Comparative pathogenesis of the entomogenous nematode Steinernema carpocapsae


Material Information

Comparative pathogenesis of the entomogenous nematode Steinernema carpocapsae
Physical Description:
viii, 121 leaves : ill. ; 29 cm.
Epsky, Nancy D
Publication Date:


bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1991.
Includes bibliographical references (leaves 110-119).
Statement of Responsibility:
by Nancy D. Epsky.
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 - 001742440
notis - AJF5156
oclc - 26181194
System ID:

Full Text







This dissertation is dedicated to my grandparents

Emery and Nellie Sandor

William and Mary Epsky


I would like to take this opportunity to express my

sincere appreciation and thanks to the Department of

Entomology and Nematology at the University of Florida, and to

all who helped with the completion of this research.

Special thanks are given to my advisor Dr. John L.

Capinera for his advice, encouragement and financial support

for this research; to Drs. G. Smart, Jr., D. Boucias and S.

Zam for serving on my research committee, and to Dr. D. Hall

for reviewing my dissertation and attending my defense.

Thanks are given to Dr. Khuong Nguyen, Dr. Frank Slansky,

Jr., Dr. Greg Wheeler, Catharine Mannion, and Marineide

Aguillera for their valuable input of information, advice and

material for this project. I would like to thank John Diem

for his assistance in the laboratory, especially the nematode

progeny production counts, and for enthusiastically pitching

in when needed.

I would like to thank the Capinera family for their

friendship and hospitality; and Marilyn Epsky, Barbara, Gary,

Steve and Laura Poleskey for listening to more nematode talk

than they ever wanted to hear. Finally, I wish to express my

great appreciation to my parents, Raymond and Ruth Epsky, for

their encouragement and support.




ABSTRACT .. . ..... vi


1 INTRODUCTION . . .... 1

Specificity of Steinernematid Nematodes 1
Pathogenesis of Steinernematid Nematodes 4
Nematode Host-Finding Behavior . 4
Nematode Invasion of the Host .. ..... 5
Host Suitability for Nematode Infection .. 9
Assessment of Nematode Efficacy .. 11


Introduction . ... .. 14
Materials and Methods . .. 16
Insects and Nematodes . .. 16
Bioassay Procedure . ... .. 17
Effect of Number of Hosts Per Arena .. 18
Effect of Host Exposure Period. .. 20
Effect of Nematode Concentration .. 20
Statistical Analysis . 21
Results and Discussion .... .. .. 22
Effect of Number of Hosts Per Arena .. 22
Effect of Host Exposure Period .. 26
Effect of Concentration ... 30


Introduction . . .. 35
Materials and Methods ... .. 37
Insects and Nematodes ...... 37
Efficacy Bioassay ..... 37
Invasion Efficiency Bioassay ... 38
One-on-one Bioassay . 40
Results and Discussion . .... 41
Efficacy Bioassay ..... 41
Invasion Efficiency Bioassay . 41
One-on-one Bioassay . 49


Insects and Nematodes ....... .57
Determination of Physiological Host Age .57
Mortality and Invasion Efficiency Bioassay .59
Results and Discussion . ..... 60


Introduction . . 75
Materials and Methods . 77
Insects and Nematodes . 77
Bioassay Procedure . ... 77
Effect of Collard as Host's Food .. .79
Indirect Effects of Host's Food Treatment 80
Direct Effects of Allelochemical in Host's Food 81
Results and Discussion . 81







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



Nancy D. Epsky

August 1991

Chairperson: John L. Capinera
Major Department: Entomology and Nematology

Pathogenesis of entomogenous nematodes was studied by

comparing various nematode-host combinations. Organisms used

in the study were Mexican strain and All strain infective

stage juveniles of the entomogenous nematode Steinernema

carpocapsae (Weiser), and larvae of three lepidopteran hosts:

Aarotis ipsilon (Hufnagel), Spodoptera fruciperda (J. E.

Smith), and Galleria mellonella (L.). Nematode pathogenesis

was examined under various conditions in bioassays using

filter paper-substrate test arenas. LCs5, invasion efficiency,

and progeny production were determined. Conditions of the

bioassay significantly affected invasion efficiency

(proportion of tested nematodes that established in the host);

invasion efficiency was positively related to host exposure

period and number of hosts per test arena, and negatively

related to substrate surface area per host. Host mortality

was less affected by bioassay conditions, and appears to be a

relatively insensitive index of nematode activity. In tests

with one larva per test arena, estimates of LCs ranged from

4 to 91 infective stage juveniles per host, and invasion

efficiency ranged from 11% to 31% among the six host-nematode

combinations. There were no obvious host- or nematode strain-

related patterns in invasion efficiency; invasion efficiency

was not significantly related to concentration-dependent host


Effects of host age on nematode pathogenesis were

examined. Generally, host susceptibility decreased and

invasion efficiency increased as Acrotis inter- and intra-

instar age increased. An exception was found in invasion

efficiency against the last larval (7th) instar; invasion

efficiency dropped to levels observed in fifth instars and

decreased significantly late in the instar. Highest invasion

efficiencies were 30% and 33% in late-sixth instar Galleria

and Aqrotis; lowest were 14% and 1% in late-seventh instars,

respectively. Newly molted larvae of both species were highly


Tritrophic effects were examined by comparing

pathogenesis in collard-fed and artificial diet-fed Aqrotis

larvae. Host resistance increased slightly, invasion

efficiency was unaffected, but progeny production was

significantly reduced by ingestion of collards. Reduction may


be attributed primarily to decrease in lipid content in

collard-fed larvae, but incorporation of the allelochemical

sinigrin into artificial diet at biologically relevant levels

also significantly reduced progeny production.



Entomogenous nematodes in the families Steinernematidae

and Heterorabditidae are insect-parasitic nematodes that are

mutualistically associated with entomopathogenic bacteria.

Gaugler (1988) listed wide host range, ability to actively

seek host insects, environmental safety, and ability to be

mass produced as some of the attributes that give these

organisms biological control potential. One limitation,

however, is lack of predictability in level of control in

field application (Ehler 1990). These nematodes are

considered to be generalist parasites, and little research has

been directed towards examining specificity in the host-

parasite interaction, or understanding the roles of host-

specific and/or nematode-specific factors that mediate

infection. However, factors in the host-parasite interaction

are important in determining host susceptibility, and an

understanding of these factors may lead to improved use of

nematodes as biological control agents.

Specificity of Steinernematid Nematodes

Many insect species were susceptible to Steinernema

carpocapsae (Weiser) infective stage juveniles (infectives) in


laboratory trials (Laumond et al. 1979, Morris 1985). Most

steinernematids and heterorhabditids were isolated initially

from holometabolous insects, indicating more recent evolution

with these forms (Poinar 1990). (An exception is Steinernema

scapterisci Nguyen and Smart, a nematode isolated from mole

crickets; it has a more restricted host range [Nguyen & Smart

1990, 1991a]). Populations of S. carpocapsae have been

obtained from different geographic locations and these

populations have been given strain designation (reviewed in

Poinar 1990). Efficacy varies greatly among different

nematode strains tested against a single target insect and

among different insects tested with a single strain (Bedding

et al. 1983, Morris et al. 1990). Adaptation to local

environmental conditions and to host location in the soil

profile, sensitivity to host-produced chemicals, ability to

invade a host and overcome host internal immune responses are

factors that may lead to specificity among localized

populations of nematodes (Poinar 1990).

Abiotic factors are well known to affect nematode

performance, and sub-optimal environmental conditions are

deleterious to nematode field efficacy (Gaugler 1988). The

range of tolerable environmental conditions varies among

different strains and may represent acclimation to local

environment (Molyneux 1985). Movement within the soil profile

differs among species of entomogenous nematodes (Choo et al.

1989), but does not differ among strains of S. carpocapsae.


Movement upward to the substrate surface and nictation, the

lifting of the forward half of the body off the soil surface

and waving it from side to side, are parts of the dispersal

and host-finding repertoire for S. carpocapsae regardless of

strain (Ishibashi & Kondo 1990).

Other factors leading to specificity encompass aspects of

pathogenesis of nematode infection. The paradigm of nematode

pathogenesis is based primarily on infection of larvae of the

greater wax moth, Galleria mellonella (L.), by S. carpocapsae

DD-136 strain infectives. Only recently have comparative

studies been conducted using a wider array of nematode strains

and insect hosts, but these comparisons are limited to studies

of a single host with multiple nematodes or, more rarely,

multiple hosts with a single nematode. Variations in many

areas of pathogenesis remain largely unexplored. Yet details

of pathogenesis are critical for uncovering potential

specificity in mechanisms used by nematodes for successful

infection. Studies on comparative pathogenesis may provide

not only an understanding of the biological basis for

differences in efficacy, but also methods for identification

of nematode-based factors and host-based factors that limit

efficacy. Nematode-based factors may be amenable to genetic

selection to improve biological control potential against

specific hosts, whereas host-based factors may be more

difficult to overcome and may require switching to a different

nematode or a different method of control. Thus, information


from studies on comparative pathogenesis may lead to improved

understanding of nematode efficacy and to better predictions

of control.

Pathoaenesis of Steinernematid Nematodes

Nematode Host-Finding Behavior

Most free-living nematodes use chemical cues in their

environment, at least in part, for host finding (Zuckerman &

Jansson 1984), and CO2 is the most commonly used general

chemoattractant for plant parasitic nematodes (Klinger 1963,

1970). Use of host-specific chemical cues, however, has been

questioned. Plant parasitic nematodes may be attracted or

repelled, show no response to host plant root exudates, or be

attracted by non-host root exudates (Croll 1970, Viglierchio

1960, Zuckerman & Jansson 1984). In contrast, the

bacteriophagous nematode Caenorhabditis elegans is strongly

attracted to its host bacteria and weakly or not at all to

non-host bacteria (Andrew & Nicholas 1976).

Steinernema carpocapsae infectives respond to CO2, frass,

cuticle-washes, bacteria and various ions (Schmidt & All

1978,1979; Gaugler et al. 1980; Pye & Burman 1981). Host-

finding ability varies among g. carpocapsae strains, but was

fairly poor for all strains in laboratory bioassay (Gaugler et

al. 1989). Level of host-seeking was strongly correlated to

the concentration of CO2 produced by the target host (Gaugler


et al. 1991). Change in host size or activity level may

similarly affect level of host chemical production and

influence the ability of infectives to locate a host (Kaya


Nematode Invasion of the Host

After locating a potential host, the infective stage

nematode must invade the host to parasitize it successfully.

Early studies on steinernematid nematodes found that

infectives primarily entered the host by way of the mouth,

moved down the alimentary tract into the crop and/or midgut

and then moved into the hemocoel. This mode of infection was

observed with S. glaseri (Glaser 1932), S. feltiae (= S.

bibionis), g. affinis (Bovien 1937), and S. carpocapsae DD-136

strain (Welch & Bronskill 1962). Spiracles may serve as

another port of entry, with infectives moving from trachea

into the hemocoel (Poinar 1979).

Poinar and Himsworth (1967) studied infection of greater

wax moth larvae by a. carpocapsae DD-136 strain infectives in

Petri dish bioassay and with per os inoculation. Infection

was followed by dissection of the host or by histological

examination, respectively, at several time intervals following

exposure to infectives. Host dissections supported the

earlier findings that the mouth was the primary site of

entrance. Fujii (1975) followed the invasion of g.

carpocapsae DD-136 strain in workers of the Formosan termite,


Coptotermes formosanus Shiraki, by histological examination at

various time intervals after initial exposure. Infectives

were found in the hindgut within one hour, in the foregut

after 22 hours, and but never in trachea. Thus, the anal

opening was proposed to be the primary site of invasion.

Infected termites were sluggish by 22 h, and this inactivity

was thought to enable additional infectives to enter through

the mouth. Infectives were observed frequently on the

exterior of the termite in the intersegmental folds, but no

direct cuticle penetration was observed.

Kondo and Ishibashi (1988) studied invasion of S.

carpocapsae DD-136 strain in larvae of a cutworm, SDodoDtera

litura (F.), by both histologic and scanning electron

microscopic examination. They found non-selective movement of

nematodes onto various external body surfaces of the host,

with entrance into the host through mouth or anus. Nematodes

were also observed in invaginated intersegmental membranes,

and many accumulated in an eversible subesophageal vesicle

that occurs in this and other noctuid larvae.

