Evaluation of the entomopathogenic fungus Beauveria bassiana on the red imported fire ant, Solenopsis invicta

MISSING IMAGE

Material Information

Title:
Evaluation of the entomopathogenic fungus Beauveria bassiana on the red imported fire ant, Solenopsis invicta
Physical Description:
xiii, 135 leaves : ill. ; 29 cm.
Language:
English
Creator:
Pereira, Roberto Manoel
Publication Date:

Subjects

Subjects / Keywords:
Solenopsis invicta -- Control   ( lcsh )
Entomopathogenic fungi -- Evaluation   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 117-134).
Statement of Responsibility:
by Roberto Manoel Pereira.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001747381
notis - AJG0204
oclc - 26371918
System ID:
AA00003308:00001

Full Text










EVALUATION OF THE ENTOMOPATHOGENIC FUNGUS
BEAUVERIA BASSIANA ON THE
RED IMPORTED FIRE ANT, SOLENOPSIS INVICTA













By

ROBERTO MANOEL PEREIRA


DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1991

























A Eduardo, Ricardo, Angela, Alberto, e Rodrigo,

que construiram a base de tudo.



A Julia, Carolina, Marcel, Denise, e outros por vir,

uma nova geraqco de esperantas.














ACKNOWLEDGMENTS


I thank Dr. Jerry L. Stimac for his guidance and friendship during

the course of my studies at the University of Florida, and Dr. Sergio B.

Alves for the support and teachings in the last 10 years. I am also very

grateful for the help offered by Drs. Drion G. Boucias, Clayton W. McCoy,

and James W. Kimbrough, and for the support from the faculty and staff of

the Department of Entomology and Nematology of the University of

Florida

My gratitude is extended to Lois Wood for her ever-present helping

hand, and for entertaining stories. I would also like to acknowledge the

help of people who in many ways contributed to the completion of this

work, namely: Lisa Huey, Robin Register, Reginald Coler, Hugh Smith, Dr.

TomAs Zoebisch, and Dr. David Oi.

I am grateful for the patience and understanding of my wife Bete,

and for the loving and nurturing home she creates for our daughters and

me.
















TABLE OF CONTENTS



ACKNOWLEDGEMENTS .................

LIST OF TABLES .......................


LIST OF FIGURES ......

KEY TO ABBREVIATIONS


. . . . ix

...............................xi


ABSTRACT .........................................

CHAPTERS

1 INTRODUCTION .............................

Geographic Distribution and Introduction of Fire Ants
into the United States .....................
Im portance .................................
Present Control Methods .......................
N natural Enem ies .............................
M icrobial Agents ........................
Use of Beauveria bassiana in Insect Control ..........
Beauveria bassiana as a Microbial Control for Fire
A n ts . . . .
O objectives ..................................

2 DOSE-RESPONSE OF THE RED IMPORTED FIRE
ANTS, SOLENOPSIS INVICTA, WORKERS TO
BEAUVERIA BASSIANA CONIDIA ...............


Introduction ..........
Materials and Methods ...
Fungus .........
Ants and Soil .....


. iii


....................
....................
....................
....................








Spray Tower and Spraying Procedure .......... 21
Dose-Response Experiments ................ 22
Mortality Assessment and Analysis of Results .... 24
R results ........................ ... .. ... .. .. 25
Direct Spray Experiments ................... 25
Soil Incorporation Experiments .............. 30
Surface Exposure Experiments ............... 38
Discussion .................................. 42
Effect of Fungus in Direct Spray Experiments ..... 42
Effect of Conidia Mixed into Soil ............. 44
Infection of Ants by Conidia in Soil and on Plate
Surface ...................... .... 46

3 TRANSMISSION OF BEAUVERIA BASSIANA
WITHIN ARTIFICIAL FIRE ANT NESTS IN THE
LABORATORY .............................. 48

Introduction ......... ... .................. 48
Materials and Methods ......................... 51
Fungus and Inoculation of Ants .............. 51
Preparation of Artificial Colonies ............. .52
Experimental Procedures .................... 52
Estimation of Number of Conidia in Cadaver Pile .54
Experiments with Conidia Formed on Ants ...... 54
Assessment of Survival and Analysis of Results ... 55
Results ............................... ..... 57
Experiments with Live Infected Ants in
Plaster-Bottom Cups .................. 57
Experiments with Live Infected Ants in Soil ...... 61
Experiments with Dead Infected Ants ........... 64
Conidia per Ant ...........................70
Experiments with Conidia Formed on Ants ...... 70
D discussion ................................... 74

4 GROWTH OF BEAUVERIA BASSIANA IN FIRE
ANT NEST SOIL WITH AMENDMENTS ............ 79

Introduction ................................. 79
Materials and Methods ...........................81
Fungus ................................81
Am endm ents ............................ 82
Preparation of Soil Plates and Inoculation ........ 83
Observations, Measurements and Analysis of
Results ............................ 85







R results .. .. .. .. . . . 86
Mycelial Inoculum .........................86
Conidial Inoculum ........................ 88
Ant Inoculum ...........................93
D discussion ................................... 94
Effects of Inoculum Types ................... 94
Effects of Amendments ..................... 99

5 CONCLUSIONS AND IMPLICATIONS OF
RESULTS ON THE MICROBIAL CONTROL OF
SOLENOPSIS INVICTA WITH THE
ENTOMOPATHOGENIC FUNGUS BEAUVERIA
BASSIANA .................................102


APPENDICES


A EFFICIENCY OF WASHING PROCEDURE IN
ESTIMATION OF NUMBER OF CONIDIA PER
ANT AFTER APPLICATION OF CONIDIAL
SUSPENSIONS ...........................

B EVALUATION OF EFFECT OF SOIL ANTAGONISM
ON BEAUVERIA BASSIANA INFECTION OF FIRE
ANT W ORKERS ..........................


... 112


114


LITERATURE CITED ............

BIOGRAPHICAL SKETCH ........


. . . 117














LIST OF TABLES


table page

1-1 Effective dose of B. bassiana against insect pests under
several experimental conditions as reported by different
authors....................................... 12

2-1 Lethal concentrations conidiaa / ml of suspension) of B.
bassiana against S. invicta, when suspensions are
directly sprayed on the ants which are maintained in
clean plastic cups for 15 days ....................... 28

2-2 Lethal times (days) for B. bassiana against S. invicta,
when suspensions are directly sprayed on the ants,
which are maintained in clean plastic cups for 15 days .... 29

2-3 Lethal concentrations conidiaa / ml of suspension) of B.
bassiana against S. invicta, when suspensions are
directly sprayed on the ants which are maintained in
plastic cups containing either sterile or nonsterile fire
ant nest soil for 15 days .......................... 33

2-4 Lethal concentrations conidiaa / g of soil) of B. bassiana
against S. invicta, when suspensions are applied to
either sterile or nonsterile fire ant nest soil, and ants are
maintained in soil for 15 days ...................... 37

2-5 Lethal concentrations conidiaa / mm2 of plastic Petri
dish) of B. bassiana against S. invicta, when suspensions
are directly sprayed on Petri dishes and ants are allowed
to walk in the dishes for 24 hours before being
transferred to clean plastic cups for 15 days ............ 40









2-6 Lethal times (days) for B. bassiana against S. invicta,
when suspensions are sprayed on Petri dishes and ants
are allowed to walk in the dishes for 24 hours before
being transferred to clean plastic cups for 15 days ........ 41

3-1 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected individuals and maintained in
containers with no soil (Experiment 1) ............... 59

3-2 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected individuals and maintained in
containers with no soil (Experiment 2) ............... 60

3-3 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected individuals and maintained in soil
(Experim ent 1) ................................ 63

3-4 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected individuals and maintained in soil
(Experim ent 2) .................. ................ 65

3-5 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected cadavers and maintained in containers
w ith no soil .................................. 66

3-6 Mean survivals of ants and mean weights of surviving
ants in colonies receiving different percentages of B.
bassiana-infected cadavers and maintained in containers
w ith soil ..................................... 67

A-1 Results of the estimation of washing efficiency of
procedure used in the estimation of dose of conidia per
ant after spraying of conidial suspension ............. 113

B-1 Percent mortality and infection of fire ant workers (S.
invicta) exposed to nest soil containing conidia of the
fungus B. bassiana ............................. 116
















LIST OF FIGURES


figure page

2-1 Mortality and infection of S. invicta treated with direct
spray of B. bassiana conidial suspensions and
maintained in clean plastic cups for 15 days ............ 27

2-2 Mortality and infection of S. invicta treated with direct
spray of B. bassiana conidial suspensions and
maintained in plastic cups containing sterile fire ant
nest soil for 15 days ............................. 31

2-3 Mortality and infection of S. invicta treated with direct
spray of B. bassiana conidial suspensions and
maintained in plastic cups containing nonsterile fire ant
nest soil for 15 days ............................. 32

2-4 Mortality and infection of S. invicta maintained for 15
days in sterile fire ant nest soil treated with B. bassiana
conidia ....................... .............. 34

2-5 Mortality and infection of S. invicta maintained for 15
days in nonsterile fire ant nest soil treated with B.
bassiana conidia ............................. 36

2-6 Mortality and infection of S. invicta exposed to Petri
dish surface treated with direct spray of B. bassiana
conidial suspensions, and maintained in in clean plastic
cups for 15 days ................................ 39









3-1 Number of conidia per cup and per gram of ants
produced on S. invicta cadavers, in population either
with 100% of B. bassiana-infected insects or with
mixture of 50-50% of healthy and infected ants ......... 71

3-2 Mortality of S. invicta treated with direct spray of B.
bassiana conidial suspensions prepared from conidia
produced either on SDAY plates or on ant cadavers from
experiments with or without soil ................... 72

4-1 Diameter of 15-day old B. bassiana colonies on sterile
and nonsterile fire ant nest soil amended with different
materials and inoculated with dry mycelium particle .... .87

4-2 Number of B. bassiana colony forming units (CFU's)
produced after 15 days of incubation on either sterile or
nonsterile fire ant nest soil amended with different
materials and inoculated with dry mycelium particle ..... 89

4-3 Number of B. bassiana colony forming units (CFU's)
produced after 15 days of incubation on either sterile or
nonsterile fire ant nest soil amended with different
materials and inoculated with conidial suspension ...... 91

4-4 Diameter of 10-day old B. bassiana colonies on sterile fire
ant nest soil amended with different materials and
inoculated with conidial suspension ................. 92

4-5 Diameter of 15-day old B. bassiana colonies on sterile
and nonsterile fire ant nest soil amended with different
materials and inoculated with with fire ant (S. invicta)
abdom ens .................................... 95

4-6 Number of B. bassiana colony forming units (CFU's)
produced after 15 days of incubation on either sterile or
nonsterile fire ant nest soil amended with different
materials and inoculated with fire ant (S. invicta)
abdom ens .................................... 96














KEY TO ABBREVIATIONS


ANOVA -Analysis of Variance.

