Quality control of mass-reared lepidoptera using the fall armyworm, Spodoptera frugiperda (J.E. Smith), as a model


Material Information

Quality control of mass-reared lepidoptera using the fall armyworm, Spodoptera frugiperda (J.E. Smith), as a model
Spodoptera frugiperda
Physical Description:
viii, 172 leaves : ill. ; 28 cm.
Fisher, William Randolph, 1948-
Publication Date:


Subjects / Keywords:
Fall armyworm   ( lcsh )
Insect rearing   ( lcsh )
Insect pests -- Control -- Biological control   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1983.
Includes bibliographical references (leaves 158-171).
Statement of Responsibility:
by William Randolph Fisher.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000372980
notis - ACB2150
oclc - 10034930
System ID:

Full Text

USING THE FALL ARMYWORM, Spodoptera frugiperda (J.E. Smith),


William Randolph Fisner




Copyright 1983

William Randolph Fisher


I thank N.C. Leppla, chairman of my committee, for

patience and guidance, and for providing the opportunities

that made my work here much more meaningful than simply an

academic experience. I thank D.L. Chambers for the support

he has shown for me and this project over the past four

years. I sincerely appreciate the participation of my

committee members and examiners who have been patient and

understanding and have contributed to my experience,

training, and philosophy. They are D.L. Chambers, D.A.

Dame, S.H. Kerr, J.E. Moore, J.L. Nation, and R.I. Sailer.

I am indebted to P.M. Achey for unselfishly sharing his

laboratory and equipment with me. I appreciate the

technical assistance provided by S.L. Barrette. I am

grateful to my wife, Jeanne Wiegand, who supported me and

offered her photographic and technical skills. I thank the

U.S. Department of Agriculture for the support that enabled

me to pursue this degree. I would also like to thank many

other individuals who have helped me during the course of

this project: H.R. Agee, T.R. Ashley, D.G. Boucias, F.D.

Brewer, C.O. Calkins, V. Chew, G. Claire, S.M. Coniglio,

J.C. Davis, V.M. Ferguson, P.D. Greany, C.W. Green, M.D.

Greenfield, G.G. Hartley, D.E. Hendricks, A.J. Hill, M.D.




Huettel, E.G. King, E.F. Knipling, M.L. Laster, F.L. Lee,

J.E. Lloyd, M.E. Martignoni, D.F. Martin, W.W. Metterhouse,

E.R. Mitchell, R.F. Moore, T.M. O'Dell, S.D. Pair, F.I.

Proshold, W.J. Repole, J.R. Rye, J.L. Sharp, J.M. Schalk,

A.C. Thompson, F.C. Tingle, E.V. Vea, J.C. Webb, and G.P.

Whitmer. Finally, I would like to thank myself for pursuing

this degree and for achieving more than I had originally

hoped for.



ABSTRACT. .. . . vii


Quality Control of Colonized Insects: A
Historical Perspective. .. 1
Factors Influencing the Quality of
Colonized Insects ............ 16
Comparison of Colonized Insects With Wild
Standards . . 25
Identifying and Interpreting Changes in
Insect Quality. . ... 31
Characteristics of Lepidopteran Rearing 35


Source of Insects and Rearing Conditions. 46
Holding Conditions for Test Insects ..... 49
Characteristics of Immature Insects .. .51
Testing of Adult Insects. . .. .52
Wingbeat frequency. . 52
Response to pheromone . ... 54
Four-minute mating observations .. 56
Two-night mating studies. . ... 57
Treatments. .............. 58
Development of Quality Control Charts ... 60

RESULTS . .. 70

Occurrence of Dietary Fungal Contamination. 70
Insect Density in Larval Rearing Containers 70
Characteristics of Immature Insects .. .73
Pupal weight .. ... 73
Pupal buoyancy. . ... 75
Abdominal rotation. .. 76
Pupal mortality . .. 77
Characteristics of Adult Insects. .. 77
Eclosion and wing deformity ... .77
Wingbeat frequency. ... . 78
Response of laboratory and wild males
to pheromone ... . 79
Four-minute mating observations ... 82
Two-night mating studies. .. 85



A Model Quality Control Program ... .121
The rearing facility. . .. .122
Procedures and quality testing. .. .124
Feedback system . .. 133


APPENDIX. . ... .. .156

REFERENCES. . . .. 158



Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

USING THE FALL ARMYWORM, SDodoptera frugiperda (J.E. Smith),


William Randolph Fisher

April 1983

Chairman: Norman C. Leppla
Major Department: Entomology and Nematology

A series of tests and observations on developmental,

morphological, and behavioral traits was developed for

evaluating the quality of fall armyworms, Soodoptera

frugiperda (J.E. Smith), reared on artificial diet in

communal containers, and for establishing a model quality

control program for mass-reared lepidopterans. Black mold

grew on the diet surface in 35% of the containers and was

especially prevalent on Mondays, Wednesdays, and Fridays,

the busiest days for insectary employees. The incidence of

mold was correlated with reduced pupal abdominal rotation

and buoyancy, increased pupal mortality, and unsuccessful

adult eclosion. Process control charts were shown to be

sensitive tools for identifying changes in insects during

colonization. Average pupal weights showed some



out-of-control points caused by high larval densities.

Downward trends were noted in sample averages for weight and

rotation but sample ranges were within limits of normal

variability. Pupal mortality averaged 1.6% and was

significantly higher when mold was present. Wing

deformities occurred in 4.3% and 5.2% of males and females,

respectively. Wingbeat frequency (WBF) for males averaged

44.7 cycles per second and remained within limits; moths

from heavier pupae had reduced WBF. About half of the

laboratory males demonstrated anemotactic flight in response

to pheromone-producing females, compared to 83% for wild

males. Only 73% of the laboratory males arrived at the

source in an average of 65 sec while all of the wild males

arrived in an average of 39 sec. Adult diet and pupal

radiation treatment altered mating behavior and success.

Untreated laboratory males attempted to clasp a female once

every 15 sec whereas males from a group of irradiated pupae

averaged once in 82 sec, and unfed males pursued females for

an average of 120 sec but never attempted copulation.

Results were incorporated into a model quality control

program based on an integrated facility designed to ensure

efficient production in a sanitary environment, procedures

that allow natural behavior of larvae and adults, a battery

of laboratory quality control tests, process control

charting, and a system that feeds data from tests directly

back into the production program.


Quality Control of Colonized Insects: A Historical

Quality control has been practiced as long as man has

maintained insect colonies. For many years, quality insects

were considered those that were merely alive. However,

during the early 1960's, published accounts demonstrated

that sublethal changes in colonized insects severely reduced

potential effectiveness of pest control strategies. For

example, Weidhaas et al. (1962) reported unsuccessful

attempts to introduce sterility into a wild population of

Anopheles auadrimaculatus Say using irradiated laboratory

males. Investigating possible causes for this failure, Dame

et al. (1964) showed that the problem resulted from an

inability of the males to disperse in the wild and seek out

mates. In 1962 sterile laboratory flies were released in

the southwestern United States to control the screwworm,

Cochliomyia hominivorax (Coquerel), which caused significant

economic losses to the livestock industry. A precipitous

drop in screwworm infestations occurred within one year and

the program was hailed as a milestone in pest control.

However, within a decade the release of laboratory flies in

numbers five times greater than in 1962, proved ineffective



in controlling the screwworm. Lack of control was

attributed to laboratory insects of poor quality. Alley and

Hightower (1966) reported that newly colonized screwworms

mated less frequently than laboratory-adapted strains and

mating frequency for small screwworm flies was significantly

lower than for larger flies no matter how long they remained

in culture. Spates and Hightower (1967) found that males

from recently colonized strains of screwworms did not harass

laboratory females as vigorously as did males from older


In 1966, C.N. Smith edited a book entitled "Insect

Colonization and Mass Production" that described procedures

and problems associated with rearing over 40 species of

insects, mites, and ticks. Most authors stressed the

importance of the laboratory environment on production

output and insect quality, although the latter term was

infrequently used. Emphasis was placed on preventing

contamination of diet and insects, especially by disease

organisms. Cleanliness of the rearing facility and its

equipment, ano sanitary techniques were discussed by Cowan,

Gast and Davich, Henneberry and Kishaba, and others. The

importance of isolating wild colonizers to ensure

disease-free stock prior to introduction into the laboratory

was recommended by Harein and Soderstrom. Difficulty in

establishing wild insects in the laboratory was a common

problem and those authors that addressed the issue concluded

that a large colonizing population was the best means to


ensure survivors. Gahan reported difficulty getting wild

mosquitos to mate in capitivity and suggested a forced

mating ritual for fertilization. De Meillon and Thomas

recommended selecting for variability in a laboratory colony

of mosquitos by using only those larvae and adults that

showed vitality.

Some authors discussed measuring and evaluating insect

quality. Lumsden and Saunders listed several parameters for

monitoring colonies of tsetse flies. They included rate of

population increase, longevity and mating frequency of

individual females, number of pupae per female, length of

larval period, and pupal weight. Baumhover et al. described

a quality control program for irradiated screwworm flies

that included measurements taken on viability of treated

pupae, sterility and longevity of adults, mortality

resulting from storage and shipment to the release site,

sexual aggressiveness, and field tests of competitiveness.

Gahan described differences in behavior between wild and

laboratory mosquitos and he cautioned against predicting

field behavior from studies conducted in the laboratory.

Lumsden and Saunders addressed the same problem with tsetse

fly and indicated that frequent gene reinforcement from wild

populations may be required to improve field competitiveness

of the laboratory colony. Smith's book illustrated the

diversity of species being reared and identified basic

problems common to the production of most of them.


During the same year, a conference of the International

Atomic Energy Agency (IAEA) updated insect rearing

information to improve sterile release technology. R.T.

Gast (1968) stated the objective of mass rearing to be the

production of an acceptable insect at the lowest possible

cost. He defined an acceptable insect as one that was

inevitably different from wild, but one that could perform

in its intended role of pest suppression. He stressed the

development of standards of acceptability. Once an

acceptable insect was being reared, emphasis should be

placed on reducing production costs. Gast also discussed

the significance of selection and microbial contamination in

reducing insect quality. Recommendations of the IAEA panel

included designing rearing facilities and equipment to

reduce the potential for contamination, defining symptoms of

nutrient deficiencies, and establishing quality control

criteria to optimize insect performance in field releases.

In 1970, a similar panel discussed the application of

induced sterility for the control of Lepidopterous pests

(IAEA 1971). It reported that mass production "represents

one of the most difficult obstacles to immediate field

testing of the sterility principle for a great many of these

pests", with disease and nutrition being of specific

concern. To help alleviate these problems, they recommended

compiling and disseminating information about the design and

construction of rearing facilities including specifications

for equipment and building materials such as wall and floor


finishes. During these and other IAEA proceedings,

discussions centered on the deterioration of insect quality

(e.g. dispersal, orientation to environmental stimuli, and

mating behavior) resulting from exposure of the insects to

sterilizing radiation.

In 1971, a symposium sponsored by the International

Organization for Biological Control (IOBC), "Implications of

Permanent Insect Production" was held in Rome. Two timely

papers resulted from the conference. The first, by E.F.

Boiler (1972), discussed changes in behavior during

laboratory colonization of wild insects. These changes,

based on genetic or environmental factors, were especially

apparent during generations one to six, after which colonies

became stabilized, albeit different from their wild

conspecifics. Boller suggested modifying the traditional

concept of mass-production--i.e., the number of fertile

insects per dollar--to the number of insects required to

achieve specified objectives. Steps to help ensure

production of quality insects included strategies for

maintaining genetic variabiltiy and monitoring insects for

quality control using various behavioral tests in the

laboratory and field.

The second paper dealt with the genetic changes that

occur during colonization. Mackauer (1972) stated that

preservation of essential attributes (adaptability,

competitiveness, and mobility) during colonization was

dependent on, among other things, the size and origin of the


founder colony. It must be large enough to assure genetic

flexibility and thus survival under laboratory conditions.

He suggested that large founder populations are less

susceptible to random genetic changes (i.e. drift) and have

greater genetic diversity, resulting in a laboratory strain

that is similar to wild insects.

At that time, examples of poor performance by

laboratory insects showed that success of programs such as

sterile insect technique (SIT) and release of beneficial

control agents was intimately related to colonization and

production processes. To investigate this relationship, a

workshop in 1974 dealt with the genetics of insect behavior

and its relation to insect production and quality control.

The workshop was organized by D.L. Chambers, W. Klassen,

L.E. LaChance, I.C. McDonald, R.L. Ridgway, and D.E.

