THIS MATERIALS
MAY NOT
BE REMOVED FRosw
,4 “THE LlBRARY

Action 8r Inacfon Levels
‘n Pest Management

 >-
 »”%i/I7§   l "#4.      i  -/ <1 /    '  a   5 l.//' a

 o» 

v/



 

TEXAS AGRICULTURAL EXPERIMENT STATION / Neville P. Clarke / The Texas A&M University System / College Station, Texas

5-1480
July 1984

           

       

 
 

   

 

 

 

 

 

I,
. y .
/
‘\ f.
I -' '
<
; .
\ .
' i
-
K
‘ .
\

     

Action and Inaetion Levels in
Pest Management

Winfield Sterling
Department of Entomology
Texas A&M University and
The Texas Agricultural Experiment Station
College Station, Texas 77843

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Systems Analysis in Decision Making . . . . . . . . . . . . 3

Ecoloical Disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Action Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

The Cotton Crop as an Example . . . . . . . . . . . . . . . . 5

Action or Action Level Models . . . . . . . . . . . . . . 5

Plant Compensation . . . . . . . . . . . . . . . . . . . . . . . 5

Plant Phenology . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Host Plant Resistance . . . . . . . . . . . . . . . . . . . . . 7

Durational Stability of the Crop . . . . . . . . . . . . 7

Other Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Arthropod Life Systems . . . . . . . . . . . . . . . . . . . . . . . . 8

Arthropod Densities . . . . . . . . . . . . . . . . . . . . . . . 8

Dispersal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Dispersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Instar, Size and Sex . . . . . . . . . . . . . . . . . . . . . . . 9
Simultaneous Damage by a Complex of Pests . 9
Action Levels for Cotton Pests . . . . . . . . . . . . . . 9

Soils and Fertility . . . . . . . . . . . . . . . . . . . . . . . . . 9

Insecticide Resistance . . . . . . . . . . . . . . . . . . . . . . . . . l0

Yield Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l0

Inaction Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ll

Key Predators . . . . . . . . . . . . . . . . . . . . . . . . . . . . ll

Colonization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l2

Switching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . l2

Economics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 

Insecticides . . . . . . . . . . . . . . . . . . . . . . . . .

Cost/Benefit Ratios . . . . . . . . . . . . . . . . . 

Production Costs . . . . . . . . . . . . . . . . . . . . y

Insect Losses . . . . . . . . . . . . . . . . . . . . . . . 7

Supply and Demand . . . . . . . . . . . . . . . . .81

Ecology and Society . . . . . . . . . . . . . . . . . . . . . I
Side Effects of Management Actions. . . i
Uncertainty in Pest Management . . . . . . >

Insecticide Decomposition . . . . . . . . . . . . L

Policy Decisions . . . . . . . . . . . . . . . . . . .. A

Strategies and Tactics . . . . . . . . . . . . . . . . . . . .

Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Community Strategies . . . . . . . . . . . . . . . .I

Pesticidal Tactics . . . . . . . . . . . . . . . . . . . . .'

Cultural Tactics . . . . . . . . . . . . . . . . . . . . 

Other Tactics . . . . . . . . . . . . . . . . . . . . . . . . ‘

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A

Computer Models and Simulation . . . . . . 

Models and Pest Management . . . . . . . . . _

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . z - "

Literature Cited . . . . . . . . . . . . . . . . . . . . . . . .. 5V

  
    
  
      
    
  

1...... g L. QQ-{gg ‘=-.-  >6

Introduction

  
 
  
  
 
 
  
 
 
 
 
  
  
  
 
 
 
  
 
  
  
 
 
 
 
  
  
 
  
 
  
  
 
  
 
 
 
  
 
  
  
 
  
 
 
 
  

 fulness of a pest control concept can be evaluated
effectiveness in solving actual problems (13). In
 d, the economic threshold concept (137, 138)
in extremely useful. Contemporary pest manage-
‘ stems that have been developed for insect pests
l» and other crops, usually use the economic
ld concept or some modification of it in the
1| -mal<ing process. However, the economic
 concept is not without some conceptual and
Y? limitations. For example, the concept as cur-
nceived and applied in pest management situa-
“I used to make decisions that call for emergency
Q primarily the use of insecticides. It is seldom
_ conjunction with biological, cultural, or any
fanagement tactics that are essentially curative.
i“ l difficulties are encountered in attempting to
(term "economic threshold” to indicate the den-
natural enemies needed to maintain pests below
le loss levels. One solution to this problem
_- to use the term “action level” as a replacement
 "economic threshold” and the term “inac-
l" for the critical natural enemy densities (134).
“rms are more fitting because economic and
 factors are both important in pest manage-
 isions. Also, the word “threshold” implies that
A population density is projected to increase to an
1e loss level when in actuality it may decrease.
 ision to take action at the threshold level may

§l%l§IJvu_-'- -

“oid some of these problems.

 factors involved in the establishment of the
‘gels have been reviewed in other reports (37,
i, 120). In this paper those factors that are
fto be of importance in calculating the most
action levels are reviewed. Although the dis-
flimited to decision-making in the management
4 o o pests of cotton, many of the same concepts
Kn to other pests such as weeds, plant dis-
I atodes, and arthropod pests of other com-

 Analysis in Decision-Making

pt s analysis and computer models can improve
‘ement systems and can unify and guide re-
g "Consider the ecosystem” is the first princi-
 control, around which all other principles
i  The key to the term ‘agroecosystem’ is the
F eco (also found in ecology, economics, and
 which is derived from the Greek oikos,
, use or household (161). In current usage, eco
if- wisdom and authority to manage, that is,
 ‘ng and decision-following in the best inter-

 or and the concepts of “action and inaction.

est of the household. In ‘agroecosystem’ the idea of
orderly household is expanded to include the managed
environment (161). Ecosystems are self-sufficient habi-
tats where living organisms and the nonliving environ-
ment interact to exchange energy and matter in a con-
tinuing cycle (84).

The cotton ecosystem is a complex ecological unit.
The cotton field is part of an ecological system that
includes associated crop systems, pastures, woods,
streams, and more. The major components of the cotton
field include the plants, the soil and its biota, the
physical and chemical environment, pest species with
their natural mortality factors including disease and na-
tive enemies, arthropod competitors for food and space,
and overall conditioning of man including his manage-
ment of the system 

However, the cotton agroecosystem (compared to
other natural systems) is a simplified ecosystem fre-
quently interrupted and prevented from undergoing
natural succession that would lead toward the climax
state. It is often a high-technology, capital-intensive
agricultural production system. In developed countries
the cotton agroecosystem often requires a continuing
energy input of fertilizer, herbicide, fungicide, de-
foliant, insecticide, and fuel for farm machinery 

Many of the components of any agroecosystem are
interdependent, so that management decisions affecting
one component have side effects that may have a detri-
mental relationship with other components. An im-
proved understanding of these side effects should be
very useful in developing more rational pest manage-
ment tactics and decision-making techniques for the
future. The warning that “we can never do merely one
thing” in ecological systems (43) might well be remem-
bered when we attempt unilateral tactics against a single
pest.

A dioristic model of the cotton agroecosystem can
be used as a guide for future modeling efforts (7). The
master model can be divided into plant growth, pest
management, grower management, and technological
and harvest-gin modules, each containing a series of sub-
modules needed to calculate the desired results of the
model. Inputs detailing energy and environmental pa-
rameters and periodic updates of information on plant,
arthropod, and economic dynamics should produce in-
formation of value to the decision-making process.

In this bulletin, it is not possible to discuss all the
components of future pest management and decision-
making models. However, a synopsis of the relationship
of particular components to this concept of decision
models is attempted.

Ecological Disasters

The need for improved pest management decisions
is most obvious where poor decisions have resulted in
total economic and environmental disasters. A recur-
rent pattern of cotton production throughout the world
has been classified into six phases (20,119) as follows: (a)

3

 

the subsistence phase, which uses minimal technology;
(b) the exploitation phase in which the use of irrigation,
fertilizers, insecticides, and high-yield varieties is
begun; (c) the crisis phase in which the elements of the
exploitation phase are used more intensely; (d) the
disaster phase in which pests develop resistance to
insecticides and the cotton industry may collapse; (e) the
integrated control phase in which there is a de-
emphasis on chemical insecticides and a search for
optimal management tactics, if the industry survived
the disaster phase; and  the deterioration phase dur-
ing which integrated control is abandoned in favor of
short-term economic advantages that then lead back to
the disaster phase. .

Major insecticide-related disasters in cotton have
occurred in a number of areas including Mexico, the
Rio Grande Valley of Texas, the Imperial Valley of
California, the Cafiete Valley of Peru, Central Ameri-
ca, and the Ord River of Western Australia (145). The
mechanisms of incipient disaster have been clearly
detailed for the Ord River area (Figure 1) (50, 162,
163). As farmers strived for higher and higher yields
and profits by increasingly greater inputs of irrigation,
fertilizers, and insecticides, the trends up to 1970 sug-
gested few limits to yield and profit potential. However,
by 1972 the American cotton bollworm, H eliothis armig-
era (Huebner), developed resistance to DDT, tox-
aphene, and methyl parathion. Up to 125 kg of insec-
ticide per hectare failed to control the pest. Thus, in only
1O years, since the start of large-scale commercial cotton
production in the Ord, a total disaster resulted, largely
due to the inability to control the American cotton
bollworm with insecticides. Since no cotton is now being
produced in the Ord, it is an example of an ecosystem
that reached the disaster phase but has yet to evolve into
the integrated control phase.

" 300

YIELD OF l.lNT (Kg/Ha)

msscr CONTROL cosrs - 5°

20- - - - - - - 5Q
I963 64 65 66 67 68 69 70 71 I972

Figure 1. History of cotton production in the Ord River
of Western Australia.

  
   
   
    
  
  
  
  
  
  
  
   
  
  
   
   
   
 
  
   
  
   
  
   
  
  
   
   
  
   
 
   
   
    
  

Van den Bosch (145) paints a rather bleak =2
the cotton agroecosystem as he contends tha
today is one of the world’s most ‘bugged’ c,
timized by an ecological backlash to heavy i 
drenching. The sad state of the cotton ecosyst
out as an example of the worst in pest control. k,
use of pesticides has created an entomologi‘
mare, bringing in its wake economic ruin, hum
and death, and gross environmental pollution",

He adds (144) “virtually no ecological cons
has gone into the development and use of i,
insecticides. In some places disregard for insect)
has resulted in a zooicidal overload which not s;
target pests but also decimates populations of - 1'
species including natural enemies of the pest
Elimination of these natural enemies creates A
vacuum that promotes the resurgence of the tar
and the eruption of previously innocuous speci

Although broad spectrum, chemical ins,
have repeatedly been implicated in these agroe -
disasters, they are not exclusively at fault. Man
elements of the technologically intense cotton '
tion system have contributed to these disasters. _
these contributing elements may include m0n_
production systems and excessive use of fertilize
gation, herbicides, fungicides, and potentially
yielding crop varieties.

é

l‘

Action Levels

The mere sight of a boll weevil, Anthonomus
dis Boheman, a cotton fleahopper, Pseudato f
seriatus (Renter), or bollworm Heliothis zea (Bi
may automatically trigger decisive action on the t
some growers and other decision makers. The --
action is often the massive use of insecticides, w 
needed or not (144). The economic threshold, v
visioned in the classical paper of Stern et al (138)-
developed as a partial solution to the problem 0f f
ranted use of chemical insecticides. Figure 2 depi
elements of the system as modified for H eliothis sp
The relationship between the general equilibrium, f?
tion (GEP), the action level (AL), and the intolerabl
level (ILL) is evident over the ten year period. The)
should be based on population models predicting
pests at the AL will reach the ILL and not 0n l_
assumption that pest populations always increas
abundance. Resurgence of Heliothis spp. after i g
ticide application results in a modified average de
(MAD) that may increase, depending largely on int‘,
ty of insecticide use. With the termination of insecti
use, the pest density may return to the GEP. Note,“ l
in the absence of insecticides the density of Heli’
spp. fluctuates about the GEP and only occasio
reaches the ILL. However, once insecticides are u r ,
the ILL is reached more frequently. s

The GEP was defined (138) as “the average den
of a population over a period of time (usually lengthy
the absence of permanent environmental change." t“
economic-injury level was defined as “the lowest pop a

2i

at g
)d

ll 
hi

TTURAL CHEMICAL NATURAL
(VNTRUL CONTROL cumrnor

ILL
In‘ ,1 l"l
I I j
I 1111*" l‘. ,". f“
l | I ‘I '
‘ |I-|—L|| IJ-FIH‘ f‘ |
\ ‘j-Ll ‘v  I_|_|_|\ '
I I I | | n '
%H—‘k--_‘|-‘q_fi
t \| I U I: n II
I ‘l u I
‘I ll u I'
\\| v v \|
J l:
a b

.””23456789l0

YEARS

i.
hematic of change in the modified average
ity 0f Heliothis species due to the use of
ecticides: (MAD = modified average den-

 ;lLL = intolerable loss level; GEP = gen-
ial equilibrium position; AL =
tion level) Modified from Stern et al. (138).