ATCC American Type Culture Collection.

CFU Colony Forming Unit, i.e. spore, group of spores or other fungal

element that will initiate a colony when plated on agar plates.

C.I. Confidence Interval.

LC50 Lethal Concentration for 50% of the treated population.

LC90 Lethal Concentration for 90% of the treated population.

LD50 Lethal Dose for 50% of the treated population.

LD90 Lethal Dose for 90% of the treated population.

LT50 Time from inoculation to death of 50% of treated population.

LT90 Time from inoculation to death of 90% of treated population.

PLSD Protected Least Significant Difference.

SDAY Sabouraud Dextrose Agar Medium + 1% Yeast Extract.

SDS Sodium Dodecyl Sulfate

sem Standard Error of the Mean














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


EVALUATION OF THE ENTOMOPATHOGENIC FUNGUS
BEAUVERIA BASSIANA ON THE
RED IMPORTED FIRE ANT, SOLENOPSIS INVICTA

By

Roberto Manoel Pereira

May 1991

Chairman: Jerry L. Stimac
Major Department: Entomology and Nematology

Solenopsis invicta, a serious insect pest in the United States, is

presently managed with chemical insecticides. Recent studies

demonstrating the infectivity of Beauveria bassiana to fire ants have

renewed interest in the development of this fungus as a microbial control

agent.

Laboratory experiments conducted under four sets of conditions

evaluated the effectiveness of B. bassiana to infect and kill ant workers. A

LC50 of 1 x 106 conidia/ml was estimated for ants directly sprayed with

conidial suspensions and subsequently transferred to cups with or without









sterile or nonsterile nest soil. A LC50 of 1 x 103 conidia/mm2 was

estimated for ants exposed to plastic surface covered with fungal conidia.

If conidia were applied to nest soil, the LC50 was 1 x 102 conidia/g of sterile

soil compared with 2 x 109 conidia/g of nonsterile soil, suggesting that

antagonistic factors detrimental to fungal infection were present in the

nonsterile soil. The estimated LD50 was 2100 conidia/ant worker.

Transmission of fungal disease from infected to healthy ants

occurred only in the absence of soil. In the presence of soil, ants removed

cadavers from the nest and prevented contact with fungus by covering

sporulating corpses with soil. Conidia harvested from ant cadavers were

as infective to S. invicta as conidia produced on an artificial medium.

However, conidial yield on cadavers exposed to live ants was less than on

cadavers not exposed to live ants. Both ant hygienic behavior and corpse

removal prevented spread of fungal disease in nests.

Beauveria bassiana was able to grow on nonsterile soil when

mycelial particles or infected ant abdomens were used as inoculum. With

conidial inoculum, no growth occurred and final fungal biomass was less

than initial inoculum. On sterile soil, more fungal growth occurred with

all inoculum types, but more on soil amended with ant carcasses or rice

powder than on soil amended with chitin or no amendment.

If sufficient contact between infectious conidia and the insects can be

achieved, S. invicta can be controlled by B. bassiana regardless of soil

antagonism and poor stability and growth of fungus in soil.















CHAPTER 1

INTRODUCTION

Geographic Distribution and Introduction of Fire Ants into the
United States
The genus Solenopsis occurs worldwide with approximately 160

species (Wilson 1986), mostly in the Neotropical region (Wheeler et al.

1922). The subgenus Solenopsis, which represents the fire ants, is

restricted to the new world (Wojcik et al. 1976). Six species and 2 hybrids

of the S. geminata species group, which includes the fire ants and some

related ants, occur in the United States (Trager in press). Of these two

species, S. invicta and S. richteri, are imported from South America

where 15 other species of fire ants occur. Native American fire ants

include S. aurea, S. amblychila, S. xyloni, and S. geminata, which occurs

in Florida.

Solenopsis richteri was probably introduced into the United States

in cargo ship ballast in 1918 at the harbor of Mobile, AL (Lofgren 1986,

Buren et al. 1974), and initially it spread over a large area. However, when

S. invicta was introduced at the same area within the period 1933-45

(Lennartz 1973), it displaced S. richteri which survives today only in a

limited range in northern Mississippi and eastern Alabama (Buren 1983).










A broad zone of interbreeding has been detected where the two

populations meet in northern Alabama and Mississippi and northwestern

Georgia (Ross et al. 1987, Vander Meer & Lofgren 1988).

Due to its high reproductive rate, large colony size, efficient

dispersal capacities via mating flights, floatation on water or man-assisted

transport in nursery stock, S. invicta has been successful in colonizing

disturbed areas and outcompeting native ant species (Wojcik 1983). From

its initial introduction point, the fire ants have spread over a large area in

the southeastern United States, now estimated at approximately 93

million ha (Lofgren 1986), with most of the area inhabited by S. invicta

only.


Importance

Fire ants assume special importance in the United States due to

several problems they cause both as an agricultural and an urban pest. It

has been estimated that in 85% of the households in the state of Georgia

control of fire ants was attempted in 1989 (Diffie & Sheppard 1990). The

same study estimated a cost per household of little more than US$ 36,

including US$ 5 for medical expenses related to fire ant stings. High cost

of repairs also occurs when fire ants damage highways (Banks et al. 1990).

Serious medical problems from allergic reactions to the sting of fire

ants are not uncommon among the human population in the infested

regions of the US. The problem assumes greater importance with








3

children, since they are generally less resistant to sting-related medical

problems and receive a disproportionally high share of stings (Clemmer &

Serfling 1975, Adams & Lofgren 1981). About 3.8 and 4.5% of the

American population has severe reactions to the fire ant stings, requiring

some type of medical attention (Apperson & Adams 1983). The symptoms

of severe reactions in humans can include nausea, vomiting, dizziness,

perspiration, asthma, and other symptoms of severe allergic reaction

(Vinson 1980).

Fire ants can also be serious agricultural pests, causing direct or

indirect damage to crops. Direct damage is caused by ants feeding on

young citrus trees (Gallo et al. 1978, Adams 1986) or soybean plants

(Lofgren & Adams 1981) which can lead to a reduced stand (Adams et al.

1983). Indirect damage can be caused by the ants interfering with the

biological control of some pests (Sterling et al. 1979, Salles 1988, Lee et al.

1990, Tedders et al. 1990), or protection of honeydew-producing aphids

(Wilkinson & Chellman 1979, Tedders et al. 1990). Interference with crop

harvesting by damaging equipment and attacking harvesting crews

decreases productivity and increases harvesting costs (Adams et al. 1976,

Gallo et al. 1978).

Because S. invicta can be an efficient predator, reports in the

literature of beneficial action are numerous. Reports document significant

predation on important pests such as the boll weevil, Anthonomus

grandis, in cotton crops (Sterling 1978), the sugar cane borer, Diatraea










saccharalis, (Negm & Hensley 1969), the lone star tick, Amblyomma

americanum, (Harris & Burns 1972, Fleetwood et al. 1984), and the

velvetbean caterpillar, Anticarsia gemmatalis, (Elvin et al. 1983). Reagan

(1986) cites 15 other pests on which fire ants will prey. Also, reports have

been made on the increase in soil fertility from nest excavation by fire ants

(Herzog et al. 1976).


Present Control Methods

Effort and large financial support has been dedicated to the control

of fire ants since its recognition as a pest problem. Until 1971 the USDA

emphasized eradication of S. invicta, but when that task proved to be an

impossible one, the program was redefined to one for control without

eradication of fire ants (Talmadge, 1980).

Control of fire ant populations can be achieved by individual

mound treatment or by broadcast application of pesticides with ground

equipment for small areas, and with aircraft for extensive areas (Lofgren et

al. 1975). Formulations used today include liquid drenches, granules, and

baits. Present control of fire ants is based on the use of chemical

insecticides, as it has been since the introduction of the ants in the United

States. The principal products used include hydramethylnon (AMDRO),

and fenoxycarb (LOGIC) as baits (Drees & Vinson 1990); chlorpyrifos

(DURSBAN, EMPIRE@), diazinon, several pyrethroids (McAnally 1990,










Collins 1990), as drenches and sprays, and some experimental compounds

not yet registered for use such as teflubenzuron (Banks 1990).

The use of toxic bait, as first formulated with the insecticide mirex

and now with hydramethylnon (AMDRO) and fenoxycarb (LOGIC@),

has been the most efficient method of control of the fire ants. Close to

100% reduction in population can be obtained (Banks et al. 1988).

However, the broadcast use of mirex baits has been implicated in the

spread of fire ant populations due to the elimination of competitor ants

(Wojcik 1983), and the use of mirex has been prohibited due to

carcinogenic properties (Jouvenaz et al. 1981). No method of biological

control has been used in large scale.


Natural Enemies

Because S. invicta is an introduced pest species in the US, it

represents the classical case of a species reaching high population densities

and pest status due the lack of natural enemies and competitors which

control populations in the homeland (Whitcomb 1980). Therefore,

surveys have been conducted in South America with the objective of

identifying natural enemies for possible importation into the US.

Surveys conducted both in the United States and in South America

by several researchers have revealed a limited number of natural enemies

of fire ants (Jouvenaz 1983). Classical biological control agents reported in

the literature include a social parasite, Solenopsis (Labauchena)











daguerrei, (Silveira-Guido et al. 1968) which was brought into the US but

did not survive in the laboratory due to death of host colony (Patterson

1990). This may have been due to the detrimental effects of the parasite on

the host colony (Silveira-Guido et al. 1973).

Endoparasites of fire ants include wasps, Orasema spp. (Fam:

Eucharidae) a common parasite of several ants (Wheeler & Wheeler 1937)

which have been observed attacking fire ants in South America (Williams

& Whitcomb 1974). The young larvae are internal parasites but pupae

occur among the ant brood and are cared for by the ants.

Phorid flies have also been observed parasitizing fire ants both in

the US (Smith 1928) and in South America (Williams et al. 1973, Wojcik et

al. 1987). Species observed in Brazil belong to the genus Pseudacteum and

do not occur in the US. The species observed attacking S. geminata in the

US were described as Plastophora crawfordi and P. spatulata. A

strepsipteran (Stichoterma wigodzinsky) has also been reported from

Argentina as a parasite of S. richteri, the black imported fire ant(Teson &

Lenicov 1979, cited by Jouvenaz 1983).