Weidhaas. Participants identified objectives for future

activities based on improving poor performance. These

included developing a systematic approach to measuring

quality in the laboratory and field, and incorporating the

results back into the production system to make required

changes; monitoring founders as soon after collection as

possible for identification of the precise changes that

occur during colonization; using data processing to simplify

quality control efforts; eliminating disease and identifying

the effects of sublethal concentrations of pathogens;

determining whether changes in behavior are induced by

genetic or non-genetic factors; establishing standards,


especially with field insects; and identifying the role of

nutrition in fitness. The conference also recommended

elevating the study of insect quality to a higher priority,

organizing research groups, workshops, and task forces

accordingly, and instituting educational courses to

familiarize production personnel with relevant principles of

ecological, quantitative, and population genetics. This

workshop provided the comprehensive and integrated base upon

which future quality control philosophy and application were

developed. Further, it educated many scientists and helped

coordinate their activities in the field.

In 1975, D.L. Chambers outlined basic concepts of

quality control and defined important terms. He defined

quality as the degree of excellence in some trait relative

to a standard, usually the wild insect, or more simply,

fitness for use. Problems with measuring traits of wild

insects, however, sometimes made the use of such standards

infeasible. Therefore, comparison with an internal standard

or the untreated laboratory insect was recommended. While

its use was not necessarily a measure of fitness in the

field, the internal standard did provide a reference value

for detecting changes in stock quality. Chambers stressed

the importance of regular measurements of performance traits

using a variety of parameters, and he stated that success of

the release program was an inadequate indicator of quality

because it did not necessarily ensure continued production

of quality insects. Further, the use of only one or two


traits to estimate field performance may be inaccurate and

misleading. Rather, he suggested a series of laboratory and

field tests to evaluate overall quality, or how well the

insect performs its function in the field. Those tests that

measure adaptedness (physiological functions, genetic

divergence), motility (capability and propensity to move),

and reproductive success are most important.

A symposium entitled "Solving Insectary Production

Problems" was held in 1975 as part of the Eastern Branch

meetings at the national conference of the Entomological

Society of America. Two papers are of special interest.

First, Huettel (1976) summarized some of the tests and

techniques used to monitor quality of colonized insects.

They will be discussed together with those suggested by

Chambers (1975). Changes in gene frequencies of colonized

populations were monitored by allozyme electrophoresis which

detects the relative movement of charged enzyme molecules in

an electric field. Life history measurements were made

directly on developmental rates, weights or size, fecundity,

fertility, longevity, and mortality. Physiological

processes were monitored by exposing samples of insects to

stressors such as extreme temperatures, relative humidity

(RH), or insecticides, and by determining uptake rates of

oxygen or production rates of carbon dioxide. Testing the

insects' ability to locate food sources and to avoid

predators measured survival potential. Periodicity of

activity was determined by actographic techniques that used


various types of physical sensors to detect movement of

caged insects. Rhythmicity was also established by

analysing carbon dioxide output or recording sounds produced

by the insects. Motility was monitored by measuring flight

ability and propensity. Flight behavior was studied using a

flight mill which consisted of a horizontal rotor attached

to a perpendicular support in such a way that it was free to

rotate. An insect was tethered with quick-setting glue at

one end of the rotor and an opaque flag was attached to the

other. As the insect "flew" in circles, the flag passed

between a photocell and a light source, opened an electrical

circuit, and produced a mark on a strip chart recorder.

Information was measured on frequency, velocity, endurance,

and distance flown. Wingbeat frequencies were determined

stroboscopically or by acoustical analysis. Flight

propensity was measured by startling insects with a stimulus

that would normally cause flight, such as light impinging on

dark-adapted adults. Orientation to visual and chemical

stimuli was also monitored. The electroretinogram (ERG)

technique, using minute electrodes implanted in the eye,

recorded stimulation of optic fibers. Thus, spectral

sensitivities and threshold response for monochromatic light

of varying intensities were measured. Preference tests

using various colors were used to evaluate behavioral

responses to visual stimuli. Olfactometers and flight

tunnels were useful in monitoring the response of insects to

chemical stimuli, such as host plant attractants and


pheromones. The qualitative, quantitative, and periodic

release of pheromone production was also monitored.

Reproductive success was measured by the ratio test where

various proportions of treated (e.g. irradiated) and

untreated males were caged together with untreated females.

Percentage fertility indicated the relative degree of

competitiveness of the treated insect.

The second paper of interest at the 1975 symposium

discussed the potential of genetic improvement of mass

reared insects (Hoy 1976). Successes in silkworm and

honeybee production were cited as examples of progress that

has been made. Hoy stated that before genetic gains are

possible, effective rearing methods must be developed,

desirable attributes must be clearly defined, and adequate

genetic variability must exist. She concluded that genetic

aspects of the laboratory colony must be considered during

development of the production program and that genetic

quality is aided by colonizing representative samples from

the field, maintaining colonies under natural conditions,

and preventing the intense genetic selection that usually

occurs in early generations of laboratory populations.

In 1977, two major publications were responsible for

the general acceptance of quality control concepts and for

their implementation into mass-production programs. First,

an article entitled "Quality Control in Mass Rearing", by

Chambers (1977), appeared in the Annual Review of

Entomology. The mechanisms of change in laboratory insects


and the need for quality control were summarized so that

entomologists, even those unfamiliar with insect production

systems, could understand the scope and importance of the

problem. As such, it may have been one of the most

important milestones in the development of insect quality


In addition, a handbook on quality control and

monitoring techniques was published for individuals involved

in the production of fruit flies. The book, "Quality

Control: An Idea Book for Fruit Fly Workers", was edited by

Boller and Chambers (1977). In it, the concept of

industrial control charting was introduced as a graphical

means to identify changes in the quality of insects. With

this aid, significant deviations in traits measured over

time could be detected and used to help identify and solve

problems in the production system. Many tests for

monitoring essential traits in fruit flies were presented.

Also, Leppla et al. (1977) discussed evaluating insect

adaptedness using life tables and other developmental data.

They emphasized the need for a thorough understanding of all

rearing operations and the effects they have on all

developmental stages. The idea book simplified the

application of quality control and provided the basis for

the development of standardized programs.

In 1979, the IOBC and other national and international

organizations sponsored an International Course on Quality

Control in Ceratitis capitata (Wiedemann) in Spain. It was


organized by E.F. Boiler, C.O. Calkins, D.L. Chambers, and

N.C. Leppla. In this training program, the RAPID Quality

Control System was introduced, evaluated, and accepted as

international test and evaluation procedures for the

Mediterranean fruit fly. RAPID was designed for quick and

simple measurement of key insect traits in laboratory and

field tests. They measured a developmental characteristic,

motility traits, and sexual activity. Pupal size was used

as an indicator of stressful conditions within the

production system and as a correlate of flight potential. A

machine divided insects into several categories according to

diameter. Puparia were placed between two sloping steel

cylinders with the distance between cylinders increasing

slightly from top to bottom. Smaller diameter insects fell

through to collection boxes near the top and larger ones at

the bottom.

The RAPID system tested both flight ability and

propensity. A test of the first trait measured the ability

of adults to fly from a container that consisted of a

cardboard tube coated on the inside with paint to reduce

friction and prevent flies from walking out. The tube was

placed on top of a petri dish containing a sample of

puparia. After emergence, able flies left the tube and

remaining insects were classified according to eclosion

success, wing deformity, and pupal mortality. Flight

propensity was measured using a startle test in which

dark-adapted flies were briefly exposed to an overhead


light. The number of insects arriving at the light

indicated their propensity to take flight. A startle

activity index was calculated giving more weight to those

flies that arrived at the light.

Studies of sexual behavior were divided into female

response to pheromone and mating propensity. Female

response was measured in an olfactometer where air was blown

across pheromone-producing males and into a large chamber

containing virgin females. Respondents flew upwind to the

pheromone source where they were captured in a cone trap and

counted. Mating propensity was assessed in plexiglass cages

where the number of mating pairs was recorded during each

10-min period for 60 min. A mating index was calculated by

giving more weight to flies that mated faster and less to

those that paired more slowly. Ratio tests using different

strains were conducted in the same manner except that

different strains or treatments were marked for


The performance of fruit flies was also evaluated in

field tests during the training course. Olfactometery and

mating tests were conducted in large walk-in cages

constructed of plastic screening. To measure female

response to pheromone, small cages of virgin males were

placed in a circular pattern on the floor of these field

cages. Females were then released and collected every 15

min from the surface of the male cage to which they were

attracted. Different strains, or treated and untreated


insects, were tested at the same time by differentially

marking each group. Periodicity of pheromone production and

of responsiveness by females was also determined. The field

mating test was similar to that conducted in the laboratory.

Virgin male and female flies (1:1) were simultaneously

released into the field cage. At 15-min intervals during

the 120-min test, observers inside captured mating pairs.

The mating index and an isolation index were calculated.

The latter was a measure of assortative mating when

different strains were used and was calculated as the ratio

of interstrain to intrastrain mating. Release-recapture

tests measured the ability of the flies to disperse in the

field. Flies were released from a central location and were

collected in trimedlure-baited traps placed at varying

distances from the release site. Estimates of the density

of natural fly populations were determined using the

resulting data. The uses of these tests during the training

session in Spain and other countries were described by

Boller et al. (1981) and Chambers et al. (in press).

A conference entitled "Advances and Challenges in

Insect Rearing," chaired by R.F. Moore, was held in 1980.

It brought together expertise in all areas of insect

production, including genetic aspects of colonization,

artificial diets, containment of insects, engineering,

automation, control of pathogens and health hazards,

management, and quality control. All papers impinged on

quality insect production to a greater or lesser degree but


several dealt specifically with quality control. Chambers

and Ashley (in press) advanced the use of process control

charting to detect abnormal changes in production processes.

A second paper, by Webb (in press), stressed that

operational quality control can occur only when pertinent

information on essential traits is returned to the

production program in a typical feedback system. In

addition, relevant production data (temperature, RH,

lighting, procedures, diet) must be recorded to complete the

feedback loop. Papers presented at this conference will be

published as a USDA Technical Bulletin entitled "Advances

and Challenges in Insect Rearing" edited by E.G. King and

N.C. Leppla.

The quality control effort culminated last year with

the formation of a Global Working Group on Quality Control

under the auspices of IOBC. The workshop, held in

Gainesville, Florida, was organized by T.R. Ashley, E.F.

Boller, C.O. Calkins, D.L. Chambers, D.A. Dame, and N.C.

Leppla. Discussions centered on pests of man and animals,

fruit flies, lepidopterans, data management, and model

systems. This workshop brought together scientists

interested specifically in quality control, provided a

comprehensive outline or frame of reference for unified

discussions rid coordination of future activities, and

enhanced the stature of the quality control effort. A

report on attempts to control the Mediterranean fruit fly in

California illustrated the need for simple, standardized


tests and for organization of quality control efforts,

especially in emergency situations. Major recommendations

for future activities in quality control focused on genetic

sexing and marking techniques; improved laboratory strains;

data analysis, storage, and retrieval systems;

distinguishing between genetic and non-genetic causes of

quality reduction; verification of laboratory measurements

of performance with field testing; correlation of easily

measured variables with behavioral traits that are more

difficult to quantify; development of reliable standards;

and closing the feedback loop to ensure operational quality

control. The conference concluded with inspection of the

screwworm and Mediterranean fruit fly mass-production

facilities in Mexico.

Factors Influencing the Quality of Colonized Insects

Quality of laboratory-reared insects may be influenced

by the collection of founders, subsequent genetic changes,

and environmental factors in the production facility.

Collection strategies are based on the geographic, seasonal,

and temporal distributions of wild insects. Their

populations are not homogeneously distributed and areas of

differential variability occur in relation to compatability

with environmental conditions (Remington 1968). For

example, in a constant environment, the optimal form of an

enzyme will be the best functioning homodimer, while in an


environment where the probability of divergent conditions is

greater, heterozygous individuals may be at an advantage

(Johnson 1974). Thus, an enzyme in Drosophila melanogaster

showed clinal variation with latitude. Alcohol

dehydrogenase (ADH), which catalyzes the reduction of

acetaldehyde to ethanol, is monomorphic in Florida where

environmental conditions are relatively constant, but is

more variable at the northern limits of the species (Vigue

and Johnson 1973). Johnson (1976) also found clinal

variation in polymorphic genes of Colias spp, a group of

butterflies in the Pieridae family. He attributed these

differences to different temperature optima for the


As a result of these differences in distribution,

founders should be derived from the target population in the

environment into which future generations will be released

and where they must survive. For example, wild corn earworm

moths, Heliothis zea (Boddie), from the area of release in

St. Croix, VI, were incorporated into existing laboratory

cultures (Young et al. 1975). The new laboratory strain was

more competitive than the old one from Georgia in finding

and mating with wild females in St. Croix. In addition,

males of the new strain remained more competitive even after

six generations in the laboratory. In another example, a

parasite of the walnut aphid, Chromaphis juglandicola

(Kaltenbach), was collected in southern France, shipped to

California, mass produced, and released throughout the


state. Those that survived did so only in areas near the

coast where the climate was similar to that of southern

France. For the hotter, drier areas parasites were obtained

from a climatologically similar area of Iran. Subsequent

production and releases in California were successful in

controlling the aphid (Van den Bosch and Messenger 1973).

