Proc. Natl. Acad. Sci. USA
Vol. 94, pp. 12243–12248, November 1997
A total system approach to sustainable pest management
W. J. Lewis*†, J. C. van Lenteren‡, Sharad C. Phatak§, and J. H. Tumlinson, III¶
*Insect Biology and Population Management Research Laboratory, Agricultural Research Service, United States Department of Agriculture, P.O. Box 748, Tifton, GA
31793; ‡Department of Entomology, Agricultural University, P.O. Box 8031, 6700 EH Wageningen, The Netherlands; §Department of Horticulture, Coastal Plain
Experiment Station, University of Georgia, P.O. Box 748, Tifton, GA 31793; and ¶United States Department of Agriculture-Agricultural Research Service, Center for
Medical, Agricultural, and Veterinary Entomology, P.O. Box 14565, Gainesville, FL 32604
Contributed by J. H. Tumlinson, III, August 13, 1997
ABSTRACT A fundamental shift to a total system ap-
proach for crop protection is urgently needed to resolve
escalating economic and environmental consequences of com-
bating agricultural pests. Pest management strategies have
long been dominated by quests for ‘‘silver bullet’’ products to
control pest outbreaks. However, managing undesired vari-
ables in ecosystems is similar to that for other systems,
including the human body and social orders. Experience in
these fields substantiates the fact that therapeutic interven-
tions into any system are effective only for short term relief
because these externalities are soon ‘‘neutralized’’ by coun-
termoves within the system. Long term resolutions can be
achieved only by restructuring and managing these systems in
ways that maximize the array of ‘‘built-in’’ preventive
strengths, with therapeutic tactics serving strictly as backups
to these natural regulators. To date, we have failed to incor-
porate this basic principle into the mainstream of pest man-
agement science and continue to regress into a foot race with
nature. In this report, we establish why a total system ap-
proach is essential as the guiding premise of pest management
and provide arguments as to how earlier attempts for change
and current mainstream initiatives generally fail to follow this
principle. We then draw on emerging knowledge about mul-
titrophic level interactions and other specific findings about
management of ecosystems to propose a pivotal redirection of
pest management strategies that would honor this principle
and, thus, be sustainable. Finally, we discuss the potential
immense benefits of such a central shift in pest management
The therapeutic approach of killing pest organisms with toxic
chemicals has been the prevailing pest control strategy for over
50 years. Safety problems and ecological disruptions continue
to ensue (1), and there are renewed appeals for effective, safe,
and economically acceptable alternatives (2). Considerable
effort has been directed toward such alternatives, and new
technology has been implemented and is still emerging. How-
ever, the major trend has been toward the use of modern
chemistry and molecular biology to replace traditional pesti-
cides with less hazardous chemicals or nontoxic biologically
based products; but these means are still therapeutics. Thus,
the classic treadmill effect in pursuit of remediation of the
symptoms persists (2) while tolls due to pests grow higher by
some estimates. Crop losses due to arthropods, diseases, and
weeds, though disputed by some as a valid measure, have
increased on a world basis from 34.9% in 1965 (3) to 42.1% in
1988–1990 (4) despite the intensification of pest control.
In this report, we argue that the central weakness in how we
think about pest management as a component of agricultural
systems has not been addressed. We must go beyond replacing
toxic chemicals with more sophisticated, biologically based
agents and re-examine the entire paradigm around the ther-
apeutic approach, including how and why those therapeutics
are used. Truly satisfactory solutions to pest problems will
require a shift to understanding and promoting naturally
occurring biological agents and other inherent strengths as
components of total agricultural ecosystems and designing our
cropping systems so that these natural forces keep the pests
within acceptable bounds. Recent discoveries in multitrophic
interactions (5) together with renewed emphasis on broader
based ecosystem management (6) indicate powerful prospects
for this direction. Although we address the subject primarily
from a perspective of arthropod pests, similar cases can
generally be made for other pests [see Cook et al. (7) for
background information important to related views for other
Premise of a Revised Approach
The underlying principle of our position is that components of
agricultural ecosystems interact, and, through a set of feedback
loops, maintain ‘‘balance’’ within functional f luctuating
bounds. Furthermore, therapeutic interventions into these
systems are met by countermoves that ‘‘neutralize’’ their
effectiveness [see Flint and van den Bosch (8) and Cook and
Baker (9) for an elegant discussion of this point]. We are taught
this basic principle from our earliest training in ecology but
often overlook it in practice for various reasons, including our
tendency in science to divide things into specialized parts, i.e.,
to apply a reductionist approach. The basic principle for
managing undesired variables in agricultural systems is similar
to that for other systems, including the human body and social
systems. On the surface, it would seem that an optimal
corrective action for an undesired entity is to apply a direct
external counter force against it. However, there is a long
history of experiences in medicine and social science where
such interventionist actions never produce sustainable desired
effects. Rather, the attempted solution becomes the problem
[See Waltzlawick et al. (10) for a discussion of this subject with
coverage of underlying mathematical principles.] We find vivid
examples to this end in the problems of addiction as a
consequence of the use of drugs for treatment of pain or
mental distress and black market crime as a repercussion to the
use of prohibition as an intended solution for alcoholism. Thus,
as a matter of fundamental principle, application of external
corrective actions into a system can be effective only for short
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked ‘‘advertisement’’ in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
© 1997 by The National Academy of Sciences 0027-8424y97y9412243-6$2.00y0
PNAS is available online at http:yywww.pnas.org.