Several predators have been observed consuming fire ant larvae,

pupae and adult workers and alates. Wojcik (1975) has observed two

myrmecophilous beetles, (Myrmecophodius excavaticollis and Euparia

castanae) eating S. invicta pupae in the US. The predatory ants

Neivamyrmex opacithorax, Paratrechina melanderi arenivaga, and










Solenopsis molesta, Conomyrma insana, Lasius neoniger,

Pogonomyrmex badius as well as S. invicta itself and other ant species

are responsible for great reduction in the number of founding queens after

nuptial flight in Florida (O'Neal 1974, Whitcomb et al. 1973). A spider

Lycosa timuga, a earwig, Labiduru riparia, and a tiger beetle, Cicindella

punctulata, have been observed causing mortality of fire ant queens on

the ground. Predators such as dragonflies and several birds also reduce the

population of fire ant queens during the nuptial flight (Whitcomb et al.

1973).

Distribution and abundance of S. invicta may be affected by

competitive ant species in South America (Whitcomb 1980). Some of the

interactions between these ants and fire ants have been investigated

(Fowler 1988). It has been suggested that control of fire ants could be

accomplished by the importation of selected ant competitors (Buren 1983).

This idea was never implemented, probably due to the difficulty of

importation of a large number of non-host specific ant competitors and

the uncertainty of possible problems that might be caused by the imported

ant competitors.


Microbial Agents

Most reports of natural enemies of fire ants are microbes found both

in the South American homeland of the fire ants and in the United States.

Nevertheless, the incidence of microorganisms associated with ants in the








8
US is apparently very low (Beckham et al. 1982). In 1974, Allen & Buren

reviewed the literature available at the time on the occurrence of diseases

in ant populations and suggested that the fungi Metarhizium anisopliae

and Beauveria bassiana are very important in ant populations in South

America.

The first observation of microsporidian infection in ants was made

in Brazil with a species of Thelohania, and possibly also Nosema (Allen

& Buren 1974). Thelohania solenopsae also has been detected in other

ant collections both in South America (Knell et al. 1977, Jouvenaz et al.

1980, Allen & Silveira-Guido 1974) and the United States (Jouvenaz et al.

1977). Other microsporidia have also been detected in fire ants (Jouvenaz

1983), including Mattesia geminata, Burenella dimorpha, and

Varimorpha invictae (Jouvenaz et al. 1980, Jouvenaz & Hazard 1978,

Jouvenaz & Ellis 1986). A European species of the Solenopsis genus, S.

(Diplorhoptrum) fugax has been long known to harbor

Myrmicinosporidium durum, a haplosporidium (Buschinger & Winter

1983).

Only one report of a pathogenic bacterium, perhaps Bacillus

finitimus, has been made to this date (Jouvenaz et al. 1980). Reports on

viruses and nematodes occurring in fire ants are limited. A virus-like

particle has been described from a Solenopsis ant, and observed in S.

geminata without any further information on infection rates or

symptoms on the living host (Avery et al. 1977).











A 3-mm Mermithid nematode has been detected in the gaster of

native fire ant workers (S. geminata) in Florida (Mitchell & Jouvenaz

1985), and another nematode, Tetradonema solenopsis, has been isolated

from fire ants collected in several locations in South America (Nickle &

Jouvenaz 1987). A third type of nematode was recently reported from

collections of S. richteri in Argentina and identified only as a larval

mermithid (Jouvenaz & Wojcik 1990).

Reports on the natural occurrence of fungi on different ant

populations can be found in the literature (Leatherdale 1970, Blackwell et

al. 1980, Papierok & Charpentie 1982, Evans & Samson 1982 and 1984,

Clark & Prusso 1986, Balazy et al. 1986, Sanchez-Pefia 1990), including one

report on the occurrence of possibly B. bassiana on a 25-million years old

fossil ant (Poinar & Thomas 1984) In fact, Evans & Samson (1982) argue

that the idea that fungal epizootics are rare in ant populations is

erroneous. However, only a limited number of reports on the natural

occurrence of fungi in the genus Solenopsis can be found in the

literature. M. anisopliae was isolated from S. invicta from Brazil by

Allen & Buren (1974), and a fungus of the form-genus Hirsutella

attacking a Solenopsis ant (Evans & Samson, 1982), and a yeast infection

(Jouvenaz et al., 1980) have also been observed.

More recently strains of Metarhizium, Beauveria, and

Paecilomyces have been isolated from field collected ants and nest soil in










Brazil after large mortality of untreated ants was observed in laboratory

experiments (Stimac et al. 1987, Alves et al. 1988). In a separate study these

authors (unpublished data) have also determined that B. bassiana occurs

naturally in fire ant populations in 6 Brazilian states and Uruguay. A

three year study has revealed a fluctuation on the prevalence of the

disease throughout the year in Mato Grosso, Brazil.


Use of Beauveria bassiana in Insect Control

Beauveria bassiana has been used experimentally against many

insect pests and has been used extensively in the field against some pests

(McCoy et al. 1988). In an effort to demonstrate efficiency of this fungus

against different pests and demonstrate its potential as microbial control

agent, researchers have established lethal doses and times for several pests

when the fungus is applied under different experimental conditions.

Lethal doses of B. bassiana for several insects have been established

using different experimental procedures. Spray experiments were used

with Galleria mellonella and Sitophilus granarius Leptinotarsa

decemlineata Ostrinia nubilalis and Acanthoscelides obtectus (Hliichy

& SamsindkovA 1989; Fargues 1972 and SamsinakovA et al. 1981, Feng et al.

1985, Ferron & Robert 1975, respectively). With Diuraphis noxia,

immersion experiments were used (Feng & Johnson 1990). Soil

incorporation experiments have been used with Elasmopalpus

lignosellus and L. decemlineata (McDowell et al. 1990, Fargues 1972).







11

Leaf disk surface treatment has been used as a standard procedure in the

evaluation of B. bassiana against Trichoplusia ni larvae E. lignosellus ,

and L. decemlineata (Ignoffo et al. 1982, McDowell et al. 1990, Ignoffo et al.

1983, respectively). Filter paper surface treatment was used in establishing

LD50's for Heliothis zea larvae (Pekrul & Grula 1979). A summary of

results obtained in the above experiments are presented in Table 1-1.

Lethal times have also been established for some insects such as

Melanoplus sanguinipes and Leptinotarsa decemlineata (Marcandier &

Khachatourians 1987, Ignoffo et al. 1983).

Practical use in the field of B. bassiana to control pests has not been

as widespread as use in laboratory experiments. However, noticeable

exceptions exist such as the use of B. bassiana in China and Europe to

control 0. nubilalis, and in the USSR to control L. decemlineata

outbreaks (Ferron 1981, Roberts & Humber 1981, Riba 1984). Other

potential uses of B. bassiana in the control of pests in soil, greenhouses

and agricultural plants have been reviewed by McCoy (1990).


Beauveria bassiana as a Microbial Control for Fire Ants

Studies on the effects of B. bassiana on fire ants have been

conducted in the past but development of a microbial control strategy

using the fungus was not pursued (Sikorowski et al. 1973, Broome 1974,

Broome et al. 1976, Quattlebaum 1980). However, recent suggestion by

Alves & Stimac that B. bassiana causes natural mortality in field














(N r-4
fs 00
Y


a) -~

P
to
>
0



*i^
(A

'(U


N a




c 0

0


0
0 b


Ta


0
W
a)
1-1








g,
(A


0

(U
.P.4







a)
Cl




.2











a)
x










a)
1-.
(2)












0
4-.

cl
i>

















r:j
a)
U,
4.6



(U


U3



ajw
U)i.
(A)









a)

(U
IH


N s

4-


U

U,

0
Uu
"flj
4-


Ut
v


-o '-4




U U
U, *a.
o
-I I-
4-






Uy *..
O Q


oo' ao



0 0



















(0 0
u u
tJ (/J
o 0





6I :
'-.3 ^s


0 0
0 u
6 ` 6





I 6fi r



ou
8u






01 No r- <









U 3 )


ar) a) P



a) a) L4

0 0

u u





















N c
o- 00




* N
a1 \o
i-iJ


--









13




Ii .4 ..-.
I 00 0









I0 0 0
M (a M

















0 e i X
0
'-4 ~I
N 34


0 0 0 1








I (U
n No tn o '











o I
x X









I d
CI








SC I

IG I
0 0
.a




X"
Sti
E0 .W *
QjU( ; 3 ^
3d ^ .










populations of S. invicta in Brazil (Carruthers & Hural 1990) renewed

interest in development of a microbial pesticide. Several isolates of B.

bassiana have been obtained and tested under laboratory (Stimac et al.

1987, Alves et al. 1988) and field conditions (Stimac et al. 1989). The

chemical defenses of S. invicta against B. bassiana have been

investigated (Storey 1990).

An isolate of the fungus Beauveria bassiana obtained in Brazil

from populations where the fungus occurs enzootically is currently being

developed as a possible microbial insecticide to be used in the control of

fire ant populations in the United States. This study includes not only the

screening of fungal strains against fire ants in standard bioassays in the

laboratory, but also the development of application methodology and

formulation of final product to be applied in the field. Experimental

results provided information which allowed the establishment of

methodology for production and application of B. bassiana formulation

on fire ant nests in the field (Stimac 1990). This methodology is

undergoing continuous improvement and has triggered large public

interest.


Objectives

The development of B. bassiana as a microbial insecticide requires

understanding of the action of the fungus on the ants, and the possible

dissemination of the disease among individuals in a nest. Also,










understanding of the role of the soil in the fungal infection and disease

dissemination is important in designing of an efficient application

method for use of the fungus in field populations of fire ants. The

objectives of the studies described herein include:

1. Evaluation of the effectiveness of B. bassiana against S. invicta

workers, and establishment of mortality curves for fire ants inoculated

with the fungus under different conditions.

2. Evaluation of the possibility of inducing B. bassiana epizootics

in fire ant populations with initial infection of only part of the population

with the pathogen.

3. Evaluation of the capacity of B. bassiana to grow, spread, and

sporulate in soil of fire ant nests, with or without some nutrient

amendments.

Objective 1 is examined in chapter 2 which describes four types of

experiments including sterile or nonsterile soil as the environment in

which ants live. Dose-mortality curves were generated for each one of the

experimental conditions and the effective doses against 50 and 90 % of the

ant population were estimated. Also, mortality curves were obtained for

two different doses in two of the experiment types, and times for 50 and 90

% mortality were estimated. Comparisons are made between the results of

each experiment type, and discussion on the role of soil and antagonistic

factors present in the soil is provided.










Objective 2 is examined in chapter 3 describing experiments in

which different proportions of live ants in small ant colonies are treated

with B. bassiana. Experiments were also conducted in which infected ants

were only added to the colony after ants had died from fungal infection.

In order to test the effect of soil on the prevention of transmission of the

disease, comparisons were made between experiments with or without

soil. Survivals of the untreated ants were determined after 3 weeks in

order to estimate possible transmission of disease from the infected to

healthy ants. Discussion is provided on the possibility of disease

transmission in fire ant colonies and on the role of soil in preventing

transmission.