In a second series of experiments, Kondo and Ishibashi

(1989) exposed S. litura and G. mellonella larvae, pupae and

adults to very high numbers (10,000-20,000) of S. carpocapsae

DD-136 strain infectives to determine non-oral infection

routes. Non-oral routes examined were spiracles, cuticle, or

wounds in the cuticle. A clump of infectives was placed on

the side of the immobilized host, near a spiracle (both

species, all stages) or a wound (. litura larvae only).

Infectives invaded through spiracles in all stages of wax

moth, but in only the adult stage of S. litura. The spiracles

of S. litura larvae and pupae were heavily armed with spines;

these appeared to mechanically prevent nematode invasion.

There was evidence of unsuccessful attempts to penetrate the

cuticle directly. There were small, dark pigmented spots in

the smooth, thin cuticle of the intersegmental membranes and

in the subesophageal vesicle in larvae exposed to infectives.

No such spots were observed in the cuticle of control larvae.

These spots were smaller in diameter than infectives. There

was successful penetration of cuticle in the subesophageal

vesicle, and nematodes were found in the hemocoel immediately

adjacent to holes in the cuticle. Nematodes readily entered

the host through wounds.

Unlike tylenchid and mermithid entomogenous nematodes,

which use a stylet to penetrate insect cuticle directly,

rhabditid nematodes have no stylets. Heterorhabditis spp.

nematodes are thought to penetrate insect cuticle using an

anterior tooth, which is exposed in exsheathed infective stage

juveniles (Bedding & Molyneux 1982, Poinar & Georgis 1990).

Kondo and Ishibashi (1989) reported that during preparation

for scanning electron microscopic examination, a. carpocapsae

infectives were subjected to physical stress by exhausting CO2

from the specimen. Most of the body of the infective shrank

and became twisted, but the anterior end remained rigid and


intact. Since the head was "hemi-spherical and had no

prominent projections," the authors speculated that this rigid

head structure enabled infectives to penetrate insect cuticle

mechanically, if the cuticle was thin enough. They found

infectives aggregated on newly healed wound tissue or in the

subesophageal vesicle. Group attack by aggregated infectives

may allow successful penetration of thin insect cuticle, such

as that observed in the subesophageal vesicle of S. litura.

Poinar and Himsworth (1967) used histological studies to

examine activities of infectives as they moved from the

alimentary tract into the hemocoel. Infectives exsheathed

(i.e., lost the second-stage cuticle) in the crop or midgut

soon after per os inoculation. Exsheathment was an active

process in which there was a transverse break in the cuticle

just behind the head of the infective. The nematode crawled

out and left the anterior end of the sheath intact.

Exsheathment was always complete before entry into the

hemocoel. What triggered exsheathment was not known;

infectives also exsheathed in a drop of water and infectives

may be exsheathed mechanically by movement through a sand

column (Timper & Kaya 1989). In the Poinar and Himsworth

(1967) study, exsheathed infectives were found "wedged between

the midgut epithelial cells working their way toward the gut

serosa and hemocoel" using mechanical pressure to penetrate

the gut wall. Infectives were able to penetrate any part of

the alimentary tract of termites to move into the hemocoel,


and penetration through the gut wall was observed in the

proventriculus, midgut and rectum (Fujii 1975). Steinernema

scapterisci infectives that entered mole crickets through

spiracles used mechanical pressure to break the tracheal tube

and enter the hemocoel (Nguyen & Smart 1991b).

Host Suitability for Nematode Infection

Within 11 hours of initial exposure, S. carpocapsae DD-

136 strain infectives were found in the hemocoel of greater

wax moth larvae (Poinar & Himsworth 1967). There they changed

from infective to parasitic third-stage juveniles. The

nematode became larger, the alimentary tract began

functioning, and the mutualistic bacterium, Xenorhabdus

nematophilus, was excreted and began multiplying in the

hemocoel. Nematodes spread rapidly throughout the host body

and movement of nematodes in the hemocoel was passively aided

by host hemolymph circulation (Kondo & Ishibashi 1988).

Nematode and bacteria work in concert to prevent or overcome

successful host internal immune response (reviewed in Dunphy

& Thurston 1990). Nematode invasion into the hemocoel may

induce an array of host immune responses (Stoffolano 1986).

Jackson and Brooks (1989) observed melanotic encapsulations of

four strains of S. carpocapsae infectives by larvae of the

western corn rootworm, Diabrotica viraifera vircifera LeConte;

but found no consistent relationship between immune response

and host susceptibility. No successful immune response has


been demonstrated in naive lepidopteran larvae (Dunphy &

Thurston 1990). Even if the host immune response is induced

by nematodes in the hemocoel, the response may not prevent

release of mutualistic bacteria and subsequent host death.

Establishment of Xenorhabdus in the host cadaver serves

a number of functions beyond killing the host. Xenorhabdus

produce an array of antimicrobials that prevent other bacteria

from colonizing the cadaver; nematodes feed on the bacteria as

well as host tissue; and nematode reproduction is dependent on

availability of Xenorhabdus as a food source (Akhurst 1980).

First generation female S. carpocapsae have extremely high

reproductive rates and produce high numbers of progeny; this

aids in colonization of the cadaver and prevents attack by

other saprophytic organisms (Poinar 1979). Steinernema

carpocapsae reproduce sexually, so presence of both males and

females in the cadaver is necessary for reproduction.

Typically, there are two generations of adults in the cadaver,

with progeny from second generation adults becoming infective

stage juveniles that leave the cadaver in search of new hosts.

Crowding and depletion of food are both factors in the

developmental switch to production of infectives in the

cadaver (Fodor et al. 1990).

Number of infectives produced per cadaver is a measure of

host suitability for nematode pathogenesis, and progeny

production has been found to vary among hosts. Steinernema

carpocapsae reproduction in cadavers of greater wax moth


larvae may yield 1700 infectives per mg, based on host weight

(Dutky et al. 1964). In other hosts, reported levels include

100 and 680 infectives per mg in larvae of the mountain pine

beetle, Dendroctonus ponderosae Hopkins, and the lesser

European elm bark beetle, Scolytus multistriatus F.,

respectively (MacVean & Brewer 1981), and 300 infectives per

mg in larvae of the carrot weevil and the armyworm,

Pseudaletia unipunctata (Haworth) (Kaya 1978).

Assessment of Nematode Efficacy

A two-step screening procedure has been recommended for

assessment of nematode efficacy against target pests (Bedding

1990). The first step was an initial screening, with an

individual insect exposed to a high concentration of

infectives (1000 per host). If host mortality exceeded 50%,

then screening continued with the determination of LCo0 (i.e.,

concentration that causes 50% mortality) using standard

bioassay procedures (Bedding et al. 1983, Woodring & Kaya


Other efficacy assessment procedures have been proposed.

Although LC50 tests have been the most common method of

efficacy determination in laboratory bioassays, measurements

of LT5s (time until 50% mortality) at a single nematode

concentration have also been employed (Woodring & Kaya 1988).

An alternative "penetration rate assay" has been proposed as

a measure of nematode virulence (Glazer et al. 1991). Hosts


exposed individually to a single concentration of nematodes

for increasing periods of exposure (e.g. 3, 6, 9 and 12 h),

were rinsed and moved to nematode-free arenas. Mortality was

recorded after 48 h and dead insects dissected 24 h after

death to determine number of nematodes in the body cavity.

Data from host mortality and number of nematodes established

in the host per exposure period were used for efficacy

comparisons. A variation of the penetration rate assay is a

nematode concentration/establishment bioassay (Fan & Hominick

1991). This assay also used dissection to determine number of

internal nematodes, but hosts were exposed to a range of

nematode concentrations. Number of nematodes established in

the host was regressed against concentration, and the slope of

the regression reflected the proportion of nematodes that

successfully invaded the host.

Quality control of nematode preparations is an important

component for comparisons among studies made in the same

laboratory over time, as well as for comparisons among

different laboratories. Larvae of greater wax moth are highly

susceptible to nematode infection by most entomogenous

nematodes and are the prime candidate for a universal bioassay

organism for nematode efficacy assessment (Hominick & Reid

1990). A one-on-one bioassay was developed in which a single

greater wax moth larva was exposed to a single infective.

This procedure has been used commercially for quality control

(Georgis 1990), but its utility in research with other hosts

has not yet been tested.

These assessments of nematode efficacy represent one-

dimensional or single factor approaches. Results are

extrapolated to other conditions or other systems, but

validity of extrapolation is questionable. Therefore, a

multiple factor approach to nematode pathogenesis, which used

direct comparisons to elucidate dynamics of host-nematode

interactions and limitations to nematode efficacy, was

employed herein. The effect of change in physical condition

of the bioassay procedure on host mortality and nematode

invasive ability was examined to determine if a single

bioassay procedure could be used to make valid comparisons

between different hosts and/or nematodes (Chapter 2). A

bioassay procedure was developed to allow direct comparisons

among different host-nematode combinations, and this was used

to examine relationships between nematode invasive ability and

host susceptibility (Chapter 3). Factors within the host

population were then examined for their effects on nematode

pathogenesis (Chapters 4, 5). Finally, results from these

comparative studies were used to develop a hypothesis on host-

specific and nematode-specific factors that mediate nematode




The entomogenous nematode Steinernema carpocapsae

(Weiser) has great promise for use as a biological control

agent; a large number of insect species are susceptible to

infective stage nematodes (infectives) in laboratory trials

(Laumond et al. 1979, Morris 1985). However, efficacy, as

indicated by LCso, varies greatly both among nematode strains

tested against a single target insect and among insect species

tested with a single nematode (Bedding et al. 1983, Morris et

al. 1990). Abiotic factors are well known to affect nematode

performance, and sub-optimal environmental conditions are

deleterious to nematode field efficacy (Gaugler 1988). Less

is known about the biological basis for differences in the

levels of host mortality. For example, Dunphy and Webster

(1986) found that differential virulence between S.

carpocaDsae DD-136 strain and S. carpocapsae Mexican strain

against larvae of the greater wax moth, Galleria mellonella

(L.), was not due to virulence of the nematode/symbiotic

bacteria complex per se or to host internal immune responses.

They speculated that the differences may be in host-invasion


The model for nematode-host interactions is based

primarily on infection of greater wax moth larvae by S.

carpocapsae DD-136 strain infectives (Poinar 1979). Few

studies have quantified number of tested nematodes that invade

the host, much less determined how variable this is between

different hosts or under different experimental conditions.

LCs5 estimate is affected by conditions of the bioassay and

invasion may be similarly affected. Therefore, prior to

making comparisons among different host-nematode combinations,

it is important to ascertain how aspects of the bioassay

procedure may affect levels of host mortality and nematode


Larvae of the fall armyworm, Spodoptera fruaiperda (J.E.

Smith), and the black cutworm, Aqrotis ipsilon (Hufnagel), are

susceptible to infection by S. carpocapsae in both laboratory

and field trials (Fuxa et al. 1988, Richter & Fuxa 1990,

Capinera et al. 1988). LCs5 estimates from laboratory

bioassays with last instar fall armyworm and black cutworm

have shown that black cutworm larvae are significantly more

resistant to both S. carpocapsae Mexican strain and S.

carpocapsae All strain (Chapter 3). There is a 10-fold

increase in LC50 estimate from the most susceptible host-

nematode combination (fall armyworm-S. carpocapsae Mexican

strain) to the least susceptible (black cutworm-S. carpocapsae

All strain). Therefore, these two host-nematode combinations

were used to determine how factors in the bioassay procedure

affect both nematode invasion and host mortality, and to

ascertain if these effects varied in hosts with different

levels of susceptibility. For comparative purposes,

additional studies were conducted with greater wax moth-S.

carpocapsae All strain.

