel approach possible?)

Assess outcome I Yes

I Take new action I

\V I Assess outcome I

Figure 2. Decision tree for minimizing economic risks of acridid management on grasslands.

LIST OF CONTRIBUTORS

1. A. BALDI Animal Ecology Research Group of the Hungarian Academy of Sciences, Hungarian Natural History Museum Ludovika sq. 2. Budapest H-J083 Hungary

2. G.E. BELOVSKY Ecology Center and Department of Fisheries and Wildlife Utah State University Logan, Utah 84322 USA

3. M.E. CHERNY AKHOVSKIY Department of Zoology and Ecology Moscow Pedagogical State University 6/5 Kibal'chicha St., Moscow 129278 Russia

4. O.V. DENISOVA Department of General Biology, Novosibirsk State University 2 Pirogova St., Novosibirsk 630090 Russia

5. F.A. GAPPAROV Uzbek Institute for Plant Protection (UzNIIZR) 700140 Tashkent Uzbekistan

6. A. JOERN School of Biological Sciences University of Nebraska-Lincoln Lincoln, NE 69599-0118 USA

7. V.E. KAMBULIN Kazakhstan Quarantine Inspection (formerly) Zheltocsan 16/73 P.O. Box 47 Astana 473000 Kazakhstan

217

218

8. T. KISBENEDEK Animal Ecology Research Group of the Hungarian Academy of Sciences, Hungarian Natural History Museum Ludovika sq. 2. Budapest H-I083 Hungary

9. A.V. LATCHININSKY Entomology Section, Department of Renewable Resources University of Wyoming Laramie, WY 82071 B3354 USA

10. M. LECOQ Centre de Cooperation Internationale en Recherche Agronomique pour Ie Developpement (CIRAD) PRIFAS - Acridologie operationnelle B.P. 5035 - 34032 Montpelier Cedex 1 France

11. J.A. LOCKWOOD Entomology Section Department of Renewable Resources University of Wyoming Laramie, WY 82071-3354 USA

12. O. OLFERT Agriculture and Agri-Food Canada 107 Science Place Saskatoon, SK . Canada S7N OX2

13. lA. ONSAGER USDA / ARS / Northern Plains Agricultural Research Laboratory (formerly) 1500 North Central Avenue Sidney, MT 59270 USA

14. M.J. SAMWAYS Invertebrate Conservation Research Centre School of Botany and Zoology University of Natal Private Bag XO 1 Scottsville 3209 South Africa

15. M.G. SERGEEV Department of General Biology, Novosibirsk State University 2 Pirogova St., Novosibirsk 630090 and Institute for Systematics and Ecology of Animals Siberian Branch of Russian Academy of Sciences 11 Frunze St., Novosibirsk 630091 Russia

16. I.M. SOKOLOV Laboratory of Agrobiocoenology All-Russian Institute for Plant Protection (VIZR) Podbelsky Street, 3 St.-Petersburg - Pushkin 189620 Russia

17. I.A. V ANJKOV A Department of General Biology, Novosibirsk State University 2 Pirogova St., Novosibirsk 630090 Russia

219

221

TABLE I. Participants in the NATO Advanced Research Workshop on Acridogenic and Anlhropogenk Hauuds to the Grasslond Biome: MtJlllJling Grasshopper Outbreaks without Risking Environmmrol Disarter

Name Afrdiatlon Couutry

Gary BeIovsky Dept. of Fisheries /I( Wildlife. Utah Stale Universi!y USA

Mikbail Chemyakhovs1tiy Dept. of Zoology /I( Ecology. Moscow State Pedagogical Universi!y Russia

Furkat Gapparov Uzbek Institute for P1anl Prou:ction (UzNlIZR) Uzbekistan

Nick lago Natural Resources Institute (NRI)

AnIbony loern School of Biological Sciences. Uoiversi!y of Nebraska-Lincoln USA

Vladimir Kambulin Kazakhstan Quarantine Inspection (formerly) Kazakbstan

Tibor Kisbenedek HWJgarian Natural History Museum HWJgary

Stepban Krall Deutscbe Gesell,chaft fiir Tcchnische Zusammeuarbeit (GTZ)

lohn Larsen Unitl:<l States Depanment of Agriculture (APHIS) USA

Alexandre laIchininskY Dept. of Renewable Resources. Universi!y of Wyoming USA

Michel Lecoq CIRAD-AMISIPRIFAS France

leffrey Lockwood Dept. of Renewable Resources. Universi!y of Wyoming USA

Annie Monanl Plant Production /I( Prou:ction Division. FAO-UN Italy

Owen Olfert Agriculture /I( Agri-food Canada Canada

lerome Onsager United StaleS Department of Agriculture (ARS) USA

Daniel one Academy of Natural Sciences. Philadelphia USA

Ludmila Psbeoitsyna Dept. of General Biology. Novosibirsk State Uoiversi!y Russia

Michael Samways Zoology /I( Entomology Dept. Uoiversi!y of Natal South Africa

Scott Schell Dept. of Renewable Resources. Universiry of Wyoming USA

MicIiaeI Sergeev Dept. of General Biology. Novosibirsk State Universi!y Russia

Igor Sokolov All-Russian Institute for P1anl Protection (VIZR) Russia

grasshoppers and grassland health: managing grasshopper outbreaks without risking environmental...

Documents