Large numbers of wild insects should be used to

establish a laboratory colony to minimize the potential for

founder effect that reduces genetic variability and results

in genetic divergence from the source population. However,

large collections do not always ensure colonization success,

especially when the insects are taken from the more

homogeneous marginal areas of their range, near the

environmental limits of the species. Those selected from a

more central location will tend to be outcrossed, with more

heterogeneity (Remington 1968). Mackauer (1976) stated that

a founder colony of about 500 individuals from a population

of high average size plus long-term persistence, collected

over a "conveniently" large area, should include a level of

genetic diversity that adequately characterizes the parental

population. In the tobacco budworm, Heliothis virescens

(Fabricius), representative allele frequencies have been

described using only 30-40 individuals, and therefore a

colonizing population of 50-100 pairs should be adequate

(M.D. Huettel pers. comm.). For establishing colonies that

will ultimately interact with conspecifics in the field,

MacDonald (1976) suggested that as many insects as possible


should be collected from the eventual target area, and

collections should be made over several generations and

during all discernable periods of daily activity.

Once founders have been introduced, other factors act

to change the colony. Inbreeding results in a greater

expression of harmful traits because of the increasing

number of homozygotes for deleterious recessive alleles.

Random genetic drift is a chance difference between gene

frequencies of a parental generation and the ones

represented by their progeny. The potential for drift

depends on the size of the population, the selective value

of the alleles (fitness), mutation pressure, and gene flow.

Small, isolated populations like those in the laboratory are

especially susceptible to drift and this may lead to

fixation of one allele while its complement is completely

lost from the population, regardless of its adaptive value

(Herskowitz 1979). Under laboratory conditions, drift can

result from a small founder population and periodic

fluctuations in colony size (Mackauer 1976).

Directed selection for individuals best suited for the

new conditions generally occurs in initial generations.

During this period, the insects are forced through

"bottlenecks" that alter and reduce the level of genetic

variability (Boller 1972; 1976). The first bottleneck

occurs in the larval stage in which mortality eliminates

those individuals that cannot survive under the artificial

regimen of diet, temperature, photoperiod, relative


humidity, density, and containment. The second occurs as

similar artificial conditions reduce the reproductive

potential of founders, as selection acts on adult behaviors

that are directly related to insect quality.

Yields of colonizers are low at first when selection is

most intense, but after about five to seven generations they

improve as the number of laboratory-adapted individuals

increases. For example, Raulston (1975a) found that newly

colonized tobacco budworms mated less frequently and had a

lower percentage of mating than insects in colonization for

many generations. This difference in mating behavior,

however, disappeared after six or seven generations in the

laboratory. Similarly, Leppla et al. (1980) reported that

an increasing number of matings and mating frequency of wild

cabbage looper moths, Trichoplusia ni (Hubner), were

indicative of progressive adaptation to laboratory

conditions. They concluded that supernumerary matings may

have contributed significantly to increased fecundity and,

therefore, to success under laboratory conditions.

To reduce the effect of selection on initial

generations, Boller (1972) suggested using rearing methods

that give relatively high yields in the first generation and

not selecting too early for a standardized strain. Further,

various natural stimuli, including extracts of natural foods

added to artificial diets, kairomones, and natural shape and

color of rendezvous sites, etc., should be used to maintain

specific behavioral traits. Finally, insects in culture


should be made to perform searching, orienting, flying, and

mating behaviors in a relatively natural environment.

Laboratory rearing may result in other types of

physiological deterioration that are often expressed as

delayed maturation, increased mortality, and reduced

fecundity (Mackauer 1972). This type of decline in quality

has been described for the predaceous green lacewing,

Chrvsopa carnea Stephens (Jones et al. 1978). Developmental

time of immatures increased, and egg viability, food

consumption, searching ability, fecundity, and longevity of

adults decreased with time in culture. No recovery phase

occurred in this species, and it was therefore recommended

that insects intended for release should not be reared for

more than six generations in the laboratory.

Genotypes or gene expression of laboratory-adapted

individuals may differ from the original populations. For

example, variation in diet can influence the allozyme

patterns in insects (Sluss et al. 1978). Two groups of

tobacco budworms reared on different artificial diets were

found to have differences in allele frequencies at three

loci, including two that code for hexokinases. These

results were attributed to differences in sucrose content of

the diets (hexokinase catalyzes the breakdown products of

sucrose, glucose and fructose, in glycolysis). Gutherie and

Carter (1972) found that larvae of the European corn borer,

Ostrinia nubilalis (Hubner), reared continuously on

artificial diet, lost the ability to survive on their


natural host. However, survival equaled that of wild borers

after they were backcrossed to native insects. Whitten

(1980) reported differences in allozymes at the alpha

glycerophosphate dehydrogenase locus in the screwworm fly.

The implication of this change is described below.

Environmental conditions in the laboratory may cause

alterations in phenotype without associated changes in

genotype. For example, density of insects in larval rearing

containers reduces pupal weights in many species (Henneberry

and Kishaba 1966; Sullivan and Sokal 1963; Barbosa et al.

1972). Peters and Barbosa (1977) have summarized the

effects of density on fecundity, cannibalism, developmental

rate, behavior, and disease susceptibility. Possible causes

for density effect include mechanical stimulation by crowded

individuals, amount and quality of food, and presence of

insect-produced chemicals such as toxic factors and growth


The diet on which insects feed can alter growth,

metamorphosis, reproduction, behavior, and defense (Hagen

1974 and references therein). Nutrient balance appears to

be an important dietary factor. House (1965) found that

larvae of the sphingid, Celerio euphorbiae (Linnaeus),

feeding on an imbalanced diet ate less, were less efficient

in food conversion, and gained less weight than larvae

feeding on a diet of normal balance. Webster and Stoffolano

(1978) reported that female apple maggot flies, Rhagoletis

pomonella (Walsh) fed a protein/sucrose diet showed greater


development of ovaries, follicles, and accessary glands

compared to a similar group given the same diet but without

protein. Adult diets have been shown to influence the

length of life for the codling moth, LasDeyresia pomonella

(Linnaeus) (Howell 1981). Materials added to artificial

diets also affect insect quality. Singh and House (1970)

tested 21 agents commonly added to diets to reduce microbial

contamination. Using larvae of a sarcophagid fly, Agria

affinis auct. nec Fallen, they found that the length of

larval development is inversely proportional to the dietary

concentration of antimicrobial compounds. Successful

pupation, adult emergence, and survivability were also

adversely affected.

The effect of temperature on developmental rates of

poikilotherms is well known (e.g. Butler and Henneberry

1976). Briese (1980) examined the effects of temperature on

fertility, weight, survivability, fecundity, and sex ratio

of the potato moth, Phthorimaea operculella (Zeller).

Melanin formation in the integument of saltmarsh

caterpillars, Estigmene area (Drury) was found to increase

at higher rearing temperatures (Fye 1979). High

temperatures were shown to cause abnormal wing development

in the cabbage looper (Grau and Terriere 1967). However,

this effect was moderated when the larval diet was

supplemented with a source of polyunsaturated fatty acids

(Dadd 1973).

Photoperiod and cyclical fluctuations in temperature

can alter developmental and reproductive characteristics.

For example, in continuous light, adults of the tobacco

budworm and black cutworm, Agrotis ipsilon (Hufnagel) showed

increased irritability, inability to separate after

copulation, a shortened life span, and a decrease in

fecundity. Two days after establishment of a 14 hr light:10

hr dark photoperiod, normal activity returned (Fisher

unpublished). Messenger (1964) reported that a laboratory

strain of the spotted alfalfa aphid, Therioaphis maculata

(Buckton), held under fluctuating temperature developed

faster and had improved fecundity and adult longevity

compared to a similar group reared under constant

conditions. On the other hand, larvae of a laboratory

strain of the corn earworm reared under constant temperature

developed faster but had lighter pupae and greater fecundity

than a group reared under fluctuating conditions (Fisher


Rearing procedures and post-production treatments can

affect adult behavior and cause reductions in insect

quality. For example, the process of agitation used to

separate puparia from their diet caused a droopy-wing

syndrome in colonies of Mediterranean fruit flies, Oriental

fruit flies, Dacus dorsalis Hendel, and melon flies, D.

cucurbitae Coquillet (Ozaki and Kobayashi 1981). The

syndrome resulted in reduced eclosion rates and increased

numbers of non-fliers as intensity and duration of agitation


increased. Five-day-old puparia were found to be the most

sensitive to agitation (Ozaki and Kobayashi 1982). The

detrimental effects are well known for treatments, such as

sterilizing doses of irradiation or chemosterilants, on the

mating competitiveness of laboratory insects (Noblet et al.

1969; Snow et al. 1972; Brower 1978; and Villavaso 1981),

sperm transfer (Holt and North 1970; Snow et al. 1970), and

flight (Shepard et al. 1968). However, to decrease rearing

costs, dieldrin was added to the larval diet of the

calliphorid, Lucilia cuprina (Wiedemann), to selectively

kill females that were of no use in an autocidal release

program. Males were protected by a Y-autosome translocation

for dieldrin resistance. Laboratory assays of flight

activity (actograph), sexual competitiveness (ratio tests

evaluated by number mating), and visual sensitivity (ERG)

showed no differences between treated males and internal

standards, but treated males did not disperse as well when

released in the field (Smith et al. 1981).

Comparison of Colonized Insects With Wild Standards

Tests have shown a wide disparity between laboratory

and wild insects. To illustrate, White and Mantey (1977)

observed that male laboratory codling moths were more

sedentary than wild ones in a field release. Similarly, the

suspected inability of laboratory screwworm flies to

disperse in the field was cited as the reason for reduced


effectiveness of a release program in the mid-1970's (Bush

1979). This problem was attributed to rearing at a high,

constant temperature to increase production efficiency.

Selection for a laboratory-adapted electromorph, GDH2,

occurred in preference to GDH1, the form that occurs almost

exclusively in wild flies. This enzyme is critical in the

transfer of reducing equivalents to the glycerol phosphate

shuttle during energy production in flight muscles. As a

result, released flies were unable to disperse and mate.

Sharp (1976) reported differences in time of flight,

distance flown, and flight velocity between a wild strain of

Caribbean fruit fly, Anastrepha suspense (Loew), and one

reared in the laboratory for more than 80 generations.

However, flight ability of the wild strain declined to the

level of the laboratory strain by the third generation.

Lab-adapted cabbage looper larvae dispersed less than wild

larvae in field cages (Leppla and Guy 1980).

The diel periodicities of laboratory insects may differ

from their native counterparts. In a field release of

tobacco budworm adults, Raulston et al. (1976) recovered

twice as many wild as laboratory males in traps baited with

virgin females because laboratory males became active 2 hr

later. Turner et al. (1977) found that wild cabbage looper

moths respired nearly twice as much carbon dioxide compared

to laboratory moths. Wild insects had a bimodal pattern of

nocturnal activity, while laboratory moths were unimodal.

It was suggested that this reduced activity by laboratory


insects during scotophase may mean reduced dispersal

capacity, host-seeking capabilities, and mating

competitiveness with the field population. On the other

hand, a lower metabolic rate in the laboratory strain may

have resulted in increased fecundity, fertility, and


Visual sensitivity may decline in colonized insects.

Agee and Chambers (1980), using ERG, found wild

Mediterranean fruit flies that developed on fruit as larvae

required 0.1 microwatt/cm20f light energy to stimulate a

response, whereas a group of laboratory flies reared for

four generations on unnatural bagasse diet required a

stimulus of about 0.65 microwatt/cm Vision declined

further if mold occurred on the diet or if agar replaced

corncob meal. Histological examinations revealed

differences in the ultrastructure and cellular organization

of the ommatidial units of the compound eyes, with the

rhabdomers of the wild flies well organized and full bodied.

Conversely, those of the laboratory flies were smaller in

diameter with larger inter-rhabdomeric spaces (Agee and

Davis unpublished). Goodenough et al. (1977) observed that

the vision of mass-reared screwworms was three times less

sensitive than that of wild flies.

Pheromone production and response to it also may be

altered during colonization. Fletcher et al. (1968) found

that females of an 11-year old colony of screwworms

responded much more aggressively to pheromone extracts than


the recently colonized strain. Selection pressure for 11

years probably favored a colony of flies that readily mated

in dark, crowded conditions. Pheromone responses of male

gypsy moths, Lymantria dispar (Linnaeus), reared for one or

more generations in the laboratory on artificial diet were

compared with moths reared from field-collected pupae

(Richerson and Cameron 1974). In laboratory bioassays, wild

males were more responsive to disparlure than were

laboratory males. In a field cage, wild males were more

successful in orienting to pheromone produced by wild

females than were laboratory males. Also, a greater number

of wild males made initial contact with the female,

regardless of whether she was laboratory-reared or wild.