Materials and Methods

Insects and Nematodes

Fall armyworm and greater wax moth larvae were obtained

from the Insect Attractants, Behavior and Basic Biology

Research Laboratory, USDA, Gainesville, FL for use in this

study. The black cutworm colony was obtained from J.C. Reese,

Kansas State University, and was cultured using methods of

Reese et al. (1972). Steinernema carpocapsae Mexican strain

and All strain nematodes were obtained from G.C. Smart,

University of Florida. These were reared in vivo in greater

wax moth larvae using standard rearing procedures (Dutky et

al. 1964). Infectives were stored in deionized water and held

at 60C until use. Infectives were used within one month of

collection from the host cadaver (Appendix A). Last instar

larvae were used for all tests. Fall armyworm was tested with

S. carpocapsae Mexican strain infectives, black cutworm and

greater wax moth were tested with S. carpocapsae All strain


Bioassay Procedure

Larvae were exposed to nematodes on filter paper in Petri

dish arenas (Woodring & Kaya 1988). Individual larvae were

tested in medium arenas (Petri dishes, 60 x 15 mm) with 2

pieces of filter paper (Whatman #1, 5.5 cm diam). Infectives

were added in 1 ml deionized water. Groups of larvae were

tested in large arenas (Petri dishes, 100 x 15 mm) with 2

pieces of filter paper (Whatman #1, 9.0 cm diam); nematodes

were added in 2 ml deionized water. The bioassay was modified

to permit exact counts of numbers of nematodes tested per

arena for some studies. To do so, the filter paper was cut in

half, and one piece was placed on top of the other in the

arena. This left half the arena uncovered. A drop of water

containing the approximate number of infectives desired was

placed in the uncovered half and exact number of infectives

tested was determined by microscopic examination. Additional

deionized water was added to the filter paper, and the filter

paper was moved into contact with the drop containing

infectives. Total volume tested was 0.5 ml in medium arenas

and 1.0 ml in large arenas.

Insects were added to bioassay arenas one hour after the

infectives. Petri dishes were enclosed in a large plastic bag

and held at 250C. Number of nematodes that invaded the host

was determined by dissection using the procedure of Kondo and

Ishibashi (1986a) (Appendix A). Invasion efficiency was


determined from the percent of tested nematodes that

successfully invaded the host (summed from all hosts in trials

with multiple hosts per arena).

Effect of Number of Hosts Per Arena

LC,5 and invasion efficiency estimates were determined for

treatments of one fall armyworm per medium arena and three

fall armyworm per large arena. Two arena sizes were used to

keep substrate area per host between treatments. Totals of 40

(one larva per arena) and 39 (three larvae per arena) larvae

were exposed to nematode concentrations of 0, 10, 20 and 40

infectives (Mexican strain) per larva. Mortality was recorded

after 48 h, and cadavers of larvae exposed to concentrations

of 20 or 40 infectives per host were dissected after 72 -

96 h. Some larvae were discarded because, by the time of

dissection, progeny production was too advanced to obtain an

accurate count of nematode invasion.

Greater wax moth larvae were used to determine effect of

multiple hosts per arena on carpocapsae All strain

infectives. Unlike the studies with fall armyworm, substrate

surface area was not kept constant and LCs5 and invasion

efficiency estimations were determined in separate trials.

LCs0 was determined for treatments of one greater wax moth per

medium arena and ten greater wax moth per large arena. The

ten per arena treatment followed the standard LCs5 bioassay

procedure (Woodring & Kaya 1988). A total of 65 (one larva


per arena) larvae were exposed to concentrations of 1, 10 and

50 infectives per host (30, 20 and 15 individuals per

concentration, respectively) and 60 (ten larvae per arena)

larvae were exposed to concentrations of 1, 5 and 10 per host.

Exact count was obtained for all concentrations. Mortality

was recorded after 48 h.

Invasion efficiency of All strain infectives in greater

wax moth was determined with four nematode concentration-host

density treatments (designated as number of infectives on

number of larvae per medium arena). Levels tested were 1-on-

1, 10-on-l, 50-on-1 and 50-on-5; with 60, 10, 10 and 50 larvae

per treatment, respectively. Exact count was obtained for all

treatments. These treatment levels allowed comparisons

between a constant number of nematodes per host (10-on-l

versus 50-on-5), between a constant number of nematodes per

arena (50-on-1 versus 50-on-5) and between a varying number of

nematodes per host or arena. A second 1-on-l treatment was

added to further evaluate the effect of substrate surface area

per host. For this bioassay, 25 larvae were tested

individually in small arenas (Petri dishes, 35 x 10 mm) with

2 pieces of filter paper (Fisher Quantitative, 2.5 cm

diameter), and exposed to a single infective added in 0.2 ml

deionized water, the total amount of water per small arena.

Mortality was recorded after 48 and 72 h. Both 1-on-l

treatments evaluated invasion by individual nematodes (Miller

1989), and invasion efficiency was determined from percent


mortality. Invasion efficiency for the 10- and 50-nematode

concentration treatments was determined from dissection of

larvae 24 h after death.

Effect of Host Exposure Period

Host exposure time periods of 24, 48, 72 and 96 h were

examined. Totals of 70 fall armyworm larvae, exposed to 50

and 100 infectives (Mexican strain), and 160 black cutworm

larvae, exposed to 100 infectives (All strain), were tested

individually in medium arenas. Mortality was recorded daily.

Ten, 10, 10, and 5 fall armyworm were sampled per

concentration per time period, respectively, and 20 black

cutworm larvae were sampled per time period. All of the

sampled fall armyworm larvae and ten of the sampled black

cutworm larvae were rinsed and dissected immediately. The

remaining ten black cutworm larvae per time period were rinsed

and moved to nematode-free arenas. These larvae and remaining

unsampled black cutworm larvae (host exposure time > 120 h)

were dissected 72 h after death. Black cutworm surviving

beyond 10 days were not dissected.

Effect of Nematode Concentration

Individual fall armyworm and black cutworm larvae were

exposed to concentrations of 10, 50, 100, 250 and 500

infectives. There were 100 fall armyworm tested (30, 20, 10,

10 and 10 per concentration, respectively) and 90 black


cutworm tested (20, 20, 30, 10 and 10 per concentration,

respectively). Mortality was recorded after 48 h and cadavers

dissected after 72 h.

Statistical Analysis

PCPOLO (Russell et al. 1977) probit analysis was used to

estimate LCso and slope and to compare statistically the

effects of multiple hosts per arena on nematode efficacy.

Invasion efficiency was analyzed by factorial ANOVA using Proc

GLM (SAS Institute 1985) to determine significance of main

effects and to test for interactions between main effects.

Results from significant ANOVAs were followed by Duncan's

(1955) multiple range test (E = 0.05) in all tests except the

study on host density effects in greater wax moth. Although

there were three factors (concentration, number of hosts per

arena, substrate surface area) tested, all factors were not

tested at all levels. Therefore, one-way ANOVA was run on the

five treatments (Zar 1974). Tukey's test (P = 0.05) was used

for one-at-a-time, pair-wise comparisons between treatment

means. For single factor experiments on effect of

concentration, factorial analysis was followed by regression

analysis using Proc REG (SAS Institute 1985) to test the

adequacy of a linear model to describe the relationship

between concentration and number of nematodes invading the


Results and Discussion

Effect of Number of Hosts Per Arena

Average mortality was slightly higher for fall armyworm

tested in groups of three than for fall armyworm tested

individually at each concentration tested (10 nematodes per

host: 67% versus 52%; 20 nematodes per host: 74% versus 73%;

and 40 nematodes per host: 90% versus 80%). However, there

were no significant differences in either LCs0 estimate or

slope obtained from probit analysis. LC5s estimates (95% CL)

for fall armyworm tested in groups of 3 or individually were

5.2 (1.0 9.2) and 8.3 (3.6 12.4), respectively; both had

slopes of 1.3 + 0.5. Similarly, LC50 estimate and slope were

not affected by greater wax moth host density. LCo0 estimate

(95% CL) for greater wax moth tested individually was 4.0 (1.8

- 9.1) and for greater wax moth tested in groups of ten was

3.2 (1.7 5.8); slopes were 1.5 + 0.3 and 1.2 + 0.3,

respectively. Molyneux et al. (1983) cited unpublished data

that stressed the importance of testing one individual per

bioassay unit, because the number of hosts per container

affected level of mortality. In my studies, LC50 estimates

were not affected significantly by a change in number of hosts

per arena.

Invasion efficiency was significantly affected by

presence of multiple fall armyworm in the arena (F = 37.5; df

= 1, 32; P < 0.001) and increased from 14 + 2% to 37 + 3% as


the number of hosts in the arena increased from one to three.

Concentration did not affect invasion efficiency (E = 0.01; df

= 1, 32; E = 0.99), and there was no interaction between

nematode concentration and number of hosts per arena (E =

1.62; df = 1, 32; = 0.21). Invasion efficiency was also

significantly affected in greater wax moth by the

concentration-host density treatment (E = 6.67; df = 4, 26; P

< 0.01) (Figure 2-1). All greater wax moth mortality occurred

by 48 h in treatments with 10 or 50 infectives, but increased

between 48 and 72 h in the 1-on-i treatments. After 48 h,

mortality was 9% and 20% in 1-on-i treatment in medium and

small arenas, respectively. However, results of the mean

comparison test remained the same whether using data from 48 h

or 72 h. Invasion efficiency in the 50-on-5 treatment was

significantly different from the 1-on-l (medium arena), 10-on-

1 and 50-on-1, but not from the 1-on-1 (small arena)

treatments. There were no significant differences among the

other treatments.

Increase in invasion efficiency observed could be

explained by increase in number of hosts alone (fall

armyworm), and/or by decrease in substrate surface area per

host (greater wax moth). Kondo and Ishibashi (1986b) found

that invasion efficiency decreased as depth of soil in the

bioassay increased. Increase in substrate volume increases

the distance infectives have to traverse to locate the host.

Steinernema carpocapsae infectives have demonstrated poor

host-finding ability (Gaugler et al. 1989) and may remain

r-1 4 $4 > Id k1
-0 0) A .4 0
4) 0

oP t) 0 -
0 r-4(0 '0 4)
)) U 0
1 W 1 0 0

W -.4 0o a
C .0 .0

o0) go .0
.a c i I
oa to Ao .0
0 o.0 0 -'
0 $4j 4 W 1
-) 1 0 *
o 4) W; r

04 0 t 4j0 C.
0 0 p -4 r.

4 4 *M C
4 ".4 0 4
*' 4

.4 04)
F) ) :4 'a c a

4) a -*4 4
to r 04
41U .,1 .4

04 0 *4 aP 4)
00 O ra r

tn U r r. c1
O < 0 ,
o kB U) 'a f

(0 4) ,-4 0
40 a) r. 0 > VO 4
4o 0 0 Or

V- -46 -4 0
OC > > a+
*1 *4 t 4 C

M 3 rla 0) 0(
4 0 0 g o 0--*
k -4 8 M 4
3 i r-41 4 .
or a0 x 4(
CO 10 & *4 r-




Ln o 0 o



inactive until stimulated by presence of a host (Ishibashi &

Kondo 1990). Several host products, including CO2 (Gaugler et

al. 1980, Schmidt & All 1978) and frass (Schmidt & All 1979),

have been shown to mediate host-finding activity. Thus, an

increase in concentration of host cues from multiple hosts in

a bioassay and/or by a decrease in substrate surface

area/volume per host may stimulate an increase in host-finding

and invasion. Successful host infection may also be limited

by number of natural body openings in the host (e.g., lower

susceptibility of pupae versus larvae (Kaya & Hara 1980)).

Multiple hosts in the bioassay arena may provide a "super

host" that gives off more host cues and has more ports of

entry than a single host. Thus, measurements of invasive

ability made with a single host per arena may underestimate

invasive potential of tested nematodes, and enhancement due to

multiple hosts may preclude making comparisons between studies

using different numbers of hosts per arena.

Effect of Host Exposure Period

After 24 h, 85% of fall armyworm larvae were infected

with nematodes, and 100% were dead by 48 h. In contrast, only

25% of black cutworm were infected by nematodes within 24 h,

and 57% were dead within ten days. Time until host death has

long been recognized as an important aspect of mortality-

response assays. While maximum host mortality usually occurs

within 48 h in greater wax moth, it may take longer in other

hosts. Capinera et al. (1988) found that high levels of black


cutworm mortality occurred within 24 h of contact with S.

carpocapsae Kapow strain, but not until 72 h with two other

nematode species tested. However, low mortality in black

cutworm after 48 h was not overcome by lengthening period of

exposure to infectives.