"that will cause economic damage” and the

' eshold as “the density at which control
 uld be taken to prevent an increasing pest
‘lbom reaching the economic-injury level.”
‘idefinitions of the economic threshold have
(35, 37, 45, 46, 84) that somewhat modify
of Stern et al (138). Most of these defini-
jor attention to economics, and relatively
4 n t0 ecological consideration. Thus, the
i mic threshold” and “economic injury level”
ed by the more inclusive terms “action
iintolerable loss level,” respectively. Also,
A has been paid to the economics of action
fthan insecticides.
'n threshold is another related concept
gas “the level of pest population at which
_jbe taken to prevent the population from
‘economic threshold where significant dam-

p visual threshold was defined (84) as “the
. el at which individuals of the pest species
'3 Since it is difficult to place a definitive
be on the aesthetics of ornamental plants,
7injury level concept was introduced (93).
'c thresholds have been divided into prag-
 ased 0n experience and field observations

the desired results and, definitive types,

_ fully planned and executed experiments
{there has been a proliferation of terms that
V“ onfusion. To avoid much of the confusion,
 n level” is suggested and is defined as the

number of pests and/or injury level at which supplemen-
tary action tactics will optimize profits and minimize
detrimental side effects. Ideally, action tactics used will
not eliminate natural enemies, which will provide sup-
plementary pest suppression in addition to that provided
by the actions taken.

Methods used to establish economic thresholds
have been summarized (68, 104, 136), as have assess-
ments of plant damage and potential crop loss (60, 118).
The empirical evidence, i.e. replicated observations and
the experience of specialists, can be used to establish a
provisional economic threshold (137). This same evi-
dence can be used in the calculation of provisional action
levels; these can be tied to a sampling method such as
sequential sampling and can be evaluated in practical
pest management programs (129, 135). Improvement in
profits, yields, environmental stability, reduced risks,
less pollution, and conserved energy will indicate the
value of the provisional action levels. When necessary,
these action levels can be modified by a definitive empir-
ical approach. The adjustment of economic thresholds
should be a continuing endeavor (35).

The Cotton Crop as an Example

The focus of any crop protection program must be
the crop, and thus a crop growth and developmental
model should be the central feature of any systems
approach to crop protection (115). Cotton crop models
have been developed (10, 38); SIMCOT II, a single plant
model (81) was modified to make it more useful in pest
management modeling efforts (38).

Action or Action Level Models

Improved cotton crop models can be employed in
predicting action levels (i.e. action level models) and in
determining the need to take action (i.e. action models).
In the evolution of these models, it would intuitively
appear that “action level” models will precede “action”
models, since action level models can be verified by field
decision sampling and can be of value in operational pest
management programs as the models are improved. Our
ultimate objective should be to develop “action” models
where pest management decisions are made without the
need for field decision sampling to verify computer
model predictions. The requirement for precision and
reliability in “action” models will be much greater than
in “action level” models, since the computer makes
decisions based on action models, whereas the field
scout or farmer makes decisions using action levels.
Thus, errors made by action models are less likely to be
detected before actions are taken. But, the modules or
sub-routines used in the “action level” models should be
useful in the development of the “action” models.

Plant Compensation
Until the sixteenth century, cotton was grown as a

5

perennial (57). Since the seventeenth century it has
been grown as an annual crop, though it retains its
perennial growth capacity (109). Thus, due to its peren-
nial, indeterminant growth, the plant is often able to
compensate for fruit and leaves lost to pests (109). The
cotton plant normally produces an excess of vegetative
and reproductive parts. For a variety of reasons, the
plant is often unable to produce sufficient photosynthate
to mature all the fruit that the plant attempts to set.
Thus, fruit loss attributed to insect damage might have
occurred naturally as a result of plant stress. Fruit that
can be lost to pests without affecting yields or profits is
referred to as “surplus fruit” (121).

In spite of a wealth of information on cotton plant
compensation, there is still a widespread inability among
growers and professional agriculturists to distinguish

A between actual crop damage and economic damage (35).

Often, any perceptible crop damage is viewed as reflect-
ing losses of yield or quality. The amount of fruit re-
tained and its contribution to harvestable yield should be
of greater concern than fruit damage 

Some pest damage to the cotton plant may actually
be beneficial (26, 40, 42, 109). Figure 3 depicts a
schematic of the effects of various pest densities on yield.
A small amount of insect feeding may stimulate the
plant, resulting in increased yields. Greater pest den-
sities may result in either plant compensation,
noneconomic loss, or economic loss (147).

   

    
    

     
  

YIELD
STIMULATION
COMPENSATION

ECONOMIC

 

NON-ECONOM IC LOSS

   

|
|
|
|
|
|
|
|
I
|
|
h
LOG INSECT DENSITY

 

Figure 3. Schematic of crop yield and pest density (after
van Emden 147).

Cultivated Acala cotton nomally sheds more than 70
percent of the fruit because the plants cannot meet their
carbohydrate demand (40, 81). However, after bolls
have reached their maximum growth rate the plant will
retain them and shed smaller fruit during periods of
carbohydrate stress. Bolls 12-14 days old are usually safe
from damage by the boll weevil (95).

The size of fruit when damaged will affect the speed
with which it can be replaced. For example, a square (a
flower bud) can be replaced much more rapidly than a
boll. A mature boll requires >2000 degree days (D°)

6

  
 
 
   
 
  
  
 
 
  
  
 
  
 
  
  
 
 
  
  
  
 
  
 
  
  
 
  
 
  
  
 
 
  
  
 
  
 
  
  
  
 
  
  
  
 
  
 
 
 
 
  
 
  
  

>53.5°F, whereas a mature square requires
D° (41). Thus, the older the fruit when u}
greater the time required for replacement an
er the probability of loss. Also, the cotton p »
experience sufficient heat during late se If
maturation of replacement fruit. i

Plant size is also an important conside 
cially in stripper cotton production. (Strip =4
mechanically harvested by stripping the lin I
from the plant along with burrs, leaves, 
whereas mechanical pickers attempt to rem I
lint and seeds.) Plant compensatory growth :0
or leaf loss by pests may result in an undes,
plant that is difficult to harvest by a stripper-
relationship between fruit loss and plant size?’
modeled in areas where stripper-harvested o‘
produced. j

Cotton plants on which pests have fed m) .
sate for damage by using the nutrients that L,
been used by the damaged plant parts to de
leaves, fruit, stems, and roots and to increase =
the remaining fruit (10). Prediction of the pl »
to compensate at any point in time during th
season should be an important component of, .
level model. If the plant has reserves of nu _ g
sufficient time remaining during the growingi a‘
replace damaged fruit, then a higher action I 
be set. The effect of leaf damage as rela 
phenological stage of cotton plant growth is ill!
Figure 4. Removal of more than 25 percent
from the cotton plant during the boll growth ‘v

10!

I18
tt‘

. - eq
result in some yield reduction, whereas 100 0-! M
removal during the lint-ripening stage may  I, t]
yield loss. a  en

Although compensation provides us with p
of latitude in pest management decision-m w
are also certain problems that must be address
we can readily accept pest damage and 
compensation in cotton production. One probl
as the cotton plant begins compensatory gro t‘ l¢
boring plants increasingly shade each other. i
ing reduces the amount of photosynthate prod I
Dependence on compensation may also delay r I
and thus expose the crop to more generation's;
and bad weather. Increased shading results w’,
drying and a greater incidence of disease, su i‘
rot. In gulf coast areas of the United States, thei
ity of heavy rains increases during Septemli
Therefore, cotton not harvested before Septe
have rain-related harvest difficulties, and the,
continue to grow and be damaged by pests.
insect pests of cotton enter diapause at high rat
September and October. Thus, a delayed h I
increase the numbers of overwintering pests 
invade future crops. 

Plant injury by insect feeding provides a i.‘
entry for pathogens so that, though the plant m)
pensate for the injury, it may become diseased‘
less vigorous. Also, plants suffering from othe?
factors such as moisture or nutrient extremes I
slower to compensate than healthier plants.

 lb

_'or

  

VEGETATIVE

 
  
  
   
  
   
  
   
   
 
   
  
    
     
   
  

PHENOLUG ICAL STAGE

 . Leaf loss causing no yield reductions = sur-
 plus leaf area (after Rahman, 109).
s
(in chemicals such as fertilizers, insecticides,
,1, nematocides, or herbicides are used in the
'1» cosystem, the quality, chemical ingredients,
 of application, and amount applied will influ-
igrowth and development of plants, animals, or
i . The use of pesticides may increase the nitro-
_  in treated plants and may lead to increased
aphid abundance (92). Thus, the direct action
 s may influence the role of plant compensa-

if nology

; m for a crop such as cotton is that pests of
i,- it cannot cause damage until these plant
ivailable. For example, the boll weevil may
, -fruiting cotton fields, but effective coloniza-
toccur until squares are available for feeding
jtion (160). Preliminary models of boll weevil
“(dynamics provided relatively poor predic-
lithe dynamics of the weevil was tied to the
‘of the cotton plant. When the weevil model
at the time the cotton crop first began to set
f wn squares, predictions of boll weevil dy-
ved considerably over what could be pre-
-~ perature alone (15). Logically, the coloni-
.~ n by both pests and the natural enemies of
fw be related to plant phenology. Sub-
ant phenology will be useful in action level
.

f Resistance
plant resistance to pests offers an ideal

method of suppressing insect pests because most resis-
tant characters are expressed under all weather condi-
tions, while some weather may hamper other tactics
(72). Maxwell (76) reviewed the major characters impart-
ing resistance for several pests of cotton.

Cultivars of cotton may have three types of resis-
tance: tolerance, nonpreference, and antibiosis. With
tolerance the GEP of a pest may not be changed but the
action level is raised. However, with nonpreference or
antibiosis, the ability of the pest to reproduce is reduced
and therefore the GEP is lowered (138).

Many varieties of cotton have been bred for pest
resistance under a chemical insecticide umbrella. As a
result, by breeding resistance characters into these cul-
tivars, unknown natural plant defense mechanisms may
have been unintentionally selected against and lost from
the gene pool. Resistance mechanisms might be more
fairly evaluated in the absence of insecticides. Resistant
cultivars will undoubtedly affect the prediction of the
action level.

Durational Stability of the Crop

Cotton as a crop may be considered to have low
durational stability since it is usually destroyed at the
end of the season and must be replanted at the begin-
ning of each growing season (with the exception of ratoon
[stub] and perennial cottons). However, long-season,
indeterminant varieties have provided a durational sta-
bility to this crop in relation to the survival of the boll
weevil (Figure 5). Use of long-season varieties has in-
creased the synchrony of boll weevil and crop phenology
so that the need for emergency action tactics, such as
insecticidal use, has been increased. With the long-
season crop, the spring suicidal emergence is minimized
and production of diapausing individuals is maximized.
The use of short-season, determinate varieties has les-
sened the durational crop stability. A short-season crop
enables the pest manager to optimize planting dates—
that is, to increase spring suicidal mortality while pro-
ducing fewer diapausing individuals in the fall. The
value of the short-season, minimal input system has
been discussed in detail in other reports(12, 32, 97, 127,
151, 153, 160).

The short-season approach also provides reduced
vulnerability to Heliothis spp. (151) and the pink boll-
worm, Pectinophora gossypiella (Saunders) (3, 164).

Other Factors

In a model of the cotton crop for use in the calcula-
tion of action levels, other factors must be considered.
Fertilizer can speed up plant development (74) if used in
optimal amounts. However, crop production systems
using excessive amounts of fertilizer and irrigation can
prolong growth and delay harvest (50). The importance
of optimal amounts of nitrogen and water in minimizing
damage from lepidopterous pests has been demon-
strated (97).