Abbreviations: IPM, Integrated Pest Management; Bt, Bacillus thu-
†To whom reprint requests should be addressed. e-mail: wjl@tifton.
term relief. Long term, sustainable solutions must be achieved
through restructuring the system so that inherent forces that
function via feedback mechanisms such as density dependence
are added andyor function more effectively.
The foundation for pest management in agricultural systems
should be an understanding and shoring up of the full com-
posite of inherent plant defenses, plant mixtures, soil, natural
enemies, and other components of the system. These natural
‘‘built in’’ regulators are linked in a web of feedback loops and
are renewable and sustainable. The use of pesticides and other
‘‘treat-the-symptoms’’ approaches are unsustainable and
should be the last rather than the first line of defense. A pest
management strategy should always start with the question
‘‘Why is the pest a pest?’’ and should seek to address underlying
weaknesses in ecosystems andyor agronomic practice(s) that
have allowed organisms to reach pest status.
Attempts for Change: No Real Change
Throughout the debate on alternative methods for controlling
pests, various ideas have been expressed and new approaches
have emerged. Three subject areas, Biological Control, Inte-
grated Pest Management (IPM), and Biotechnology, have
achieved particular importance in our quest for better pest
Biological Control. Biological control has a long history of
use in pest management and has gained renewed interest
because of problems encountered with the use of pesticides.
The term ‘‘biological control’’ has been used, at times, in a
broad context to encompass a full spectrum of biological
organisms and biologically based products including phero-
mones, resistant plant varieties, and autocidal techniques such
as sterile insects. The historical and more prevalent use of this
term is restricted to use of natural enemies to manage popu-
lations of pest organisms.
Biological control has been spectacularly successful in many
instances, with a number of pest problems permanently resolved
by importation and successful establishment of natural enemies.
These importation successes have been limited largely to certain
types of ecosystems andyor pest situations such as introduced
pests in perennial ecosystems. On the other hand, this approach
has met with limited success for major pests of row crops or other
ephemeral systems. In these situations, the problem is often not
the lack of effective natural enemies but management practices
and a lack of concerted research on factors that determine the
success or failure of importation attempts in the specific agro-
ecosystem setting. Thus, importation programs, to date, are
largely a matter of trial and error based on experience of the
individual specialists involved.
Conservation of natural enemies received more attention as
part of a cultural management approach before the advent of
synthetic pesticides. Since that time, this realm of biological
control has been neglected. The term ‘‘conservation’’ tends to
limit one’s vision to a passive approach of acknowledging that
natural enemies are valuable and should be harmed no more
than necessary. It is important that we develop a more active
approach that seeks to understand natural enemies and how
they function as a part of the ecosystem and to promote their
effectiveness by use of habitat management (landscape ecol-
ogy) and other cultural management approaches.
Augmentation through propagation and release of natural
enemies is an area of biological control that has received much
attention in recent years. These efforts include research on in vitro
and in vivo mass rearing technology and on transport and release
methodology for area-wide population suppression and for field-
to-field therapeutic treatments. Although the development of this
technology is valuable, it is an extension of the treat-the-
symptoms paradigm. In principle, natural enemies used in these
methodologies are biopesticides, and the general approaches
differ from conventional pesticidal applications only in the kind
of products used. From this ‘‘product formulation’’ perspective
and from our existing infrastructure, the major emphasis in
augmentation schemes becomes focused on how to produce and
transport a large number of natural enemies at a low cost. Less
emphasis is placed on how natural enemies function and how we
can promote their natural effectiveness.
In keeping with the historical therapeutic-based attitude and
existing infrastructure, most concentrated efforts for biological
control appear to be directed toward the ‘‘rear and release’’
augmentation, followed by importation and thirdly by conserva-
tion. This order of priorities should be reversed. First, we need to
understand, promote, and maximize the effectiveness of indige-
nous populations of natural enemies. Then, based on the knowl-
edge and results of these actions, we should fill any key gaps by
importation. Finally, therapeutic propagation and releases should
be used as a backup to these programs when necessary.