Objective 3 is examined in chapter 4 which describes experiments on

the growth of B. bassiana in soil containing fire ant cuticle, chitin, rice or

no amendment. Growth curves were generated and total fungal

production on soil plates with different amendments was estimated.

Discussion is provided on the effect of different amendments and the

possibility of fungal growth in soil containing naturally or artificially

added ant carcasses or starchy materials.

Chapter 5 provides a general discussion of the results obtained and

possible connections between results in the different experiments. Also, it

examines the implications of the obtained results on the future use of B.

bassiana for microbial control of fire ants.














CHAPTER 2


DOSE-RESPONSE OF THE RED IMPORTED FIRE ANTS,
SOLENOPSIS INVICTA, WORKERS TO
BEAUVERIA BASSIANA CONIDIA


Introduction


Even though Allen and Buren (1974) suggested that both

Metarhizium anisopliae and B. bassiana were very important natural

control agents of fire ant populations in South America, these fungi have

received little attention as microbial control agents of the fire ants until

recently. Earlier studies on the use of B. bassiana on fire ants (Sikorowski

et al. 1973, Broome 1974, Broome et al. 1976, Quattlebaum 1980) showed

that this fungus could cause mortality of ants both in laboratory and in the

field but further investigations to develop the fungus as a microbial

control were not pursued.

More recently, preliminary field studies demonstrated control of fire

ant nests in Brazil (Stimac et al. 1989). Several fungal species from ants

collected in different regions of South America (Stimac & Alves, unpubl.

data). Several of these isolates have been proven to be pathogenic to fire

ants (Stimac et al. 1987, Alves et al. 1988) and B. bassiana has been shown










to kill Solenopsis colonies in the field (Stimac et al. 1989). Based on these

results, methodology for production and application of B. bassiana

formulation on fire ant nests in the field has been reported (Stimac 1990).

The pathogenicity of B. bassiana to fire ants varies with fungal

isolate and experimental techniques. Reports of worker mortality due to

the fungus in laboratory experiments vary from 35 % (Broome 1974) to 94

% (Quattlebaum 1980). The juvenile form of the fire ant is more

susceptible than the adult to B. bassiana as well as other fungal and

bacterial pathogens (Broome 1974). The LD50 of B. bassiana against

reproductive larvae of S. richteri was 36 conidia for topical application

compared to 390 for per os application, and slopes of the dose-response

lines were estimated at 0.52 to 0.75. Mortality occurred soon after

inoculation and the LT50's for the same application modes were 2.1 and

4.4 days, respectively, when the dose per larva was 5 x 104 conidia.

Alves et al. (1988) determined that different isolates of B. bassiana

obtained from infected fire ants produced mortality on fire ant workers

between 60 and 82 %. The estimated LT50 was greater than 9 days for 2

isolates, whereas a third isolate showed LT50 of c. 7 days in experiments in

which the dose per ant was estimated at 3000 conidia per ant. In another

study (Stimac et al., 1987) with similar experimental methods for 5 isolates

of B. bassiana, fire ant mortality varied between 62 and 84%, and the

estimated LT50 between 7 and 14 days. One of the isolates used by Alves et











al. (1988) was # 447 which has been selected as an efficient pathogen of fire

ants, and is used in the experiments described herein.

The development of any agent as a pesticide requires as a

preliminary step the establishment of dose-response curves and

estimation of lethal dose. The objective of this study was to establish the

dose-mortality response data for conidia of B. bassiana, isolate 447, against

workers of the fire ant, S. invicta, under different conditions. Four

experiments were used to determine the effect of soil in the infection

process, and the role of contaminated surfaces as an inoculum source.

Two experiments involved direct spray of conidial suspensions on ants:

one in which insects were maintained in clean plastic cups, while in

another experiment ants were maintained either in sterile or nonsterile

fire ant nest soil. In a third experiment, conidial suspensions were added

to the soil prior to addition of ants, and the fourth experiment involved

spraying of a Petri dish surface with conidial suspensions and exposure of

ants to this contaminated surface.


Materials and Methods


Fungus

The B. bassiana strain # 447 (ATCC 20872) was originally isolated

from adult workers of the fire ant, S. invicta, in Mato Grosso state, Brazil.

The fungus was reisolated from inoculated adult fire ants from Florida








20

and maintained in Sabouraud dextrose agar + 1% yeast extract (SDAY) in

the laboratory. Third passage of the fungus on SDAY after reisolation

from ant cadavers was used for all experiments. Conidia were harvested

from the medium plates and passed through a 100-mesh sieve (opening of

0.120 mm). Conidial suspensions (105 109 conidia/ ml) were then

prepared with sterile deionized water and used in the inoculation of the

ants, soil, or Petri dish surface. Conidial viability was above 93% for all

experiments.


Ants and Soil

Ants used in experiments were obtained from field populations and

maintained in the laboratory in plastic trays. The method described by

Jouvenaz et al. (1977) was used for ant collection. In summary, colonies

were removed from the soil and put in 5 gallon buckets lined with

Fluon which prevented the escape of ants. Water was then dripped into

the buckets forcing the ants to the surface of the rising water from where

they were collected and transferred to plastic trays. In the lab, the ants

were fed a variety of foods including pureed vegetables and fruits, insect

larvae and diluted honey. Usually ants were fed 3 times a week, but, prior

to experimentation, colonies were fed extra quantities for 1-3 days. For

experiments, only larger workers (c. 5 7 mm) were selected and separated

in groups of 10 in plastic cups.










Mound soil was allowed to air-dry in the shade and then sifted to

eliminate plant materials and larger rocks. Further sifting through 60-

mesh sieve eliminated most residual ant cadavers.


Spray Tower and Spraying Procedure

The spray tower was constructed using a turntable as a rotating

stage. A wooden frame which slid on vertical poles suspended the

spraying nozzle and the spray chamber consisting of plastic container of 25

cm height and 22 cm diameter. The nozzle, mounted through an opening

on the top of the spray chamber, had two intake openings on the top, one

for compressed air and the other for liquid inoculum. Compressed air

(pressure of 15 lbs/inch2) entered the tower after depression of a trigger.

Conidial suspensions were added to the tower with a Pipetman, the tip

of which was inserted in the liquid intake opening on top of the spray

nozzle.

Spray arenas consisted of plastic cups with c. 6 cm diameter and 3

cm-high walls painted with Fluon to prevent escape of the ants. A layer

of plaster was added to the bottom of each cup to serve as holding surface

for the ants, preventing them from being blown away during spraying.

The spray arenas with ants or Petri dishes were placed on the center of the

rotating platform of the spray tower and sprayed for about 10 seconds with

0.2 ml of the conidial suspensions.










Estimation of dose per ant

Ants treated with conidial suspensions of 3 x 107 to 108 conidia per

ml were used to estimate the dose per ant applied in the spray tower. Ant

selection and spraying were as described above, but within 1 minute of

being sprayed, ants were transferred to test tubes with 10 ml of chilled

sterile deionized water. Ants were then vortexed in water and then

exposed for 30 s to ultrasonic waves (300-Watt Fisher sonic dismembrator

with 19 mm probe and maximum setting). Aliquots of 0.1 ml of the

conidial wash from the ants were inoculated onto plates containing

oatmeal-Dodine agar medium (Beilharz et al. 1982) as modified by Steve

Krueger (pers. comm.)

Plates were incubated in the dark at 2520 C for 7 days after which

the number of colonies formed on the plates were counted. For

estimation of dose per ant, a total of 90 ants were used on three different

occasions and a total of 26 different operations of the spray tower.

Efficiency of the washing procedure in removing the spores from the ants

was calculated by applying a known number of spores on ants, submitting

them to washing procedure described above, and calculating the recovery

rate (Appendix A).


Dose-Response Experiments

Each experiment was repeated twice, and each treatment was

replicated 10 times, with 10 ants in each replicate. Controls were used in











each of the experiment and consisted of ants treated in a similar way to the

treatment ants, but with sterile deionized water.

In both experiments involving direct spray of ants, 7 conidial

suspensions ranging from 106 to 109 conidia/ml were used. In direct spray

experiments without soil, ants were then transferred from the spraying

arenas into clean 1-ounce plastic cups. In direct spray experiments with

soil, ants were removed from the spray arenas into clean plastic cups

where they stayed for 2.5-3.0 hours before being transferred to cups

containing c. 10 g of either sterile or nonsterile nest soil mixed with 10 %

(v/w) of sterile deionized water.

In the soil incorporation experiment, sterile or nonsterile dry nest

soil was mixed with 10% (v/w) of conidial suspensions in order to obtain 6

conidial concentrations between 10 and 106 conidia/g of dry sterile soil,

and 5 concentrations between 106 and 108 conidia/g of dry nonsterile soil.

Also concentrations of 109 and 3 x 109 conidia/g in nonsterile soil were

obtained by mixing dry conidia into the soil and then adding deionized

sterile water (10% v/w of dry soil). The soil was then separated into 10 g

aliquots in 1-ounce plastic cups, and ants were added to each of the cups.

An additional 10 replicates each of the sterile soil treatment with 105

conidia/g soil, and nonsterile soil treatment with 3 x 107 conidia/g soil

were prepared, and ants were transferred from soil to ocean cups on day 3

(transfer treatments).










For surface exposure experiments, 60-mm plastic Petri dishes were

sprayed in the spray tower with 7 conidial suspensions ranging from 106 to

109 conidia/ml, which resulted in concentration of 1.4 x 101 to 1.4 x 104

conidia/mm2 of the plate, as determined by direct count under microscope

of conidia on sprayed surface. The suspensions were allowed to dry in a

laminar flow hood before ants were added to each plate, where they stayed

for a period of 24 hours. After this period the ants were transferred to

clean plastic cups, where they stayed for the duration of the experiment.


Mortality Assessment and Analysis of Results

Ants were maintained in incubators at 2520C and checked daily in

cups with no soil, and on days 10 and 15 in cups with soil. Cadavers were

removed, washed in 95% ethanol and allowed to dry on paper towels.

Surface-sterile ant cadavers were individually placed into wells in micro-

titer plates which were then incubated for 7-15 days in moist chamber to

allow sporulation of infected cadavers. Ants were marked as infected if

signs of B. bassiana growth were detected during or after incubation.

Total mortality at 15 days, and cumulative daily mortality were used in

analysis of results.

POLO, a probit and logit analysis software package for IBM-

compatible computers, was used to analyse the results and obtain

estimates of the dose-mortality and time-mortality curves. These were

then used to estimate lethal concentrations necessary to kill 50 and 90 % of







25

the populations (LC50 and LC90, respectively), and the time to kill 50 and

90 % of the populations (LT50 and LT90, respectively) when mortality was

checked daily. Comparisons of mean estimates were made taking in

consideration 95% confidence intervals. Mean values for estimates from

the two experiments of the same type were obtained by testing the

hypothesis that dose-response or mortality curves from the experiments

had similar slopes and intercepts. The hypothesis of parallel lines with

different intercepts was also tested.