Soil types may affect the action level through their
effect on plant growth and development. The economic

7

OVERWINTERED
WEEVIL
EMERGENCE

CUMULATIVE
COTTON WEEV“

FRU|T|NG D|APAUSE
PROHLE

 
 
 
 
 
 
   

SHORT
SEASON
SUICIDAL
EMERGENCE

NUMBERS —>

LONG
SEASON
SUICIDAL
EMERGENCE

Figure 5. Schematic 0f boll weevil phenology in short
and long-season cotton production systems.

returns due to boll weevil control were related t0 soil
type and the soil’s effect 0n plant growth (113). Plants
growing on black soils are excellent producers of plant
nectar; alluvial soils result in good nectar production
most years; grey or red soils result in good nectar
production only under favorable conditions; and sandy
soils only occasionally result in nectar production (30).
Nectar flow of the cotton plant influences the abundance
of red imported fire ants, Solenopsis inoicta Buren, (4) so
soil type may indirectly influence the abundance of the
fire ants. These ants are key predators of Heliothis spp.
(78-80) and the boll weevil (58, 130). However, nec-
tariless cotton is less attractive to some insects pests (76).
Thus, the abundance of both phytophages and en-
tomophages may be affected by nectar production of
cotton on various soil types.

Weather factors generally used in modeling the
cotton plant include daily rainfall, maximum-minimum
temperature, solar radiation, and pan evaporation,
which is an index of humidity (10). Most of the yearly
fluctuations in cotton yield are caused by weather pat-
terns and not by insects (41). Cotton farmers tend to
overestimate yield potentials because they remember
high production years without remembering the as-
sociated weather. In years of lesser yields, insects are
often blamed instead of the weather. Weather patterns
based on historical climatological data are currently be-
ing used as the best estimate of future weather in insect
predictive models (44, 133).

Secondary plant substances have evolved as de-
fenses against enemies of plants (107). These products
are often toxic or repellent to animals and other plants.
The term “allelopathy” is used to describe the biological
inhibition of feeding or toxicity to plants or animals by
the secondary plant substances such as gossypol or tan-
nins in cotton plants.

8

 
  
 
  
 
 
  
 
 
  
 
 
  
  
 
  
 
  
 
  
  
 
  
 
  
  
  
  
  
  
 
 
 
  
  
 
  
 
 
  
 
 
 
 
 
  
 
  
 
 
  
 
 
  

Crop life tables can provide a firm foun)
the analysis of pest damage and cost/benefit.
management (71) and should be useful for cotto
associated arthropods.

Arthropod Life Systems

Arthropod Densities

The primary thrust in decision-making for 
ted pest management programs has been the 
nation of pest densities and pest dynamics. The;
ics of both phytophages and entomophages if
need to be understood if reliable action levels 0r 
levels are to be calculated. Whether these poi
densities are increasing or decreasing will grea '
ence the pest management decision. A pest pop
at the action level should require no manage _
tions if pest abundance is declining naturally, es n,
if key entomophages are present in numbers a:
inaction level. Thus, the ability to predict the pop‘
dynamics will be critical in future pest mana‘
systems. I

Simulation models have been developed for _
weevil (15, 59, 154), Heliothis spp. (44, 96, 140),’_
spp. (41), and the cotton fleahopper (133). These  _
simulate the dynamics of these insects through I
their seasonal cycles, usually during the cotton 9 
season. 1

Pest management decisions made for one 
may affect future seasons (54). In cotton this con
very important because both the boll weevil - 
pink bollworm can often be managed by tactics.
during one year that effect pest numbers in future
Both insects have few host crops in he United 
other than cotton on which they can rapidly repr l‘
Thus, management tactics designed to minimize
winter survival have been very effective. Destruc-
the cotton crop before these insects enter diapa
the fall is a key management tactic in some areas. s
case of the boll weevil, supplementary suppressiv
tions in the fall using insecticides have been an acc
and effective tactic (116). 
Pest management systems will benefit from of
that consider the dynamics of insects throughout-
year, and from year to year. Winter mortality c
especially critical for the suppression of some in,
below the action level during the following season.

Dispersal g

The importance of dispersal in pest manage
systems has been reviewed by Babb and r
(108,148). Single night flights of Heliothis spp. up
miles and 4-day movement of 45 miles have been
served (126). Computer simulation models of Hel'
spp. (44, 140) include the dispersal of moths be
crops as a function of crop attraction. p.

Predictions of pest and entomophage dispersal
crop colonization will be of great value in develo
future pest management models. '

   
   
  
  
   
  
   
  
   
  
   
 
  
  
  
    
 
 
 
   
  
   
  
 
   
    
   
  
   
 
  

gr rsion

_ distinction can be made between macro-
frsion (the geographical distribution of arthropods)
Fimicro-dispersion (the distribution of arthropods
I a limited area such as a field or plant). The macro-
rsion may affect the importance of a pest, since it
7 sually not occur in a uniform density throughout its
 A key pest for one area may be only sporadic or
 ous in other areas (134). One of the attributes of a
l ‘ st is that it should have a high historical probabili-
l current economic damage over many years. Once
gjhistorical probabilities have been determined,
hould be very useful in calculations of probabilities
nomic loss for use in action levels. However, the
‘ility of historical probabilities would be question-
g based on the frequency of years when insecticides
(fused for pest control unless supplemented by
sampling of pest densities.

A e macro-dispersion of H. zea throughout the USA
Er iewed by Snow 8: Copeland (122). The micro-
{lion of cotton arthropods has received consider-
ttention because it is important in developing
ed research or pest management sampling tech-
f‘ (63). Also, “spatial aspects of populations are
Lt t0 model, and it is very rare to find both spatial
 poral dynamics included in a single model (114).

0', Size and Sex

iditionally, emergency actions taken for H eliothis
 e been designed to prevent the development of
7 mbers of large larvae. These large larvae cause

ae remove larger fruit that require a longer
j ent time than smaller fruit. Removal of large
_= results in a greater loss of plant photosynthate
oval of smaller fruit.

0' and sex of Lygus hesperus Knight may also
j§lerent amounts of damage. Adult females cause
 ately twice as much damage as males, whereas
ge for third to fifth instar nymphs is near that of

1|
.
.
.

neous Damage by a Complex of Pests

in damage caused by two or more pest species
T synergistic, or antagonistic? What should be
.en the crop is infested by species A, B, C, and
 ofwhich has reached the action level but each of
y be within one-half to three-fourths of it (35)?
l‘: culating an action level for a single pest when a
30f pests is present, a cotton plant model may be
‘ for calculating the cumulative damage from all
A. that can be tolerated and the amount of fruit
ébe expected to survive the attack. When two
as the boll weevil and Heliothis spp. feed on
wthe cumulative damage might possibly be less
 ditive damage of either alone, since H eliothis
A nsume weevil-infested squares, which results
l ‘ eous damage to some squares. Definitive data
bject isnot available but one simulation study
F at reduction in yield from simultaneous dam-

l’ ore fruit damage than small larvae (143). Also, A

age by two pests was larger than the sum of the yield
losses when the damage was caused by each species
separately (77). The interactions of the boll weevil and
the bollworm have been considered in establishing an
economic threshold based on the ratio of fruit production
to the rate of cumulative fruit damage caused by both
pests (38).

Each species may not affect the action level of the
other species if each species feeds on a different plant
part and if their cumulative plant damage does not result
in plant stress. Also, the decision to take action against a
group of pests, all below the action level, is complicated
by the need for different action tactics for each pest
species (134). This helps account for the popularity of
broad spectrum insecticides among cotton farmers. Also,
the presence of the tobacco budworm, Heliothis vires-
cens (Fabricius), in the H eliothis complex will dictate the
need for more toxic insecticides or higher dosages, since
H. virescens is generally more resistant to many insec-
ticides than H. zea (88).

Action Levels for Cotton Pests

The action levels for Heliothis spp. on several crops
and alternate host plants have recently been reviewed
(131). Four chapters in SCSB 231 relate to Heliothis spp.
on cotton (10, 99, 104, 151). Graham et al (37) also
reviewed pioneering research leading to the establish-
ment of economic thresholds of ca. 0.5-0.6. larvae per
meter (2, 111). These thresholds provided guidance for
making pest management decisions for Heliothis spp.,
but many modifications have been made in the various
cotton-growing states. A major improvement was the
recognition that the threshold must be dynamic during
the growing season (47, 54). For example, after broad
spectrum insecticides have been applied, few en-
tomophages will survive, which often results in a pest
resurgence. Thus, the economic threshold decreases
after insecticide applications have been initiated and/or
as the season progresses  The need for dynamic
thresholds was demonstrated by determining that rela-
tively heavy infestations in mid- and late-season were
required to reduce yield and that 1.2 larvae/m through-
out the season significantly reduced yields (61). Cotton
was able to tolerate seasonal densities of ca. 0.6 lar-
vae/m. In Texas a seasonal average of 0.9 larvae/m
significantly reduced yields 

Action levels are reviewed for other pests of cotton
such as the boll weevil (134), Lygus spp. (41, 137), pink
bollworm (32), and the cotton leafworm, Alabama argil-
lacea (Huebner) (31). Current pragmatic action levels for
cotton arthropod pests may usually be obtained from
local Extension entomologists.

Soils and Fertility

The effect of boll weevil control on yields is related
to soil type and soil fertility (113). Also, fertile soils such
as river “bottoms” may force determinant cotton
varieties into indeterminant growth patterns. Thus, ac-
tion levels established on soils of a certain level of
fertility may not be applicable for soils with other levels
of fertility (134). Infertile soils with a low yield potential,

9

, \- s-aa .1.“
i- iiia“iiii;.a

in the absence of pest control, may show only a neglig-
ible profit increase when control measures are applied
(14). Ecosystems enriched with plant nutrients often
result in the development of dominant pest species,
whereas systems with lower concentrations of plant nu-
trients have no or fewer dominant species, and thus the
the system is more stable (98, 128). Cotton fields that
receive moderate amounts of nutrients might be expect-
ed to have fewer dominant pests than fields that receive
heavy concentrations of plant nutrients.

Contamination of soils with pesticides, salts, and
even over use of fertilizers will have some influence on
the soil microflora, soil fauna, and their relation to soil
and plant health and may be important in the calculation
of action levels.

Insecticide Resistance

In modeling for predictive purposes, time frames of
greater than a single season should be considered. One
reason is that the gene pool of the various pest and
entomophage populations changes over time. This
change is apparent in the development of insecticide
resistance, but it may also be expected in other physio-
logical and behavioral adaptions to changes in cotton
production practices. Thus, if insecticides are to be used
as a pest management tactic in future years, predictive
models of resistance should be developed for key ar-
thropods of the cotton agroecosystem.

Prediction of susceptibility to insecticides will be
useful in decisions regarding choice and dosage of insec-
ticides. The total susceptibility of a particular species to
currently used pesticides has been referred to as “biolog-
ical capital” (54). A genetic model dealing with the
developmental rate of insecticide resistance suggests
means of minimizing selection pressure (105).

Insects may also develop resistance to other man-
agement tactics such as resistance varieties and preda-
tors (55). Since insects can develop a resistance to their
own hormones (149), it would not be surprising if they
could also develop resistance to autocidal methods,
pheromones, cultural controls, mechanical controls, host
plant resistance, and others. Thus, there is a need to
determine the risks of these events and the rate of
development.

Yield Losses

Definitive evaluation of crop losses is one of the
most complex problems facing the researcher. The liter-
ature of economic entomology is replete with examples
of yield increases resulting from the use of various pest
management tactics. Often these experiments have been
conducted in small-plot, randomized plot designs where
some replicated plots were treated with broad spectrum,
chemical insecticides while certain randomized plots
were left untreated as a control. The differences in yield
between the treated plots and the control were used as
an estimate of pest damage. Experiments of this type
have often demonstrated cotton yield “increases” of 100-
500 percent. However, the difference in yields is often
due to the destruction of natural enemies in the control
plots by the insecticides drifting from the treated plots.

10

 
  
   
  
   
    
   
  
  
   
  
  
  
  
  
  
  
   
  
 
   
 
  
  
   
  
   
   
 
   
    
  
   
 
  
  
   
  

Without their natural enemies, and with ins
amounts of drifted pesticide to kill the pests,‘
damage was realized from the pests. Also, these ‘p,
were and are still frequently being conducted i
munities where large amounts of insecticides ar
applied to neighboring cotton fields, upwind fr
test plots. 