IPM. Throughout our quest for alternative pest control
measures, the IPM concept has by far received the most
attention as a comprehensive pest management approach. IPM
has had a varied history, has been defined in many ways, and
has been implemented under an array of different connota-
tions. The term was first used as ‘‘integrated control’’ by
Bartlett (11) and was further elaborated on by Stern et al. (12)
in reference to the concept of integrating the use of biological
and other controls in complementary ways. The term was later
broadened to embrace coordinated use of all biological, cul-
tural, and artificial practices (13). Subsequently, under the
term ‘‘IPM,’’ various authors have advocated the principle of
incorporating the full array of pest management practices
together with production objectives into a total systems ap-
proach. See Flint and van den Bosch (8) for a comprehensive
and ecologically based discussion of this concept and the
potential benefits of its implementation.
The principles discussed by Flint and van den Bosch (8) are,
in our opinion, solid and on target. They make a thorough case
for a comprehensive long term pest management program
based on knowledge of an ecosystem that weighs economic,
environmental, and social consequences of interventions.
However, as translated into practice, IPM has been primarily
a monitoring program in which thresholds are established and
chemicals are used only on an as-needed basis. Much less
emphasis has been placed on understanding and promoting
inherent strengths within systems to limit pest populations
through use of approaches such as landscape ecology. In other
words, IPM programs have been operated with pesticide
management objectives rather than pest management objec-
tives. We hasten to add that their use has been of major benefit
and has greatly reduced the quantity of pesticides used.
Furthermore, activities remain underway to refocus IPM
toward the achievement of its full objectives (14, 15). However,
our point is that, again, the tendency has been to remain
centered on a monitor and treat-the-symptoms approach vs.
the more fundamental question of ‘‘Why is the pest a pest?’’
Biotechnology. Although biotechnology is not a pest man-
agement approach as such, we include it because it is receiving
major emphasis and is being geared to provide a wave of new
products for pest management. In fact, many seem to view
biotechnology as an innovative means for providing safe and
effective tools that will essentially resolve pest management
problems. Major technological advances in chemistry, bio-
chemistry, behavior, neurophysiology, molecular genetics, and
genetic engineering have resulted in an array of biorational
products and materials that are less toxic and hazardous to
humans and the environment than conventional pesticides.
These products include genetically engineered plants for stron-
ger resistance to pests, plants, and natural enemies with high
tolerance to pesticides and sophisticated formulations and
delivery methods for biopesticides, semiochemicals, and other
new tools. The biorationalybiologically based materials pro-
vided are potentially valuable advancements that have an
12244 Perspective: Lewis et al. Proc. Natl. Acad. Sci. USA 94 (1997)
appropriate place in modern pest management. However, the
strategy for development and use of these ‘‘high tech’’ tools has
been dominated by a continued search for ‘‘silver bullet’’
solutions that can be easily deployed in a prescription-like
manner to remediate pest outbreaks or to exclude the pest’s
presence. As spectacular and exciting as biotechnology is, its
breakthroughs have tended to delay our shift to long term,
ecologically based pest management because the rapid array of
new products provide a sense of security just as did synthetic
pesticides at the time of their discovery in the 1940s. Also,
industry focuses on using genetic manipulation and other tech-
niques to increase the virulence and host range of biopesticides
instead of designing them as complements to natural strengths.
Thereby, the manipulated pathogens and the crops engineered to
express toxins of pathogens are simply targeted as replacements
for synthetic pesticides and will become ineffective in the same
way that pesticides have. It will be unfortunate if these powerful
agents are wasted rather than integrated as key parts of sustain-
able pest management systems.
The four major problems encountered with conventional
pesticides are toxic residues, pest resistance, secondary pests,
and pest resurgence. The latter three of these are fundamental
consequences of reliance on interventions that are both dis-
ruptive and of diminishing value because of countermoves of
the ecological system. Therefore, a mere switch to nontoxic
pesticides, such as microbials or inundative releases of natural
enemies, although helpful in reducing environmental contam-
ination and safety problems, still does not truly address the
ecologically based weakness of the conventional pest control
approach. Such tools used in this manner, whether chemical,
biological or physical, are extensions of the conventional
approach that leaves us in a confrontation with nature. Also,
this operational philosophy tends to promote the development
and adoption of the more disruptive products because, within
this paradigm, they work better than softer, less obtrusive
What, then, would represent a meaningful fundamental shift
in our pest management strategy? Furthermore, what should
be the components of such a strategy, and how can we
crystallize this strategy into programs that result in effective
and lasting pest management systems? Clearly, the central
foundation should be approaches that appreciate the interac-
tive webs in ecosystems and seek solutions with net benefits at
a total ecosystem level. Therefore, the approaches should
focus on harnessing inherent strengths within ecosystems and
be directed more toward bringing pest populations into ac-
ceptable bounds rather than toward eliminating them (Fig. 1).