Finally, in choice tests, both laboratory and wild males

overwhelmingly selected wild females for mates. In another

study, however, Waldvogel et al. (1982) found that male

laboratory gypsy moths responded to a synthetic pheromone

source in the same way as wild moths. Lab females, however,

showed no periodicity of emission and a large number failed

to release any detectable quantities of pheromone.

Conversely, wild females showed a diel periodicity of

pheromone emission (Richerson and Cameron 1974).

Reduced pheromone production was observed by Miller and

Roelofs (1980) using the red-banded leafroller, Argrotaenia

velutinana. They found that females from a strain reared

for 30 years in a greenhouse produced less pheromone than

did wild females, although body weights were the same.


Also, the percentage of (E)-11-tetradcenyl acetate in the

pheromone blend was lower in the laboratory strain. Minks

(1971) also found alteration in pheromone production in an

inbred stock of the summerfruit tortrix moth, Adoxophyes

orana (F.v.R.). Two strains were colonized simultaneously

and reared on similar artificial diets. One strain was

infused annually with 100-200 wild insects, produced

500-1500 moths per month, and received ascorbic acid in the

larval diet. The second strain was inbred, was reared at

greater densities to produce 500-1000 moths per day, and did

not have vitamin C in the diet. In field assays using traps

baited with virgin females from each strain, more wild males

were captured by females from the first strain than by those

of the inbred strain. Results of laboratory bioassays

indicated that the extractable pheromone content in the

in-bred line was much lower than that of the other females.

Legget and Moore (1982) found that the number of female boll

weevils, Anthonomis grandis grandis Boheman, responding to

grandlure traps was a function of the larval or adult diet.

Either high protein or low sugar or the combination of high

protein-low sugar was associated with increased response to

the traps.

Laboratory rearing may affect mating behavior and

reproductive physiology of colonized insects. Fye and

LaBrecque (1966) demonstrated that wild and laboratory

female house flies, Musca domestic Linnaeus, mated more

frequently with males from their own strain. LaChance et


al. (1975) found that native male pink bollworms,

Pectinophora gossyoiella (Saunders), transferred normal

amounts of eupyrene nucleatedd) sperm more often than a

strain reared for nearly eight years on artificial diet.

The duplexes of wild males contained more eupyrene sperm

bundles which apparently left the testis and descended to

the duplex region more rapidly than in laboratory males.

However, laboratory males were still able to inseminate

females with adequate amounts of sperm. Codling moth males

reared in the laboratory for four years and released in the

field were observed to mate less frequently with wild

females than did wild males (White and Mantey 1977). Males

of another lepidopteran, Adoxophves orana (F.R.) reared in

the laboratory for six years and irradiated as adults were

only 0.58 as competitive as wild males in tests conducted in

cages placed over apple trees (Denlinger et al. 1973).

Because no difference was demonstrated between laboratory

and wild males reared for one generation on artificial diet,

the authors concluded that the laboratory males were not

genetically inferior, but that the rearing procedure,

probably the artificial diet, caused the decline in

competitiveness. Henneberry and Clayton (1981) reported

that irradiated and unirradiated laboratory-reared male pink

bollworm moths failed to mate as frequently with native

females as with laboratory females. On the other hand,

laboratory females, treated or untreated, mated with equal

frequency to either laboratory or wild males.


Oviposition behavior of insects may be modified in a

laboratory environment. Greany and Szentesi (1979) compared

ovipositional responses in wild and laboratory Caribbean

fruit flies and found that wild flies selected black domes

almost exclusively, but laboratory flies showed no

preference between black and white ones. Wild flies showed

an almost absolute preference for dome-shaped substrates,

whereas laboratory flies deposited 12% of their eggs onto

flat discs. This unnatural behavior may have been caused by

selection for females that oviposit on flat surfaces

provided for them in the rearing program. Because many

tephritid flies use fruit as a rendezvous site for mating, a

change in fruit-finding capability by laboratory flies could

adversely affect their mating competitiveness.

Longevity of males from a colonized strain of tsetse

fly, Glossina moristans orientalis Vanderplank, was

significantly less than that of a wild strain (Dame et al.

1970). The differential mortality apparently resulted from

reduced feeding activity of laboratory flies in holding

cages that were larger than those in which they had been

reared. Wild flies fed normally in the large cages.

Identifying and Interpreting Changes in Insect Quality

To monitor the changes in colonized insects, the use of

industrial quality control procedures has been proposed.

The most important of these is the process control chart,


developed by Shewart (1931), which employs statistical

methods to evaluate changes over time. Such charts are

based on separating variation in sample populations into a

random component, or unidentified variation inherent in

anything being measured, and a component that results from

specific assignable causes. To construct a quality control

chart, enough samples are taken to determine random

variability (standard deviation = s.d.) and distribution.

With this information, the significance of a change in

quality can be determined. Even if the distribution is not

normal, constants used to determine chart characteristics

are so stable that those for normal curves may be used

unless the distribution is extremely skewed (Burr 1967).

The chart consists of a center line, the overall mean of the

samples, and the control limits, usually 3 s.d. units,

which are marked above and below the center line. The

ordinate is the value of the trait being measured and the

abcissa is successive sample numbers.

If only chance variation is present, no definite

patterns appear on the chart over time and values are

balanced above and below the center line. But, if an

assignable cause is altering the degree of variability,

sample values will fall outside the distribution limits and

it can be assumed, with a high degree of confidence, that

they came from populations treated differently. Such values

are said to be out of control.

Several types of charts are commonly used in industry.

Average (X) and range (R) charts show variation in sample

averages and ranges. These charts, often displayed

together, utilize variable data and require relatively small

sample sizes to give good results. Measurement of variable

data, however, may be time consuming. P-charts are used for

discrete or attribute data and show the percentage of sample

that is defective or that must be rejected. These charts

use data that are easily measured by a yes or no response

but they do require large numbers of observations for

accuracy. Operating characteristic, or OC charts define the

sensitivity of a control chart for detecting changes in

processes or product quality.

Because standard deviation varies inversely with Vn,

larger sample sizes increase the sensitivity of the chart by

moving limits closer to the control line. But large samples

are often expensive and sampling strategies should be

balanced between economy and the cost of an undetected shift

in processes. Generally, if it is more costly to sample and

test than it is for a change in product quality to go

undetected, it is better to take small samples more

frequently. Samples must be taken from production output

that has been treated similarly. Thus, if assignable causes

are present, the differences will be apparent between

samples rather than between observations within a sample

(Duncan 1959).

Interpretation of control charts begins with

identification of out-of-control points and upward or


downward trends. If this occurs on an X-chart, it means a

general change in processes has affected all samples.

Similar points on an R-chart show a change in process

uniformity. A run of non-randomness is indicated by four

out of five successive points beyond 1 s.d., two out of

three points beyond 2 s.d., or eight successive points on

one side of the center line (Bicking and Gryna 1974).

Process control charts provide information as to when, and

to what degree, modifications of processes should be made,

and they assist in identifying factors in production that

affect product quality (Charbonneau and Webster 1978).

Chambers and Ashley (in press) emphasized that quality

control is an active approach to maintaining and improving

insect quality through control of processes. A process is

the interaction of people, materials, equipment, and

facilities required to make a product. Thus, diet

preparation, egg surface-sterilization, larval development,

and pupal harvest procedures are examples of processes in a

rearing program. Processes can also include colonization

and infusion strategies designed to ensure adequate genetic

variability. Total quality control is based on a systematic

approach to solving insect production problems. It works

best when thoroughly integrated with the production program

and its effectiveness is dependent on communication between

producers and users.

The basic steps in establishing a quality control

program have been outlined by Boiler and Chambers (1977) and


Leppla et al. (1977). First, objectives of the rearing

program are defined by characterizing the target population,

prioritizing critical behaviors, and determining production

capacity and schedules. Next, standards are established

against which insect quality is measured. They should be

precise and descriptive of both the average and range of

acceptable performance levels for important performance

traits. For more complete analysis of overall quality,

standards should be established for various components of

physiology, morphology, genetics, and behavior. Then,

methods to quantify these traits are developed. They should

be simple, economical, reproducible, and able to be

performed by anyone. Finally, the quality control program

is implemented with regular monitoring and the use of

process control charts to provide the feedback necessary to

ensure production of quality insects.

Characteristics of Lepidopteran Rearing

Mass production programs for lepidopteran species can

be characterized by the composition, preparation, and

methods of dispensing artificial diets; containerization;

procedures for handling insects; facility design; and

quality control. Traditionally, larval diets have consisted

of a mixture of primary ingredients such as wheat germ,

casein, bean meals, Brewer's yeast, and secondary materials

such as carbohydrate and unsaturated fatty acid sources,

vitamins, minerals, and antimicrobial agents in an agar

suspension (Singh 1977). The expense of many of these

ingredients, however, limited large-scale rearing, but the

elimination, substitution, or reduction of the expensive

ones has allowed more economical production. For example,

casein was replaced with wheat germ (Raulston and Shaver

1970) or soy flour (Stewart in press), corncob grits reduced

the agar requirement (Brewer and King 1979), and the

concentration of vitamins was reduced without detrimental

effects on insects (Raulston 1975b; Brewer and Tidwell

1975). The use of tetracycline and formalin in two

artificial diets for the beet armyworm, Spodoptera exigua

(Hubner), and tobacco budworm was eliminated and the

concentration of primary ingredients was significantly

reduced without altering insect quality (Fisher

unpublished). Toba et al. (1970) proposed a rating system

to evaluate the suitability of diets modified to reduce

costs. Criteria such as time to pupation, pupal yield,

percentage of emergence, fecundity, and fertility were used

to calculate an average suitability value expressed as a

percentage of the results achieved with the control diet.

Artificial diets are prepared in mixers or steam

kettles with capacities from 4 to 775 liters. Ingredients

may be sterilized prior to mixing to kill microbes, render

certain toxic components inactive, reduce moisture content

to improve storage properties, and concentrate nutrients

(Vanderzant 1974). Other systems employ a


flash-sterilization technique in which diet is moved through

small-diameter tubes, quickly heated to 145-1600 C, and

sterilized within two to three minutes (Griffin et al.

1974). The diet is then pumped directly to a station where

it is aseptically dispensed into rearing containers. This

system is in use at the USDA facilities, Stoneville, MS, for

the production of Heliothis spp. (Gantt et al. 1977).

Requirements for larval containers are determined by

the species being reared and its habits (Burton and Perkins

in press). Cannibalistic species must be reared

individually or at low densities. Glass shell vials,

disposable 30-ml plastic cups, or 0.2 to 0.5-liter cardboard

cups are often used and eggs or neonate larvae are usually

transferred by hand to each container. Burton and Cox

(1966) used a food packaging machine to dispense 30-ml cups,

fill each with diet, add larvae, and cap them at a rate of

2,000 to 4,500 per hour. But, because of the cost of cups,

caps, and labor, these containers were unsuited for

mass-production. Thus, larger, reusable containers were

employed. For example, a 6-liter, round, plastic container

was developed to rear 8,000 to 10,000 codling moth pupae per

day. To reduce dessication, the diet was lightly brushed

with melted paraffin that was perforated with small holes to

allow neonates to reach the diet. Pupation occurred in

corrugated cardboard strips taped to the perimeter of the

container (Fisher in press, a). Rye et al. (1981) described

a similar, but rectangular container for weekly production


of about 6,000 cabbage looper pupae. To further improve

production efficiency of cannibalistic tobacco budworm

larvae, Raulston and Lingren (1972) used polystyrene

light-diffusing louvers imbedded in diet in fiberglass

trays. Each rearing unit consisted of individual 1.2-cm

rearing cells and could accommodate over 900 larvae. An

aluminum template, with rows of small depressions

corresponding to the location of the cells, was used to add

eggs. Loose eggs poured onto the template were gently

shaken until three to four filled each depression. After

excess eggs were removed, the cell unit was inverted over

the template, both were reinverted together, and the eggs

were gently tapped into the cells. This process has been

automated using a conveyor belt to move diet trays under

several small tubes aligned over the rows of cells. A

hopper located above, metered eggs into each tube at a rate

of two to three eggs per cell. Polypropylene cloth was then

placed on top of the cell unit to allow air flow and prevent

larval escapes (Harrell and Gantt in press). This type of

tray rearing has been used to produce over 35,000 tobacco

budworm pupae (Raulston and Lingren 1972) and 60,000

Heliothis spp. pupae (Hartley et al. 1982) per day. In the

latter program, the material covering the cell unit was

changed to a porous, autoclavable polypropylene sheet.

However, mold contamination was a problem with tray rearing

because steam autoclaving melted the cell units and

disinfection solutions were inadequate to achieve

sterilization. Therefore, a cell unit was made of

heat-stable casting resin that was easily cleaned and

sanitized (Fisher in press, b).