The effects of patchy applications of insectici
community on the abundance or change in beh
natural enemies in the untreated areas is larg
known, but it would be foolish to assume no A
Though the randomized experimental method *-
fair evaluation of the relative efficacy of the _i
chemical insecticides in the experiment, it is n
evaluation of yield losses without insecticides:
result, yield losses due to insects tend to be i;
timated. . .

Estimates of crop losses to pests can only I
evaluated in areas where minimal insecticides .
sent in the soil, on the plant, or in the air and whe
have little chance of drifting into the area. Isol 
several miles from treated areas may be essen’
these loss estimates. Also, insecticides used over
areas may reduce the abundance of natural enem’
that 2-3 years may be necessary for the effective i
ery of these natural enemies (6, 87, 100.) Thus, r‘
estimates should be obtained in areas where insecy
have not been used for several years, so that th
effectiveness of natural enemies can be realized. In '
experiments, the optimal cultural and biological.
management tactics should be employed to m 
the probabilities of producing acceptable yields
profits in both the control and treatment plots. _

A distinction can be made between direct and '
rect damage (118). Direct damage results when the”
directly attacks the part utilized by man, whereas -
age to other plant parts results in indirect damag
cotton, higher levels of indirect damage can be tole
than direct damage. The key pests of cotton, i.e.
bugs, boll weevils, bollworms, and tobacco budwo;
all cause direct damage by attacking the fruit. 
secondary pests are generally leaf feeders. Howi
this distinction is simplistic because those pests ._
cause direct damage may also cause indirect damage?
the key pests of cotton will occasionally cause dama
plant parts other than the fruit. For example, Hel’
spp. feeding on pre-fruiting cotton terminals can d _
plant growth (52). .

Despite increased use of insecticides in the Un
States, crop losses continue to increase. The chang
our agricultural production system that have contrib
to this increased crop loss are (a) use of crop vari _ ‘
susceptible to pests, (b) continuous culture of ce ,
crops with less rotation and diversification, (c) red l_ l A
crop sanitation, (d) reduced tillage, (e) growth of cro
areas where they are more susceptible to pest attack
increased pesticide resistance in pests, (g) destructiol
natural enemies of pests. (h) use of pesticides that . 7
the physiology of plants, and (i) reduced tolerance by '
Food and Drug Administration, and increased cosm
standards by processors and retailers for crops (102).

    
    
   
  
   
 
  
   
   
     
  
   
   
   
  
   
  
   
   
    
  
 
  
  
 
   
   
   
   
  

p losses due to insect pests have increased from
int in 1904 t0 a present high of 13 percent (101).
’"these increased loss trends are due t0 high
c standards” in fruits and vegetables now sold

n- j _ y
Hg. JUS market. They also result from planting
he "that are more susceptible t0 pests and eliminat-

"rotations and other sound ecological practices
i» culture.

= etic standards are of little or no importance in
i; ept for the spotting of the lint that may result
f t feeding in the green boll. For example,
“y the green stink bug, Nezara oiridula (L.), or
‘bollworm can cause reductions in lint quality.
', feeding by Heliothis spp. or boll weevil does
7: - ly reduce lint quality, because if only a small
 the boll is damaged, the lint from the damaged
w‘ not harvested (152).

Inaction Levels

 ‘gement decisions should be based not only on
'0 =1 ce of pest species but also on the abundance
lenemies of the pest. Obviously, greater num-
‘ts can be tolerated if an abundance of effective
 emies are present. Thus, the abundance of
emies will affect the action level. However,
liance can be placed exclusively on natural
‘a working understanding of the most efficient
l" d the numbers needed to maintain pests
Zaction level is needed. The presence of great-
 s of a complex of predators than of pests does
 the maintenance of pests below economic
_  The term inaction level for the density of
(emies sufficient to maintain the pests below
(level is suggested (29). McDaniel 8t Sterling
'_d an example of an inaction level. They
,¢ ratio of one key predator for each Heliothis
lqWhen 15-20 percent of the terminal buds of
a 'n a predaceous black fleahopper, Rhinacloa
a Renter, 80-100 percent of H eliothis spp. eggs
ilsumed (53). No chemical insecticides were
gded if “beneficials” were present at 20 or
 0 row feet of terminals  Inaction levels for
 tor of the boll weevil have been established
 idence supporting these levels is available
Qlnaction levels have not been established for
A mies of most insects pests.

jators

‘B edator species or stage of a predator species
f? s predictive value for forecasting future prey
1 trends and is capable of providing irreplace-
i. ensable, sensu Southwood [125], p. 374)
V, ading to prey population regulation may be
a key predator. Irreplaceable mortality is
 of the total generation mortality contributed
Q-lar agent and not replaced by other natural
= us, removal of an agent providing irreplace-
"ty would result in greater survival of the
_ s. The concept of the key predator should be

distinguished from the concept of an index predator. An
index predator provides value in forecasting future prey
population trends but may or may not regulate the prey
population. For example, if the numbers of prey are
highly inversely correlated with the abundance of a
predator species, then the predator may be included as
an index predator. If later evidence supports the conclu-
sion that the index predator is causally related to the
decline of prey numbers and actually provides irreplace-
able mortality, then it is clearly a key predator.

Thus, an index predator may also be a key predator,
but the difference is that there is no field evidence that
the index predator causes any irreplaceable mortality.
An index predator may be an excellent predictor of the
numbers of the real, but unknown, key predators. A
hypothetical example of an index predator might be a
ladybeetle that does not feed on a bollworm egg, but
numbers of the ladybeetle increase at the same time as
the real key predators that actually regulate the boll-
worm eggs. In this case the abundance of the index
predator might be correlated inversely with the abun-
dance of the bollworm eggs, but the correlation is
spurious.

A key predator might be expected to fit a modified
definition of a key factor as proposed by Solomon (123),
i.e. one of the main controlling factors affecting a popula-
tion, implying that key factors cause the mortality but is
not spuriously correlated with it. An index predator
would more closely fit Morris’s (83) definition of a key
factor, i.e. any biological or environmental condition
associated with mortality that is useful in predicting a
future trend in a population but lacks evidence of causa-
tion. Spurious correlations may have predictive value
even though changes in one may not cause the changes
in the other.

More than 600 species of predators have been ob-
served in Arkansas cotton fields (159), but only a few
species are responsible for most of the predation. In field
studies, the order of importance for Heliothis spp. eggs
was as follows: ants, lady beetle larvae, adults of the
spotted ladybeetle (Coleomegilla maculata DeGeer),
adults of Hippodamia conoergens Guerin-Meneville,
spiders, predaceous mites, and green lacewing larvae

(Chrysopa spp.). In later studies (158) Orius spp. and ‘

Geocoris spp. adults were added to the list. Similar
results were obtained in east Texas using 32F labeled
Heliothis spp. eggs (78, 79). Dominant egg predators
were ants (mostly Solenopsis inoicta), Orius insidiosus
(Say), Geocoris spp., H. convergens, Cycloneda san-
guinea (Linnaeus), C. maculata, P. seriatus, and the
spiders C hiracanthium inclusum (Hentz), Phidippus au-
dax (Hentz), and Oxyopes salticus (Hentz),. The average
seasonal egg mortality due to predators was 93 percent
after 72 hrs. Many of these same species are important
larval predators of H eliothis spp. with the notable excep-
tion of the large lady beetles (80). As a result of this high
level of predation, Heliothis spp. seldom caused
economic damage except where insecticidal perturba-
tions had occurred.

11

The efficiency of many of these predators when
reared 0r field collected and released has been evaluated
(69, 70, 110, 111, 146). A maximum of 9O to 99.5 percent
reduction of Heliothis spp. was observed as a result of
these releases, the efficiency depending 0n the numbers
of predators released. A review of inundative releases
(139) is a valuable reference 0n this subject.

N0 more than a dozen key natural enemies of pests
are of importance on cotton (112). The species and their
importance may vary between geographical areas. Thus,
their efficiency should be evaluated in many cotton
production areas and for all key and secondary pests.

Of course, predators are not the only natural
enemies of cotton pests. Parasites and pathogens may
also play a major role in the regulation of pest abun-
dance. The identification and relative importance of
Heliothis spp. parasites has recently been reviewed
(112), and a manual is provided for pathogen identifica-
tion (106).

Colonization

To predict the colonization of entomophages and
pathogens in cotton, their sources should be known. For
example, a major source of predators in south Texas is
grasses (34). Certain entomophagous species exhibit su-
perior dispersal-colonizing abilities and are characteristi-
cally dominant in disturbed habitats such as agroecosys-
tems (22). However, there is usually a trade-off between
the reproductive-dispersal abilities of a species and their
competitive abilities. Species exhibiting high reproduc-
tive-dispersal abilities have been labelled r-strategists
and those with high competitive abilities, k-strategists
(73). Ehler & Miller (22) contend that r-pests may be
controlled by r-strategist natural enemies in disturbed
systems and in habitats of low durational stability.

Since the cotton field is essentially a pauperized,
ecological desert before the crop is planted, most en-
tomophages must disperse into and colonize the field
before they can effect control of pests. To predict the
efficiency of natural enemies in the management of
pests, a knowledge of their dispersal and colonization
rates may be essential.

Switching

Polyphagous predators may switch from feeding on
a key pest species to secondary pests or innocuous
species, resulting in greater survival rates of the key
pest. In field cages, bollworm eggs survived at a rate 16
times greater when Orius spp. and aphids were both
present as compared to eggs alone (25). However,
noneconomic densities of aphids, thrips, spider mites,
lepidopterous eggs and larvae, and innocuous species

attract, hold, and increase predator and parasite abun-
dance (89).

Models

Preliminary models of natural enemies that could
be used in action level models are available (44, 142).

12

economics of alternate action tactics, which have Q

  
   
   
    
  
   
   
   
 
  
    
   
  
   
  
  
   
   
  
  
  
   
  
  
  
   
  
   

Economics

Mid-season chemical insecticide applications t
often necessitate subsequent treatments latef
season to protect the crop from resurgence of the
pest or an outbreak of a secondary pest (137). 
first insecticide application perturbs the co
roecosystem by causing a decrease in the effectiv;
natural enemies. Frequently, this perturbation r_
a virtual ecosystem addiction to the insecticides g]
that a “treadmill” of applications is essential to;
the crop. Each additional application of insectici
not ensure the protection of the crop at poten
present when the application was made. Muc
crop may remain susceptible to pest damage,“
very large losses may be suffered if the applicati
terminated. Economic analysis of cotton o?’
should definitely consider this “treadmill” effect.)

Insecticides

Of the pest management tactics available for.
sis by economists, pesticides have been univer 1'
tractive because they are “easier to cost” and 
put into practice than other methods (94). Howe“
nine separate, large-scale pest management pro =
Texas, an average 44.2 percent reduction in the -;
insecticides resulted (64) when greater dependen
placed on native entomophages for pest contro 
returns increased by an average of 77 percent. Th t

been neglected in the past, should receive a fairl
research priority in the future. "

Cost-Benefit Ratios

Cost-benefit ratios are of value because the,
easily understood by growers (35). In practical a 5
ture, a simple cost-benefit ratio is the best “first ap”
mation” of potential crop yield (137). The grower?
understand this ratio more easily than marketing flu
tions and commodity prices on a regional, nation
international basis. 1

When the concept of social cost/social benefit 
is introduced, the modeling problem becomes r
complex. The concept of social benefit (welfare) is v
and controversial (66). As in other areas of mode _
there is a need for definitive data in order to calc 
the costs and benefits to all individuals in the soci
rather than basing calculations on estimates or 
The impact of pest management decisions on the he 5,
wealth, and nutrition of individuals is one of the imj
tant elements that needs to be quantified. 

Production Costs I

Costs of contemporary pest management system
Texas have been reported (12, 32, 91, 97, 127). .
evaluating costs of pest management systems, prod
tion, management, sampling, set up, and application
tactics should not be overlooked (134).