In constructing time-mortality curves used to estimate LT's, it was

assumed that independent groups of insects were checked at each date,

although only one group of insects was used and accumulated mortality

used in the analyses of results.

A 1-tail t-test was used to evaluate differences between the transfer

treatments and the treatments with similar conidial concentration in

which ants were maintained in soil for the duration of the experiment.


Results


Direct Spray Experiments

Estimation of dose applied per ant during spray procedure provided

a basis for calculation of LD50 and LD90 from the LC50 and LC90 estimated

with direct spray experiments. Calibration of the spray tower in three

independent estimations produced similar results for mean number of

conidia ( sem) washed from fire ant workers (2448 294, 1950 155, and








26

2570 581), after application of conidial suspension with 108 conidia/ml.

Considering a recovery rate of the washing procedure of 1%, as calculated

by the calibration procedure (Appendix A), the mean of the numbers

above correspond to a dose of c. 2.3 x 105 conidia/ant.

Experiments without soil

Results from two experiments were similar. Both mortality and

infection of fire ants increased with an increase in conidial dose in the

spray suspensions. Virtually 100% mortality and infection was reached at

108 conidia/ml (Figure 2-1). Infection by B. bassiana was confirmed in c.

88% of the ant cadavers found in the fungal treatments compared to only

1% in the control treatments.

Probit analysis of the dose-response curve estimated the mean LC50

of 0.9 x 106 conidia/ml (Table 2-1). Mean LC90 was approximately 1 log

higher at 1.9 x 107, with a slope of the dose-response line equal to 1. No

significant difference in activity was observed between the two

experiments.

With the highest conidial concentration (108), the LT50 was

calculated at 5.7 days (Table 2-2) and LT90 was 10 days. With 107

conidia/ml suspension, the LT50 increased to 7 days and LT90 increased to

12.7 days, but the comparison of the confidence intervals revealed no

significant difference in the LT's for these two doses.















0,









L a



Soo
W CA



k -
tv 1tr s0
04


Sjuv Jo IuaoxaG

















~ I
Y,
i s
u
..-. U

.S .
(8 (U
(U

(U (
0)
w -^

E ^3
UI
CQ ES
0)




O 6d
I ^
0;




4-a
S3 <






(U,
t3




U) *a

oU
(U
En Li
\ 0
1(




S U
il=
b C^



h u~





U)U
0) (
(" ^

r I
<- (

c3 G


0 I



7IN
Iw
* I
VI






1- I

















U I
LI
I



I


c5
' I



00






o

00 I















x
LI
cm
-4 I


(I









o I
'-4



X



I-I


r-I



I
0
I
I

rIl
.X. I


'-4

oI\

r-.


o
I




^ I
'4 I


0x
o
I


r-I
I
I



oI
I

x I
(Q

'4 I

r-I



' I

N

6g

r-I

o\I
(S





I
rI
I s
I0

I&

I-
X
I


0


0.

VI
.4-'
(U




'4>
d





0)
C




'4
60D





a
, *0

O4O
+ O
2 0

S-,

s U)


(U nj'
g S

o

Js^
o


a O


0a)


0iU

- 0)
e.l S
ra

Ig







t ^















O
3 -0



I 0 I I u
'c ir i 6 in i


I I I
(a N o6 a;
U I VI












r-4I I I
I i C I Q N
u mt a 2 .o6


\I It I r O






to

-U) II 00
o O
Cu IO u s I. >.O -
II N






0 4 0B C >
^Iu r *










l1I f" III 5!
Id I pau CQ 2
Cu Cu I 0
~! ~Z!clCu
E-U)d I I










Experiments with soil

When ants were maintained in soil after being sprayed with

conidial suspensions, the doses necessary to produce mortalities were

similar to those needed in previous experiments without soil. Mortality

increased with an increase in dose between 105 and 108 conidia/ml

(Figures 2-2 and 2-3). One hundred percent mortality was reached with

concentrations of 108 3 x 108 conidia/ml. In most cases, infections were

low when compared with mortalities. In some cases, such as for the 108

conidia/ml dose in sterile soil in experiment 1 (Figure 2-2), less than 50%

of the ant cadavers developed fungal growth after being surface sterilized

and plated.

Dose-response lines had slopes just above 1 for all combinations of

soil type and experiment and were not significantly different (Table 2-3).

In both experiments, the LC50 for ants placed in sterile soil (c. 2 x 106) was

not significantly different from LC50 for ants placed in nonsterile soil (c. 3

x 106). The difference between the LC50 and LC90 was approximately 1 log

for both soil types.


Soil Incorporation Experiments

Large differences in the effects of the sterile and nonsterile ant nest

soil were observed when fungal inoculum was added directly to the soil.

A dose of 106 conidia/g soil caused almost 90% mortality in the sterile soil

(Figure 2-4), compared to mortality not significantly different from control





















































sjuv Jo jualad


I"

0
O


"Iof
cd

14
-M
0
Q)
.t
*P4


M M

EI







00
* *



B&
a A








HO






a a


cc



*J4
WC




a) a
-s
,g 01








'" .~
-4
(U
ba
.U)









U) 'S5u
20m
,'.D -
Vtis










U)
0o
2 0
i3^
>h tfl

a &<
,cu
g -^


07


0$













sa>
a ^


0 |


, .:~~ sXli.~


s4uv jo WuGaead


*f4


"M




*P4
r


C

1N N
0
a d


5 c




on








0


60


000



OO

%1mo






&0 L
1 20






of

Ti.

10 C5
0 b E; *- C
<0 c s *

5 M*o -
co^ I -
(0 ( 5^ c!
^ S
?|^1



















*a I
o o
.4 I O < < < I I


mm a ai q N
)C I O II r I


I 0 N0y4 I 0 Ic
4-i r" I V- r-i4 "
u x xx x I xx '






o o o 0
O-i r-4 C-- r-C-

O a I CO

6 I ei


rn 0
do XX XX XX' u

o h0 ) 16 r-! 0C
a Q) Q) 'g ;S \g 50 \
a' i Ir I9 I10



I) 0 r- o (iA Ucr Z




SI ~I
I I CgCO ( U 5
I I 0 I '0
S2 I aI
.C 3O O


I (U3 (U0)0 0)0 103






oI .S I I (U
0[*** c0 I 0 I>












































sIuV Jo IuQ3Jzd


(0 *
4)


I
14 N












s U
*
M~t


WE
.f.
.-E











.?1
t jj:
I.-.











14


(I)b




0r










in the nonsterile soil (Figure 2-5). In nonsterile soil, maximum mortality

obtained was just above 50% with the highest concentration of spores (3 x

109), and only concentrations above 108 conidia/g soil caused mortality

significantly greater than the control. A relatively low proportion of the

ant cadavers produced external fungal growth after being surface sterilized

and plated, especially in experiment 2 in sterile soil but many cadavers

when removed from the soil already showed signs of fungal growth.

When ants were transferred out of sterile soil treated with 105

conidia/g, mortality was 85 % in both experiments. When ants were kept

in the soil for the entire duration of the experiment, comparable

mortalities of 90 and 86% were achieved for the same treatment levels.

The ants in transfer treatments with nonsterile soil had mortalities of 13

and 17% respectively for experiments 1 and 2, compared with 2 and 6%

when the ants were maintained constantly in soil.

Due to the low mortality, a LC90 could not be estimated for

nonsterile soil in either experiment and LC50 could not be estimated for

the first experiment (Table 2-4). In sterile soil, the calculated LC50 had

large variation between the two experiments, but the 95% confidence

intervals overlapped. The LC90 was estimated at more than 3 logs higher

than the LC50 with an estimated slope of only 0.4. In nonsterile soil, a

slope of 1.1 was estimated based on the few concentrations that produced

mortalities different from control mortality, and this value was














































sjii jo jUolGad


*?4
0
(0)


CEI EN
* *


a a
Q o




- o




A A
14 N




l a

0 0
U.S


0 0
'-4



c
.ri
,-


-a-
15






c:.





sr


0
%;
o

,



















0 c
o


( I







(C

(U
I g











0 0)
bsa














r -
a c


























u-i
9 0



0 *3)
(


















* 0)
a ^
0)rsU








F4u

`ZjC
Vi tf
















(-* Wl


-I
04
Lo

o i







0
)
S~


o
00
0




x
0 i








0 %
r-
P u


0O




0


i If
00 1
S00



- I





x
" I
'-4
r-







x I








X X




co- c


X X




x x


0)
>
0

0











4-
to















.c
U









Z >,











0)















a II
> 3









cs










significantly different from the slope for dose response curve in sterile

soil. An estimate of the LC50 was observed only for 2.3 x 109 conidia/g of

nonsterile soil.


Surface Exposure Experiments

Concentrations of 1.4 x 102 conidia/mm2 and higher caused

increasing mortality of fire ant workers exposed to plastic Petri dish

surfaces treated with B. bassiana conidia (Figure 2-6). Virtually 100%

mortality was reached at concentration of 1.4 x 104 conidia/mm2. Also,

infections were higher with higher doses of the fungus and almost all

cadavers in the highest dose treatment showed signs of fungal

development.
Estimates of the LC50 in both experiments were close to 1.0 x 103
conidia/mm2, and LC90 estimates were 1 log higher than the LC50 (Table
2-5). The slope of the dose-response lines were not significantly different
in the two experiments.

A concentration of 1.4 x 103 conidia/mm2 of Petri dish caused 50%

mortality by 12.3 days (Table 2-6), whereas at a concentration of 1.4 x 104

conidia/mm2 the LT50 drops to 7 days, and the LT90 is only 11.1 days. The

slope of the mortality line which was estimated at 3.7 for the lower

concentration of conidia per area, increases to 6.4 with a log increase in the

dose, reflecting a much faster kill of fire ant workers.
















ti

&~~I ^ n"s3
e3 cU



aI

'-4 ei ii -
N ^ N Q i ^^ i^^j^






8 00 0 0~ 0
04 0 0E
O~C cd se

4~P4
eli Ih





.-4
A0 A
EQ "4 ;j f


8e 0 0h' 0
CDro '4'


s-4uv jo jueozad

















v; .f5



( P0

(0 S



0(,




.P-4
*- 0





0.4 (v ~


. (U
to








~0

.4 0)
CC
(U('.4



O 4)I
0 jS n









U,~L
















0) % 0)
'- ^ 0
~UU
u C..' t.

















P, 4)
= 19c
cJ g '?;















(U ,(
H ., .a..
.2 2 a
*O G--^
q2 "> (f
^^
rs ^
(f S"
g^-
\fl a> O
<8 t< a




5 '(f "
o c h
C- I 3


0 I
I -;




U o



S 0













c
'.0






I x
r<




o


U x
0

















I (


I
00




r 4










'4 I
x I
NI
* I













x
- I



'-4





r'-



CM.