Nematode invasion efficiency was significantly affected

by fall armyworm exposure period (E = 16.66; df = 3, 62; E <

0.001), and there was an increase in invasion as exposure

period increased (Table 2-1). There was no effect due to

concentration (E = 1.9; df = 1, 62; E = 0.17) nor was there an

interaction between concentration and exposure period (E =

2.12; df = 3, 62; P = 0.11). Invasion efficiency against

black cutworm, determined from cadavers of larvae that died

within 96 h of initial exposure to nematodes and contained

established nematodes, increased from 2.2% + 1.0 after 24 h to

6.3% + 1.9, but the differences were not significant (E =

1.36; df = 4, 21; P = 0.28). Black cutworm that were dead by

48 h tended to have higher numbers of internal nematodes than

larvae that died later (Table 2-2) and few individuals that

died after 96 h contained nematodes.

Within the first 24 h of contact, Kondo and Ishibashi

(1986c) found a positive correlation between length of time

Spodoptera litura (F.) larvae were exposed to S. carpocapsae

DD-136 strain and both host mortality and number of nematodes

invading the host. Similar increases were found within the

first 12 h of contact for larvae of the lepidopteran

Spodoptera littoralis Boisduval and both S. carpocapsae All

Table 2-1. Effect of host exposure period on invasion
efficiency of Steinernema carpocapsae Mexican strain
infectives in Petri dish bioassay with fall armyworm larvae.

Host exposure period n % Invasion efficiency + SEM


24 20 5.7 a + 1.2

48 20 18.0 b + 2.4

72 20 24.2 bc + 2.6

96 10 26.3 c + 3.0

aMeans followed by the same letter are not significantly
different (DNMRT, P = 0.05).

> .r4
U4 0

ON 0

0 -







O 1


0 u


1 -1-4
E0 C

c. 10


P c.44

-I de Wr

t O

4 ,4

4 01
C ,

4 '.
o a


0 0
a cz

t A

O 0
M z



Cn 0 0
co n) 0

%D co M** %o

co N'. %0~

r CO % 0 ON


0 0 0 a
%D L


0 0



'd c

.41 (
>o r

0 0
o Z

4.04 r
r -P4


0 4 0

0 *0

0w c
r g E

0 "^I
u~l B


strain and Mexican strain infectives (Glazer et al. 1991).

Infectives were found in fall armyworm and black cutworm by 24

h, and most mortality occurred within 48 h of initial

exposure. Number of nematodes per host increased from 24 to

72 h, and leveled off after 72 h. Although greater numbers of

Mexican strain infectives invaded fall armyworm over time,

period of maximum invasion was similar for both host-nematode


Effect of Concentration

As expected, percent mortality was affected by nematode

concentration. It increased from 37% at the 10-nematode

concentration to 90% at the 250-nematode concentration and

from 10% at the 10-nematode concentration to 90% at the 250-

nematode concentration after 48 h in fall armyworm and black

cutworm, respectively. Concentration, however, did not

significantly affect invasion efficiency in either fall

armyworm (F = 1.75; df = 4, 42; P = 0.17) or black cutworm (E

= 0.54; df = 4, 67; P = 0.70). Average invasion efficiency

per concentration was variable, however, and ranged from 8.6%

+ 2.4 (100-nematode concentration) to 16.2% + 2.4 (500-

nematode concentration) in fall armyworm, and from 14.8% + 1.5

(500-nematode concentration) to 21.3% + 3.4 (250-nematode

concentration) in black cutworm, but the differences were not

significant. There was a direct relationship between

concentration of nematodes tested and number invading the host

and this relationship was best fit by a simple linear


regression model for both fall armyworm (Figure 2-2) and black

cutworm (Figure 2-3). The slope obtained from the regression

model represents the proportion of tested nematodes that

established in the cadaver. Thus, slope multiplied by 100

equals the invasion efficiency tested over a range of


Although there were no significant differences among

measurements of invasion efficiency determined at different

concentrations, there was a great deal of variability in

average percent invasion per concentration. Thus, comparisons

between different host and nematode combinations made at a

single concentration level may be misleading. Hominick and

Reid (1990) suggested that a concentration/establishment

bioassay, which uses slope of the regression line to indicate

proportion of infectives that invade the host, be used for

comparative studies. Results of the study with fall armyworm,

black cutworm and greater wax moth support this suggestion and

show that nematode invasion into a host is very sensitive to

bioassay conditions, probably more so than host mortality.

Therefore, care should be taken in choosing bioassay

conditions for comparative studies so that differences

obtained are due to host-parasite interactions, not to

bioassay effects.

In summary, nematode invasive ability was generally poor,

with 10 50% of tested nematodes successfully infecting the

host. Number of infectives invading the host was

significantly affected by bioassay conditions. Invasion

L3H 0S
C1J 0 10




o Z

co W


0 to 0

S~iaao.Lvu 3N

> (iU

HC 41
Q) *e-B

O 4) *rI


) 4.)

44O ) 0

*0 0

S tM *4

r. 0 .o

.4 O-4
r-C Cx

a)( a) -,
ro 0 (41
0 Ct o Q

10 4) 4


ON r-4 10

OW 0 4 -4

tO 0 tI
CM o rI.


0 Ct


-0 3 w

c > U



-r4 A ,

O.4 4

0 00

0 $4
WM 4 0

C 0 C U

4 *

V U 1
06 HO

o tol
$4 r

0q 1
bCi P

r 0 AQ -4


efficiency was positively related to increases in host

exposure period and number of hosts per arena, and negatively

related to increases in substrate surface area per host.

Changes in bioassay conditions had less effect on mortality;

mortality appears to be a relatively insensitive index of

nematode activity.

Invasive abilities reported were determined under what

should be optimal conditions for infection. In such

laboratory trials, there were few of the environmental

constraints that limit efficacy in field applications.

Clearly, other factors of the host or the nematode population

are present that limit successful attack, even under optimal

laboratory conditions. One possibility is that some

infectives enter a quiescent phase upon emergence from the

host cadaver (Ishibashi and Kondo 1990) and may not enter an

active host-seeking phase for several months (Hominick & Reid

1990). Further studies are needed to understand factors that

limit both host susceptibility and nematode invasion. An

understanding of invasive abilities of nematodes and bioassay

operational constraints may provide the key to differentiating

these factors.



Degree of insect susceptibility to entomogenous nematodes

of the family Steinernematidae varies among different insect

hosts, and is also influenced by nematode species and nematode

strains. Host mortality is the measure of successful nematode

attack that is commonly of interest in studies on nematode

biological control potential. A positive relationship between

nematode concentration and host mortality enables estimation

of LCs,, the nematode concentration that causes 50% host

mortality. LCs5 is a relative measure of susceptibility of the

host population. However, actual number of nematodes that

invade the host and cause mortality in 50 % of the tested

hosts is generally not known.

Kondo and Ishibashi (1986a) quantified invasion of

several Steinernema spp. by dissecting the host and counting

number of infectives established in the host. They found that

invasion by different nematode species varied quantitatively,

with greatest invasive ability demonstrated by species with

greatest efficacy (based on resultant percent host mortality).

There is a positive relationship between nematode


concentration and host invasion by infectives of Steinernema

carpocapsae (Weiser) (Chapter 2). This relationship is the

basis for determining invasion efficiency, the percent of

tested nematodes that invade a host. Invasion efficiency has

been proposed as an alternative to LC50 estimation as a

measurement of nematode efficacy (Hominick & Ried 1990).

Larvae of the fall armyworm, Spodoptera fruciprerda (J. E.

Smith), and the black cutworm, Agrotis iDsilon (Hufnagel), are

potential targets for field use of the entomogenous nematode

S. carpocapsae. Black cutworm larvae are soil insects, and

have been shown to be susceptible to S. carpocapsae in both

laboratory and field trials (Capinera et al. 1988). Fall

armyworm larvae feed in the whorl of young corn plants, and

this microenvironment that may be suitable for nematode

infection. Susceptibility has been observed in both

laboratory (Fuxa et al. 1988) and field (Richter & Fuxa 1990)

trials. Larvae of the greater wax moth, Galleria mellonella

(L.), are highly susceptible to nematode infection, and are

used as an in vivo rearing host. Thus, optimal host mortality

and invasion efficiency should be obtained with this insect.

The following study was undertaken to explore invasion

efficiency as a measure of nematode efficacy by comparing

invasion efficiency among different nematode-host

combinations, and determining correlation between invasion

efficiency and LC,0 estimate.

Materials and Methods

Insects and Nematodes

Fall armyworm and greater wax moth larvae were obtained

from the Insect Attractants, Behavior and Basic Biology

Research Laboratory, USDA, Gainesville, FL for use in this

study. The black cutworm colony was obtained from J.C. Reese,

Kansas State University, and cultured using methods of Reese

et al. (1972). Steinernema carpocapsae Mexican strain and All

strain nematodes were obtained from G.C. Smart, University of

Florida. These were reared in vivo in greater wax moth larvae

using standard rearing procedures (Dutky et al. 1964). Both

nematode strains were tested against larvae of all three host


Efficacy Bioassay

Estimates of LC50, and LCg0 were determined using filter

paper-substrate Petri dish bioassays (Woodring & Kaya 1988).

Last instar larvae were tested individually in medium arenas

(Petri dishes, 60 x 15 mm) with two pieces of Whatman #1

filter paper (diameter 5.5 cm). Nematode concentrations were

adjusted according to the susceptibility of the host species,

and always included a control (0 nematodes) and at least three

concentrations that resulted in less than 100% mortality.

Infectives were added in 1.0 ml of deionized water 1 h prior

to addition of the host. Arenas were enclosed in plastic bags


and held at 250C. Mortality was recorded after 48 h. PCPOLO

probit analysis was used to estimate LC5s and LCg9, and to

determine appropriate statistical measures (Russell et al.


Invasion Efficiency Bioassay

Invasion efficiency was determined using the procedure

developed in Chapter Two. The procedure was similar to that

used in the efficacy bioassay (above), but three nematode

concentrations (10, 50 and 100 infectives/larva) were used for

all host-nematode combinations and the arena was modified to

allow exact counts of number of nematodes added per arena for

the 10 and 50 nematode concentrations (Appendix A). Host

mortality was determined after 48 h. Number of nematodes

established in the host was determined by dissection of these

cadavers (susceptible larvae) after an additional 24 h.

Larvae that survived beyond 48 h (resistant larvae) were not

dissected, and number established was assumed to be zero.

Although this may have underestimated invasion in larvae that

succumbed after 48 h, previous studies (Chapter 2) indicated

that very few infectives (< 2 per cadaver) were found in

larvae that survived beyond 48 h. Most or all mortality

occurred within 48 h, thus studies limited to susceptible

larvae allowed more direct comparisons among these host-

nematode combinations. The study was replicated over time,

using different batches of nematodes and different generations


of host larvae. There were at least ten larvae tested per

concentration for each replicate, and at least three

replicates for each host-nematode combination. Because

mortality was so much lower in black cutworm than in the other

species, more were tested to obtain sufficient numbers of

cadavers for dissection.

Regression analysis using Proc REG (SAS Institute 1985)

was used to determine the relationship between exact

concentration of nematodes tested and number establishing in

the host. The slope of that regression estimated the

proportion of tested nematodes that invaded the host.

Therefore, invasion efficiency was determined from slope

multiplied by 100. Residual analysis using Proc REG (SAS

Institute 1985) was used to assess the adequacy of a linear

regression model to describe the relationship between number

of nematodes tested and number that invaded the host for each

host-nematode combination (Appendix B). Comparisons among

regression models obtained from each host-nematode combination

were made with a homogeneity of slopes model using Proc GLM

with Tukey's test for pairwise comparisons among host species

or between nematode strains (SAS Institute 1985). This mixed

model tested statistical significance of effects of nematode

concentration, host and nematode species on number of

infectives invading the host. Independent analyses were

conducted for susceptible larvae (those dead by 48 h), and for

total larvae tested (susceptible plus resistant larvae).

One-on-one Bioassav

Invasion was also determined with a one-on-one bioassay

(Miller 1989), a test of one infective against one host. The

procedure was the same as the invasion efficiency bioassay,

except that individual infectives were removed with 0.01 ml

water and placed in the unobscured half of the arena.

Presence of a single infective was verified with microscopic

examination. There was no selection of infectives based on

viability or activity level; rather, the one-on-one bioassay

tested the same population of infectives that was tested in

the invasion efficiency bioassay. Number of hosts per

replicate ranged from 20 to 33 individuals, and there were at

least two replicates for each nematode-host combination.