  
   
  
   
   
  
 
 
 
 
   
  
   
  
   
 
   
    
  
   
   
  
 
  
   
 
   
    
    
  

 Losses

 due t0 cotton insect pests have been re-
f? (17, 39, 85). Care should be observed in evaluat-
estimates. If pest damage were eliminated, the
ue of the product would decline on a per unit
 cept in the case of price supports) because of the
p, 'elds and supplies that would result. Thus, loss
3w re often exaggerated. For example, boll weevil
0n benefits t0 the cotton producer are nebulous.
ers benefit from large-scale management pro-
 ch as boll weevil eradication because of lower
 'ces, but landowners lose because land values

f‘

 and Demand

‘rder to calculate an action level of value to the
 , some estimate of the short-term value of the
"Trequired. The supply-demand relationship of
 d’ its value has been reviewed (47). Information
 rrent status of world cotton production, US
lent of Agriculture (USDA) loan levels, and
f-y payments may be obtained from several
lsuch as the National Cotton Council, Cotton
 ted, and the Economic Research Service of the

Ecology And Society

,; ly sound management practices should also be
F ly sound. However, ecological measures such
(Id natural enemy conservation, or ecosystem
n, may be of economic value primarily over
'rm, which may extend to several human gen-
 awareness of the need to preserve non-
» resources and to leave the earth as healthy as
‘it is an ethical legacy that we must pass on to
kerations (92). In the case of ecologically sound
i optimal profits often may not coincide with
"jprofits. In the search for maximum profits,
Nd may be given to such concepts as conserva-
Ability of systems. Thus, boom or bust cycles
fnplace in cotton production. When the price
“. high every possible acre of farmland may be
=cotton. From the air the gullied and eroded
‘the cotton belt of the USA lie as a testament
_ re of this strategy. The top soils of much of
ertile land now lie in river beds and ocean
,1 s, this potentially rich inheritance of fertile
,asted and of no value to future generations
‘ers attempted to make quick profits, that are
f, to their offspring.
i, osion is not the only example of the deteriora-
 valuable natural resources. Reduction or
I of natural enemies of plant-feeding insects
»misuse of broad spectrum insecticides and
f soils with salts of irrigation waters, insec-
‘i lizers, and herbicides may have greatly af-
stability and long-term profitability of many
e history of cotton production around the
eplete with examples of financial disasters

."",‘ 9'. "..

resulting largely from a failure to understand certain
basic ecological and economic principles (145).

Side Effects of Management Actions

Economists refer to some of the side effects of
actions taken as “externalities,” or the benefits and costs
accruing to persons other than the producers or consum-
ers of the agricultural commodities involved (66). The
identification of externalities is nebulous (66) because
these secondary effects are not signalled to the decision-
maker either from the market or from the ecosystem and
are therefore not considered in his decision-making (67).
It is possible for externalities to cause a major divergence
between management strategies optimal for the grower
and those optimal for society. A decision model has been
developed to estimate the external costs and quantify the
costs of pesticide side effects (21).

According to N orgaard (90) the misuse of pest man-
agement actions is a social problem. He claims that social
objectives such as nourishment, health, and environ-
mental quality are not being met and that solutions to
this social problem lie in changing the behavior of men.
Thus, farmers need to use fewer and narrower spectrum
pesticides and biological controls consistent with the
dynamics of the agroecosystem. Or, farmers should plant
less vulnerable plants in better patterns and places at
better times.

Because of the externalities inherent in pesticides,
their use should be reduced to the extent that their
marginal social costs (including all side effects) approxi-
mate their marginal benefits (27). Marginal costs or
benefits are the total costs or benefits of cotton produc-
tion that result from each additional input such as each
additional application of insecticide.

Some of the side effects of pest management actions
have been summarized (92). These side effets relate
primarily to chemical pesticides and their detrimental
effects on the crop and nontarget organisms; they in-
clude long- and short-term effects on human and ecosys-
tem health. A review of the side effects of insecticides,
herbicides, and fungicides on some of the flora and fauna
of plant, soil, and aquatic ecosystems has been provided
(9)-

Management tactics such as cultural practices, hor-
monal regulation of plant growth, host plant resistance,
mechanical control, pheromonal control, pathogens,
classical biological control, and the augmentation and
conservation of natural enemies of plant pests are gener-
ally considered to have less detrimental side effects than
broad spectrum insecticides. This may be because man-
agement tactics other than insecticides have been poorly
evaluated for these side effects. For example, foreign
parasites introduced for the classical biological control of
pests are usually screened for hyperparasites but not for
new pathogens that could be a hazard to native natural
enemies of pests. However, even with pesticides, “in-
adequate data and the necessity of subjective evaluation
of damage to wildlife and threat of risks subvert detailed,
quantifiable, cost-benefit analsysis of overall liabilities
and benefits of pesticide use” (92).

13

Uncertainty in Pest Management

Uncertainty because of lack of information 0n the
part of the farmer is an additional factor encouraging
pesticide use. F eder (27) investigated the impact of
uncertainty on farmers’ decisions regarding pesticide use
and the way it affects reaction to various changes. Given
risk aversion on the decision-makers part, Feder’s im-
proved optimization model introduced random elements
into several components of the pest-pesticide-crop sys-
tem and used Bayesian decision rules and dynamic
programming to reproduce the farmefs decision-making
process. F eder (27) also evaluated the impact of im-
proved information regarding old and new technologies,
as well as information acquisition. The cost of informa-
tion and its effect on pesticide use was evaluated, and a
market for management information was established.

Insecticide Decomposition

A simulation model of the rate of insecticide loss
from a terrestrial ecosystem was developed (150) that
was based largely on soil surface temperature and mois-
ture fluctuations. This type of model could prove useful
in predicting the residual side effects of pesticide use.
Also, the model ecosystem approach, along with the use
of radiolabeled insecticides, is an informative and conve-
nient experimental technique for studying the environ-
mental effects of insecticides (82).

Policy Decisions

Some farm policies and practices have been coun-
terproductive. Nearly four billion dollars was spent by
the US government to retire productive land from culti-
vation (103). This land restriction encouraged high crop
yields on fewer acres, resulting in some cases in in-
creased reliance on pesticides. Thus, one trend in crop
production has been the use of more pesticides and less
land. Any program that indirectly encourages the sub-
stitution of pesticides for land should be critically ex-
amined (103).

A computer model predicted that future costs of
cotton production without insecticides would be slightly
greater than the cost of producing cotton with insec-
ticides in the absence of a government land retirement
program (103). With a land retirement program, the cost
of producing cotton would be greater with insecticide
use. Restrictions on land use appear to play a greater
role in price increases than do restrictions on insecticide
use.

The farmers income is often supported through a
guaranteed government price; thus his gross income is
proportional to his output. This is a form of insurance
that is of value to the farmer only if there is no crop
failure. Thus, the farmer often sees the use of chemical
pesticides as a means of reducing the risk of being unable
to take advantage of price supports. This type of policy
thus tends to subsidize the use of chemical pesticides
(92). Davidson & Norgaard (16) have suggested both
income or output insurance as a solution to this problem.

14

  
    
    
 
   
    
  
    
    
   
  
   
   
  
   
  
 
 
   
   
   
 
   
   
 
   
   
  
  
  
   
 
  

Marketing standards are determined by _"g
ment policy. High marketing standards dictated u
latory agencies may have little bearing on the nu
value of the product. Thus, for largely cosmetic _
action levels have been reduced to near zero, __
in a greater use of pesticides, higher productio
higher environmental costs, and an increased c0 if
final product (35). Consequently, the flame of i ' ,
which is often fueled by policy decisions and is 
ably predictable, should be considered in esta‘
action levels. f

The passage of the Occupational Safety and
Act, which requires a hazard-free environment f0,
ers, has transferred some of the costs of pestici _
effects from society to the farmer. Farmers may
legally liable for damage or contamination of nei
farms from spray drift (92). 

The federal government compensated bee ‘_
an average of $1.5 million per year in the early 1 if
for losses due to “pesticide accidents. ” “When the
picks up the tab for negligence, there is little in
to be careful” (92). 3

Thus, the effects of government policy art
counterproductive to the minimization of the def
tal side effects of pesticides. Policy decisions wi"
ably play an important role in the calculation 0 _,
levels in future pest management models. The 
the quantification of externalities resulting fro l»
ticide use as they might affect policy decisions is s 4
ed by Headley and others (48, 67). For an e‘
review of economic research on pesticides for 
decision making see the proceedings of the 197
posum of the Economic Research Service (11, 19)

Strategies and Tactics

Pest management decisions can be made at eithi
strategic or tactical level. A pest management (i;
a scientifically determined plan in which potential »
agement tactics and other pertinent contingencies
been optimized for the management of pests r
ecosystem [modified from Encyclopedia Brita a
(23)]. The three basic strategies are (a) prevent‘
eradication, (b) containment, and (c) doing nothing"
Other factors that should be considered in develop
strategy of cotton production include the selection
planting date, crop variety, and methods of weed 
trol, fertilization, irrigation, etc. These decisions MU,
made before the crop is planted and thus are part H
crop production strategy. 
Obviously, anyone who produces cotton has“
veloped a production strategy. However, one obje,
of optimized strategies may be to avoid the creati
new pest problems and to try to remedy those situal _
from which present-day problems have arisen 
Koenig and Tummala (62) contend that systems sci
techniques should be used to redesign agroecosyst ,
Crop production practices can be evaluated by syst
science to produce improved ecosystem strategies 
susceptible to pests. "

   
   
  
   
   
  
   
   
 
   
 
   
   
  
 
   
   
   
 
  
  
   
  
 
 
 
    
  
   
   
   
     
  
 
   
  

j ics were defined as the “specific methods 0r
l 'ng techniques required to carry out a basic
if’ (56). Pest management tactics include plant
f} and cultural, biological, mechanical, pesticid-
 idal, pheromonal and plant growth regulators (8,
. These tactics were categorized as follows: (a)
q- pesticides and (b) bioenvironmental controls
fioenvironmental control was defined as “any
in ical control method utilized to reduce pest
'ns by environmental manipulations and biolog-
trol" (102). Bioenvironmental controls are em-
lion more acres of crops (9%) than insecticidal
;.(6%) in the United States (102).

"g efficiency of a management tactic may affect the
l. of the action level. A tactic yielding 98 percent
J of a pest should generally have a higher action
 a tactic yielding only 50 percent reduction. If
(n level has erroneously been set too high, or
g fails to accurately detect the action level, a
 ective tactic can still be used to “clean up” and
“from these mistakes. However, a less effective
uld be of less value in recovering from decision
_ us, the tendency is to set the action level at a
tively low point (134).

T

 ers often appear to be risk-adverse. To lessen
wledge is needed not only of the expected
gewm for decisions but also of the variance of
 of possible outcomes. This information can be
lodels that yield not only the expected outcome
..the probability of all possible outcomes (18);
.4 until these predictive models are available,
 ods of minimizing decision risks will depend
Aditional experimental methods.

,production method that has shown much pro-
l educing the risk of pest damage to cotton is the
 on, reduced input systems that many of the
lg: Texas have been quick to adopt (151). Where
rtilized, indeterminant cottons are grown in
vestment risks may be greater than with the
 on systems.

,major risk in cotton insect pest management is
l will cause unacceptable losses to the crop.
I , there are many other risks associated with
, made in pest management. The risks of taking
ent actions when they are not needed and the
y ing no action when actions are needed (i.e.
 type II errors, respectively) must both be
1-: (135). Current concerns regarding the risk to
i alth due to agricultural chemical pesticides are
p‘ in legislative sanctions against their misuse.
4e risk appears to be to those who apply the
 (103). Thus, the farmer-applicator should have
cerns for the risk he takes in using certain

a =3 Y’ 5Y,"5F"~<‘.“".

I

l_ ity Strategies
_'nction can be made between local (producer
_» regional (community-wide) strategies (125).

Considering the vagility of both phytophages and en-
tomophages, it is small wonder that community-wide
strategies have some distinct advantages over individual
producer stategies. Community-wide strategies
minimize the problem of recolonization of pests from
fields not included in a management program to fields
included in such a program. However, in community-
wide insecticidal applications there is a high risk that
some fields will be treated where no treatment is
needed, unless each field is sampled separately.

Pesticidal Tactics

Of the currently available and effective tactics,
chemical insecticides remain an important component of
pest management systems. Thus, pest control models
such as the one by Talpaz and Borosh (141) are needed to
evaluate strategies for pesticide use. They applied a
mathematical-numerical optimization that selected fre-
quency of applications and dosages designed to minimize
control costs and crop damage. Watt (156) also used a
computer to evaluate alternative insecticidal programs.

Insecticides should only be used on an “as needed”
basis. Sampling is generally recommended to determine
the need for management actions. However, preventa-
tive actions have been suggested as a valid option if the
risk of economic losses is almost certain every year when
actions are not taken. In south Texas, the risk of unac-
ceptable boll weevil loss was so high every year that a
preventative strategy was incorporated into the pest
management program (51).