An automated containerization system, developed for the

corn earworm was described by Sparks and Harrell (1976). An

in-line, form-fill-seal machine pressure formed high-impact

polystyrene film into a continuous sheet of rearing cells,

dispensed artificial diet (Harrell et al. 1973) and insect

eggs (Harrell et al. 1974) into each cell, sealed the top

with plastic film, and cut the sheet into 4 by 8-cell

sections for stacking in holding rooms. The capacity during

continuous operation was projected to be 160,000 cells per

8-hour day. For the production of pink bollworm, a

non-cannibalistic species, in yields of two million per day,

diet was dispensed into trays and stored for three days to

dissipate volatiles and reduce the moisture content (Stewart

in press). The diet was then shredded in commercial

equipment to increase its surface area for feeding and

separating larvae and placed into larval rearing containers.

Various techniques were used to harvest insects from

larval rearing containers. For example, late instar pink

bollworm larvae crawled out of the containers and dropped

below to honeycomb cells where they pupated. In other

species, pupae are harvested directly from the containers.

In small and medium-sized containers, workers usually

collect pupae with forceps. However, Harrell et al. (1968)

designed a machine to mechanically harvest fall armyworm


pupae from 30-ml cups. The cups were placed on a conveyor

belt and crushed to remove lids and dislodge pupae which

were separated from the debris and collected. About 4,000

cups were harvested in an hour, an improvement of 600% over

manual collection. This machine was slightly modified to

harvest corn earworm pupae as well (Harrell et al. 1969).

Polystyrene and resin cell units were simply inverted and

pupae fell into a tray. Pupae produced by the

form-fill-seal machine were also mechanically harvested.

The plastic film covering the cells was stripped away and

the sections were inverted over a wire conveyor belt that

allowed pupae to fall through to a collection tray and that

deposited debris in a trash receptacle. The capacity was

20,000 to 25,000 cells per hour. Harvesting other

lepidopteran pupae may require additional steps. For

example, the cabbage looper pupates in silken cocoons on the

top and sides of the container. After harvest, the pupae

are placed into a dilute solution of sodium hypochlorite to

remove the silk (Henneberry and Kishaba 1966).

Adult eclosion, mating, and oviposition usually occur

in a single cage, its characteristics being determined by

availability of materials, ease of handling, and

reusability. Cardboard ice-cream cartons (3.8 liters) are

commonly used for adults (Shorey and Hale 1965).

Oviposition materials such as nylon organdy, cheesecloth, or

paper toweling are placed over the top of the cartons with

narrower pieces dropped down the sides. Other cages are


made of wooden, stainless steel, or aluminum frames covered

with screen or clear plastic (Ignoffo 1963). Automated

cages have been developed to minimize disturbance to moths

and reduce time required to harvest egg sheets (Knott et al.

1966; Carlyle et al. 1975). For the mass-production of

100,000 Heliothis spp. eggs per week, windowscreen mounted

in an aluminum frame was inserted through a slot in the cage

and used as an oviposition substrate. Its replacement does

not disrupt adults or require that they be anaesthetized

(McWilliams et al. 1981). In pink bollworm rearing,

emergence occurs in a container separate from the one used

for mating and oviposition. Newly-eclosed adults are

attracted to a black light and are picked up in an airstream

that transports them to collection centers in a walk-in

refrigerator where cool temperatures reduce their activity

and permit easy transfer to new cages for mating and


Air-borne contaminants such as wing scales and hairs

represent health hazards to insectary employees and help

disseminate disease organisms (Stewart in press; Wirtz in

press). Thus, various scale-collecting devices have been

constructed. Raulston and Lingren (1972) placed adult cages

on top of a hollow shelf attached by plastic piping to a

blower which gently pulled air through the cages and into a

filtration unit. Pink bollworm cages were attached to a

series of pipes that pulled air from the cages to cyclone

dust collectors located outside the building. Leppla et al.


(1982) described cabinets that enclosed adult cages and

removed scales using an industrial dust collector. Absolute

filters in the air handling system of the gypsy moth

facility removed scales from the air while individual work

stations were equipped with exhaust hoods and a vacuum to

further reduce this hazard (O'Dell et al. in press).

The production facility is the primary component of a

successful rearing program (Leppla and Ashley 1978). Its

design is intended to maintain environmental conditions,

confine insects, control pathogens and dietary contaminants,

isolate specialized areas, allow efficient movement of

materials and products, and provide a proper working

environment (Fisher and Leppla in preparation). Design

criteria are especially important for lepidopteran

facilities because most species are susceptible to a variety

of insect pathogens. In fact, Stewart (in press) stated ".

. .in every instance, failure to meet production quotas [of

pink bollworm] was directly related to the dominating,

detrimental effects of microorganisms". Also, Sparks and

Harrell (1976) indicated that the inability of a facility

design to maintain disease-free insects is a major cause for

failure when the production capacity is increased. The most

effective means to eliminate disease outbreaks is to isolate

critical areas. Thus, preparing diet in an area completely

separated from other rearing activities was the only

practical way to eliminate virus and bacterial diseases from

colonies of tobacco budworm and beet armyworm (Fisher in


press, a). Raulston and Lingren (1972) designed a facility

for tobacco budworm that completely isolated the brood

colony from the mass-production area, provided showers for

incoming employees, and incorporated pass-throughs and a

double-door autoclave for the movement of materials and

products without a backflow of contaminants. In addition,

all rooms except the one for pupal harvest, the dirtiest,

were maintained under positive pressure to exclude


To complement facility design, procedures have been

developed to reduce disease outbreaks. These include

addition of antibiotics to larval and adult diets,

surface-sterilization of eggs and pupae with dilute

solutions of sodium hypochlorite or formaldehyde, regular

sanitation of the facility, regulations restricting

unnecessary movement of personnel, and maintenance of

environmental conditions to reduce stress on the insects

(Sikorowski 1975; in press).

The need for quality control in lepidopteran rearing

programs is clear but usually only consists of monitoring

pupal yields and weight, fecundity, and temperature and RH

in insect holding areas. Other variables, although more

meaningful in quality assessment, are measured in only a few

programs. For example, the quality of mass-reared gypsy

moths was evaluated in laboratory and field tests (O'Dell et

al. in press). During production, routine measurements were

made of survival, eclosion, deformity, and fertility. A


flight tunnel was used to determine male response to

pheromone by measuring the length of wing-fanning periods

and the time of flight in a pheromone plume. Data from

actographs, used to identify activity periods and assess

propensity to fly, were correlated with the

pheromone-response rhythm of released moths. Observations

of mating behavior were made to determine the optimal pupal

age and radiation dosage for obtaining sterile males while

maintaining sexual competitiveness. In the field,

measurements were made of eclosion success, the periodicity

of eclosion, and longevity. Also monitored was the number

of laboratory males responding and dispersing to

pheromone-baited traps. The production of Heliothis spp.

hybrids was monitored at the production facility using

fertility of untreated eggs, surface-sterilized eggs, and

the weights of pupae before and after harvest (Brewer in

press). Other variables measured included pupal diameter,

abdominal rotation, mortality, percent adult emergence and

wing deformity, and longevity. At St. Croix, VI,

postdistribution tests of the hybrids included periodicity

of activities such as feeding and flight, dispersal, mating

competitiveness using mating tables (Snow et al. 1976), and

oviposition (Proshold 1982). For the pink bollworm, tests

of mating ability and longevity identified detrimental

effects of handling and shipping and evaluated the potential

usefulness of moths in the field (Stewart in press).

Virtually all of the quality control tests for


lepidopterans were conducted on a short-term research basis

and were not used to continuously monitor production or

product quality. In fact, procedures for lepidopteran mass

rearing are so time consuming and labor intensive that they

generally leave little time for quality assessment. Thus, a

system is needed that will continuously provide basic data

on insect quality in an economical and reliable manner.

Accordingly, the objectives of this study were to: 1)

identify developmental, morphological, and behavioral traits

that could be used to monitor production processes and to

indicate the relative quality of a lepidopteran species

reared in the laboratory; 2) design and construct equipment

to measure these traits; 3) extract information on behavior

patterns as well as the results of behavior; 4) determine

the efficacy of the testing system; 5) modify rearing

conditions to determine if resultant changes in quality can

be detected; and 6) incorporate these and other data into a

model quality control system for mass-reared Lepidoptera.


Source of Insects and Rearing Conditions

The fall armyworm, Spodoptera frugiperda (J.E. Smith),

was used for this study. It is a significant economic pest

in the southeastern United States beginning in April that

affects much of the rest of the eastern part of the country

after July (Sparks 1979). Average losses exceed $300

million annually and $500 million during years of severe

outbreaks (Mitchell 1979). Fall armyworm larvae are

polyphagous, feeding on a diversity of crops such as

grasses, soybeans, corn, millet, alfalfa, rice, cotton, and

peanuts (Young 1979). This species lacks a diapause

mechanism and its only overwintering sites in the United

States are in the milder climates of south Texas and Florida

(Sparks 1979). Thus, it is an ideal candidate for a

regional approach to early season control using sterile

insect technique (Knipling 1979, 1980).

Fall armyworm eggs are laid in masses on the underside

of leaves. However, in high population densities they have

been deposited indiscriminately on plants, buildings, and

other objects (Thomson and All 1982). Masses are covered

with scales and setae from the body of the female. After


hatch, neonate larvae eat their eggshells and, in response

to positive phototaxis and negative geotaxis, climb to the

uppermost portions of the plant where they secrete a line of

silk to aid wind dispersal or begin feeding on tender

terminal growth (Luginbill 1928; Morrill and Greene 1973).

Early-instars skeletonize leaves of host plants while later

ones eat the entire leaf, often stripping the entire plant.

Feeding occurs at night and larvae hide at the base of

plants under debris during daylight hours. Prepupae are

positively geotactic and usually pupate in the soil,

although pupae have been found in stalks, ears, tassels, and

whorls of corn (Burkhardt 1952). Adult eclosion begins

shortly after sunset and continues until midnight and

feeding on nectar is apparently the only activity performed

during the first night (Sparks 1979). On subsequent nights,

activity of moths begins about dusk. After an initial

feeding period, females begin calling from a location near

the top of the plant canopy. In response to the pheromone,

males orient their antennae toward the source, vibrate their

wings, and fly in the direction of the female (Sekul and Cox

1965). In other cases, males orient to the attractant on

the wing as they fly obliquely upwind. Often several males

respond to the same female, rejected males returning to

flight, sometimes in large groups (Sparks 1979). Mating

behavior of fall armyworm moths has not been reported but,

apparently, males and females mate only once each night and

most mating occurs by midnight. Eggs are deposited during


the early evening. In summer, and at 270 C in the

laboratory, the life cycle is completed in about 30 days.

Lab fall armyworms were produced at the Insect

Attractants, Behavior, and Basic Biology Research

Laboratory, U.S. Department of Agriculture (USDA),

Agricultural Research Service (ARS) in Gainesville, Florida,

where they had been in production for six years without

infusion of wild stock. Original founders consisted of

about 300 insects from the Southern Grain Insects Research

Laboratory, USDA, ARS, Tifton, GA, where they had been in

culture for 15 years (Perkins 1979).

At Gainesville, larvae were reared in 30.5 x 30.5 x

14-cm high plastic containers with snap-on lids modified to

improve air exchange (Rye et al. 1981). Each container

received 1000 g of a pinto bean-based artificial diet (see

appendix for ingredients). A polystyrene grid of 1.3 x

1.3-cm squares was pushed into the diet before it cooled.

This grid separated developing larvae and facilitated the

harvest of pupae. Paper toweling, covered with fall

armyworm egg masses, was glued to the inside of each

container lid. Four containers set up each work day were

dated and labeled according to rack 1 or 2 and location on a

shelf. Locations were designated A, B, C, or D with "A" on

the left end of each shelf.

Larval rearing containers were maintained at 260 + 2 C

and 50 5% RH with a 14 hr light:10 hr dark cycle. After

21 days, pupae were harvested and the approximate amount of


diet surface covered by mold and the number of insects was

determined. The number of insects was divided into three

categories: usable pupae, those that were normal looking and

from which normal adults would probably eclose; abnormal

pupae, identified by some morphological deformity such as

larval exuvia attached to the venter or a constricted

abdomen; and non-pupae, defined as larvae, or insects with a

pupal-like anterior portion and a larval abdomen. Total

number of insects was the sum of all insects in these

categories at the time of pupal harvest. After yields were

determined, usable pupae were held for quality control

testing. Twenty-five pupae were randomly selected from 151

containers between 24 August and 5 November 1981, separated

by sex, and held for limited testing (pupal weight only),

and between 70 and 150 pupae of each sex were taken from

each of two containers randomly selected on 36 harvest days

between 20 April and 30 July 1982 and held for comprehensive


Holding Conditions for Test Insects

In 1982, 30 to 90 pupae from each larval container were

set up in 3.8-liter cylindrical cardboard holding cages with

nylon organdy covers. When more than 70 pupae were set up,

they were divided approximately equally into two

replications. Males and females were placed in separate 107

x 58 x 46-cm-high plywood cabinets while pairs used in


mating studies were held in a third. Conditions within

cabinets were 28 + 20 C and 67 12% RH. Humidity was

maintained individually for each cabinet by centrifugal

atomizers. Lighting, provided by one 30-watt cool white

fluorescent tube per cabinet (480 590 lux), was on a

reversed cycle of 14 hr light:10 hr dark with scotophase

beginning at 10:30 am EST. This enabled testing and

observations of insects during their most active period,

between 2 and 4 hr after initiation of the dark cycle

(Leppla et al. 1979). Twilight conditions were provided by

a poultry timer (Grainger Model 2E023) that operated small

0.25-watt nightlights for 20 min at the end of photophase.