I -0







c il I
r1


I I





01






'-I
d i
I





I- I
* I
6"






30


x

S
x o

I C



e >
I U
G
0)



*cO

2( 1
x S

oo -





'-45
0 0
%I U


0 I 2


I U^



(U |





0, I
.4-




















P,
0)t4
It 'o
1.4


(U0)
1.4 '44
(8 U





Q5


t"

00)
12







oS


sa ^
10









p h
U) 3



U(1









ki^
.- 0)








0c
U) -*
^ ffl


O 0)


?>^?
^3f
.5 1-.

4O (U

11
(1


U

0) .
(U0)


SII6
Cf)










IN I
hcn









* '










I-<
in \o
















1 i



C~-~









x x
^ ^>(
r^ r-


0)
tis




0
o




m -a


> o
0O
c i





1-4




o 4.
z-, 0)





















4-4
.4 4-<
,cj t
Cl" .2

^^
'i:L p
^ 6





s 2
5o








illd
2; ?
E; +


en
r"-
X
>t v
cn
'(U
0~










Discussion


Effect of Fungus in Direct Spray Experiments

Treatment of ants with equivalent doses caused very similar

mortalities in experiments in which ants were directly sprayed with

conidial suspensions regardless of the post treatment environment.

Whether ants were transferred into clean plastic cups or into cups with

either sterile or nonsterile soil, estimates of LC50's, LC90's and slopes of

the dose response lines were not significantly different. These results

demonstrated clearly that the influence of the soil environment on the

infection process is negligible if conidia make contact with ants directly.

The calculation of LD50 should enable us to make direct

comparisons between results of experiments using different procedures

which may result in similar conidial suspension providing different doses

to the target insect. The lethal doses can be calculated taking the lethal

concentrations and relating them to the calculated dose received per ant

when suspensions with such concentrations are used in the spray tower.

This calculation assumes that there is a linear relationship between

conidial concentration in suspensions and doses applied on ants, since the

application methodology was constant for all conidial suspensions. Thus,

calculated LD50 and LD90 per ant were 2.1 x 103 and 4.4 x 104 conidia/ant,

respectively, based on 1% washing efficiency in estimation procedure

(Appendix A). These values are high if compared with LD50 values










between 15 and 500 conidia/larvae for Heliothis zea Ist-instar larvae

(Pekrul & Grula 1979). However, the values estimated for H. zea may not

be directly comparable because it is unclear whether they represent the

actual dose received by insects, since we have no information on the

method for estimation of dose per insect. The dose applied per ant in

experiments herein is similar to that calculated by Alves et al. (1988) who

used 108 conidia/ml suspension from 3 B. bassiana strains (including 447)

but obtained lower mortality of ants.

The low values obtained for slopes of the dose-response curves are

similar to those obtained for other insect pathogens (Maddox 1982), and

indicate a great variability in the susceptibility of the ant population to B.

bassiana. Low slopes also indicate that high doses are necessary to kill

high percentage of the population at the time of application of the fungus.

The calculated lethal times are shorter than those calculated by

Alves et al. (1988) who obtained LT50 between 7 and 9 days using

inoculation with a conidial suspension containing 108 conidia/ml. Our

results with similar suspension (108) showed an LT50 of only 5 to 6 days,

and even a dose 1 log lower had LT50 between 6 and 8 days. The calculated

lethal times seem also to be in agreement with studies by Sikorowski et al.

(1973) who reported B. bassiana-caused ant mortality occurring between 5

and 10 days after application of the fungus.










In view of the similar mortalities and lethal doses discussed above,

it can be concluded that ant deaths after treatment occur at a similar rate

whether ants are put back into soil or maintained in clean containers. By

the time of the mortality checks in the experiments with soil (10 and 15

days) many of the ants in the higher doses would have been dead for a

period of 1-4 days. This period of time was sufficient for some fungal

growth and sporulation to occur on the cadavers prior to their removal

from the experimental arenas. The subsequent surface-sterilizing process

may have adversely affected the fungus which was already in the

sporulation stage, preventing renewed growth after plating of the

cadavers. This would explain some of the low infections as calculated

after surface sterilization and plating of the cadavers.


Effect of Conidia Mixed into Soil

When the conidia are mixed into soil prior to addition of the ants,

there is a clear indication of the detrimental effect of soil antagonists on

the performance of B. bassiana. Large differences in ant mortality in both

sterile and nonsterile soils with similar conidial concentrations

demonstrate the detrimental effect of the soil antagonists. Differences may

exist between soils, and a lesser antagonistic effect of nonsterile soil in

experiment with extract of B. bassiana-amended soil (Appendix B) serves

a reminder that the soil environment has to be further studied before

efficient use of entomopathogens in soil can be made.











The results of the transfer treatments in the soil experiments

further demonstrate of the antagonistic effect of nonsterile soil. When

ants were transferred out of sterile soil into clean plastic cups, the

mortality did not significantly increase in relation to mortality of ants

maintained permanently in soil. In contrast, ants transferred out of

nonsterile soil had a significant increase in mortality (p 0.05). Therefore,

despite prolonged contact with inoculum source, ants maintained in soil

had low mortality compared to ants which were taken away from the

inoculum source. This demonstrates that the soil antagonism can affect

infective units which have already made contact with the insect. This is

in apparent contrast with results of the previous experiments in which no

effect of soil was detected when ants were directly sprayed with B.

bassiana conidia.

Possibly, the period of time the ants were maintained away from

soil in the direct spray experiments was sufficient to provide protection to

the attached conidia from soil effects. Further investigation of effects of

the different periods of contact of conidia with soil is necessary, but there

are several possibilities for the antagonistic effect of the soil. It may affect

conidia in the soil, prevent attachment to insects, or prevent

development of infection by inhibiting germination and penetration.

The strong antagonistic effect of nonsterile soil can be explained in

view of results obtained by other authors who studied similar systems.










For example, Penicillium urticae, commonly found in soil, has been

determined (Shields et al. 1981) as the producer of water soluble inhibitor

of B. bassiana. Other common soil saprophyte, Aspergillus clavatus, also

produces metabolites (Majchrowicz et al. 1990) which are fungicidal to B.

bassiana. Antagonistic effects can also be produced by ant-derived

materials. Fire ant venom has been demonstrated to be a weak fungistatic

agent, and to accumulate in fire ant nest soil (Storey 1990). Although that

author determined that the venom is heat stable and is not eliminated by

autoclaving of the soil, its role in the nonsterile soil may be synergized by

the presence of other antagonistic materials. The action of venom

alkaloid in the sterile soil may be hindered by the absence of complements

eliminated by soil sterilization. Furthermore, autoclaving can modify

organic composition of soil and provide nutrient and other substances

which may be used by the fungus


Infection of Ants by Conidia in Soil and on Plate Surface

The results for surface exposure experiments appear to differ from

other reports (Ignoffo et al. 1982 and 1983, McDowell et al. 1990), but

important differences in the experimental techniques prevent direct

comparison of data. The results presented herein demonstrate that

foraging fire ant workers could pick up fungal infection from the

environment such as a leaf surface, other organic material to which B.

bassiana conidia adhere, or various surfaces not in soil upon which







47

conidia can be deposited by climatic factors, e.g. the bark of trees and shrubs

(Doberski & Tribe 1980).

Beauveria bassiana is well known as a common soil-insect

pathogen (Carruthers & Soper 1987) which can be isolated from most soil

environments. However, in view of the results demonstrating soil

antagonism, the possibility of infection of fire ant workers from conidia

deposited directly in the soil does not seem to be very high. Although

there was high mortality when conidia were mixed into sterile soil,

demonstrating that contaminated soil can serve as inoculum source for

ant infections, nonsterile nest soil was very detrimental to the fungal

infection. Unless large amounts of conidia are deposited in localized

pockets in the soil which are then visited by the ants, conidia distributed

in the soil are more likely to be affected by antagonistic factors before they

have a chance of contacting an ant host and producing infection.

Individual fire ant workers are susceptible to B. bassiana infection
under artificial laboratory conditions. The doses required to kill fire ants
are not excessively high and once ants contact a lethal dose the presence of
soil is not detrimental to the infection process. The lethal doses

established herein are important base values to which field doses can be
compared. The demonstrated antagonistic effect of fire ant nest soil will
also be important when considering field application technology which
should provide for maximum protection of the fungal conidia from soil
antagonism and maximize contact with ants.














CHAPTER 3


TRANSMISSION OF BEAUVERIA BASSIANA
WITHIN ARTIFICIAL FIRE ANT NESTS IN THE LABORATORY

Introduction

Dissemination of a pathogen within the host population or

throughout the environment is important for successful transmission of

insect diseases (Andreadis 1987). Epizootics commonly occur in insect

populations and are heavily influenced by the environmental conditions

which can affect both the host and the pathogen (Benz 1987).

Understanding of mechanisms of transmission of diseases will increase

the possibility of success in the manipulation of natural occurrences or

artificial applications of pathogens for control of insects (Harper 1987).

Since social insects live in close proximity to each other, the contact

frequency and subsequent dissemination of pathogens should be increased

(Alves 1986). However, hygienic behavior can eliminate pathogens from

the colony environment. Grooming, on the other hand, can increase

transmission of infective units of pathogens (Watanabe 1987).

Transmission of fungal diseases in some ant populations may be

aided or hindered by the behavior of moribund and live ants. Ants of the

genus Cephalotes infected with Cordyceps fungi moved to a single tree











before dying, creating a concentration of inoculum which could be

avoided by healthy ants (Evans & Samson 1982). However, these authors

observed live ants attempting to clear the dead ants from the tree. Evans

(1989) presents evidence from several authors in support of his argument

that ants infected with fungal pathogens tend to move away from the

colony to prevent dissemination of the disease. In Solenopsis invicta

populations, injured individuals have been observed moving away from

the nests (Hblldobler & Wilson 1990).

Few examples of diseases have been studied in depth in fire ant

populations in the United States or in South America. Thelohania

solenopsae infections could not be transmitted either per os or by

moving brood from infected to healthy colonies, but another

microsporidium, Burenella dimorpha, was transmitted per os and by

inquilines or other vectors between colonies (Jouvenaz 1983).

Quattlebaum (1980) utilized diseased Heliothis larvae to transmit

Beauveria infections to field populations of fire ants and was successful

in demonstrating a higher mortality of treated nests. The dissemination

of the disease internally in the nests was not addressed by that author, but

it has been suggested (Storey 1990) that fire ants may limit transmission of

diseases internally in the fire ant colony by removing cadavers and

covering sporulating ants with soil, and by possible utilization of

antifungal properties of venom alkaloids as desinfectant.