The use of insecticides where they are not needed
can be very expensive insurance. For example, control-
ling Lygus spp. with insecticides tended to reduce yields
of Acala cotton in simulation studies (41). Thus, “farmers
were spending money to lose money” (41). The use of
pesticides rarely increases yields; rather, use prevents
loss of yields (71).

Of all the chemicals used to produce cotton, insec-
ticides have the most serious side effects (92). About 80-
90 percent of human cancer may be caused by chemical
contamination of the environment and food ('75). Also,
there is evidence that aldrin (24) and DDT (49) cause
cancer and possible birth defects and genetic mutations.

The history of chemical pesticide use in a field or
community may indicate the current availability of natu-
ral control agents. Up to three years may be required for
a normal balance of predators to return after the use of
persistent insecticidal chemicals over broad areas (6).

Cultural Tactics

Some of the cultural tactics used in cotton pests
management include planting and harvest dates, tillage,
crop rotations, water management, and trap crops (155).
Cultural control is defined as “the use of farming or
cultural practices associated with farm production to
make the environment less favorable for survival, growth
and reproduction of pest species” (155).

15

Other Tactics

Tactical decisions can be separated into emergency
actions 0r preventative actions. Emergency actions are
taken as a result of failure to accurately predict events
leading to emergency pest situations and requiring im-
mediate amelioration. The unexpected development of
economic injury because of an outbreak of pests, result-
ing in emergency applications of insecticides, is an exam-
ple. The tactics frequently used in emergency action
situations are insecticides and plant growth regulators.

Curative actions are designed to prevent pests from
reaching the action level where emergency actions are
needed. Plant resistance, cultural, biological, autocidal,
and regulatory tactics are all examples of curative ac-
tions. Usually two or more of these tactics are used
simultaneously. For example, certain varieties and
planting dates designed to minimize pest density are
used together with chemical methods.

Action levels are used primarily to determine when
to take emergency actions but probably find little use in
determining when to implement curative actions.

Discussion

Definitive predictions of action levels are likely to be
based on a clear understanding of the relation between
the plants, arthropods, economy, and ecology. The fail-
ure to consider any of these factors and their interactions
may result in erroneous management decisions that are
not in the long-term best interests of the producer,
consumer, or others in our society.

Computer Models and Simulation

Brown et al. (10) reviewed the use of computer
simulation in establishing economic thresholds. Simula-
tion is the process of designing a model of a real system
and conducting experiments with this model for the
purpose of either understanding the behavior of the
system or evaluating various strategies for the operation
of the system (117).

Adequate economic thresholds and economic injury
levels have been established for relatively few of the
most important pests of the world, in spite of the long
recognized need (35). None of the action levels currently
used are based on precise estimates that integrate the
host, the pest, enemies of the pest, the economy, and
the ecology. This multiplicity of factors needed in cal-
culating precise action levels is too complex for most
human minds to comprehend simultaneously. Compu-
ters can store, recall, evaluate, simulate, and calculate
rapidly using large quantities of data; thus their potential
for calculating precise dynamic action levels is excellent.
Some of the variables used in the calculation of the
action level are more important than others. The initial
models may include a wide range of variables that can be
evaluated through the simulation process; unimportant
variables and relationships can be eliminated, leaving
only the variables of key importance for prediction.

We should also be able to predict when occasional

16

   
   
  
  
   
  
  
   
  
  
  
   
  
  
  
  
   
 
   
   
 
   
   
    
  
  
 
  
   
  
    
   
   
  
  

pests will be present in outbreak numbers. g
eliminate unnecessary and environmentally dis 1
“insurance” treatments (138). According to Clar
(13), “Forecasts can be required for a variety 5
poses, such as: evaluating the variability of a;
injuriousness in time; preparing for possible incr’
the injuriousness of a pest; and timing the appli
recurrent control measures.” All of these objecti o
be used in either “action” or “action level” mod)

An optimization model for the calculation
economic threshold of L. hesperus was develops
Even programmable calculators have been used
late economic injury levels (157). User inputs in
program are the approximate growth stage, esti’
price of crop, control costs, and an estimate of
free yield. 7‘

Models and Pest Management v

Figure 6 depicts how models might be used [I
management programs. Updates of weathel
economics together with current plant and ar
dynamics would be used to calculate fore
strategies, tactics, and their side effects. From v
action levels could be predicted and used in d’;
sampling, or action decisions could be calculated 
by the computer with or without field validation
decision to sample or not to sample might be basi
the calculated reliability of action decisions. If thei
sion to take action has a high degree of reliability,
validation sampling may not be necessary. Howe 
the action level cannot be calculated because ‘<1;
reliability, then either decision or validation sam
may be needed. Also, if the areawide update sam,
reveals a very low or very high risk situation, then 
decision or validation sampling may be unnecei
When decision or validation sampling becomes 1
sary, the sampler will enjoy the benefit of posse
computer tactic recommendations, calculated 
levels, and computer management decisions to assi
making the final, in-field, decisions. 1

There is the need to make a careful distin
between decisions that should be made in the field
decisions that can be made by computer. As the relia
ty of computer models increases, more and more if
sions should be made by computer to decrease the 
and expense of field sampling. However, some i?’
validation or decision sampling may be needed for 5 _
foreseeable future. l"

Conclusions

The essential value of the action level and the inacf
level is that their use improves the probability a '
increases in yield, as a result of decisions maderegar .
the choice of tactics and strategies employed, will
the costs of pest management. Yield may also include i
value of the commodity to agribusiness as well as t0-
consumers of the commodity. A change in the 
level may be detrimental or beneficial to either gro 3

The costs of pest management are not simply »_

-{ WEATHERI lsusnmess wonun]

    
 
  
   
 
  
   
  
  
 
    
 
    
   
   
   
   
   
  

ntrol tactics (e.g. chemicals, parasite releases,
es, application costs, etc.); they also include
tzgfcal costs that are the result of deterioration of
item which in turn is associated with monocul-
ing, pesticide application, and the misuse of
‘land irrigation.

iver, the long-term goal of pest management
hould be to develop nondisruptive, preventa-
 that totally eliminate the need for action
1e the latter are useful only in conjunction
igency action tactics rather than with the pre-
 entative strategies.

 L. 1973. Relating economic analysis t0 cotton. Pre-
fiConf. Econ. Injury Eval., Dallas.

. L., Bailey, C. T., Hanna, R. L. 1964. Effect of the
l eliothis zea, on yield and quality of cotton. ] . Econ.
2448-50

 L., Gaines, I. C. 1960. Pink bollworm control as
 a total cotton insect control program in Central Texas.
 Stn. Misc. Publ. 444. 7 pp.

I. ., Sterling, W. L., Dean, D. A. 1982. Influence of
-T'ral nectar on red imported fire ants and other
 nviron. Entomol. l1: 629-34

YKimbrough, I. I., Wall, M. L. 1977. Cotton insect
 program. Ark. Agric. Ext. Serv. Leafl. 52. 12 pp.
B. 1964. Integration of chemical and biological con-
igical Control of Insect Pests and Weeds, ed. P.
"489-511. New York: Reinhold. 844 pp.

 ., Longworth, I. W., Evenson, I. P. 1975. Manage-
cotton agroecosystem in southern Queensland: a
‘modeling framework. In Managing Terrestrial
. . I._Kikkawa, H. A. Nix, 9: 230-49. Proc. Ecol. Soc.

1%

c, ecomowuc DATA

as

m, new
2g MODULES SAMPLING
§E 1. PEST-NATURAL ememv ACTION I0 MAKE
== FORECASTS ‘Q LEVELS "4 PEST

2. PLANT FORECASTS g Mfigéfglléfigl
—13. ecomorvuc FORECASTS g enowen n
4. STRATEGY AND mom: 3 LOCAL
OPTIMIZATION E ACTION
, DEUSIUNS VALIDATION
s. sme eeeects 0e SAMPUNG

u, TACTICS

E

Qua

‘Be

5; N0 new SAMPLING
E3

E

lit-n. .
| AGROECOSYSTEM r

Literature

Figure 6. Schematic of the basic components of an action level system with
options (Modified from Ruesink 114).

Acknowledgments

Ithank G. Teetes, S. Gravena, A. Gutierrez, T. Wilson,
and A. Dean for their criticisms of earlier drafts of this
manuscript. The support of the Environmental Protec-
tion Agency and the National Science Foundation is
gratefully acknowledged.

Cited

8. Bottrell, D. G., Adkisson, P. L. 1977. Cotton insect pest manage-
ment. Annu. Rev. Entomol. 22: 451-81

9. Brown, A. W. A. 1978. The Ecology of Pesticides. New York:
Wiley. 525 pp.

10. Brown, L. G., Hartstack, A. W., Ir., Parvin, D. W., Skieth,
R. W. 1979. Computer simulation for establishing economic
thresholds. See Ref. 131, pp. 75-84

11. Carlson, G. A. 1971. The microeconomics of crop losses. See Ref.
19, pp. 89-101

12. Casey, I. E., Lacewell, R. D., Sterlin, W. L. 1975. An example
of economically feasible opportunities for reducing pesticide use
in commercial agriculture. ] . Environ. Qual. 4: 60-64

13. Clark, L. R., Geier, P. W., Hughes, R. D., Morris, R. F. 1967.
The Ecology of Insect Populations in Theory and Practice. Lon-
don: Methuen. 232 pp.

14. Coad, B. R., Cassidy, T. P. 1920. Cotton boll weevil control by
the use of poison. US Dep. Agric. Bull. 875. 31 pp.

15. Curry, G. L., Feldman, R. M., Smith, K. C. 1978. Stochastic
model of a temperature dependent population. Theor. Popul.
Biol. 13: 197-213

16. Davidson, A., Norgaard, R. B. 1973. Economic aspects of pest
control. Europ. Plant Prot. Org. Bull. 3:63-75

17

i i i 

0 -4|_-__.._.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

18

DeBoard, D. V. 1977. Cotton Insect and Weed Loss Analysis.
Cotton Found. 184 pp.

DeMichele, D. W., Bottrell, D. G. 1976. Systems approach to
cotton insect pest management. In Integrated Pest Management,
ed. I. L. Apple, R. F. Smith, 8: 107-32. New York: Plenum. 200
PP-

Economic Research Service. 1971. Economic Research on Pes-
ticides for Policy Decision Making. Proc. Symp. Washington,
DC, April 27-29, 1970. US Dep. Agric., 172 pp.

Doutt, R. L., Smith, R. F. 1971. The pesticide syndrome. In
Biological Control, ed. C. B. Huffaker. pp. 3-15. New York:
Plenum. 511 pp.

Edwards, W. F. 1969. Economic externalities in the agricultural
use of pesticides and an evaluation of alternative policies. PhD
thesis. Univ. Fla., Gainesville. 223 pp.

Ehler, L. E., Miller, I. C. 1978. Biological control in temporary
agroecosystems. Entomophaga 23: 207-12

Encyclopedia Britannica. 1976. Chicago: Encycl. Br. 9: 603
Environmental Protection Agency. 1975. Production, distribu-
tion, use and environmental impact potential of selected pes-
ticides. EPA 540/1-74-001. Washington DC: US Environ. Prot.
Agency.

Ewing, K. P., Ivy, E. E. 1943. Some factors influencing boll-
worm populations and damage. ] . Econ. Entomol. 36: 602-6
Falcon, L. A., van den Bosch, R., Gallagher, I., Davidson, A.
1971. Investigations on the pest status of Lygus hesperus in
cotton in central California. ] . Econ. Entomol. 64: 56-61

Feder, G. 1978. Pesticides, information and pest management
under uncertainty. In Discussion Papers. Dev. Res. Cent. World
Bank No. 27. 42 pp.

Fillman, D. A., Sterling, W. L., 1983. Killing power of the red
imported fire ant [Hym.: Formicidae]: a key predator of the boll
weevil [col.: Curculionidae] Entomophaga 28: 339-344

Fillman, D. A., Sterling, W. L., 1984. Inaction levels for the red
imported fire ant: a predator of the boll weevil. Prot. Ecol. (In
Press).

Fletcher, R. K. 1929. The uneven distribution of Heliothis ob-
soleta (Fabricius) on cotton in Texas. ] . Econ. Entomol. 22: 757-
60

Flores G. R., Bodegas, P. R.B., de Coss Flores, M. E. 1978.