These solutions would avoid undesirable short term and long
term ripple effects and would be sustainable. Moreover, for
adoption of such approaches, they must reasonably meet
production demands and be cost-competitive on the short
term. We suggest three lines along which approaches can be
developed: (i) ecosystems management; (ii) crop attributes and
multitrophic level interactions; and (iii) therapeutics with
minimal disruptions. However, with all of these approaches, it
is important to keep in mind the objective of balance vs. undue
selective pressure by any single tactic. Recent experiences with
insect pest management for cotton in the southeastern United
States will be used for key examples in the discussion.
Ecosystem Management. Understanding and managing an
ecosystem within which we farm is the foundation upon which
all the farming strategies, including pest management, should
be designed. This foundation has become the victim of reduc-
tionist approaches. Because of political and funding channels,
scientific teams typically are assembled around commodities
across geographical areas. Therefore, the informational base
relative to a particular crop as an interactive component of a
farming ecosystem is very limited. For example, cotton spe-
cialists focus their interactions toward other cotton specialists,
often within their own discipline, across the cotton belt.
However, both vegetable and cotton production are increasing
in the same area and sometimes on the same farms in the
southeastern United States. These crops share many of the
same pests and natural enemy fauna. Therefore, pest manage-
ment practices on one crop can directly or indirectly affect the
other. A redirection of pest management is needed to incor-
porate year-round soil, weed, cropping, water, and associated
practices at farm and community levels and to consider the
effects of these practices on the overall fauna, nutritional state,
and balance of local ecosystems (16).
Recent studies demonstrate that such a redirection would be
highly fruitful. For example, problems with soil erosion have
resulted in major thrusts in use of winter cover crops and
conservation tillage. Preliminary studies indicate that cover
crops also serve as a bridgeyrefugia to stabilize natural enemyy
pest balances and relay these balances into the crop season (17,
18). Crimson clover and other legumes, into which cotton can
be strip tilled in the Southeast, appear to be good winter and
spring reservoirs for predators and parasitoids of cotton pests
FIG. 1. Illustration of a shift to a total system approach to pest management through a greater use of inherent strengths based on a good
understanding of interactions within an ecosystem while using therapeutics as backups. The upside-down pyramid to the left reflects the unstable
conditions under heavy reliance on pesticides, and the upright pyramid to the right reflects sustainable qualities of a total system strategy.
Perspective: Lewis et al. Proc. Natl. Acad. Sci. USA 94 (1997) 12245
(19, 20). The green cloverworm in clover serves as a good
alternate winter and spring host for the parasitoid Cotesia
marginiventris that limits subsequent outbreaks of armyworms
and loopers in the cotton (21). Also, aphid, thrips, and
budwormybollworm populations in clover appear to provide
reservoirs for establishing earlier balances between these pests
and their natural enemy guilds. On the other hand, when fields
are fallow during winter and spring, natural enemy buildups
cannot begin until a crop is available. Integrating appropriate
cover crops with conservation tillage can have a number of
agronomic benefits: reduced soil erosion, enhanced levels of
organic matter, improved soil drainage and moisture retention,
restoration of important nutrients, and weed control (20, 22, 23)
while restoring and strengthening natural pest control. Other
preventive measures including crop rotations, avoiding large scale
monocropping, leaving unsprayed strips, and planting field mar-
gins with appropriate year-round refugia for natural enemies will
contribute to prevention of pest outbreaks (24–26).
The growing of clover andyor encouragement of certain
weeds along field margins and other unplanted areas can also
provide important refugia for developing natural enemyypest
balances during a cropping season. For example, two common
weeds in the southeast, fleabane and horsetail, are important
hosts for plant bugs and their natural enemies. In fact, they are
preferred over cotton by the insects, and data indicate that
these plants act as effective decoys to coax plant bugs away
from cotton (19). Serious infestations of plant bugs occur
primarily where cotton is planted ‘‘ditch bank to ditch bank,’’
along with clean cultivation apparently caused by exclusion of
such preferred alternate host plants.
Obviously landscape ecology practices exert a variety of
desired or undesired effects on cropping systems (27). Thus, it
is vital that we assemble appropriate teams to elucidate
interactions at the ecosystem level to establish the knowledge
base for ecologically based pest management systems.
Crop Attributes and Multitrophic Level Interactions. Con-
sideration of crop plants as active components of multitrophic
level interactions is crucial to a total systems approach to pest
management. We have known vaguely for a long time that
plant traits have important impacts on both herbivores and
their natural enemies. But, again, the reductionist approach
has caused us to manipulate plant traits in ways detrimental to
long term balance in the cropping systems.