Mortality within 48, 72, and 96 h was recorded. Mortality in

bioassay arenas without nematodes (e.g. in LC;5 bioassays) was

very rare, so all mortality was assumed to be due to nematode

infection. Therefore, percent invasion was indicated by

percent mortality of all larvae tested.

Correlations between and among measurements of nematode

efficacy (as indicated by LC50 estimate) and measurements of

nematode invasion (as indicated by invasion efficiency and

one-on-one percent invasion) were determined using Proc CORR

(SAS Institute 1985).

Results and Discussion

Efficacy Bioassav

A range of LC50 estimates was obtained from the host-

nematode combinations tested (Table 3-1). Statistically

significant differences in efficacy, based on non-overlap of

95% confidence intervals, were found for both strains when

comparing black cutworm to the other two hosts. There were no

significant within-host differences between the two strains,

although LC50 for All strain was at least two times higher than

for Mexican strain for both fall armyworm and black cutworm.

Slopes obtained from probit analysis ranged from 1.1 to 2.4

(Table 3-1), and these values are typical of slopes obtained

from entomopathogens without a toxin (Burges & Thompson 1971).

Analysis of Mexican strain against greater wax moth yielded

the highest slope, indicating a homogeneity of response among

tested greater wax moth. Slopes from the other host-nematode

combinations were fairly similar. There were no significant

differences among LCg0 estimates (Table 3-1) because of large

95% confidence limits. However, ranking of the estimates for

the different host-nematode combinations remained the same.

Invasion Efficiency Bioassay

Nematode invasion efficiency in susceptible hosts (i.e.,

in hosts dead within 48 h of initial exposure to nematodes)

ranged from 11% to 31% (Table 3-2). Examination of

A) a
-' *4 4
a 4 m cO N cO m N

S00 0 0 0 0

^ )

1. O .* P* *O! C* H

NM 0 0 0 N
S4 N m 4 H N m

.$4 .O44
C 0

U) U 1.4 r


0 HiH

C 4 -W 0 t C v

r 0-
H 0 CJ 'O O I; UC

Cl0 0 H IU 0
V. clo
>*1 =r8c Co 0 0 rq
"a N m % -

0 Mi 0

4 toa r."

41 -r-4 E

A 0 4J C
(V -r4 0 L
rl 91 r( FrE4 Cd
E4 A go m



0 0' 0

. .
0 S 0 o e oo 0 ro o
3 4) z LO N (I V-4 C9 m
4J 4J0 0 0 0 0 0
0 0

'.0 + + +
?A ,4 I0 0q II II II II 4J
t V 1 o H H
r4 to m (9 r-I 0-%
*u-3. 0

0 0
H 000 0 0

0o o to I,00

0 t
E0 W 4 q V
U % 0 0 00

0 Q$4 r00
i'O S a
4 -C to to

PCflU)0 0 00*
U r- U U 4- C
d 0 "4 H*4 *' H -r4 0 00o
to 4) x x 04 x P r 4
0 V2 cn


to H04) 44

o no 5I0

e-4 l O H U O -
E 0 H) 04
N 0wl o Lo u

.0 ?a .-4 w

E-1 OH- 3

standardized residuals plotted against predicted values

indicated the need to transform the data to stabilize variance

(Figure 3-1) (Zar 1974). Therefore, number of invading

nematodes per cadaver was transformed to log (y + 1) prior to

analysis. Examination of standardized residuals showed that

the assumption of homogeneity was met (Figure 3-2). As

expected from Chapter 2, nematode concentration was the most

significant factor in predicting number of nematodes

established in the host (E = 299.46, df = 1,407; P < 0.001).

Regression was affected significantly by host species tested

(F = 32.43, df = 2,407; P < 0.001), but not by nematode strain

(E = 1.7, df = 1,407; E = 0.19). There was, however, a

significant interaction between host species and nematode

strain (E = 3.04, df = 2,407; P = 0.05). Therefore,

comparisons were made among all six host-nematode combinations

by converting the original data to percent invasion efficiency

(number of nematodes established divided by number of

nematodes tested, multiplied by 100), and determining

differences between means with Tukey's mean separation test.

Invasion efficiency of Mexican strain in greater wax moth was

significantly higher than all other host-nematode

combinations, and invasion efficiency of All strain in greater

wax moth was significantly higher than either Mexican strain

or All strain in fall armyworm. There were no differences

between fall armyworm and black cutworm for either nematode

strain tested.

0 W44 -0 .C
4) 0 4) to0 a 4)
uc )0 4)0000

0 04C
c ai o a-o o
*" (, 3 <--IlC ... S

r. 4a 0 a
a) 0 V)

> -' 4)- 0
> 4 4) $4
a)1 .4 5 0 w

to 0 tS 41
0 m t to

to, ) 0 ,, U) t.0 0
.0 44 O N
m a :4 04

N) ~ U Q

tQ "aP ,$ 4 'V "0

-, ) r
0l. U)4a Fl

to)to to Uq.4

ra I) $4 u
*MA! !

0 $4 r4 0; > -

ca g g d
0 *4' r 4*

0 O Q# O t

-4 a o t t
-rl : 5 IIg|r
& wr a" uollr

4 mm rmmd

0 me 66.e0...




_? J





*in 0..




0 N -.0 N



.0 **

0 N 0 N 0
rivnisa cmdua oi

a 0

* *4

o 0) .*4 e-4 x
.0 w to 0 0-I t0

-ri 4J 144 <40

4) 4
4C 0
^ir4 0

a) Q 4 t
4. 0t H
10k 0 4
a3 m r i nI
m a 4) 43. r. 4

.r4 4) to r.4 M ON
c, as .x*

of V a 0 -'ro E

c r-I 43 I U

I(E oz g q

-0 V -4 0 0a

go 01

N 44-f.4 O

45 .q4 u g
.go 0 I

m 4".,4 V0
.) .0 ,t X to

O C4) a > 0

0 t O 44 to
k 0 0 to

4 0 0- 4 i
lx4)14J 0 r
44444VM 0

oi o.. 0mI **

(C) J *- 0 '- C C)


& *

*, 0

0 .

-emmm 0



:** .

(C) Ni 0 *- Nt C
-ivnaisau aaziauva~vis

CT N 0 N C,



41 I :

o *
*j .*

I i I I



0* *

m m U 0

* m

* *~


+ I




0 W



o W



o I






A separate analysis was run on total larvae tested, to

determine if comparison of invasion efficiency was biased by

excluding zero counts from resistant hosts. This resulted in

an apparent decrease in invasion efficiency (Table 3-3);

however, there was a significant correlation between the two

measures of invasion efficiency (r = 0.92, E < 0.01).

One-on-one Bioassay

One infective was capable of causing mortality in all

nematode-host combinations tested, with some mortality

occurring by 48 h (Table 3-4). The 72 h mortality was

significantly correlated to 48 h mortality (r = 0.80, P =

0.06) and to 96 h mortality (r = 0.93, P = 0.02), but there

was little correlation between 48 h and 96 h mortalities (r =

0.75, l = 0.15). There was a two-fold increase in 72 h

mortality in both greater wax moth and fall armyworm when

tested with All strain versus Mexican strain infectives.

Mexican strain infectives were slightly more effective than

All strain infectives against black cutworm, however.

The 72 h mortality was used for analysis of correlation

of one-on-one bioassay results to measurements of invasion

efficiency. There were no significant correlations between

this measure of invasion and either invasion efficiency

measured in susceptible hosts (r = 0.20, F = 0.70) or in total

tested hosts (r = 0.44, 1 = 0.39). Invasion efficiency of All

strain nematodes in both greater wax moth and fall armyworm

U) Ii )-r4
W 0 0

0 o

0 00

4- 4
S40) 0 r-

PU .4

4M0 1


0 0
*rl .4 to

. "-' V

. 4
0I w0 0 1
>1O 0

.0 r
0 6 *9.4
o r

4 1 4
Sfl 0

0) 4)
P4 to rq

0.- 0

H 4.)
0 to

V,4 c4 )

r-I V-f

o 0 0
S m 0
'~ ~












U) 0
.P4 H

0 044
0) 4-4
1 0 0

-4 0

0 V 0

04 w 0
a .0 u


r) CIO 0
Cn cN 0

0 *
41 (a




d c

00 '


4 4)

0 -c

to 0

I r

04.) 0
0U t


E-4 CA Z










I o 0

r, Ln

m 0 CO 0 01 '0

0. M u-I co %D (n

(4, 0 0 0 0 0
co W 0r v M M

( 0 -4 4
U-4 r-4 q F-I U r-4
x X 4 X 4

ON 0




4 .

A1 (

V 0


6 0I




0 q

*" 0



E- t
C3 r6
c u^*



larvae was equivalent to percent mortality in one-on-one

tests. In all other combinations, invasion efficiency was

greater when nematodes were tested as a group than when tested


Invasion efficiency, whether determined from susceptible

hosts, from total tested hosts, or from one-on-one tests, was

negatively related to LC50 estimate, but correlations were not

significant (r = -0.22, P = 0.68; r = -0.58, = 0.23; r = -

0.49, P = 0.31; respectively). Surprisingly, LCs5 estimates

of All strain were similar for greater wax moth and fall

armyworm, but invasion efficiency in susceptible hosts was

similar for greater wax moth and black cutworm. There were no

obvious by host or by nematode strain pattern in one-on-one


Lack of correlation between levels of host mortality and

levels of nematode invasion has been recorded in other

studies. Kondo and Ishibashi (1986b) compared infectivity of

three steinernematid species against larvae of a cutworm,

Spodoptera litura (F.), in soil-substrate Petri dish

bioassays. In trials with 1000 infectives per host, S.

carpocapsae DD-136 strain caused 100% host mortality and had

22% invasion efficiency. Steinernema alaseri (Steiner) caused

only 5% host mortality and had < 1% invasion efficiency.

Steinernema feltiae (Filipjev) (= S. bibionis), however,

caused 70% host mortality, a level similar to that caused by

S. carpocapsae, but had < 1% invasion efficiency, a level


similar to S. alaseri. The same trends were obtained in

parallel studies done in filter paper-substrate Petri dish

bioassays (Kondo & Ishibashi 1986a). Gaugler et al. (1990)

compared three S. carpocapsae strains against greater wax moth

and found no differences in LCso estimates from standard

efficacy bioassays or in percent host mortality from one-on-

one bioassays, but did find significant differences in

invasion efficiency. The authors speculated that greater wax

moth is too susceptible to accurately reflect increase in

pathogenicity due to increase in invasion efficiency. This

may be true for even the comparatively resistant black

cutworm, since invasion by one infective is sufficient to

cause host death. Thus, there would be little correlation

between levels of host mortality and levels of nematode

invasion when mortality is ensured by invasion of the first

infective into the hemolymph. Other factors would determine

nematode attack and invasion efficiency.

Infectives may respond to quantitative differences in

host-produced chemicals. Greater wax moth and black cutworm

may produce more host cues then fall armyworm, so a greater

proportion per concentration attacked those two hosts.

Success of attack, however, may be limited by other factors,

i.e. number of natural body openings (Kaya & Hara 1980, Kaya

1985), spiracular morphology (Bedding and Molyneax 1982,

Gaugler 1988) or size of host (larger larvae more resistant

because there is more surface area for infective to traverse)


(Kondo & Ishibashi 1987). Last instar black cutworm larvae

are much larger than greater wax moth (approximately 800

versus 250 mg), which may be a factor in higher resistance to

infection. Fall armyworm are intermediate in size

(approximately 400 mg).

Invasion efficiency was highest in greater wax moth for

both strains, confirming the assumption that invasion

efficiency would be optimal in this host. Invasion

efficiencies of All strain infectives was equal in greater wax

moth and black cutworm even though LC5s estimates against these

hosts were significantly different. The lack of correlation

between percent host mortality and percent invasion efficiency

indicates that these are mediated by separate factors. The

high level of invasion when multiple hosts are present in the

arena (Chapter 2) indicates that decrease in nematode invasion

measured against individual greater wax moth, fall armyworm,

and black cutworm was not due to limitations on number of

infectives capable of infecting the host. Thus, factors in

the host, even in highly susceptible hosts, restrict ability

of infectives to enter the host.