Determinacion del nivel economico del dafio causado por Alaba-
ma argillacea al cultivo del algodonero en la zona del soconuesco,

Chiapas, Mexico. Bol. Inf. CIES 5:3-5.

Frisbie, R. E., Sprott, I. M., Lacewell, R. D., Parker, R. D.,

Buxkemper, W. E., Bagley, W.E., Norman, I. W., Ir. 1976. A
practical method of economically evaluating an operational cotton
pest management program in Texas. ] . Econ. Entomol. 69: 211-
14

Fry, K. E., Kittock, D. L., Henneberry, F. I. 1978. Effect ofthe
number of pink bollworm larvae per boll on yield and quality of
pima and upland cotton. ] . Econ. Entomol. 71: 499-502

Fuchs, T. W., Harding, I. A. 1976. Seasonal abundance of ar-
thropod predators in various habitats in the Lower Rio Grande
Valley of Texas. Environ. Entomol. 5: 288-90

Glass, E. H. 1975. Integrated pest management: rationale, po-
tential needs and implementation. Entomol. Soc. Am. Spec.
Publ. 75-2. 141 pp.

Glass, E. H. 1976. Pest management: principles and philosophy.

See Ref. 18, 4: 39-57

Graham, H. M., Lingren, P. D., Lincoln, C., Adkisson, P. L.
1972. The economic threshold of infestations of Heliothis spp. on
cotton. In Distribution, Abundance and Control of Heliothis
Species in Cotton and Other Host Plants, 3: 7-15. Southern
Coop. Ser. Bull. 169. 92 pp.

Gutierrez, A. P., Daxl, R., Leon Quant, G., Falcon, L. A. 1981.
Estimating economic thresholds for bollworm, Heliothis zea
(Boddie), and boll weevil, Anthonomus grandis Boh., damage in
Nicaraguan cotton, Gossypium hirsutum L. Environ. Entomol.
10: 872-79

Gutierrez, A. P., Falcon, L. A., Loew, W., Leipzig, P. A., van
den Bosch, R. 1975. An analysis of cotton production in Califor-
nia: a model for Acala cotton and the effects of defoliators on its
yield. Environ. Entomol. 4: 125-36

40. Gutierrez, A. P., Leigh, T. F., Wang, Y., Cave, R. 1977. 

analysis of cotton production in California: Lygus hespe"
(Heteroptera: Miridae) injury-an evaluation. Can. Entomol. 1
1375-86
41. Gutierrez, A. P., Wang, Y. 1979. An optimization model
Lygus hesperus (Heteroptera: Miridae) damage in cotton: A
economic threshold revisited. Can. Entomol. 111: 41-54 A
42. Hamner, A. L. 1941. Fruiting of cotton in relation to c0
fleahopper and other insects which do similar damage to squ ,
Miss. Agric. Exp. Stn. Bull. 360. 11 pp. t
43. Hardin, G. 1969. Not peace, but ecology, In Diversity r
Stability in Ecological Systems, ed. G. M. Woodwell, H.
Smith, 14: 151-61. Brookhaven Symp. Biol. 22, Upton, I
Brookhaven Natl. Lab. 264 pp. '
44. Hartstack, A. W., Ir., Witz, I. A., Hollingsworth, I. P., Rid'
R. L., Lopez, I. D. 1976. MOTHZV-2: A computer simulati 7
Heliothis zea and H eliothis virescens population dynamics,
manual. US Dep. Agric. ARS-S127. 55 pp. -
45. Headley, I. C. 1971. What is the economic threshold? Agric .
Age 14. s-9 i
46. Headley, I. C. 1972. Defining the economic threshold. In l
Control Strategies for the Future, 2: 100-8. Washington If
Natl. Acad. Sci. 376 pp.
47. Headley, I. C. 1975. The economics of pest management. -
Ref. 71, 3: 75-99
48. Headley, I. C., Lewis, I. N. 1967. The Pesticide Problem;
Economic Approach to Public Policy. Baltimore: Iohns Hop f
144 pp.
49. Health, Education and Welfare. 1969. Report of the Secret
Commission on Pesticides and Their Relationship to Environ "

tal Health, Pt. 1, 2. Washington DC: US Dep. Health, E

Welfare, GPO.

50. Hearn, A. G. 1975. Ord valley cotton crop: development ,
technology. Cotton Grow. Rev. 52: 77-102 I

51. Heilman, M. D., Lukefallr, M. 1., Namken, L. N., No 1 '

I. W. 1977. Field evaluation of the short season produ
system in the Lower Rio Grande Valley of Texas. In ’
Beltwide Cotton Prod. Res. Conf., pp. 80-83. Memphis, T
Natl. Cotton Counc. 258 pp. »
52. Heilman, M. D., Namlcen, L. N., Dilday, R. H., Sterling, I
1978. The effect of seedling terminal damage by Heliothis 
cens (Fab) on fruiting characteristics and earliness of two r i
genotypes. See Ref. 51, p. 84 l
53. Herrera Aranguena, I. M. 1965. Investigations of mirids of
genus Rhinacloa, important agents in the control of Hel‘
virescens on cotton. Rev. Peru. Entomol. 8: 44-60 ,
54. Hueth, D., Regev, U. 1974. Optimal agricultural pest m 1'
ment with increasing pest resistance. Am. ]. Agric. Econ.
543-52 I

55. Huffaker, C. B. ed. 1971. Biological Control. New i.’
Plenum. 511 pp. f
56. Huifaker, C. B., Shomaker, C. A., Gutierrez, A. P. 1978. l,
rent status, urgent needs and future prospects of integrated ii
management. See Ref. 102, 15: 237-59 1
57. Hutchinson, I. B., Solow, R. A., Stephens, S. G. I947.

Evolution of Gossypium, pp. 81-82. London: Oxford Univ. P i;

58. Iones, D., Sterling, W. L. 1979. Manipulation of red im 

fire ants in a trap crop for boll weevil suppression. En '

Entomol. s; 1073-77 y, -

59. Iones, I. W., Bowen, H.AD., Stinner, R. E., Bradley, R.,
Bachelor, I. S. 1977. Simulation of boll weevil populati "
influenced by weather, crop status and management pr '
Trans. Am. Soc. Agric. Eng. 20: 121-31

60. Iudenko, E. 1965. The assessment of economic elfectiven,
pest management in field experiments. PAN S (A). 11: 35 p
61. Kincade, R. T., Laster, M. L., Brazzel, ]. R. 1967. Da /--
cotton by the tobacco budworm. ] . Econ. Entomol. 60: 11
62. Koenig, H. E., Tummala, R. L. 1972. Principles of ecosyp
design and management. IEEE Trans. Syst. Man Cybern. 2:‘
59 1
63. Kuehl, R. O., Fye, R. E. 1972. An analysis of the sam
distribution of cotton insects in Arizona. ] . Econ. Entomol-
855-6O

  
  
  
  
    
  
  
  
  
  
  
 
  
   
   
    
   
   
  
   
  
   
  
     
   
 
  

   
   
   
 
    
  
   
  
  
    
  
  
  
  
   
 
   
  
    
  
  
 
 
    
   
  
  
 
 
  
   
    
 
   
 
  
  
  

Ewell, R. D. 1977. Statement on IPM cotton production
A_ ems in Texas. In Integrated Pest Management Hearings
ore the 95th Congress, pp. 152-57. Washington DC: GPO.
l‘ PP-

well, R. D., Taylor, C. R. 1978. Economic implications of
I mate boll weevil control strategies. In The Boll Weevil:
i ement Strategies, 9: 104-5. Southern Coop. Ser. Bull. 228.

ipp-

"g; ‘am, M. R. 1971. Potential for applying the Dade County
_'cide model to a wider geographic area. In Economic Re-
 h 0n Pesticides for Policy Decisionmaking, ed. V. W. Davis,
$71-79. Proc. Symp. Washington DC. 172 pp.

i; am, M. R., Edwards, W. F. 1969. Externalities in pes-
 use. Am.]. Agric. Econ. 51: 1195-1201

plerg, E. L. 1971. Field experiments for assessment of crop
’ s. In Crop Loss Assessment Methods: FAO Manual 0n the
_' tion and Prevention of Losses by Pests, Diseases and
f» , ed. L. Chiarappa, 2: 2.1/1-2.2/5. Oxford: Alden. 169 pp.
j en, P. D., Ridgway, R. L., Iones, S. L. 1968. Consumption
I eral common arthropod predators of eggs and larvae of two
‘I his species that attack cotton. Ann. Entomol. Soc. Am. 61:
.118

, ,  D., Ir., Ridgway, R. L., Pinnel, R. E. 1976. Compara-
efliciency of four insect predators of the bollworm and
y» budworm. Environ. Entomol. 5: 1160-64

4| ann, W. H., Metcalf, R. L. 1975. The pest management
' t. In Introduction to Insect Pest Management, ed. R. L.
Q: , W. H. Luckman, 1: 3-25. New York: Wiley. 587 pp.
W r, M. I., Shaver, T. N. 1972. The potential of resistant
. ~ in the suppression of Heliothis populations. See Ref. 37,
. -s2

___ ur, R. H., Wilson, E. O. 1967. The Theory of Island
graphy. Princeton: Princeton Univ. Press. 203 pp.

; M. N. A., Evenson, I. P., Edwards, D. C. 1978. The
v of level of nitrogen on earliness in upland cotton. Aust. ] .
 R88. 29: 1213-21

 L. 1976. Briefing: Clearinghouse for chemical carcino-
, ‘Science 192: 242

_ll, F. C. 1977. Plant resistance to cotton insects. Bull.
fl. Soc. Am. 23; 199-203

jg; don, R. W., Brown, L. C., Iones, I. W. 1977. Modeling
sect pest ecosystems for studying conflicting alternatives
 by species interactions. In Proc. Int. Fed. Auto. Con.,
Control Mech. Bio- and Ecosystem, Leipzig, CDR, , 5:

‘é-“el, S. C., Sterling, W. L. 1979. Predator determination
 'ency on Heliothis virescens (F.) eggs in east Texas using
fve phosphorus. Environ. Entomol. 8: 1083-87

n l, S. C., Sterling, W. L. 1982. Predation of Heliothis
 i  eggs on cotton in east Texas. Environ. Entomol. 11:

Tel, S. C., Sterling, W. L., Dean, D. A. 1981. Predators
y» budworm larvae in Texas cotton. Southwest. Entomol.

Ton, I. M., Iones., I. W., Hesketh, I. D. 1974. Analysis
photosynthesis and growth. See Ref. 51, pp. 118-24

. R. L. 1977. Model ecosystem approach to insecticide
'on: a critique. Annu. Rev. Entomol. 22: 241-61

‘ . F. 1959. Single factor analysis in population dynamics.
. 40: 580-588

 Academy of Science. 1969. I nsect-Pest Management and
Publ.1695, Washington DC: Natl. Acad. Sci. 508 pp.

l‘ Cotton Council. 1974. Boll weevil losses: value and
of losses caused by the boll weevil. Natl. Cotton Counc.

I Research Council. 1975. World food and nutrition
,8 hancement of Food Production for the US Natl. Acad.
1' . Res. Counc. Washington DC.

M. M., Martineau, R., Rose, A. H. 1971. Diprion her-
? artig), European spruce sawfly (Hymenoptera: Dip-
. In Biological Control Programs Against Insects and
 Canada 1959-1968, 4: 136-43. Commonw. Inst. Biol.
ech. Commun.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

Nemec, S. I., Adkisson, P. L. 1969. Laboratory tests of insec-
ticides for bollworm and boll weevil control. Tex. Agric. Exp.
Stn. Prog. Rep. 2674. 5 pp.

Newsom, L. D. 1972. Theory of population management for
Heliothis spp. in cotton. See Ref. 37, pp. 80-92

Norgaard, R. B. 1976. The economics of improving pesticide use.
Annu. Rev. Entomol. 21: 45-60

Norman, I. W., Ir., Henson, I. L., Wolfenbarger, D. A., Hard-
ing, I. A., Cage, E. V., Parker, R. D., Ir., Collins, C. 1979.
Effects of water management practices on economics of insect
control on cotton. Lower Rio Crande Valley, Tex. ]. Econ.
Entomol. 72: 367-70

Oelhaf, R. C. 1978. Organic Agriculture-Economic and Ecologi-
cal Comparisons with Conventional Methods. Montclair, NI:
Alanheld, Osmum. 271 pp.