Recent discoveries of tritrophic level interactions among
plants, herbivores, and parasitoidsypredators have demon-
strated how tightly interwoven these components are and
illustrate the importance of multitrophic perspectives for
effective and sustainable pest management strategies. Plants
have long been known to possess toxins and other chemicals
that serve to discourage herbivore feeding. The discipline of
host–plant resistance directed toward breeding plants resistant
to pest attack was developed around such knowledge and
contributed greatly to pest management. Recent studies also
show, however, that plants play an active and sophisticated role
in their defense against insect activities, and their defense
responses often are customized for certain, interactive, mul-
titrophic situations (5). For example, some plants respond to
insect herbivory by releasing volatile chemical cues that attract
predators and parasitoids that, in turn, attack the herbivores
(28–30). These volatiles are released only in response to
herbivore damage, not by mechanical damage similar to her-
bivory, and are released from the entire plant (31, 32). This
effect enables the natural enemy to distinguish infested plants
from uninfested neighbors. For example, cotton fed on by beet
armyworm larvae releases terpenoids that attract the parasi-
toid C. marginiventris. Furthermore, a certain naturalized
variety of cotton releases '10 times more of these chemicals
in response to the insect herbivore damage than do commercial
lines (33). By understanding the mechanisms governing such
important defense attributes, they can be restored to domestic
cultivars, and their incidental loss, while breeding for other
traits, can be prevented in the future.
Crop plants also provide vital food resources for certain key
natural enemies. Floral and extra floral nectaries, for example,
provide necessary food for foraging parasitoids. Extra f loral
nectar increases the attraction, efficiency, and retention of the
key parasitoids C. marginiventris, Microplitis croceipes, and
Cardiochiles nigriceps, important to the control of armyworms
and bollwormsybudworms in crops such as cotton (34, 35).
However, extra floral nectar also serves as food for certain
pests such as the adult moths of the caterpillar pests just
mentioned. Based on information regarding the role of the
extra floral nectar as food for moths, nectariless cotton
varieties were released a few years ago without regarding their
importance as a food for natural enemies of cotton pests. These
facts emphasize the need to broaden our base of information
upon which we design pest management strategies.
Also, there is a rapidly expanding body of knowledge about
similar signaling that enables injured plants to produce toxins
and antifeedants that are directed specifically toward herbi-
vores. For example, feeding activities of certain caterpillars on
the leaves of tomatoes and potatoes induce a systemic pro-
duction of protease inhibitors expressed throughout the plant
that interfere with the digestion process and feeding behavior
of insects (36).
Even greater than our limited knowledge of the mechanisms
regulating these important plant attributes is the void in our
knowledge of how factors like soil properties, nutrition, andyor
water stress affect their expression. Inadequate availability of
a key soil element for example could make a major difference
in the effectiveness of one or more of a plant’s interactions with
herbivores or natural enemies, thereby influencing a plant’s
vulnerability to herbivore damage in a major way. Greater
understanding of the factors that regulate these interactions in
cropping systems can allow us to deal with plant health at an
entirely different level.
There is a tendency within the traditional paradigm to use
toxins, attractants, or other plant attributes as products and to
intervene in ways that are out of harmony with natural system
interactions. For example, we identify, synthesize, and formu-
late herbivore toxins and natural enemy attractants as sprays
to kill herbivores and lure the natural enemies, respectively.
Also, we breed and engineer plants for constitutive expression
of traits in ways that maximize immediate deterrence of pests
or attraction of natural enemies without regard to pest density
or plant damage. Natural systems provide evidence that this is
not always an appropriate approach for plant defense. In the
case of the protease inhibitor in tomato and potato cited above,
these materials are constitutively expressed in the fruit but only
induced by damage in leaves (37). We suggest that this system
has been selected in nature because it is the most durable
strategy. A system of constitutive expression in fruit but only
inducible in leaves experiencing damage by feeding insects
provides maximum protection of the fruit. Leaves serve as a
decoy alternative for feeding by caterpillars but possess a
mechanism that limits feeding damage. This strategy also
provides hostyprey resources that allow participation by a
plant’s parasitoidypredator allies. We must observe and con-
sider natural systems when developing strategies for novel
traits such as a gene for producing Bacillus thuringiensis (Bt)
toxin, i.e., plant engineering [see Gould (38) for an excellent
reference in this regard]. For example, cotton cultivars with a
full constitutive expression of Bt toxin have been introduced
commercially. This practice amounts to a continuous spraying
of an entire plant with the toxin, except the application is from
inside out. Various methods for resistance management, in-
cluding pestynatural enemy refugia and limiting acreage
planted with a cultivar, are being used. However, we urge more
concerted efforts toward breeding and engineering plants with
12246 Perspective: Lewis et al. Proc. Natl. Acad. Sci. USA 94 (1997)
traits such as tissue-specific and damage-induced chemical
defenses that work in harmony with natural systems.
Genetic engineering and other such technologies are pow-
erful tools of great value in pest management. But, if their
deployment is to be sustainable, they must be used in con-
junction with a solid appreciation of multitrophic interactions
and in ways that anticipate countermoves within the systems.
Otherwise, their effectiveness is prone to neutralization by
resistance in the same manner as with pesticides.
Therapeutics. Therapeutics have a valuable role in ecolog-
ically based pest management strategies, but they should be
viewed as backups rather than as primary lines of defense.