Last instar larvae of the greater wax moth, Galleria

mellonella (L.), have been commonly used to bioassay most of

the entomogenous nematodes in the family Steinernematidae;

standard bioassays using this host have been developed

(Woodring & Kaya 1988, Georgis 1990). Greater wax moth larvae

in this stage and late last instar larvae of the black

cutworm, Aqrotis ipsilon (Hufnagel), were previously used in

studies described in the previous chapters on the

pathogenicity of the entomogenous nematode Steinernema

carpocapsae (Weiser). Concentration-dependant host mortality

and invasion efficiency were fairly consistent between trials;

however, there was a large amount of variation between

individuals within trials. Factors in the host were

hypothesized to primarily restrict the ability of S.

carpocapsae to infect (Chapter 3). One potential host factor

is age. Host age affects susceptibility of insects to

entomopathogenic viruses and bacteria (Tanada 1964).

Susceptibility to entomogenous nematodes may be affected

by host stage, and by instar or chronological age of


individuals in the same stage (Boivin & Belair 1989; Bracken

1990; Fuxa et al. 1988; Geden et al. 1985; Kaya 1985; Kondo &

Ishibashi 1986c; MacVean & Brewer 1981). Preliminary studies

of invasion efficiencies of S. carpocapsae All strain and S.

carpocapsae Mexican strain infectives indicated that nematode

invasion was more successful in younger greater wax moth

larvae (NDE, unpublished). Thus, variation in host age may

account for some of the variation in invasion efficiency

recorded in Chapters Two and Three. Size of the host may also

affect number of nematodes invading a host (Kondo & Ishibashi


The purpose of this study was to examine the influence of

variation in host age on nematode pathogenicity. There are

several difficulties in determining physiological age in

greater wax moth larvae, e.g. small size of early instars and

brevity of larval stadia. Therefore, additional studies were

done with black cutworm larvae. Attributes of black cutworm

included ease in differentiating newly molted larvae, large

size of larvae, and long larval developmental time

(approximately 4 wk from egg hatch until pupation), which

facilitated tests of inter-instar and intra-instar age

effects. Therefore, the effects of host age on host

susceptibility and nematode invasive ability were tested with

S. carpocapsae All strain infectives and larvae of both black

cutworm and greater wax moth.

Material and Methods

Insects and Nematodes

The black cutworm colony was obtained from J.C. Reese,

Kansas State University, and cultured using the methods of

Reese et al. (1972). Greater wax moth larvae were obtained

from the Insect Attractants, Behavior and Basic Biology

Research Laboratory, USDA, Gainesville, FL. Steinernema

carpocapsae All strain nematodes were obtained from G.C.

Smart, University of Florida. These were reared in vivo in

greater wax moth larvae using standard rearing procedures

(Dutky et al. 1964).

Determination of Physiological Host Age

Molt by an individual indicates start of a stadia and

size of the head capsule indicates instar number. Direct

observation of molting by greater wax moth larvae is

difficult. Early instars are very small, and larvae remain

hidden in their food until late in the last larval instar. At

that time (about 21 d after initial egg placement) they move

up, out of the diet and form a loose cocoon. These wandering

greater wax moth are used for in vivo rearing and were used

for previous studies on host mortality and nematode invasion

(Chapters 2, 3). In the standard greater wax moth rearing

protocol, eggs are set up daily in mass rearing containers

with food following the procedure of King et al. (1972).


Containers set up on the same day contain larvae that are

fairly synchronous in development. Therefore, chronological

age was determined from the number of days since initial egg

placement (0 d), and a range of larval ages was obtained by

sampling consecutively dated containers. A combination of

factors including larval weight, color, wandering-cocoon

spinning behavior (Bean & Silhacek 1989), and head capsule

measurements were used to estimate physiological age of tested

individuals. Greater wax moth were too small to test before

15 d, and by 23 d larvae had formed solid, interlocked cocoons

which made it difficult to obtain larvae without damaging

them. Therefore, 15- to 22-d larvae were tested.

Direct observation of evidence of a molt was used to

determine host age of black cutworm, although some larvae may

burrow into the artificial diet as they feed, they usually

remain on or near the surface of the diet and are generally

visible. Black cutworm larvae were checked daily to detect

individuals that had molted within the previous 24 h, and

instar number was determined from measurement of the shed head

capsule (Archer & Musick 1977). Newly molted larvae were

transferred to fresh diet, and instar and date of the transfer

were recorded. Instar number and intra-instar chronological

age were used to indicate physiological age. There are seven

larval instars, and stadia length increases with instar

number. Intra-instar age groups tested were third and fourth

instar: 0-, l-d; fifth instar: 0-, 1-, 2-, 3-d; sixth instar:

0-, 1-, 2-, 3-, 4-d; and seventh instar: 0-, 1-, 2-, 3-, 4-,

5-, 6-d. Most of the 7-d seventh instars were prepupae and

were not included in this study.

Mortality and Invasion Efficiency Bioassay

Host mortality and nematode invasion efficiency were

determined concurrently using filter paper-substrate Petri

dish bioassays. Preweighed larvae were tested individually in

Petri dish (60 x 15 mm) arenas with two aligned semicircular

pieces of Whatman #1 filter paper (radius 2.75 cm). The

nematode concentration tested was 50 or 25 infectives per

black cutworm or greater wax moth larva, respectively. After

1 h, larvae were added to the bioassay arena. Weight, color

and behavior were recorded for each individual. Arenas were

enclosed in a plastic bag and held at 250C. Totals of 300

black cutworm and 80 greater wax moth larvae were tested.

Mortality was recorded after 48 h. Measurements of head

capsule and shed head capsule, if a molt occurred during the

bioassay, were recorded. Actual number of black cutworm per

age group varied from 3-24 because final instar identification

of black cutworm was based on post-experiment head capsule

measurements, and some individuals had been initially assigned

to the incorrect instar. Actual number of greater wax moth

per age group varied from 5-23. Number of invading nematodes

was determined by dissection of cadavers after an additional

24 h. Invasion efficiency was calculated from the number of

internal nematodes divided by the exact number of tested

nematodes, multiplied by 100. Cadavers without internal

nematodes upon dissection were deleted from analysis.

One-way analysis of variance (ANOVA) was used to test the

effect of host age on host weight and nematode invasion

efficiency using Proc GLM (SAS Institute 1985). Significant

ANOVAs were followed by Duncan's multiple range test (Duncan

1955) or by Tukey's mean separation test (P = 0.05). Data

were square-root transformed prior to analysis to stabilize

the variance. For greater wax moth, comparisons were

conducted using estimated physiological age. For black

cutworm, an inter-instar comparison was done using data pooled

by instar, and independent intra-instar age comparisons were

conducted for each instar. The relationship between host

weight and nematode invasion was determined with Proc REG (SAS

Institute 1985).

Results and Discussion

There was 90-100% mortality by 48 h for all greater wax

moth chronological age groups, with the exception of 50%

mortality for the 21-d sample. Head capsule measurements of

greater wax moth showed that fifth, sixth and seventh instars

were tested. A single fifth instar was obtained from the 15-d

sample, sixth instars were obtained from the 15-, 16- and 17-d

samples, and seventh instars were obtained from all samples

except 15-d. Physiological age, based on color and body

weight, could be determined accurately for sixth-instar

greater wax moth. The sixth stadium lasts 48 h, with weight

increasing from 16 mg to a maximum of 55 mg within the first

24 h (Bean & Silhacek 1989). Thus, sixth instar greater wax

moth were classified as newly molted (dark body color) or

late-instar (light body color). Except for dark-colored newly

molted individuals, physiological age could not be accurately

determined for seventh-instar greater wax moth. The seventh

stadium lasts six days, with newly molted individuals weighing

53 mg (Bean & Silhacek 1989). Weights of males and females

diverge after one day, reaching maximums of 183 and 269 mg,

respectively, and begin wandering-spinning behavior after

three days (Bean & Silhacek 1989). Therefore, based on weight

and behavior, seventh instar greater wax moth were divided

into four groups. These groups were: newly molted larvae;

small larvae (wgt < 200 mg); large larvae (wgt > 200 mg); and

wandering-spinning larvae. The small-larvae group included

pre-maximum weight males and females as well as maximum weight

males, and thus represents the greatest mix of physiological

ages. The large-larvae group was composed of maximum weight

females. Nematode invasion efficiency and host weight were

analyzed for the six physiological age groups, and effect of

age group was significant (Table 4-1). Invasion efficiency

was the highest in the late-instar sixths, and it was

significantly higher than invasion in the large-larvae


- A C% -

) m




44 4

- 0

0 0
0 (0
(f~ Q

* a




$ 4

0-l 0
r-4 0
A v xr
to 9 t
E-4 )



C E:


> >1
0 41
4j C
a 0
*P -,-

a 0
M -- 0

0 .4J

to 0


O > *

Q) (4

.4 "w w
S40 0 0
C0 0 V

0 >H 0
0 to 3;
0 (0 U)

to 0 t-4 0
*-4 0 8 4
0 t

0 OH O


a) 0'44' 01

D00 v'

Black cutworm mortality was affected by host instar, and

susceptibility decreased with host age increase (Table 4-2).

Invasion efficiency, however, did not mirror trends in

susceptibility. The highest invasion efficiency occurred in

sixth instars. Among the earlier instars, there was a slight

trend of an increase in invasion efficiency with increasing

age, but the differences were not significant.

No significant differences in host mortality or nematode

invasion efficiency due to intra-instar host age were observed

among third or fourth instars (Table 4-3). There was an

increase in nematode invasion as age of fifth instars

increased, but the differences were not significant. However,

intra-instar age did affect host mortality for sixth and

seventh instars (Table 4-3). For both instars, the newly

molted larvae (0-d) were more susceptible than older larvae of

the same instar. Nematode invasion efficiency increased as

age of sixth instars increased, with significant differences

between 0-d and 4-d, and this increase continued the trend

observed among the fifth instar age groups. This trend was

not continued in seventh instars. Instead, for most of the

seventh instars, invasion efficiency dropped back down to the

level of fifth instar. There was a significant decrease in

invasion efficiency in 6-d seventh instars compared to the

other seventh instars.

Regression analysis was used to determine the

relationship between invasion efficiency and host weight for

o -
OH ,

9 4






0 -4

0 0




H 4)
4 r



E-4 r0.



*H r



141 63




0 '

H 9(0




Q r,,

4 41 41
C9, q49 LA +J

Cr >4

41 ( r.

C3 0

r. 0

*4) 4

e < e 5

CC 0c

0 o tn

>1 04.)
4 ) W 4

4l) 4


CO 0 -41
C 0

.0 go0

g0 c 4 U(
0 "c ~ >1
440 >1HJ
440O~~ aU~


n r-- o r- -
m 0CCI

0 r-4 H-4 0 O
H- m m Ln w

*M -
0 V
0. 4)





4w 0

0 1




*4 0

r 4)


4 1
) 4
- r. 0
m r1ff.
A e in


-P C
A 4)


0 %



a >
0 -4
*M C
< u t
k *4


to 0

c k

to 4
4f) V

C< 1
a *4 .4
V 04) 0

0 a 4) '

44 g Q)
0 C 0 *
4j 0<1^

od 0
0.1 0 &

e" O 04 C1

N )C ) b
go.4.ca4 ) a

to r. r.
U -51 e
fl 05 41S-
.r4 Q V 05

o co

rI- c

u- *r
"4i "4

0 "

0 H

Q 44

C0 0

"- m N r "
S 0 H0
S .H II II 4

>o V

*r ^ '0 eQ .
4 rl n I

II ,1 .4
*) i n
O v V0

V.Q 'v

1B 0Oto 0 0
.) >

4 I 4
COco a

> a04

r g
1 o o0

.r4 FA C)-I-I
H IA :j

0 H to0 pi4











4 R






CM >,
X 1.04 4-4

4J 4J

*H *u

3 U
*'o *a

41(a 0t


0 **

r-4 M E
o4 .4)

O 0

.> 4

u 5 o

0 wQ (0 44 M

^ Gi co r



f-i r-o e w
0 m 9
4) 90 44 1

cn O 0 C 0 o

O0 LA m. %0 N

o r-I C4 m









c .