Small fans, used for cooling electronic equipment, operated

during the light cycle to remove heat produced by lights.

All cabinets were located in a controlled environmental


Insects in holding cages were checked twice daily for

adult eclosion and aging of adults began when 25% had

emerged. Adults received 10% sucrose/water solution in

cotton placed on top of the organdy covering the cages.

When all testing was concluded, cages were removed from

cabinets and counts were made of the number of dead pupae,

the number of unsuccessfully closed adults (those unable to

completely separate abdomen or wings from exuviae), and

adult deformities (wing curl).

Characteristics of Immature Insects

All additional pupal data were taken on 5-day-old

insects. In 1981, average weights were calculated using

variable sample sizes while, in 1982, individual weights

were determined for 20 male and female pupae from each

container. Abdominal rotation was determined from another

group of 20 pupae using a protractor-like device (Fig. 1).

To support the pupae, a rubber trough was attached by mastic

to an 11-cm diameter clear plastic disk. Lines, in

increments of ten degrees radiated from the midpoint of the

end of the trough. A pupa was placed in the support trough

with its ventral side up and its abdominal segments

projecting over the edge. Slight pressure on the thoracic

region stimulated the abdomen to move in a circular motion

and the maximal amount of movement to the left and the right

in ten rotations were added to get the degree of abdominal


The inverse time of pupal descent in water, a measure

of buoyancy, was determined for 20 male and female pupae

from each container. A 250-ml graduated cylinder was filled

with 240 C water to which one drop of surfactant (liquid

dishwashing detergent) had been added. Using forceps, a

pupa was placed in the water and slightly shaken to remove

air bubbles adhering to it. The pupa was then released and

allowed to descend to the bottom of the cylinder. Timing

with a stopwatch started when the pupa passed the 230-ml


mark and ended when it crossed the 70-ml line, a distance of

15.3 cm. Pupae were caught at the bottom in a 3.0 x 5.0-cm

cylindrical cage made of 3.0-mm-mesh hardware cloth. A wire

attached to the cage was used to pull it out of the

graduated cylinder.

In preparation for all adult tests, male and female

moths were transferred from holding cages to vials 1/2 to 1

1/2 hr before the end of photophase. At the beginning of

scotophase, vials were placed in the dark where they

remained until testing. Temperature was 25 10 C and RH

was uncontrolled at about 50%. Lighting varied and is

discussed under individual procedures.

Testing of Adult Insects

Wingbeat frequency

Wingbeat frequency (WBF) was determined using a

stationary vacuum tethering device (Fig. 2). Moths were

attached to the distal end of a 100-microliter pipet.

Surgical tubing connected the proximal end of the pipet to a

vacuum source via two needle valves. Valve A, closest to

the vacuum source, remained open a precise amount to ensure

constant pressure. Valve B was used to shut the vacuum off

after each insect was tested. Vacuum pressure at the distal

end of the pipet was about 339.9 g/cmi The vacuum tethering

device was a non-destructive means to determine WBF. That

is, unlike quick-setting glues, vacuum did little more than


more than remove a small patch (about 1.5 mm in diameter) of

hairs and scales from the area of attachment, thereby

enabling the insect to be used for other quality control

tests. Illumination during this test came from a yellow

25-watt incandescent bulb mounted 30.5 cm above the tethered

insect and was controlled by a rheostat to about 13 lux. A

stroboscope (General Radio Strobotac Model 1538A) was used

to determine WBF.

In preparation for testing, 5- to 7-day-old moths were

placed in 20-ml glass vials, chilled in chopped ice for 2.5

min to slow their activity, emptied into a 5-cm-diameter

watchglass, and oriented with the dorsal side up and the

anterior end facing the observer. After the vacuum was

turned on at valve B, the watchglass was lifted and the moth

was slowly brought into contact with the pipet. Attachment

occurred at the center of the metathorax so the longitudinal

and transverse axes of the insect were perpendicular to the

pipet. A plumbob was used to ensure that the pipet was

perpendicular to the ground. Timing with a stopwatch began

when the moth started beating its wings, usually immediately

after tarsal contact was lost when the watchglass was

removed. After 30 sec, the stroboscope was turned on and

WBF was taken 1.0 min later. Then, valve B was closed, the

moth dropped into an empty vial, and another moth was

tethered. Replications consisted of five to ten moths.

Response to pheromone

This test was conducted in a flight tunnel made of

clear acrylic similar to that described by Miller and

Roelofs (1978) (Fig. 3). An 11-cm-diameter fan pushed air

through a baffle system consisting of two partitions of

76-mesh nylon organdy material. These diffused the air and

provided a uniform flow down the length of the tunnel at a

rate of 24 + 2 cm/sec. Females in wire cages were placed in

a chamber between the last baffle and the flight tunnel

proper, being separated from the latter by a 20-mesh screen

to keep males from entering the chamber. A 12.0 x 12.0-cm

plastic target was located on the screen to frame the

pheromone plume produced by the females. At the opposite

end of the tunnel, a 14.0-cm-high stand held the release

cage in front of a 10.2-cm hole through which

pheromone-laden air was exhausted outside the building via

flexible hose. A clear plastic starting gate, attached to a

control rod, was used to keep males from leaving the release

cage before the observer was ready to begin the test. Males

were removed from the tunnel with a suction device at the

completion of each run. Lighting was provided by two

15-watt red (greater than 600 nm) fluorescent tubes located

behind the tunnel. The back wall and ends of the tunnel

were covered with light-diffusing material (Armstrong 420A)

that provided backlighting to silhouette insects, enabling

precise observation without altering their behavior. Light

intensity was about 11 lux.


Four- to 5-day-old wild and laboratory males were

placed individually into release cages made from 30.0 x

55.0-mm plastic dilution vials each with a 19.0-mm-diameter

hole drilled in the bottom and covered with aluminum window

screen to allow air to flow through them. Prior to testing,

the fan was turned on in the tunnel and two females were

placed in each of two wire cages that were mounted on

11-cm-high screen pedestals in the center of the chamber.

Females began calling within ten minutes when a release cage

was placed on the stage with its open end directed upwind

and covered with the starting gate. The gate was pulled

away from the cage when the observer was ready to begin the


Response to pheromone was first indicated by clasper

extension inside the release cage. Timing by stopwatch

began as soon as the male left the cage. Moths were

observed for directed flight (anemotactic flight for more

than 3 sec in the pheromone plume) and were scored as to

whether or not they landed inside the plastic target. The

amount of time it took a male to arrive at the source,

whether by flying or walking, was recorded. Also, the time

spent in flight was recorded using a second stopwatch.

Tests were terminated when a male arrived at the target or

at the end of three minutes. Eight to ten males were tested

per larval rearing container. After each test day, all

components of the tunnel system were cleaned with Windex" to

remove residual pheromone molecules.



Four-minute mating observations

Observations of mating behavior were made in a clear

plastic box partitioned into five cubicles (5.3 x 5.3 x

5.0-cm each). Five, 2-cm-diameter holes were drilled in the

box lid over the center of the corresponding cubicles.

Organdy material (20 mesh) was placed between the top of the

cubicles and the lid to keep moths from escaping. A moth

was transferred to a cubicle with the aid of an injector,

consisting of a cork plunger inside a 13.0 x 120.0-mm glass

tube. The insect was prompted to crawl into the barrel of

the injector, which was then placed in one of the holes in

the lid of the box. The organdy was gently moved aside and

the moth pushed out of the injector with the plunger.

Lighting was the same as that described for flight tunnel


Three- to 4-day-old male and 3- to 5-day-old female

moths were transferred from holding cartons to 20-ml glass

vials. Females were taken from vials and injected into the

box, one per cubicle. After females began calling (about 10

min), a male was put into the first cubicle. Timing of

several behaviors began immediately. Time prior to sexual

activity was the period from introduction of the male into

the cubicle to extension of claspers, increased wing

movements, and initiation of searching behavior. Time in

pursuit of the female was the amount of time that the male

spent within 1 cm of the female, trying to mate with her.



Time to successful clasp was the amount of time between

introduction of the male and a clasp that resulted in the

pair remaining in copula for at least one minute.

The number of clasp attempts per male was determined

using a hand-held counter. An attempt was counted only if

the male was oriented properly vis-a-vis the female, if he

attempted to clasp the tip of the female's abdomen, and if

his claspers came to within 5 mm of her abdomen.

Observations were terminated when a successful clasp was

made or at the end of 4 min when another male was put into

the next cubicle. If the preceding pair mated during the

next 4-min observation period, it was noted, but time to

successful clasp was not recorded. For each larval rearing

container, five to ten replications were run in one or two

partitioned boxes. Males and females tested together were

from the same containers.

Two-night mating studies

Two-night mating competitiveness studies were conducted

in 11.0-cm-high x 8.5-cm-diameter cylindrical cages made of

3-mm-mesh hardware cloth and capped on the ends with plastic

petri plates. Moths were given a 10% sucrose solution for

food. Each cage held one female, one "untreated" male, and

one "treated" male. In most cases, the "treated" males were

merely those from a different rearing container which was

set up on the same day as that of the female and the


"untreated" male. The other treatment consisted of adult

males exposed to gamma radiation as pupae (see below).

Controls consisted of one female and two males from the same

container, or one female and two unirradiated males,

depending on the test.

To begin a test, 3- to 4-day-old moths were chilled for

15 to 20 min at 4.40 C in a walk-in cold room. During this

time, females were taken from 3.8-liter cartons and placed

into test cages. Males were marked on the tip of one wing

with a red or black felt pen to indicate treatment before

being placed in cages with the females. Cages were held in

a controlled-environment cabinet and transferred after two

nights to a freezer. Females were dissected and the number

of spermatophores transferred by males was counted. Both

males were also dissected, and the mated status evaluated by

the color of the simplex duct near the base of the aedeagus.

Dark brown or black indicated the male had not mated, while

a clear, cream, or yellowish color meant the male had mated

at least once (Snow and Carlysle 1967). Four replications

of each treatment and controls were set up.


Insects exposed to conditions unlike those in the

laboratory rearing program at the USDA were considered to be

treated and were used to determine if 1) the testing system

was sensitive enough to detect differences between treated


and untreated insects, 2) variables in the rearing program

could be manipulated to improve insect quality, and 3)

laboratory insects were comparable in quality to wild ones.

Larval diet was modified by changing the concentration

of vitamin suspension (see Appendix for composition).

Concentrations of 3.3 ml and 12.0 ml were compared to normal

amounts of 10.0 ml per 1000 g of diet. Adult diet was

modified for other insects by feeding males ad libitum on

10% sucrose solution (standard adult diet for this study),

water only, or nothing.

Another treatment consisted of exposing 5-day-old male

pupae to between 20 and 35 krad of gamma radiation from a

cobalt source. Two groups of 20 30 male pupae from the

same larval container were placed in separate large test

tubes. Tubes from the first group were placed in a circular

rack so that each tube was the same distance from the

center. The rack was then placed into the irradiator and

the source lowered into the center. Dosage was determined

by the time of exposure and the distance of the tubes from

the source. The second group of pupae served as

unirradiated controls. After treatment, both groups were

returned to the controlled-environment cabinet with the

other insects.

Wild fall armyworm larvae, collected in May and August

from corn fields near Alachua, FL, were placed in 3.8-liter

cartons containing about 4 cm of sand for pupation. Cartons

were covered with 76-mesh organdy and placed outdoors under

a bench



and placed

where they


that protected them from rainfall and direct

Larvae were fed daily with corn leaves until

Pupae were harvested when two to three days old

in the walk-in chamber with laboratory pupae,

remained until testing.

Development of Quality Control Charts

An attribute (P) control chart for percent non-usable

pupae per container was developed. To establish process

capability, the average line on the chart (P) and the upper

(UCL) and lower (LCL) control limits were calculated from

the first 20 containers harvested using the following


P = total number of non-usable pupae

total number of insects

UCL = P + 3 P (1.00 P)/n


LCL = P 3 VF (1.00 P)/n


For example, if the number of non-usable pupae is 528 out of

6892 total pupae, then P = 0.077 and the limits are:

0.077 077 3 0.077 (1.000 0.077) or, 0.087, and 0.067.