50
In the state of Mato Grosso, Brazil, where natural occurrence of high

levels of B. bassiana infection of fire ants have been observed (Stimac &

Alves cited in Carruthers & Hural 1990), mechanisms must exist for

transmission of disease, including transmission between colonies. Study

of these mechanisms has not yet been conducted but would provide

valuable information for the development of B. bassiana as a microbial

insecticide against the fire ants.

The utilization of pathogens in the control of insects requires either

application methodology which guarantees delivery of a lethal dose to the

pest population or dissemination of the pathogen and transmission of the

disease among the treated population. The objective of this study was to

evaluate the possibility of inducing B. bassiana epizootics in fire ant

populations by artificially infecting only part of the population with the

pathogen and allowing transmission to occur in artificial colonies in the

laboratory. Presence or absence of soil in the colony containers were used

to evaluate the effect of soil in the transmission of disease. Also, addition

of dead ants infected with the fungus but not yet showing external signs of

infection was used to evaluate possible differences in disease transmission

between live and dead carriers.










Materials and Methods


Fungus and Inoculation of Ants

Fungal isolation and culture were as previously described (see

Chapter 2, page 19). Conidial suspensions with concentrations from 3 x

108 to 109 (> 95% viability), as determined by hemocytometer counts, were

used to inoculate ants. These concentrations were expected to cause close

to 100% mortality of the ant population since previous experiments

demonstrated high ant mortality when similar suspensions were sprayed

on the insects. High doses were used to guarantee that the infected

portion of the experimental colonies would indeed die and serve as

substrate for fungal growth and sporulation. The conidia formed on the

ant cadavers could then serve as inoculum for the untreated portion of

the ant population and therefore allow the occurrence of disease

transmission to be tested under experimental conditions.

To inoculate the ants, approximately 300-500 ml of the conidial

suspensions were prepared and the ants were placed into and shaken in

the suspension for 1-3 minutes. Ants and conidial suspension were then

poured on a sieve to allow the excess suspension to drain. Ants were then

transferred to trays containing a nest cell and provided with honey water.

The ants were held in trays for 24 hours when live infected ants were used

as an inoculum, or for a period of one week when dead infected ants were

used.










Preparation of Artificial Colonies

Ants and soil were obtained as previously described (see Chapter 2,

page 20). Ants were separated from the stock colony, avoiding large

amounts of brood and alates, and other foreign material that could

interfere with weighing of ants. Ants were then separated into cohorts to

be treated with conidia (infected ants) or deionized water (untreated ants).

Colonies of 3 g of ants were prepared by mixing infected and healthy ants.

Treatments consisted of adding proportions of the infected ants to artificial

colonies ( 0, 5, 15, 50, 85, and 100% of the total weight). Six replicates of

each treatment were prepared in each experiment. Artificial colonies were

maintained in humid chambers in incubators at 2520C.

At the time the experimental colonies were prepared with the

mixed populations, weighed samples of infected and untreated ants were

counted to determine the number of ants present per gram of insect mass.

These estimates were used for calculation of total initial population as

well as of numbers of untreated and infected ants in the various

treatments. Total number of ants in the 3 g artificial colonies varied from

a minimum of 3949 to a maximum of 7361 ants according to the size

distribution of individuals in the ant population used in each experiment.


Experimental Procedures

Four types of experiments were conducted. In two of them live

infected ants were mixed with healthy population and in the other two











types dead infected ants were used. In the live infected ant experiments,

infected and untreated ants were weighed and mixed in the

predetermined percentages 24 hours after the inoculation of the ants.

Mixed populations were transferred to 16-ounce plastic containers either

with a moist plaster layer at the bottom or to similar cups containing 300 g

of ant nest soil + 30 ml water. The container lids had a 1-inch-diameter

opening covered with 80-mesh wire screen, which allowed gas exchange

and through which the ants were fed.

In the dead infected ant experiments, infected and healthy ants were

weighed and mixed in the predetermined percentages one week after the

fungal inoculation of the ants. At this time > 95% of fungus-treated ants

were dead from fungal infection and freezing was used to kill the

remaining live ants. Only live untreated ants were used to prepare the

experimental colonies. Mixed populations were transferred to either

plastic containers with plaster bottom or containers with 300 g of ant nest

soil + 30 ml water.

Feeding and observations

Ants were fed 3-5 times a week with drops of honey-water solution

which were placed on top of the screen on the lid of the containers. At

feeding time, the remaining honey-water from previous feeding was

removed from the screen with a paper tissue. At the same time, ant

behavior and mortality, as well as position and other characteristics of the








54

cadaver piles were observed. Also, observations were made on the growth

and sporulation of the fungus on the dead ants and the possible spread of

the fungus from the cadaver piles through the surrounding soil (when

present) into areas used by live ants.


Estimation of Number of Conidia in Cadaver Pile

An experiment was conducted to estimate the effect of the presence

of live ants on the conidial yield of B. bassiana-infected cadavers. The

number of conidia produced per gram of ant cadavers was estimated in

cups in which the entire population was infected with the fungus and

compared to yield in populations with only 50% of infected ants. Two

treatments were used consisting of ant populations containing: a) 1 g of B.

bassiana-treated ants only; b) 1 g of B. bassiana-treated ants mixed with 1

g of healthy ants. Two replicates of each treatment were prepared and the

ants were maintained in cups with plaster bottoms for c. 30 days until total

mortality was observed. After completion of B. bassiana sporulation on

the ant cadavers, water with 0.5% Tween 800 was added to each of the

cups. Numbers of conidia per ml of the suspensions were estimated by

using hemacytometers and a microscope. The numbers of conidia per cup

and per gram of ant cadavers were then calculated.


Experiments with Conidia Formed on Ants

Experiments similar to the direct spray experiments described in the

previous chapter were conducted with conidia produced on the ants and










conidia produced on artificial medium (SDAY), in order to compare their

relative virulence and verify the infectivity of conidia produced on ants.

Ant cadavers from experiments with and without soil with live infected

ants were used in separate experiments. The concentrations in the

suspensions used to spray ants were 106, 107, and 108 plate or ant-derived

conidia/ml. Control ants were sprayed with sterile deionized water only

and otherwise treated as the ants in other treatments.

Mortalities were checked daily and cadavers were removed, washed

in 95% ethanol and individually plated in wells of micro-titer plates.

Plates were maintained in humid chambers to allow for sporulation of

infected ants. After an incubation period of 7-10 days ants were observed

and marked as infected if signs of B. bassiana growth were present on the

cadaver.


Assessment of Survival and Analysis of Results

Comparisons of treatment effects were made on the percentage of

untreated ants surviving after 3 weeks, because the total number of ants

in the initial population as well as the number of untreated ants varied in

different treatments. Also, the infected ants were not expected to survive

the fungal dose used, and the mortality of untreated ants was an

indication that transmission of disease occurred in the population.

Two experiments were conducted in which mixed populations of

live infected and untreated ants were maintained in clean plastic cups










with a plaster bottom. In the first experiment, the 85% treatment was not

used, but this additional treatment was included in the second experiment

in an attempt to determine if the total population could be eliminated by

treating some percentage that was less than all of the ants in the colony.

Two experiments were conducted in soil with control, 5, 15, 50 and 100%

treatments.

Two experiments were conducted in which infected ants that died

were added to the healthy populations, one in which no soil was present

in the experimental arenas and the second with soil. These experiments

included the 85% treatment but did not include the 100% treatment since

only dead infected ants were added to the healthy populations.

At the end of a 3-week period, the containers were opened and the

live ants were separated, counted and weighed, and final observations

were made on the sporulation of the fungus on cadavers. Number and

weight of survivors for each treatment replicate were transformed into

percent survivals in relation to the total initial population and to the

initial untreated population. Arcsin-transformations of mean survivals

were then compared using analysis of variance and treatment means were

grouped using a Fisher's Protected Least Significant Difference (PLSD) at

p=0.05. Average weight of surviving ants in the different treatments were

also compared using ANOVA and multiple range tests.

Results of infectivity experiments were analysed by using ANOVA

and Fisher's PLSD after arcsin-transformation of total mortalities at 15










days. A Student t-test was used to compare the number of conidia

produced on cadaver piles of infected only versus infected + healthy ant

populations. In all cases the significance level used was p=0.05.


Results


Experiments with Live Infected Ants in Plaster-Bottom Cups

After one day of mixing healthy and B. bassiana-treated ant

populations, ants in the highly infected percentage treatments (85 and

100%) were observed to be distributed evenly over the internal surface of

the experimental arena. In the 50% treatment only part of the population

was spread out in the cup. This behavior was different from that of ants in

other treatments (0, 5, 15 and 50%) in which the ants were mostly clumped

in large piles in one side of the plate.

In all treatments, cadaver piles were observed starting on day 1 of

the experiment, that is 2 days after inoculation of B. bassiana, but only on

day 3 did they start increasing in size at a faster rate in the treatments with

high percentages of infected ants. In general cadavers were piled on one

side of the experimental arena, while the live ants mostly stayed on the

opposite side, even though some ants were always observed wandering on

top of the cadaver piles. By day 3, the size of cadaver piles in the different

treatment cups were apparently proportional to the treatment received.










Growth of the fungus on the cadaver piles was evident in the 100%

treatment but only limited fungal growth was observed in the other

treatments. Mycelial growth on the cadaver piles could be observed

starting on day 3, i.e. 4 days after B. bassiana inoculation, and sporulation

started on day 4. In other treatment cups the growth of the fungus always

seemed to be more abundant on the bottom of the dead cadaver pile than

on the top.

In both experiments, the survival of untreated ants decreased with

increasing percentage of infected ants in the populations (Tables 3-1 and 3-

2). By day 4 most of the 100% treatment ants were already dead, and just a

few ants (< 30) were still surviving, and by day 11 in the first experiment

and day 8 in the second, there were no more survivors in the 100%

treatment cups. Therefore, survivors in the other treatments can then be

assumed to be from the untreated part of the population.

In both experiments the survival of the control ants was c. 58% of

untreated ants after the 3 weeks of the experiment. The decrease in

survival of untreated ants with increase in the percentage of infected ants

in the colony appears to be similar in both experiments, even though the

statistical separation of the treatments is not the same. In the first

experiment the 5% treatment was not significantly different from the

control, but both these treatments were significantly different from the 15

and 50% treatments (Table 3-1), whereas in the second experiment, all















6I 6 6 6 I4

II I
o9J I I +I I I
-.3 I I- I (



t*S13 ui "d o d o 2


k I











I I










II CD N I
.41 F 5 I
I 0 Q\ \


Ia O



(I I I g I C

IZ I 10 I
I o a
Ic* 10 Nn 0


a) Io ( Ud
Z" 3) 0 0





| I flu 8















c(1 nr














0)U



.i 3
S14
0)








0)~

4-b
.1 c











%4 -4

-0)
< -<












>
p c;




C; t5







S 0





(Ur
0
a


























P.C
tc
(a




Ha
a .5a



s; W





I ^a





a -s


.0
I I


(U


10) .
I~
I I



II


I I I

I 10) I
I I~ I
I 10) I
(d











I
C)











I
I
















I I I
n I ZI
















I )I I
I I








I 0 I0
I I
w




5 5





3O 6

I, i. O Is

I




I I





ti0<


11
F

er

II->
I I
I IJ
i 10I
I .