0 O N A

A LA 1- A A A

in tO M N N O

w O 0 l Hm C
LA 'o LA ^ C4 (N

(0 N O Cl 0

N cO N 0


nH H o
3 S ^ <^ S *

O N rN m m LA LA
m v mA v N N

O r-4 N Ml V L W0
o *-Mn 4in


SC a)

a0 r
> t0o

a 4)

Eu c

*^ 4)
^*< M

H :3 C4 0
r-4 0

0 .-4 0 U
.4)4) E-4 rEu

. r X H


S II o

o II I V



* l o



both greater wax moth and black cutworm. The polynomial model

y = wgt + wgt2 best described the relationship between

invasion efficiency and host weight for greater wax moth (F =

4.2; df = 2, 71; E = 0.02; y = 15.2 + 0.13 wgt + 0.001 wgt2,

R2 = 10.6) and for black cutworm (F = 6.3; df = 2, 194; P <

0.01; y = 34.5 + 0.04 wgt + 0.0001 wgt2, R2 = 6.1), when all

ages were combined. Invasion efficiency in sixth instar

greater wax moth was also described by a polynomial model (F

= 26.4; df = 2, 11; P < 0.001; y = 37.7 1.71 wgt + 0.03

wgt2, R2 = 82.8). The curvilinear regression line predicted

by the polynomial model reiterated the pattern observed in

effect of age on invasion efficiency. Invasion efficiency

increased as weight increased until early in the seventh

instar, and then decreased as weight increased (Figure 4-1).

A simple linear regression model was adequate for sixth instar

black cutworm (F = 4.4; df = 1, 53; P = 0.02; y = 23.5 + 0.18

wgt, r2 = 14.5). There was no relationship between host

weight and invasion efficiency in the other instars of black

cutworm or greater wax moth. Kondo and Ishibashi (1987) found

that smaller larvae contained fewer invading nematodes than

did larger larvae. This direct relationship was found in

sixth instar black cutworm, but increase in size was not

directly related to increases in invasion efficiency for most

of the ages tested.

Increased susceptibility to nematodes with increasing age

has been observed in some hosts. Small size of the oral


+ 0

+ D 00

II 0 0



00 0 D

0 0 0

0> 0
0 0 ,

0 ro

co In m C C -

(%) AON3101dd3 NOISVANI

* >i4 C

-4,44-I C
0 0,

0 I

"4 10

-0 0U

$,--4 0 j 0
r-I ( 0.

0 4
C (0

0 0 0
fa o in

V U 0 0
-P0 4 34P
J) 0 0

i::00 H

0 M-
tr # 0 w!

Wi 04

) u00
0 *? 00

r 0 "F 0 .0

3I4 10 1


^ L


cavity of young mosquito and black fly larvae (in which

ingestion is the only mode of infection) physically excluded

or damaged infectives; infectives were less damaged by

ingestion in older larvae and were able to infect the host

successfully (Dadd 1971, Gaugler & Molloy 1981). Behavioral

differences between neonates and older larvae were responsible

for neonate resistance observed in noctuid larvae, as neonates

moved away from infectives placed on diet in the bioassay

arena and thereby avoided contact with nematodes (Kaya 1985).

Small size of young nymphs (pronotal width < 3 mm) of the

tawny mole cricket, Scapteriscus vicinus Scudder, was thought

to prevent infection by S. scapterisci Nguyen and Smart

infectives, as larger nymphs were susceptible (Hudson & Nguyen

1989). Adult S. vicinus were three-times more susceptible

than large nymphs, although there was little difference in

size (Hudson & Nguyen 1989), thus size may not be the only

factor limiting infection in nymphs.

Decreased susceptibility with increasing host age, at

least within a stage, is found more commonly in studies with

nematodes. Fuxa et al. (1988) found a decrease in

susceptibility of fall armyworm, Spodoptera fruciperda (J. E.

Smith), to S. carpocapsae as larval age increased from 1st to

3rd to 5th instar. In that study, 100% of 1st instar, 38% of

3rd instar and 10% of 5th instar succumbed to 1 infective.

Similarly, mortality of Heliothis armicera Hubner larvae

decreased as age, indicated by larval weight, increased


(Glazer & Navon 1990). Second instar cabbage maggots are more

resistant than first instar (Bracken 1990).

Fewer studies have looked at intra-instar variation.

Susceptibility of carrot weevil adults decreased with

increasing age (Boivin & Belair 1989). Teakle et al. (1986)

found intra-instar fluctuations in susceptibility in larvae of

the lepidopteran Heliothis punctiger. There was no intra-

instar variation until the third instar, and susceptibility

generally decreased as inter-instar and intra-instar age

increased. However, an increase in susceptibility was

observed late in the third and fourth instars, just prior to

molt to the next instar. Thus, they concluded that

susceptibility to virus was highest at hatching or molting.

Concurrent studies found no corresponding difference in

susceptibility to injected virus in third or fourth instars,

but newly molted fifth instars (ultimate larval instar) were

more resistant to injected virus. Therefore, they speculated

that change in susceptibility was mediated by the gut in third

and fourth instars, but by a different mechanism in fifth

instars. The mechanism for the gut-mediated resistance

proposed was a higher rate of food passage in these larvae

that prevented viral attachment in the midgut. High rate of

food passage has also been proposed as a mechanism for

resistance to nematodes by scarabid larvae because infectives

are unable to contact the gut wall (Bedding & Molyneux 1982).

Increased resistance of fifth instar H. punctiger to both per

os and injected virus was thought to be due to host changes

related to pupation (Teakle et al. 1986).

Unlike the studies on viral susceptibility, in which

virus was given pe os on diet, the bioassays with black

cutworm and greater wax moth did not include food. Ingestion

is not a prerequisite for infection to occur in lepidopteran

larvae. Larvae do chew up the filter paper in the bioassay

arena, however, either to feed on or to burrow into. This

feeding activity may aid infectives by increasing passive

movement into the gut, or be deleterious by physically

damaging the infectives. Rate of feeding and gut condition

(empty versus full) may be inferred from the average initial

weight recorded for the intra-instar age group. Results from

black cutworm indicate that these larvae actively feed early

in the instar. Newly molted sixth and seventh instar black

cutworm were highly susceptible to nematode infection. These

larvae have fairly empty guts and may more actively consume

the filter paper. Thus, both factors may contribute to the

increase in susceptibility of newly molted larvae. Increased

resistance in 5-d and 6-d seventh instar black cutworm appear

to be due to changes in the host that are related to pupation.

Black cutworm prepupae are fairly resistant to S. carpocapsae

and few nematodes are found established in cadavers of

prepupae (NDE unpublished). Prepupae that molt within 24 h of

exposure to infectives, however, are highly susceptible to

nematode infection (NDE unpublished).


As observed in Chapter Two, there was no apparent

relationship between invasion efficiency and percent

mortality. Larvae appear to be most susceptible to nematode

infection immediately after or immediately prior to a molt.

The susceptibility after the molt may be explained by lack of

food in the gut and/or increased feeding activity. The

increase in invasion efficiency in larvae late in the instar

may be due to lower activity level of these larvae or

vulnerability of newly molted individuals to direct cuticular

penetration at ecdysis. It is not known why there was such

high invasion efficiency into the penultimate larval instar.

Additional studies are needed to see if this is true in other

hosts and to determine the underlying mechanisms.



Previous chapters have examined factors in host-nematode

interactions that directly affect nematode pathogenesis.

Understanding these factors may improve our ability to predict

biological control potential of entomogenous nematodes. These

studies did not consider tritrophic interactions, that is,

factors in the first trophic level (e.g., host's food) that

affect host growth and development and may, in turn, affect

entomophagous organisms attacking that host.

Insects reared on artificial diet are often used in

laboratory studies, but plant foliage that insects consume in

the field is often lower in nutrient content and may contain

secondary chemicals that are not present in artificial diet.

Chemical content, especially allelochemical content, can

negatively affect herbivore growth and survival. House and

Barlow (1961) demonstrated that parasitoid success was

influenced by nutritional quality of the host's food.

Secondary chemicals in the host's food may be deleterious to

parasitoids if these allelochemicals are present in the host's

tissue (Thurston & Fox 1972), or if size of the insect host is


reduced (Beach & Todd 1986). These results have led to

questions on compatibility of host plant resistance and

biological control in insect pest management (Bergman & Tingey


Although deleterious tritrophic effects have been

documented, biological control may be enhanced by concurrent

use of host plant resistance. Starks et al. (1972) found that

combined use of parasites and resistant plants yielded

improved control of the greenbug, Schizaphis graminum

(Rondani). Decrease in food quality has been linked to

increased susceptibility of grasshoppers to insecticides

(Hinks & Spurr 1989), and susceptibility of Colorado potato

beetle, Leptinotarsa decemlineata (Say), to the

entomopathogenic fungi, Beauveria bassiana (Balsamo) (Hare &

Andreadis 1983).

The potential for deleterious tritrophic effects on

pathogenicity of nematodes has been largely unexplored. Black

cutworm, Aqrotis ipsilon, (Hufnagel) is a polyphagous

herbivore, and will feed on a wide variety of plants (Apple

1967). Black cutworm will readily feed on collard, Brassica

oleracea, foliage in the laboratory. Collards belong to the

Cruciferae, members of which contain glucosinolates, among

other allelochemicals, that may be toxic to non-specialist

herbivores (Van Etten & Tookey 1979). Therefore, a collard-

black cutworm-Steinernema carpocapsae All strain model was

used for an examination of influence of host's food on


pathogenesis of entomogenous nematodes. Both indirect effects

of changes in host nutritional or physiological condition, and

direct effects of the allelochemical, were examined.

Materials and Methods

Insects and Nematodes

The black cutworm colony was obtained from J.C. Reese,

Kansas State University, and cultured using methods of Reese

et al. (1972). Steinernema carpocapsae All strain nematodes

were obtained from G.C. Smart, University of Florida. These

were reared in vivo in greater wax moth, Galleria mellonella

(L.), larvae using standard rearing procedures (Dutky et al.


Bioassay Procedure

Filter paper-substrate Petri dish bioassays were used for

all studies (Woodring & Kaya 1988). Black cutworm larvae were

tested individually in Petri dish (60 x 15 mm) arenas with two

pieces of Whatman #1 filter paper (diameter 5.5 cm).

Infectives were added in 1.0 ml of deionized water.

Preweighed larvae were added to the bioassay arenas after 1 h.

Arenas were enclosed in plastic bags and held at 250C.

Mortality was recorded after 24, 48, 72, 96 and 120 h.

Nematode invasion was quantified for some larvae that died

within 48 h of initial exposure to infectives. Number of

nematodes established in the host was determined by dissection

of these cadavers after an additional 24 h. By this time,

invading nematodes were late stage juveniles or adults and

were counted more efficiently.

Other cadavers were transferred to modified White

emergence traps (White 1927) to allow nematode reproduction

and infective emergence. Modified traps consisted of a medium

Petri dish (60 by 15 mm) secured (with glue from a hot glue

gun) open-side up in the center of a large Petri dish (100 by

20 mm). Two pieces of water-saturated Whatman #1 filter paper

(diameter 5.5 cm) were placed in the medium dish, water added

to the outer dish to form a moat, and a single cadaver rinsed

and placed in the inner dish. Progeny production was

determined only for larvae; production in pupae was not

tested. Cadavers were kept in emergence traps at 250C for at

least five weeks. During this time, additional water was

added to the center dish, and infectives harvested from the

moat weekly. After five weeks, all remaining nematodes were

infective stage juveniles and those still in the cadaver were

collected by dissecting the cadaver, teasing apart remaining

cuticle and other host tissues, and vigorously shaking the

dissected tissue in water in a plastic vial (12 or 20 dram).

Total number of infectives produced per host was estimated for

each individual from subsample counts. The total number of

infected produced was divided by initial host weight to


determine number of infectives produced per mg as an

adjustment for host weight.

PCPOLO probit analysis was used to estimate LC5, and to

determine appropriate statistical measures (Russell et al.

1977). Other statistical procedures included two-sample t

tests using Proc TTEST, factorial analysis of variance (ANOVA)

and regression analysis using Proc GLM (SAS Institute 1985).

Results from significant ANOVAs were followed by Tukey's mean

separation test (P = 0.05).

Effect of Collard as Host's Food

Collards were obtained from a local market. Newly molted

fourth instar black cutworm were transferred individually to

plastic cups (25 ml) containing pieces of collard foliage.