Because the number of insects per container (n) was

variable, separate limits were calculated for each container

and plotted on a quality control chart (Charbonneau and

Webster 1978). P-values and control limits were then

calculated for the remaining 16 containers and added to the


Average (X) and range (R) control charts were prepared

for developmental and behavioral traits. Pupal weight will

be used to illustrate the development of these charts.

Three random samples of five pupae each were selected and

weighed for each of eight larval rearing containers. The

average and range were determined for each sample and used

to calculate the overall average (X) and average range (R).

Range, determined by subtracting the smallest from the

largest value, was used in place of standard deviation

(s.d.) because it is easier to calculate and because it

closely approximates the standard deviation with a sample

size of 15 or less. X was used as the center line for the

average chart and the limits were calculated by using R:


UCL = X + A2R


LCL = X A2R.

For example, if n = 5, X = 218.6 mg, R = 66.2 mg, and A2 =

0.577, then the limits are:

218.6 + (0.577)(66.2), or 256.8 and 180.4 mg.

A2 is a multiplication factor that varies inversely with

sample size (see Juran et al. 1974, Appendix II, p. 39).

Therefore, as sample size increases, A2R, which approximates

3 s.d., will be reduced, and the sensitivity of the chart

will be increased. Range charts were developed similarly





Using the same example, and D4 = 2.114, then,

UCL = (2.114)(66.2), or 139.9 mg.


D3 and D4 are multiplication factors similar to A2 and are

derived from the same chart. The value for D3 when n = 5 is

0.0 so the lower limit for such a control chart is 0. The

center line for the R-chart is R. Control charts for

abdominal rotation and WBF were prepared in the same manner

as those for pupal weight, except that the sample size for

WBF was four instead of five.

An operating characteristics (OC) chart was prepared to

complement an X-chart for pupal weight to determine how

sensitive it was for detecting changes in central tendency.

An OC chart is based on the z-distribution and represents

the probability that the average of a single sample of size

n will fall within control limits. Hence, it is an aid in

identifying the optimal sample size that will balance the

cost of looking for a problem when it doesn't occur (type I

error) and the cost of not looking for a problem when it

does exist (type II error). The abscissa of the OC chart is

the degree of shift from the expected mean of the process (X

k a- where k is the number of s.d.'s and a = s.d.//n;

the ordinate is the probability, from 0 to 1, of detecting

such shifts. The probability of any given shift being

detected was determined by using an expression, k-3, and

z-tables for areas under a normal curve. For example, when

no shift has occurred, the probability of a sample point

falling within control limits is 0.9987 (k=0, 0-3=-3, thus

the probability z> -3=0.9987). But, if the process average

shifts to X + 3 s.d., then the probability of a sample point


falling within control limits is 0.5000 (k=3, 3-3=0; z-value

for 0 is 0.50).

Analysis of variance and chi-square tests were used to

determine significant differences (p = 0.05) and means were

separated using Duncan's new multiple range test (DNMRT)

with harmonic means calculated when sample sizes were

unequal (Chew 1977). Significant values for correlation

coefficients and regressions are reported at p = 0.01 unless

otherwise noted.

Fig. 1. Device used to determine degree of abdominal
rotation for fall armyworm pupae.

Fig. 2. Vacuum-tethering equipment for determining wingbeat
frequencies of fall armyworm moths. A=valve used to
maintain constant suction pressure; B=valve for turning
vacuum on and off; C=illumination source; D=rheostat to
control light level; E=pipet for tethering moths;
F=stroboscope; G=watchglass for orienting and mounting
moths to tether.



4.) ai C1)
'-4 S)

co ci (1

rE C:

'- 4 r-

S r0 0

o- q-4 ba


ad CdJ

Cd II r


L) c)

ci0 0

ba-) 0 i
L.a W~

bO a ) :3
C i: c


.,A CO
r-- r 4_4

b ci o t


a0C) cq)
I r. 4-) .

4) 0 rbOO

bO 5 "-4 ci)
-4 C..C..C~
Cl-.Ci)Cd 4~
*~4~) Cd
WL) Cz




Occurrence of Dietary Fungal Contamination

Mold, presumably Asoergillus sp., growing on the

surface of the artificial diet, occurred in about 35% of all

larval rearing containers (n=184). Incidence of

contamination was not significantly different among the four

container locations or between holding racks, but it varied

with set-up day as 38.9, 16.7, 45.0, 25.0, and 47.2% of the

containers established on Monday through Friday,

respectively, were moldy. The amount of surface area

covered by mold in contaminated containers did not show

daily differences and averaged 42.8 6.6%.

Insect Density In Larval Rearing Containers

Mean values for densities of normal and abnormal

insects per container at the time of pupal harvest differed

significantly between 1981, when the rearing containers were

first placed into production, and 1982. The total number of

insects, percentage of usable pupae, and the percentage not

pupating increased from 285 + 6.9, 88.5 + 0.4%, and 2.1

0.1%, respectively, in 1981 (n=156) to 338.0 + 14.0, 91.6 +


0.5%, and 2.9 0.3% for 1982 (n=72). The upward trend for

total number of insects accelerated during the end of the

1982 period (Fig. 4). On the other hand, the proportion of

deformed pupae declined from 9.4 + 0.3% of all insects

harvested in 1981 to 5.6 + 0.3% in 1982. Thus, insect

yields increased as experience was gained with the new


The trend in percentage non-usable pupae produced

during 1982 was monitored on a control chart (Fig. 5).

Limits were calculated individually for each harvest date

because of unequal sample sizes. Significantly excessive

numbers of non-usable pupae occurred in about 22% of the

harvest days (circled points). The average density for

these days was 479 41.7 insects per container, much

greater than the overall average of 338 + 14.0. All

containers with excessive numbers of non-usable pupae were

set up on Monday, Wednesday, or Friday with 63% set up on

Friday. Mold occurred in 38% of these containers, not

significantly more than the overall average of 35%. On the

other hand, those containers yielding significantly more

usable pupae (points enclosed in squares) had an average

density of 318 52.3 insects per container, and were all

set up on Tuesday. None of the containers set up on Tuesday

in 1982 had any fungal contamination of the diet.

Total number of insects was correlated with the

percentage of usable pupae, with most efficient yields

between 250 and 400 insects per container (Fig. 6). The day

containers were set up influenced insect density in 1981 and

1982, and the percentage of usable pupae in 1982, but the

daily pattern was not the same as that produced by dietary

fungal contamination. Rather, yields from containers set up

on Thursday in 1981 were greater than any other day;

Monday yielded the lowest numbers (Table 1). Variability

among containers from all days averaged 13.0 but more than

doubled to + 29.5 in 1982 when greatest yields occurred on

Friday, and the average percentage of usable pupae was

lowest on Mondays and Fridays (Table 2).

The concentration of vitamin suspension in the

artificial diet was strongly correlated with insect yields

(r= -.991). Between 3.3 and 12.0 ml per 1000 g of diet,

yields increased by 50 insects per container for every 1 ml

reduction in vitamin suspension. In addition, at 3.3 ml,

the number of usable pupae increased by 24% over the highest

yield with standard, 10-ml amounts. Standard concentrations

of 66 I.U. of vitamin A per gram of diet may have been

toxic, as it is known to be in mammals, resulting in

degeneration of organs, poor growth, and loss of weight

(Maynard et al. 1979).

Characteristics of Immature Insects

Pupal weight

The weights of male pupae were similar for both years

and averaged 220.0 1.5 mg (n=3535). However, weights of

female pupae declined significantly from 222.3 1.4 mg

(n=1660) in 1981 to 216.0 + 1.8 mg (n=1420) in 1982. The

day of container set up influenced pupal weights for males

and females in 1981 (Table 1). Containers set up on Monday

yielded the heaviest and those from Thursday, the lightest.

This corresponded with lowest insect density on Monday and

the highest on Thursday; generally, pupal weight was

inversely correlated with insect density per container (r=


Quality control charts were prepared for pupal weights

in 1982 of males from container locations B and D, and

females from location D (Figs. 7 9). R-charts indicated

little change in process uniformity over time, the number of

points above and below the average range line generally

being in balance. Conversely, X-charts for pupal weight

showed downward trends, indicating a general change

affecting all samples. Non-randomness in the X-charts can

be identified in the following manner (Juran et al. 1974):

Fig. 7 has 12 consecutive points below the average; Fig. 8

has four of five successive points beyond 1 standard

deviation; Fig. 9 has two of three successive points beyond


2 standard deviations. Some of the out-of-control points

were caused by sampling error, such as in Fig. 7 in which

the average weight of the 6th sample was influenced by a

relatively large range; the weight of one pupa was much less

than average, probably because it was near death.

Conversely, other points indicated that production processes

were legitimately out of control. For example, values for

points 14 and 17 from Fig. 9 showed a real decline in weight

because the ranges of these samples were very near normal.

This fact, in addition to the obvious downward trend in

weights from point 13 on, indicated a significant change was

occurring. Comparison of these charts showed that average

weights and trends were similar, but the degree of

variability differed due to container location and sexes

from the same location.

An operating characteristic (OC) curve associated with

the average chart for pupal weight from Fig. 9 is given in

Fig. 10. This curve shows the probability of not catching a

shift in the process average on the first sample taken after

the shift has occurred. With successive, independent

samples, the probability computed from the OC curve can be

used to determine the chance of not observing a shift of a

specified amount within two, three, or x samples taken after

the shift has occurred (Duncan 1959). For example, if the

average pupal weight shifted from 222.8 to 256.5 mg, the

probability that the average of a single sample of size five

will fall within control limits is about .50, or there is a


50% chance that such a change will go undetected. On the

other hand, there is only about a 15% chance of this

occurring when the sample size is ten. If the average

weight drops only 11 mg, a sample size of five results in a

98% chance of not detecting the change; the larger sample of

ten, about a 96% chance. The probability of the change

going undetected for five groups of five samples each is

(0.98)" or 90.4% and for five groups of ten samples is

(0.96)5 or 81.5%. For most entomological needs, a sample

size of five is sensitive enough to detect anomolous trends

before harm is done to insect quality. However, at least

one variable in a rearing program (e.g. pupal weight) should

be monitored regularly with a sample size of ten or more.

This reference variable should be easy to measure and

correlate with others that are more difficult to quantify.

Sensitivity of the control chart for the reference variable

will more quickly alert the manager to developing problems.

Pupal buoyancy

Buoyancy for male pupae was 0.297 0.004 sec and for

females 0.318 + 0.004 sec (n=1420 for both sexes). It was

positively correlated with pupal weight (r = .302), making

it a possible alternative measurement when taking weights is

infeasible or too time consuming. Buoyancy also was

negatively correlated with the percentage of diet surface

covered by mold (r= -.309).

Abdominal rotation

Abdominal rotation of male pupae averaged 76.0 0.5

degrees and that of the female, 73.9 0.5 degrees (n=1420

for both sexes). For male pupae, rotation correlated

negatively with the percentage of diet surface covered with

mold (r= -.342) and, for both sexes, it was negatively

correlated with density in larval rearing containers (avg.

r= -.451).

Quality control charts for male pupae are shown in Fig.

11. Point 10 is out of control in both the range and

average charts, indicating that sampling error and not an

assignable cause was the reason for the anomoly. However, a

slow downward trend in degree of rotation is apparent on the

average chart, and it becomes more significant because point

16 is out of control while the corresponding point on the

R-chart is near normal. There is some question, however, as

to whether point 16 is indeed a warning of declining

rotation, especially because the previous five points are so

close to average. Sampling error can almost be ruled out

because the probability of selecting a subsample of five

pupae with abnormal rotation out of hundreds of normals is

negligible. Thus, the probability that a problem does not

exist (type II error) is small, and a search for its cause

should be initiated. The anomoly may be characteristic of

one container or it may be part of a general trend.

Production records showed that the container from point 16


did not have abnormally high insect density (364) but that

it did have 40% of its diet surface covered with fungal

contamination. Verification of a significant downward trend

requires further data.

Pupal mortality

The average mortality of insects in the pupal stage was

1.8 0.3% for males (total n=2981, 40 reps) and 1.4 + 0.3%

for females (total n=2667, 40 reps). Increasing diet

surface covered with mold increased mortality for both male

and female pupae (r= .438 and .385, respectively). Percent

pupal mortality for males was fairly constant over time

(Fig. 12), but the chart indicated out-of-control values for

samples 2 and 17 when excessive mortality was caused by

fungal contamination covering 100% of the diet surfaces.

Characteristics of Adult Insects

Eclosion and wing deformity

Of all pupae that appeared to be normal at the time of

pupal harvest, 7.3% of the males and 7.8% of the females did

not develop into functional adults because of pupal

mortality, unsuccessful eclosion, or wing deformities. Of

both male and female fall armyworm moths, 1.2 + 0.3% were

unable to separate completely from pupal exuviae (total

n=2927 and 2629, respectively, 40 reps. ea.). In males,


unsuccessful eclosion was positively correlated with

increased diet surface covered with mold (r= .613), and with

reduced buoyancy in water (r= .412). As expected, it was

also correlated with increased pupal mortality (r= .565).