Also, therapeutics should be recognized as potentially disrup-
tive and used as unobtrusively as possible. The key principle is
that they should be geared toward bringing a pest organism
into acceptable bounds with as little ecological disruption as
possible. Synthetic products, natural products, and living or-
ganisms can be effective as therapeutics, and the fact that a
product is natural andyor nontoxic does not necessarily mean
it is less disruptive than synthetic products. The important
thing is to work as much in harmony as possible with the
system’s inherent defenses.
A wide array of therapeutic products are available, and more
are being developed with modern technology. A vast arsenal
of natural products identified from plants, insects, and micro-
organisms is being synthesized and formulated for use as
biopesticides. Semiochemicals such as sex pheromones and
natural enemy attractants can be used as baits and lures to
disrupt pest activity and promote natural enemy presence.
Pathogens, parasitoids, as reared in vivo or in vitro, are
available and are being touted as therapeutic tools. All of these
organisms andyor their by-products are important biofriendly
alternatives to toxic, broad spectrum, conventional pesticides.
Still, our primary pest management tactic should be maximi-
zation of built in pest reduction features of an ecosystem.
Therapeutic tools should be used as secondary backups.
Overreliance on them will return pest management strategies
to a treadmill situation (Fig. 1).
Another problem is the tendency to seek therapeutics that
give us the quickest effect. Sales of biological insecticides
amount to about $110 million annually, and Bt is the main
product ($90 million). Generally, microbial organisms work
slowly relative to synthetic pesticides. Therefore, industry has
as first priority formulation of microbials to obtain faster kill
and is less interested in long term pest reduction effects. Thus,
the role that microbials could play in orchard and forest pest
management, as well as in programs like control of grasshop-
pers in Sahelian, Africa, is neglected (39).
Retarded development of pests may be more desirable than
quick kill in certain situations. For example, Bt products are
considered unacceptable for controlling beet armyworms in
cotton because of their slow killing action. Yet, some studies
indicate that a slow kill may be more preferable when exam-
ined from a larger perspective. As indicated above, C. mar-
giniventris is a key parasitoid for managing the beet armyworm
and interventions should avoid disrupting this natural enemy.
Beet armyworm larvae intoxicated by sublethal dosages of
MVP (Mycogen, San Diego) (a Bt-derived biopesticide) ex-
perience retarded development and feeding and are subject to
higher parasitism than nontreated beet armyworm larvae (40).
In other words, an effective, nondisruptive way to manage a
moderate beet armyworm outbreak may be to retard its
development and damage while giving the parasitoids time to
work, thereby strengthening the parasitoids’ effect during
subsequent generations. A similar effect was reported earlier
for Bt and a parasitoid of gypsy moths (41). A quick kill may
provide more immediate results but destroys a resource for
parasitoids and limits their presence with subsequent genera-
tions of pests, thus leading to resurgence.
We must remember—our primary objective in pest man-
agement is not to eliminate a pest organism but to bring it into
acceptable bounds. The role of therapeutics is not to replace
natural systems. Rather, their role is to serve as complements
while the system is temporarily out of balance. From that
perspective, it is clear that interventions that interfere with the
restoration of balance are counterproductive. Waage (39)
suggests that biopesticides could form the ‘‘methadone of
IPM,’’ helping agroecosystems to recover from the habit of
calendar spraying while we are redesigning and nurturing them
to a more self-renewing capacity.
The benefits of a total system approach would be immense,
directly to farming and indirectly to society. The approach
takes into account impacts on our natural resources such as the
preservation of flora and fauna, quality and diversity of
landscape, and conservation of energy and nonrenewable
resources. Long term sociological benefits would also emerge
in areas of employment, public health, and well being of
persons associated with agriculture (42, 43).
In The Netherlands, prototypes of various multidisciplinary,
arable farming systems have been evaluated on a semi-
practical scale (44). In 1979, a national experimental farm for
the development and comparison of alternative farming sys-
tems was set up in Nagele (one of the ‘‘polders’’). The size of
the farm was 72 hectares (almost 300 acres). Among other
studies, integrated and conventional farming practices were
compared for seed potatoes, dry peas, carrots, onions, sugar
beets, and winter wheat. Crop protection and other manage-
ment practices with the integrated approach followed the basic
principles discussed herein.
Over a 15-year period, pesticide use on these integrated
farms was reduced over 90% (Fig. 2). They found that pesti-
cides, and fertilizers, can be decreased through implementa-
tion of alternative practices based on intensified knowledge of
the ecosystem. Artificial fertilizers are replaced by organic
manure and effective use of crop residues. Insect, weed, and
disease problems are reduced through natural control by the
enriched natural enemy fauna, the use of weed-competitive or
disease- and pest-resistant varieties (with an emphasis on
durable systems for resistance), reduction of nitrogen fertili-
zation, and judicious use of chemical pest control based on
careful population sampling and decision thresholds. Results
from these demonstration farms have been so encouraging that
implementation of integrated farming is being enforced by the
FIG. 2. Average use of pesticides (kilogram active ingredienty
hectareyyear) in conventional and integrated farming demonstrations
in The Netherlands (1986–1990); after Wijnands and Kroonen–
Backbier (43). ha, hectare.