Control larvae were maintained on black cutworm diet (BCWD)

(BioServ #9240), the artificial diet used for black cutworm

rearing. Food was changed daily or as needed. Molts were

recorded and instar was confirmed by measurement of the shed

head capsule. Seventh instars weighing 0.6-1.0 g were used

for all trials, unless otherwise stated. There were four

nematode concentrations, ranging from 20-160 nematodes per

larva. The entire experiment was replicated four times, using

two generations of black cutworm and nematodes. A total of

320 larvae (10 larvae per concentration per host food per

replicate) were tested. LC,0 estimates and nematode progeny

production were determined from these larvae. Finally, 27


seventh instars per diet were tested with a single

concentration of 150 nematodes per larva, and nematode

invasion efficiency was determined for larvae that died within

48 h.

Indirect Effects of Host's Food Treatment

Based on results of comparisons between progeny

production in cadavers of collard-fed larvae versus artificial

diet-fed larvae, a series of experiments was initiated to

examine the indirect effects of changes in host physiological

and nutritional condition on nematode progeny production. The

relationship between host weight and nematode progeny

production was determined for sixth and seventh instar larvae

from collard and BCWD. Prepupae were not tested. Larvae were

exposed to concentrations of 40, 80, and 160 nematodes per

larva, and there were 180 (90 per host food) larvae tested.

An additional 20 (10 per host food) larvae, 600 to 1000 mg in

size, were killed by freezing. Lipid composition of the oven-

dried cadavers was determined from loss in weight after 3 h

Soxhlet extraction with petroleum ether (Slansky & Wheeler


Effect of host age was tested in an experiment using

larvae from artificial diet only. Intra-instar chronological

age of seventh instars was determined from the date of the

molt to seventh instar. Ages tested were 3-, 4-, 5-, and 6-d


old seventh instars. Larvae were tested with a 250 nematode

concentration, and 57 larvae (10-20 per age) were tested.

Direct Effects of Allelochemical in Host's Food

A commercially obtained glucosinolate, sinigrin, was

added to BCWD to differentiate between effects of

glucosinolates in collard foliage from effects of nutritional

differences between plant foliage and artificial diet. Newly

molted fourth instar black cutworm were transferred to BCWD

containing 0, 0.01, 0.05, or 0.1% sinigrin (wet weight).

Larvae were checked daily for occurrence of a molt and 170

(35-45 per sinigrin diet) seventh instars, either 4-d or 5-d,

were exposed individually to 150 infectives per larva. Larvae

that died within 120 h of initial exposure to infectives were

transferred to modified White traps and progeny production was

determined for these larvae.

Results and Discussion

There was less average cumulative mortality in collard-

fed black cutworm larvae than BCWD-fed larvae, but level of

mortality of BCWD-fed larvae varied greatly between replicates

(Figure 5-1). Due to lack of concentration-dependent response

in larvae from both diets in one replicate, that replicate was

deleted for LCs5 analysis. LCs5 estimates were obtained from

pooled data of the remaining replicates, and LC,, estimate

obtained from collard-fed larvae was higher than estimate from

to 0

- M


- CM

* *

U, o

Ai.lVitHOI IJSOH 3AlV-nInflvnO

1 ,*<1*.***


\ -

(.)\w\ <

$4g 0. r-4
0 4-) 0
in 3 u
-r 4) 44 44

+ 0IV

o U$4

1 a0 W

0 >

> 0*

o 0

. 0$4

44 .0

0 r-

_ 0
0 $4

%04 0

.0 0

k A4 0 to
3 U P.-) V


BCWD-fed larvae, but the difference was not significant (Table

5-1). There was no difference in nematode invasion in

collard-fed versus BCWD-fed larvae (_ = 1.05, df = 16, P =

0.31). Invasion efficiency was 17% and 15%, respectively.

There was, however, a significant difference in progeny

production in larvae from the two host food treatments (t =

5.19, df = 97, P < 0.001). Progeny production decreased from

an average of 230 infectives per mg in cadavers from BCWD-fed

larvae to 126 infectives per mg from collard-fed larvae.

Progeny production decreased as time period until host death

increased beyond 72 h, but number of infectives from collard-

fed larvae remained significantly lower than from BCWD-fed

larvae (Figure 5-2).

Total progeny production was affected by host age of

BCWD-fed seventh instars (Table 5-2), and there was a

significant decrease between 5-d and 6-d seventh instars.

However, change in total progeny production with host age was

due to change in host weight primarily, as host age did not

affect significantly the number of infectives per mg host

(Table 5-2). Larvae reared on collard foliage grew slower and

were smaller than same chronological-age larvae reared on

artificial diet. When stadia length differs, within-instar

chronological age may be an inadequate indicator of

physiological age. Thus, collard-fed larvae may be older

physiologically than BCWD-fed larvae, even though size and

within-instar chronological age of larvae were similar. This



0 g
V 1


.0 0




43 L

L 41


Q r
(0 1
0 0l










-) ,-4

p.4 g~4


0 *4


t0 4

a>r- ^
o eIL

10 ral'
> 10
I P 10
(0 0 *

C-4 >
* -4 4 r 0
4.10 to U
O 0- 4 *-I
4J 0 4. 44
Co r41 44 -H
0 *4 o -Ic

4J 4j rq
C r ..-4

r-400 0

0 *0
U O >'
S4J 4 -0
gm to
>0 0 0n-4
u o c. *-'
SC --4Q


*4 to

O 0
> -4 to
W 0 W'q


u P

n E0-4

w c Q)
O > tMa P 04-4
-4 Q) 0 0 W -*
r4 W X: 4-4 Z r

(000 x) NOILOnaOlUd AN39OlUd

S w

to 0

i t


O3 r




n C


o r

14 Ci-I
4J.) H
*.0 0

t Q


Q0 V
. C

< 0



a "











CO 0

m LA 'n t

C M >i
0 N r-4
OW e4
0 4 041
$*4 -r
o ( 4 .

41 1

Ot 00

10 4
S' flu

) 0400

UA V 0
0) 4)0 W

-0 E U 0 .

( l(M t,

14g EU 0 W

g >

4 r-1 14 4
So) -ro aV
r 0 4) 01 0

0) aS44

aE a
(0 (0

-4 i N
r-I r- ,-4

age difference may account for difference in percent mortality

between collard-fed and BCWD-fed larvae. Older black cutworm

tend to be more resistant to nematode infection (Chapter 4).

Increase in age may explain decrease in percent mortality, but

not decrease in progeny production in collard-fed larvae.

In the initial comparison between collard-fed and BCWD-

fed larvae, seventh instars tested were within the weight

range of 600-1000 mg. A curvilinear relationship was found

between black cutworm weight and invasion efficiency (Chapter

4), i.e., invasion efficiency increased with increasing weight

towards the end of the sixth instar, but decreased towards the

end of the seventh instar. A similar curvilinear relationship

between host weight and progeny production might prevent

separation of effects due to host's food from effects due to

decreased growth in collard-fed versus BCWD-fed late seventh

instar larvae. Maximum weight of collard-fed larvae was

approximately 1.0 gm, but many BCWD-fed larvae were in the

1.0-1.2 gm weight range. Total progeny production, however,

was directly related to the weight of the host, and increased

as weight increased for both BCWD-fed and collard-fed larvae

(Figure 5-3). The slopes obtained from regression equations

of larvae from the two diets, however, were significantly

different (E = 15.25; df = 1, 72; P < 0.001). Therefore,

differences in progeny production were obtained over the full

range of larval weights and results were not affected by the

differences in larval growth on the two diets.

I- I -.** o

ii o

o o 0
o .

8 8 8 8 o
,It x) NOd N

(000 L x) NOilonaGOId AN3E90UJd

00 ) Q) ) )S

r$4 0. 0

S0 0S

Sr0. 4)

4 Or4 e
4 r 0 4

4W 0 O3*
0 0
4H 00 IV

U 0 14

.A 040
O o
) '^r- 4 0 VX

0e*0 8
41 415

M "_ U M j
0 04 $4
$4 O 0
0 1 (a o

S0 0 4
S0- 0
oto o

00 > 00

(0 0

.m O
M 0 0.4- 4
01 0 44V04)


r 0 n M4
$r4 r .0 6

0 W V 4)

1 U 0 4 4-i $
W: U) 04 () 0
X! (0 0i-l

3 % 44 0
Wo r 0 1 o:

10) U (141 r-I 4
O N > it to 0r c

There are nutritional differences in plant-fed versus

artificial diet-fed larvae that may contribute to suitability

for nematode reproduction. Cookman et al. (1984) found that

lipid levels in velvetbean caterpillars fed artificial diet

were significantly greater than in caterpillars fed plant

foliage. Lipid content of collard-fed larvae was

significantly less than that of BCWD-fed larvae (t =6.22, df

= 20, P < 0.001). Lipids accounted for 16% of total dry

weight in BCWD-fed larvae, but only 8% in collard-fed larvae.

The two-fold difference in lipid level is directly correlated

with the two-fold difference in progeny production, and may be

the major factor in explaining effect of collard diet on

progeny production. Lipid content also differs qualitatively

in plant-fed versus artificial diet-fed larvae (Cookman et al.

1984), but these differences may be less important for

nematode reproduction because S. carpocapsae is able to

utilize a wide array of host sterols for normal development

(Morrison & Ritter 1986).

A representative glucosinolate, sinigrin, was

incorporated into BCWD to examine allelochemical effects

separately from nutritional effects of plant foliage

consumption. Results of two-way ANOVA on effects of diet and

time period until host death indicated that time period until

host death significantly affected progeny production (E =

15.88; df = 3, 108; P < 0.001). Mortality occurred within

72 h for 83% of the tested larvae, so comparisons were limited

to this subset of data. (See below for analysis of time period

until host death).

Among larvae that died within 72 h of exposure to

nematodes, sinigrin concentration in BCWD significantly

affected progeny production per mg (E = 3.25; df = 3,97; P =

0.025). There was no effect on progeny production per mg in

the comparisons among larvae fed diets with 0 to 0.05%

sinigrin, but there was a significant decrease for larvae from

the 0.1% sinigrin diet. Progeny production decreased from 265

infectives per mg in the 0% sinigrin treatment to 220

infectives per mg in 0.1% sinigrin treatment. This

concentration of sinigrin is biologically relevant;

concentration in collard foliage is approximately 0.07% fresh

weight (Blau et al. 1978). Unlike the trials with collard

foliage as host's food, addition of 0.1% sinigrin to

artificial diet had no apparent effect on larval growth or

development. Effect of sinigrin diet on host lipid content,

however, was not tested.

The reduction in progeny production in collard-fed larvae

may be attributed to either consumption of plant foliage

and/or presence of allelochemical in host's food. Beach and

Todd (1986) found that parasitoid production decreased two-

fold in a comparison between larvae of the soybean looper,

Pseudoplusia includes (Walker), fed foliage from a

susceptible soybean variety versus artificial diet, and

decreased two-fold again when larvae were fed foliage from a

resistant soybean variety. They did not, however, separate

effects of host size from effects of host's food, nor did they

indicate the resistance mechanism in soybean foliage. Mueller

(1983) found that superior foods for the host insect resulted

in superior parasitoid progeny production. This appears to be

true for S. carpocapsae also.

The relationship between rate of nematode infection, as

indicated by time period until host death, and host

suitability for reproduction was examined for larvae from the

0%, 0.01% and 0.05% sinigrin diets. Sinigrin at these

concentrations did not significantly affect progeny

production, so all individuals from these diets were pooled

for subsequent analysis. There was a significant negative

relationship between time period until host death and

subsequent progeny production per mg (E = 50.45; df = 1, 87;

P < 0.001) (Figure 5-4). Tukey's mean separation test was

used to compare mean progeny production per time period, and

reproduction in hosts that died within 72 h was significantly

higher than in hosts that died after 72 h. A direct

relationship between decreased host susceptibility (indicated

by percent infection or percent mortality) and decreased

suitability for S. carpocapsae reproduction was observed for

larvae of the lesser European bark beetle, Scolytus

multistriatus F., and the mountain pine beetle, Dendroctonus

ponderosae Zimmermann, and for different stages of both the

mountain pine beetle (MacVean & Brewer 1981) and the carrot

c /
1 i -
" /


S< <]t


9ON U83d

o m
o cm


0 u


ra 10 >4
0 4)4A
' 0

C0 4




' *



04) V4 *
*MC1 &

1 4
Q)i o I4

V 4 ) )
-14- 4
Pw :3 tI


fn a