Wing deformity occurred in 4.3 + 0.5% of all adult males

(total n=2892, 40 reps) and in 5.3 0.8% of females (total

n=2596, 40 reps). In females, it was positively correlated

with pupal buoyancy (r= .415) and pupal mortality (r= .504).

Adult deformity was not influenced by dietary contamination.

Wingbeat frequency

Average wingbeat frequency (WBF) for males was 44.7

0.4 cps (n=358, 32 reps). Males that received no

carbohydrate source as adults (n=10) had significantly

reduced WBF compared to those that fed ad libitum on 10%

sucrose solution (41.3 vs. 44.7 cpm, respectively). WBF was

negatively correlated with pupal weight (r= -.650).

Control charts for WBF showed consistency over the time

period monitored (Fig. 13). Points 10 and 12 on the average

chart were very close to the lower limit line and an

increase in frequency of low values, without an apparent

downward trend, would indicate that all containers or

samples were not handled similarly. Investigation of

assignable causes would therefore center on factors which

affect individual containers such as their location in the

larval holding room and number of eggs placed on the diet.

But identification of causes may be difficult because WBF

did not correlate with incidence of mold, insect density,

abdominal rotation, or pupal buoyancy.

Response of laboratory and wild males to pheromone

Variables of pheromone response for laboratory and wild

fall armyworm males are given in Table 3. Nearly all males

responded positively to pheromone by extending their

claspers inside the release cage. Of those not extending,

83.3% made no attempt to leave the release cage; those that

left, did not find the pheromone target. About half of the

laboratory males demonstrated anemotactic flight directed to

the pheromone source; 83% of the wild males performed this

behavior. The remainder of the males either did not fly or

flew in a random pattern within the flight tunnel. Only 73%

of the laboratory males that arrived at the target did so by

flight, while the remainder walked to the source. On the

other hand, all wild males reached the target by flight. As

a result, the average time required to reach the target was

nearly twice as long for laboratory males as it was for wild

insects. Only 37% of the laboratory males that arrived did

so in less than 30 sec, compared to 71% for wild males. Lab

males flew for 20% of the time they were in the tunnel,

compared to 46% for wild ones.

Subjective observations also revealed differences

between strains. Most laboratory males would leave the


release cage and drop directly to the tunnel floor, 14 cm

below. On the floor, they would often hop around in what

appeared to be an attempt at flight. Most of those that

took flight, however, first climbed to a high surface such

as the wall of the tunnel or the platform holding the

release cage. Generally, flight was erratic with little

hovering and of very short duration. Conversely, wild moths

left the release cage and, after an initial flight period,

arrived at the source target or landed in the tunnel. These

males did not hop and were able to take flight easily from

any surface. They flew more precisely, were capable of

hovering and slower flight, and had longer periods of

sustained flight.

The trend over time for the percentage of laboratory

males showing direct flight and the percentage arriving at

the target is shown in Fig. 14. Because the variability

among samples was so great, and the sample size small (10),

the calculated lower control limits would have allowed

acceptance of insects that showed no direct flight and that

did not arrive at the source target. As a result,

subjective limits for percentage showing directed flight

(LCLf) and arriving (LCLa) were placed above the values for

the four samples with the worst performance for each

variable. These limits represent realistic standards which

can be used to evaluate strain improvement or deterioration

as variables within the rearing program are changed. For

example, increasing the percentage showing directed flight


might involve placing insects in mating and oviposition

cages that require males to fly to food sources and to

females for mating. Moths would be selected for during

several generations, and as improvement occurred, control

limits would be reduced until the colony became stabilized.

The two variables in Fig. 14 have been plotted together

because there is a strong correlation between them (r=

.719). Consequently, it's expected that males detecting

pheromone and orienting to it on the wing, will eventually

arrive at the target. This can be evaluated by comparing

corresponding points on a given harvest day. For example,

on days 9 and 11, the percentage of males that failed to

reach the target was higher than the percentage that showed

anemotactic flight, indicating that successful orientation

during flight did not ensure that the moth would find the

pheromone source. This may indicate an inability to sustain

anemotactic flight. Conversely, if more males reached the

target than showed directed flight, then some moths located

the pheromone target by chance, through random flight, or by

walking to it, as occurred with half of the males for the

3rd sample.


Four-minute mating observations

Mating behavior of fall armyworm moths was divided into

discrete events. Males placed into mating cages containing

females did one of two things. They remained motionless

with their wings folded rooflike over their abdomens and

with their antennae flush against the sides of their bodies

or, they immediately assumed an "alert" stance with the

costal area of the wings at about a 45 degree angle to the

longitudinal axis of the body. Most motionless males became

active at some point during the 4-min test period and

assumed the "alert" stance. Wings were then vibrated

dorso-ventrally through an arc of about ten degrees while

the antennae, extended anteriorly, remained motionless. In

the next event, males began moving their antennae in a

circular pattern. Within 1-2 sec, the antennae were stroked

with the epiphysis of the forelegs, followed by extension of

the claspers. Wing fanning then increased to an arc of

about 80 degrees, thus ending the time period prior to

sexual activity. At this point, the male began searching

for the female, and upon finding her, he oriented his body

parallel to hers, abdomen next to abdomen. The male then

extended his claspers and thrust them toward the tip of her

abdomen in a copulatory attempt. After clasping the female,

the male would turn around and align the length of his body

to her longitudinal axis and begin pumping his abdomen,

apparently to insert his intromittent organ into the female.


This was accompanied by the male bending his forelegs close

to his body and moving his head from side to side. If the

pair remained together longer than 1-2 min, the probability

of a successful mating was very high. Unsuccessful matings

resulted when the male oriented himself improperly with

respect to the female, directed his claspers to

inappropriate areas, or was rejected by the female. These

breakdowns in mating behavior were similar to those

described for Heliothis virescens (Teal et al. 1981).

The care and treatment of fall armyworm males directly

affected their mating behavior (Table 4). Time to

initiation of sexual activity was similar for males fed 10%

sucrose solution, unfed, or irradiated as pupae.

Seventy-five percent of the fed males became sexually active

while only 30% of the unfed and 56% of the irradiated ones

did so. Time spent attempting copulation was twice as long

for irradiated males than for the other groups. But, during

this time, the number of clasp attempts was much higher for

fed males than for the other treatments, resulting in 60% of

them mating compared to 0% of the unfed and only 11% of

those that were irradiated. In a preliminary study, males

fed only water mated as successfully (55.6%) as those

receiving 10% sucrose solution. Thus, the prevention of

dehydration appears to be more important than the uptake of

sucrose. The average time to mating was greater for the fed

than for the irradiated males and averaged nearly 2 min.

However, about 1/2 of the time was consumed before males


began any type of sexual activity. Of the remaining time,

the fed males spent only about 30 sec in pursuit of the

female. Therefore, the most accurate measure of mating

efficiency was the ratio of time in pursuit of the female

and the number of mating attempts (Fig. 15).

A trend of increasing variability among samples is

apparent at the end of the R-chart. However, the factor

responsible is only affecting one insect in each sample

because the average chart remains in control, except for

observation 13 for which the range was so great that it

resulted in an out-of-control point. In the upper half of

the X-chart, points were determined by dividing the number

of seconds in pursuit by the number of clasp attempts. In

some cases no clasp attempts were made and the time in

pursuit was averaged and plotted below the zero line. Thus,

the frequency of samples with males actively pursuing

females, but not attempting copulation in 4 min, was 33%.

The average time of pursuit for this group was 14.9 sec,

which is about the same as for those that attempted to

clasp. The ratios for unfed and irradiated groups deviated

significantly from normal.

The average number of matings remained in control for

all samples of untreated laboratory moths (Fig. 16).

Because the number of observations differed among harvest

dates, an average number was used to determine control

limits. This moved the limits slightly further from the

center line ( .51), reducing the sensitivity of the chart.

This technique is justified

insufficient data. However,

used, and when sufficient

collected, new control limits

if charts are developed from

equal sample sizes should be

amounts of data have been

should be calculated.

Two-night mating studies

The average number of spermatophores transferred was

1.7 0.08 (n=176 pairs) with 92.0% of the females mating.

The number of matings were in control except for point 10

which represents males that were treated with 20 kr of gamma

radiation (Fig. 17). Two-night ratio tests, conducted with

irradiated pupae showed that only one treated male (4%)

mated, and that the transferred spermatophore was badly

deformed. Seventy-six percent of untreated males mated.

M -3- )n -

: 0.0
0 ( 1 .0 c.0 Ic

*rl r 5
3 bO 2 CO

(+I +1 +1 +1 +1
C 4 -
3 ,- LC-- T- I. .-6

S rM (M '- l- <\j
(0 0)

o > ( *
40 EE \ Jn > DE-

0- .. 0 N. Vd
C > C', (ClD0 CZ

SL (No -N-' 00
oL a *H (\

oV,. >HI N' N C 0 W

S V o +1 +1 + +1+1 0

V 00 a co LC-
olo ra xi (u

ad m OCO

O c l o-I

>0 0 a)
<4 C+4n> o () my cu S
C-j C\I C\1 *tC: a 1 -14)

,- C -,. --I
SE< ) w o +1 +1 +( +10 +I o

L.3O n3|- 1+*

C4 Cl E E-C 00
&34(Cal *o h

S?> < 0 0 0 *= 3 (0
oj<\j~jmm CT W


Average number of fall armyworms per
container and percentages of usable
pupae in 1982.

Average Number
Set-up of Insects Average Percentage
Day per Container a/b/ of Usable Pupae

Mon 315.7 38.2 b (12) 89.7 1.3 bc(12)
Tue 299.9 24.1 b (12) 94.2 0.7 a (12)
Wed 357.8 +25.5 ab(20) 92.4 + 0.8 ab(20)
Thu 291.9 +28.4 b (16) 92.8 + 0.7 a (16)
Fri 430.1 31.2 a (12) 87.8 + 1.6 c (12)

a/ Means + S.E. and (number of containers).
b/ Means in columns not followed by the same
letter differ significantly (p=.05) as
determined by DNMRT.


TABLE 3. Behavioral comparisons between laboratory
and wild fall armyworm males responding
to pheromone in a flight tunnel. a/


% extending

% not leaving

% showing

% arriving
at target

% landing

Avg. time in
flight (sec)

Avg. time to
target (sec)

Lab b/

97.0 + 0.9

2.9 + 0.8

55.4 + 2.5

65.7 + 2.4

48.3 2.5

8.1 0.5 c/

65.4 + 4.1

Wild d/



83.3 7.6

70.8 + 9.3

70.8 9.3

13.5 + 1.8

38.7 +10.3

a/ Mean + S.E.
b/ 40 replications of 10 observations each.
c/ 28 replications of 10 observations each.
d/ 28 observations.

TABLE 4. Comparisons of mating behavior among fed,
unfed, and irradiated fall armyworm males. a/


Fed b/

Unfed c/

Irradiated d/

1 Avg. time
activity (sec)

2 % extending

3 Avg. time in
persuit of
female (sec)

4 Avg. number
of mating

5 % males

6 Avg. time to
mating (sec)

60.9 + 7.0

75.0 + 2.6

29.2 3.2

2.7 + 0.2

59.8 + 3.0

113.6 7.4

54.3 +21.7

30.0 14.5

18.7 +10.7

0.2 --



54.8 24.5

56.0 +16.5

49.3 +24.6

0.6 --

11.0 --


a/ Mean + S.E.
b/ For variables 1 and 3, n=119 with 13 replications;
for 4, n=178 with 32 replications; all others, n=274
with 32 replications. All males fed ad libitum on
10% sucrose solution.
c/ n=10, unreplicated; adults were starved.
d/ n=9, unreplicated; 5-day old pupae were irradiated
with 20kr ionizing radiation; adults fed ad libitum
on 10% sucrose solution.


0 a



c :3





0 L
-H) C"
o bc







:3 a


a) 0





S2 2SNI 101

/ '

S i ,3SN! **fil

4) e r-4
Ca a, c1)r
04 () r-I Cli

L. r-) Nm

caL- -4
o no) c1
3( Ca
rCI2 *
Ld Nd (
c~a *.3 o -

-4 -4:

1) Ca C -4
:3 0
z C\1 0 r.

i Ca

crC a~
1 00 co

Ca -
(1) m2 co r
Ca W) S-4
4-)r-0 0
a *-HS Lt.
Q) CaJ r4 )
Q4.) C
rC 0
a) 0 0

(1) bO* Lr-

L) H C) I
0 Q)C 4)

> (1) 3:~
Ca 0

I r-i 0 U2L
0C 4-. ) 0
0 C) -4 -
4) -4 0 (..

q~a-4 -
I -4 Cl)
Pr L~
24) C
*0 C:r ~-4-
LCn ..0 C
e4( 0 0..