Perspective: Lewis et al. Proc. Natl. Acad. Sci. USA 94 (1997) 12247
Dutch Ministry of Agriculture to reduce environmental pol-
lution and to create a firmer basis for survival of agriculture in
the longer term.
Yields were somewhat lower on the demonstration farms but
were compensated for by cost reduction through lower pesti-
cide and fertilizer inputs. Thus, the net short term profits of the
demonstration farms were equal to those of the conventional
farms. We emphasize the short term economic aspect of
sustainable farming because immediate profitability figures,
along with the environmental concerns, are crucial to adoption
of the practices. However, the eventual consequences of
conventional farming are so severe, environmentally, socially,
and economically, that it is wise to initiate changes even under
situations in which short term economic benefits are marginal.
Bio-friendly agriculture and good economics, over the long
term, clearly go hand in hand.
Recent quests for effective, safe, and lasting pest management
programs have been targeted primarily toward development of
new and better products with which to replace conventional
toxic pesticides. We assert that the key weakness with our pest
management strategies is not so much the products we use but
our central operating philosophy. The use of therapeutic tools,
whether biological, chemical, or physical, as the primary means
of controlling pests rather than as occasional supplements to
natural regulators to bring them into acceptable bounds
violates fundamental unifying principles and cannot be sus-
tainable. We must turn more to developing farming practices
that are compatible with ecological systems and designing
cropping systems that naturally limit the elevation of an
organism to pest status. We historically have sold nature short,
both in its ability to neutralize the effectiveness of ecologically
unsound methods as well as its array of inherent strengths that
can be used to keep pest organisms within bounds. If we will
but understand and work more in harmony with nature’s
checks and balances we will be able to enjoy sustainable and
profitable pest management strategies, which are beneficial to
all participants in the ecosystem, including humans.
We are grateful to Karen Idoine and Drs. Charlie E. Rogers (now
deceased), G. T. Fincher, John Garcia, Bryan Beirne, Thomas Eisner,
R. James Cook, and Fred Gould for providing important advice and
suggestions throughout the drafting and revision of this paper.
1. Wright, L. (1996) New Yorker, Jan. 15th, 42–48, 50 –53, 55.
2. Benbrook, C. M. (1996) Pest Management at the Crossroads
(Consumer’s Union, Yonkers, NY).
3. Cramer, H. H. (1967) Bayer Pflanzenschutznachrichten 20, 1–523.
4. Oerke, E. C., Dehne, H. W., Shoenbeck, F. & Weber, A. (1994)
Estimated Losses in Major Food and Cash Crops (Elsevier,
5. Tumlinson, J. H., Lewis, W. J. & Vet, L. E. M. (1993) Sci. Am.
6. Altieri, M. (1995) Agroecology: The Science of Sustainable Agri-
culture. (Westview, Boulder, CO).
7. Cook, R. J., Gabriel, C. J., Kelman, A., Tolin, S. & Vidaver, A. K.
(1995) BioScience 45, 354–357.
8. Flint, M. L. & van den Bosch, R. (1981) Introduction to Integrated
Pest Management (Plenum, New York).
9. Cook, R. J. & Baker, K. F. (1983) The Nature and Practice of
Biological Control of Plant Pathogens (Am. Phytopathol. Soc., St.
10. Watzlawick, P., Weakland, J. & Fisch, R. (1974) Change: Prin-
ciples of Problem Formation and Problem Resolution (Norton,
11. Barlett, B. R. (1956) Agric. Chem. 11, 42–44.
12. Stern, V. M., Smith, R. F. van den Bosch, R. & Hagen, K. S.
(1959) Hilgardia 29, 81–101.
13. van den Bosch, R. & Stern, V. M. (1962) Annu. Rev. Entomol. 7,
14. Board on Agriculture, National Research Council, Committee on
Pest and Pathogen Control Through Management of Biological
Control Agents and Enhanced Natural Cycles and Processes
(1996) Ecologically Based Pest Management: New Solutions for a
New Century (National Academy Press, Washington, DC).
15. van Lenteren, J. C., Benuzzi, M., Nicoli, G. & Maini, S. (1992)
in Biological Control and Integrated Crop Protection: Toward
Environmentally Safer Agriculture, eds. van Lenteren, J. C., Minks,
A. K. & de Ponti, O. M. B. (Purdoc, Wageningen, The Nether-
lands), pp. 77–89.
16. Edwards, C. A., Lal, R., Madden, P., Miller, R. H. & House, G.
(1990) Sustainable Agricultural Systems (St. Lucie, Delray Beach,
17. Blumberg, A. Y. & Crossley, D. A. (1982) Agro-Ecosyst. 8,
18. McPherson, R. M., Smith, J. C. & Allen, W. A. (1982) Environ.
Entomol. 11, 685–689.