Olkowski, W. 1973. A model ecosystem management program.
In Proc. Tall Timbers Conf. on Ecol. Anim. Control by Habitat
Manage. 5: 103-17

Ordish, C. 1962. Economics and pest control. World Rev. Pest
Control. 1: 31-38

Parker, R. D. 1979. Interactions of cotton genotype with boll
weevil, Anthonomus grandis Boheman, populations as influenced
by planting pattern, irrigation level and insecticide treatment.
PhD thesis. Texas A&M Univ. College Station. 225 pp.

Parker, M. W., Brown, L. C., Harris, F. A., Andrews, C. L.
1975. A simulation of field populations of insects (Heliothis in
cotton). Proc. Winter Sim. Conf. Assoc.for Comput. Mach. pp.
247-50

Pate, T. L., Hefner, I. I., Neeb, C. W. 1972. A management
program to reduce cost of cotton insect control in the Pecos area.
Tex. Agric. Exp. Stn. Misc. Publ. I023. 7 pp.

Patrick, R., Strawbridge, D. 1963. Variation in the structure of
natural diatom communities. Am. Nat. 97: 51-57

Phillips, I. R., Clower, D. F., Hopkins, A. R., Pfrimmer, T. R.
1979. Economic thresholds of Heliothis species on indeterminant
cottons. See Ref. 131, pp.44-50

Pickett, A. D. 1959. Utilization of native parasites and predators.
]. Econ. Entomol. 52: 1103-5

Pimentel, D. 1976. World food crisis: energy and pests. Bull.
Entomol. Soc. Am. 22: 20-26

Pimentel, D. 1978. Socioeconomic and legal aspects of pest
control. In Pest Control Strategies, ed. E. H. Smith, D. Pimen-
tel, 5:55-71. New York: Academic. 334 pp.

Pimentel, D., Shoemaker, C. 1974. An economic and land-use
model for reducing insecticides on cotton and corn. Environ.
Entomol. 3: 10-20

Pitre, H. N., Mistric, W. I., Lincoln, C. C. 1979. Economic
thresholds: concepts and techniques. See Ref. 131, pp 12-30
Plapp, F. W., Ir., Browning, C. L., Sharpe, P. I. H. 1979. Anal-
ysis of rate of development of insecticide resistance based on
simulation of a genetic model. Environ. Entomol. 8: 494-500
Poinar, C. 0., Thomas, C. M. 1978. Diagnostic Manual for the
Identification of Insect Pathogens. New York: Plenum. 218 pp.
Price, P. W. 1975. Insect Ecology. New York: Wiley. 514 pp.
Rabb, R. L., Kennedy, C. C. 1979. Movement ofHighly Mobile
Insects. Raleigh, Univ. Graphics, NC State Univ. 456 pp.
Rahman, M. A. 1977. Response of the cotton plant (Cossypium
hirsutum Cv. Rex. S. L. Okra) to defoliation and disbudding.
PhD thesis. Univ. Queensland, Australia. 261 pp.

Ridgway, R. L. 1969. Control of the bollworm and tobacco
budworm through conservation and augmentation of predaceous
insects. In Proc. Tall Timbers Conf. On Ecol. Anim. Control by
Habitat Manage. 12: 127-44. Tall Timbers Res. Stn. Bull. No. 1.
244 pp.

Ridgway, R. L., Iones, S. L. 1968. F ield-cage releases of Chryso-
pa carnea for suppression of populations of the bollworm and the
tobacco budworm on cotton. ] . Econ. Entomol. 61: 892-98
Roach, S. H., Smith, I. W., Vinson, S. B., Craham, H. M.,
Harding, I. A. 1979. Sampling predators and parasites of
Heliothis species on crops and native host plants. See Ref. 131,
pp. 132-45

Robinson, I. M., Arant, F. S. 1932. Eight years of experimental
work in boll weevil control on plots receiving different rates of
fertilizer. ] . Econ. Entomol. 25: 759-66

19

 

i. aiiaafiiiznaa-i

I I Ijbpl»

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.
127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

20

Ruesink, W. G. 1975. Analysis and modeling in pest manage-
ment. See Ref. 71, pp. 353-76

Ruesink, W. G. 1976. Status of the systems approach t0 pest
management. Annu. Rev. Entomol. 21: 27-44

Rummel, D. R., Frisbie, R. E. 1978. Suppression of potentially
overwintering boll weevils as a pest management practice. In The
Boll Weevil: Management Strategies, 4: 39-49. Southern Coop.
Ser. Bull. 228. 130 pp.

Shannon, R. E. 1975. Systems Simulation, The Art and Science.
Englewood Cliffs, NI: Prentice-Hall. 387 pp.

Smith, R. F. 1967. Principles of management of crop losses
caused by insects. In FAO Symp. 0n Crop Losses, Rome, pp.
205-24

Smith, R. F. 1969. The new and old in pest control. Proc. Acad.
Naz. Lincei, Rome, 1968. 366: 21-30

Smith, R. F. 1969. The importance of economic injury levels in
the development of integrated pestlcontrol programs. Qual.
Plant. Matero. Veg. 17: 81-92

Smith, R. F., Falcon, L. A. 1973. Insect control for cotton in
California. Cotton Grow. Rev. 50: 15-27

Snow, I. W., Copeland, W. W. 1971. Distribution and abun-
dance of the corn earworm in the United States. US Dep. Agric.
Plant Pest Control Div. Coop. Econ. Insect Rep. 21: 71-76
Solomon, M. E. 1949. The National Control of insect popula-
tions, ]. Anim. Ecol. 18: 1-35

Southwood, T. R. E. 1978. Ecological Methods. Halsted. 524 pp.
Southwood, T. R. E., Norton, G. A. 1972. Economic aspects of
pest management strategies and decisions. Ecol . Soc. Aust. Mem.
1: 168-84

Sparks, A. 1972. Heliothis migration. See Ref. 37, pp. 15-17
Sprott, M. I., Lacewell, R. D., Niles, G. A., Walker, I. K.,
Gannaway, I. R. 1976. Agronomic, economic, energy and envi-
ronmental implications of short-season, narrow-row cotton pro-
duction. Tex. Agric. Exp. Stn. Misc. Publ. 1250. 23 pp.
Stephenson, S. N. 1973. Fertilization responses in six southern
Michigan old field communities. Bull. Ecol. Soc. Am. 54: 33
Sterling, W. L. 1976. Sequential decision plans for the manage-
ment of cotton arthropods in south-east Queensland. Aust. ].
Ecol. 1: 265-74

Sterling, W. L. 1978. Fortuitous biological suppression of the
boll weevil by the red imported fire ant. Environ. Entomol. 7:
564-68

Sterling, W. L., chairman. 1979. Economic Thresholds and Sam-
pling of Heliothis Species on Cotton, Corn, Soybeans and Other
Host Plants. Southern Coop. Ser. Bull. 231. 159 pp.

Sterling, W. L., Dean, D. A., Fillman, D. A., Iones, D. 1984.
Naturally occurring biological control of the boll weevil [Col.:
Curculionidae]. Entomophaga. 29: 1-9.

Sterling, W. L., Hartstack, A. W. 1979. Emergence threshold
with validations for forecasting the spring emergence of the
cotton fleahopper. Environ. Entomol. 8: 649-54

Sterling, W. L., Lincoln, C. 1978. Survey, detection, and
economic thresholds. See Ref. 116, pp. 4-14

Sterling, W. L., Pieters, E. P. 1979. Sequential decision sam-
pling. See Ref. 131, pp. 85-101

Stern, V. M. 1966. Significance of the economic threshold in
integrated pest control. In Proc. FAO Symp. Integrated Pest
Control, Oct. 11-15, 1965, Rome, pp. 41-56

Stern, V. M. 1973. Economic thresholds. Annu. Rev. Entomol.
18: 259-80

Stern, V. M., Smith, R. F., van den Bosch, R., Hagen, K. S.
1959. The integrated control concept. Hilgardia 29: 81-101
Stinner, R. E. 1977. Efficacy of inundative releases. Annu. Rev.
Entomol. 22: 515-31

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.

162.

163.

164.

  
   
   
    
   
 
   
   
    
 
 
  
   
   
  
   
 
   
    
   
  
   
  
   
 
  
   
  
  

Stinner, R. E., Rabb, R. L., Bradley, I. R. 1974}
dynamics of Heliothis zea and H. virescens in North’
simulation model. Environ. Entomol. 3: 163-68

Talpaz, H., Borosh, I. 1974. Strategy for pesticide
and application. Am. ] . Agric. Econ. 56: 769-75

Tamaki, G., Long, G. E. 1978. Predator complex ~g
peach aphid on sugarbeets: expansion of the preda v ‘l?
efficiency model. Environ. Entomol. 7: 835-42 
Townsend, I. R., Ir. 1973. Economic threshold‘:
Heliothis spp. on cotton. MS thesis, Miss. State Un‘?

van den Bosch, R. 1972. The cost of poisons. Environ '
36
van den Bosch, R. 1978. The Pesticide Conspiracy. 
NY: Doubleday. 226 pp. ‘i
van den Bosch, R., Leigh, T. F., Gonzalez, D., Sti’
1969. Cage studies on predators of the bollworm in'
Econ. Entomol. 62: 1486-89 V.
van Emden, H. F. 1974. Pest control and its w k
Institute ofBiologiés Studies in Biology No. 50. Lond 1
59 pp. " .
Vaughn, C. R. 1978. Radar, Insect Population Ecolo 
Management, NASA Conf. Publ. 2070, Wallops Isla 

PP- . ~.
Vinson, S. B., Plapp, F. W., Ir. 1974. Third generation
the potential for the development of resistance to 
Agric. Food Chem. 22: 356-60 '
Walker, A. 1974. A simulation model for prediction of
persistence. ] . Environ. Qual. 3: 396-400.

Walker, I. K., Frisbie, R. E., Niles, G. A. 1979. Helio 11'
in short-season cottons in Texas. See Ref. 131, pp. 31 “I
Walker, ]. K., Hanna, R. L. 1966. Effects ofthe boll 
bollworm on cotton quality. ] . Econ. Entomol. 59: 2 t A
Walker, I. K., Niles, G. A. 1971. Population dynamics ~j_
weevil and modified cotton types. Texas Agric. Exp. S
1109. 14 pp. a
Wang, Y., Gutierrez, A. P., Oster, G., Daxl, R. 1977. A
tion model for plant growth and development: couplinf
herbivore interaction. Can. Entomol. 109: 1359-74 »
Watson, T. F., Moore, L., Ware, G. W. 1975. Pract‘
Pest Management. San Francisco: Freeman. 196 pp.
Watt, K. E. F. 1961. Mathematical models for usein
control. Can. Entomol. Suppl. 19. 62 pp. A
Welch, S. M., Poston, F. L. 1979. The programmable w}
as a pest management tool. Bull. Entomol. Soc. Am. 25:.
Whitcomb, W. H. 1967. Field studies on predators of 7
ond-instar bollworm, Heliothis zea (Boddie) (Lepidopte "
tuidae). ]. Ga. Entomol. Soc. 2: 113-18 ‘
Whitcomb, W. H., Bell. K. 1964. Predaceous insects, 
and mites of Arkansas cotton fields. Ark. Exp. Stn. Bull. i‘

‘:-

h

PP- .
White, I. R., Rummel, D. R. 1978. Emergence profile I
wintered boll weevils and entry into cotton. Environ. Ent V,
7-14 .
Wilhelm, S. 1976. The agroecosystem: a simplified plant u

nity. See Ref. 18, 6: 59-70 
Wilson, A. G. L. 1974. Resistance of Heliothis armigera to_
ticides in the Ord Irrigation Area, North Western Aus w
Econ. Entomol. 67: 256-58 I
Wilson, A. G. L. 1975. Troubles aplenty in our cotton 
ECOS, CSIRO Environ. Res. 5: 10-12
Wilson, F. D., Wilson, R. L. 1979. Earliness, Pubescen
pink bollworm resistance in cotton. Agron. Ahstr. 67 v

   

             

 

 

. - , .- nqals-samimaaaazuwrumfiz, ~._.,.=2."~‘ i usual

Mention of a trademark or a proprietary product does not constitute a guarantee or a warranty of the product by the Texas Agriculturi

Experiment Station and does not imply its approval to the exclusion of other products that also may be suitable.

All programs and information of The Texas Agricultural Experiment Station are available to everyone without regard to race, color, religion, se
age, handicap, or national origin. .

2M—7-84