19. Fleischer, S. J. & Gaylor, M. J. (1987) Environ. Entomol. 16,
20. Phatak, S. C. (1993) in Cover Crops: Cotton Relay Cropping (2nd
Organic Growers Conference, Calif. Inst. Rural Studies, Davis,
CA), pp. 269–274.
21. Ruberson, J. R., Herzog, G. A., Lambert, W. R. & Lewis, W. J.
(1994) Florida Entomol. 77, 440–453.
22. Bugg, R. L., Wa¨ckers, F. L., Brunson, K. E., Dutcher, J. D. &
Phatak, S. C. (1991) J. Econ. Entomol. 84, 408–416.
23. Griffith, D. R., Mannering, J. V. & Box, J. E. (1986) in No-Tillage
and Surface-Tillage Agriculture: The Tillage Revolution, eds.
Sprague, M. A. & Triplett, G. B. (Wiley, New York), pp. 19–57.
24. Altieri, M. A. (1994) Biodiversity and Pest Management in Agro-
ecosystems, (Food Products Press, New York).
25. Luna, J. M. & House, G. J. (1990) in Sustainable Agricultural
Systems, eds. Edwards, C. A., Lal, R., Madden, P., Miller, R. H.
& House, G. J. (St. Lucie, Delray Beach, FL), pp. 157–173.
26. Olkowski, W., Daar, S. & Olkowski, H. (1991) Common Sense
Pest Control (Taunton, Newton, CT).
27. All, J. N. & Musick, G. J. (1986) in No-Tillage and Surface-Tillage
Agriculture: The Tillage Revolution, eds. Sprague, M. A. & Trip-
lett, G. B. (Wiley, New York), pp. 347–387.
28. Dicke, M. & Sabelis, M. W. (1988) Neth. J. Zool. 38, 148–165.
29. Dicke, M., Sabelis, M. W., Takabayashi, J., Bruin, J. & Posthu-
mus, A. (1990) J. Chem. Ecol. 16, 3091–3118.
30. Turlings, T. C. J., Tumlinson, J. H. & Lewis, W. J. (1990) Science
31. Turlings, T. C. J. & Tumlinson, J. H. (1992) Proc. Natl. Acad. Sci.
USA 89, 8399– 8402.
32. Ro¨se, U. S. R., Manukian, A., Heath, R. R. & Tumlinson, J. H.
(1996) Plant Physiol. 111, 487–495.
33. Loughrin, J. H., Manukian, A., Heath, R. R. & Tumlinson, J. H.
(1995) J. Chem. Ecol. 21, 1217–1227.
34. Lewis, W. J. & Takasu, K. (1990) Nature (London) 348, 635– 636.
35. Stapel, J. O., Cortesero, A. M. Demoraes, C. M., Tumlinson, J. H.
& Lewis, W. J. (1997) Environ. Entomol. 26, 617– 623.
36. Ryan, C. A. & Farmer, E. E. (1991) Annu. Rev. Plant Physiol.
Plant Mol. Bio. 42, 651–743.
37. Pearce, G., Ryan, C. A. & Liljegren, D. (1988) Planta 175,
38. Gould, F. (1991) Am. Sci. 79, 496–507.
39. Waage, J. K. (1994) Biodiversity, Business and Biotechnology:
Shaping the Future of Biological Control (Wageningen Agricul-
tural University, Wageningen, The Netherlands).
40. Soares, G. G. Jr., Lewis, W. J., Strong-Gunderson, J. M., Waters,
D. J. & Hamm, J. J. (1994) Proceedings of the Second Canberra
Bacillus thuringiensis Meeting (Commonwealth Scientific and
Industrial Research Organization, Canberra, Australia).
41. Weseloh, R. M., Andreadis, T. G., Moore, R. E. B., Anderson,
F. F., Dubois, N. R. & Lewis, F. B. (1983) J. Invert. Pathol. 42,
42. Vereijken, P., Edwards, C., El Titi, A., Fougeroux, A. & Way., M.
(1986) Proceedings of Workshop on Integrated Farming. Bull.
I. O. B. C./W. P. R. S./IX 2, 1–34.
43. Wijnands, F. G. & Kroonen-Backbier, B. M. A. (1993) in Modern
Crop Protection: Developments and Perspectives, ed. Zadoks, J. C.
(Wageningen Pers, Wageningen, The Netherlands), pp. 227–234.
44. van Lenteren, J. C. (1997) in Modern Agriculture and the Envi-
ronment, eds. Rosen, D., Tel-or, E. & Chen, Y. (Kluwer, Dor-
drecht), pp. 17–31.
12248 Perspective: Lewis et al. Proc. Natl. Acad. Sci. USA 94 (1997)