(U (U (U (U (U I j
ca mS -~I 0
0--,0 I C

+ +0 +1 + +CL




lot
N 0 e n 0 0 C







o o c4 to -
r- 0 N a RR
C N r n %D








-14







r CDI CD
% D0 0 0 C*1 4 j f < 1







I a I





too3
UE






















Uo a' 00 00 0 '
\d ~ ~ cr ca










treatments were different from each other, except for 50 and 85% (Table 3-

2). The differences in the results of the two experiments were basically on

the survival of ants in the 5% and the 50% treatments. In both cases the

survival was lower in the second experiment. The 85% treatment in the

second experiment had only an average of 11 survivors, or less than 2% of

untreated ants.

In both experiments, the survival as determined by the weight of

the live ants was in all cases lower than the survival in terms of the

number of ants. This fact was due to a decrease of 8 to 46% in the mean

weight of the individual ant during the course of the experiments. In

general, the mean weights of the surviving ants in the different

treatments were not significantly different. The only statistically

significant difference detected was between the mean weight of ants

surviving in the control and 15% treatment in the first experiment.


Experiments with Live Infected Ants in Soil

Ants deposited cadaver in piles on one side of the cup and could be

seen from the outside through the clear cup wall. Cadavers were

deposited starting on day 2 in all treatments and although no quantitative

measurements were taken, size of piles appeared to be related to

percentage of infected ants in the populations. While the cadaver piles in

the lower treatments were all well defined, with very few if any cadavers

seen outside of these piles, in the 100% treatment ant cadavers were not in










very well defined piles and were observed throughout the soil, in the

galleries or on the soil surface. Cadaver piles in lower treatments were

always surrounded by soil with galleries that went along the borders of the

cadaver piles.

By day 4, very few ants could be observed in the 100% treatment

cups, but some ants remained alive in this treatment for the duration of

the experiments. In comparison, all infected ants left in cup without soil

were dead by day 12.

Fungal growth was observed on the ant cadavers in all treatments,

but with more intensity in the 100% treatment. Sporulation of the fungus

was evident by day 7 of the experiment on cadavers in 100% treatment.

Activity of ants in all treatments except 100% seemed to be normal

and galleries were observed over most of the surface visible through the

cup walls. Treatments that had large numbers of survivors seemed to

consume the honey-water given as food, but the ants in treatments with

fewer survivors did not always consume all the liquid diet.

In the first experiment, the survival of the untreated ants was

around 70% in all treatments that received untreated ants (Table 3-3).

Four out of 6 replicates of the 100% treatment had surviving ants at the

termination of the experiment, with an average of 4 ants per cup, but

survival was less than 1% of the initial population. Therefore, the

survivors in the other treatments can be assumed to be derived from the














SI
SI CN O
ZVI 9 9 9



C d d o
.j *t t
S 1++


ka) LO I
s L- i (u (U (0
I" I I i<


00 -





%0 0













IN C












I N
1 C

























S(U I I I
- Kn I
(I C I
I 6 C n

I I
n 0 C00 0 N











I I











I I H.


9 I II
o
oQ o
S III

0 %

I I
Sf I



i


-1

N






D












-D
I..c


IB








I
i W

I S(

fl .i










untreated portion of the populations. When the survivals as percentages

of the untreated ants were compared, all the treatments were statistically

equal with survivals between 67 and 75%.

In the second experiment very similar values for survivals were

obtained and the survival of untreated ants was between 63 and 68%

(Table 3-4), with no statistical differences among the treatments. All the

replicates of the 100% treatment had survivors with an average of 43 ants

per cup.

As in the experiments in plaster-bottom cups, the survival in terms

of weight of ants was less than the survival in terms of number of ants.

The decrease of mean ant weight was between 15 and 22%, with the

exception of the 100% treatment in both experiments. In this treatment,

the mean weight of ant survivors was 20-50% less than the weight of

survivors in the other treatments. Also, in both experiments the weight

of survivors in the control treatments was c. 7% less than that of ants in

other treatments, and a statistical difference was detected between this

treatment and the 50% treatment in the second experiment (Table 3-4).


Experiments with Dead Infected Ants

In both experiments with dead infected ants, the survival of ants in

terms of numbers of survivors was low, with maximum of less than 44%

of the initial population for both experiments (Tables 3-5 and 3-6).















g I 1d u
SD 0 0 C 0 C






S. 3o a
-H ti -H +1t -




S0 0
0J- I .
o I N I Cu




S0 I
IW
I






S( < a 4

.3 1a .5 I
A) I II i
to I 3 o m
c-b










(B ga o
0" t< ~ B


)I I
cuC
CD Cf) UI 00






1D -7' z I

I I I1





7 S O I ; ^
V4 Q






H IH I I 0
c2 H-~

















hI I
.0 .






0
< I



* I )

:I 0e

















Scto
0 10

I t I
I I
























IPC I
I I I I.
I I I


(u I IE
I I I
3 t>0 k
en (Q a>

1~ I


tv X) -09 -MO u



+1 -HI -+
0 0 6 0 0 dI







o -H, Zo I o
(U



C\ as aN C-4

1 I 1 1 hi
0i 0 t 6 g 6 6 I
IRV "r C4bo
*^ ca P CM .
(U I8^ U T
_~ --./~ -- -~. I -




0 00 "0 O\ T4
rsi coj d enj d c
14 Ci 5 i 6
I00 (U M m t
cnn
Ld cd


.0 (U 0 Cf)



04 00 C6 m



o6i tr> ci 1i o 10.
TC< O CO r.5.0


Cf)

I u
I -1

If) LO O 9
00 00 0 OJ' ^ I
C0 CO CO CC 6 0
CO ^< CO r-i I


i Q


L N



^1
co co ; ^< o o '





0- rU r< 0 ) >"
U a0
F3 S
NM '0 CO 00 0 I4 "
CO 0 '. N' NM !
NY N 9

Ia S


0 ^ IU) 0 10 c>
Q 10 '-4 10 00 Di
flu .0













1 l I <-4 1 I 0

I 0 0 0 60 D
S+1 +1 +1 +1 +1 +1 I







; 6




I I
C. c-i 4 i d









I II H
SAo vi L6 ci 6 1
0 I N CO CO N I w




0I 4
1- 4

I I I s '
0 ) I5 '> ~ I0 00 004

I 0 N1 COI -s








S0 C4 CI co CQ r4


>I 10 I 1o Co


I v" N



ID 0 Iv


z S I 0 in 0 in Lo
o Iu U iU- IO 00











Unlike the previous experiments in which the survival in terms of

weight of ants was always lower than the survival in terms of number of

ants, the results of these experiments showed better survival in terms of

weights. This result was a consequence of an averages of 10 and 2% weight

gain by the ants (excluding ants in the 85% treatment) in the first and

second experiments respectively.

Behavior of the ants was similar to that observed in the previous

experiments, but ants in treatment with 85% of infected ants did not pile

cadavers in a very organized fashion, and especially in the experiment in

which soil was used, cadavers were not moved from the place were they

were dumped when the experiment was initiated. Therefore, while

cadaver piles were mostly buried under soil in other treatments, in the

85% cups with soil, the cadavers could be observed on the soil surface.

Fungal growth was observed on the ant cadavers starting 3 days after

B. bassiana inoculation, and by day 4 vegetative growth (not confirmed to

be B. bassiana) was extensive without observable sporulation. Also,

fungal growth was more restricted to the lower layers of cadavers in the

piles and the top layer had little fungal material.

In the experiment with no soil, both the 50 and the 85% treatments

had significantly lower survival of untreated ants than the control

treatment (Table 3-5). Less than 3% of ants in the 85% treatment survived

the 3 weeks of the experiment, compared with almost 40% survival in the







69

control treatment. Ants in the 5% treatment actually showed significantly

better survival than the control ants, but this was not observed for the

weight of the surviving ants.

In the experiment with soil, ant survivals in the 50 and 85%

treatments were closer to that of ants in the control treatment (Table 3-6).

The 85% treatment was distinct from the control, but in this case the ant

survival in the 50% treatment was not significantly lower than in the

control. In this experiment, the best survival of untreated ants occurred in

the 15% treatment which was significantly higher than control survival if

number of survivors is considered, but not if weight of survivors was

used for comparison.

Even though in general the survival in this experiment was lower

than in the previous one, more than 10% of ants in the 85% treatment

survived the treatment, compared with 26% in the control and a

maximum survival of 36% in the 15% treatment.

The final mean ant weight showed an increase or very small

decrease in relation to the mean weight of untreated ants at the start of the

experiments. Maximum increase in weight occurred in the control

treatment in both experiments, with other treatments showing either

smaller increases or even decreases in the mean ant weight during the

experiment. In both experiments, the mean ant weight in the 85%

treatment was significantly lower than weight in other treatments, with










losses of 15 and 30% respectively in first and second experiments in

relation to the average starting weight of an individual ant.


Conidia per Ant

Cups with only infected ants had a mean conidial production of 3.1
x 1010 ( sem = 0.15 x 1010) conidia per cup, whereas cups with 50/50

healthy + infected ants produced significantly fewer conidia (1.9 0.16 x

1010) (Figure 3-1). Differences between treatments were larger when the

yield was considered in terms of number of conidia produced per g of ant

host. Only 0.9 0.80 x 1010 conidia were produced per gram of ant in the

healthy + infected populations, compared with the 3.1 x 1010 value for the

infected only population. The number of conidia being produced on an

average ant cadaver can be calculated by dividing the total conidial

production per gram of host by the number of individual ants per gram of

host. Considering the estimate of total number of ants/g in populations

with 100% of infected ants (1316 ants), and in 50/50 populations (1410

ants), an average of 2.3 x 107 conidia were produced per ant in the infected

only population, compared with only 0.66 x 107 conidia per ant in the

mixed population.


Experiments with Conidia Formed on Ants

Experiments using conidia harvested from ant cadavers from the

100% treatments (Figure 3-2) confirmed the virulence of the ant-produced




















Lu


0M
(U






(U
Lu






0
(U)


-4 a
o o
rl r-


vTpTuo3 jo zjqmnM


a?

S3
.0


*s x

os
fa i


U '- II
o "
0'4-. 1.4^
re Qj C




0ia


63





*0)
P8 ^







'-45
11.


a o -



Z .5 5



670















F64


H-


rd
0) o
Ien
oa O
O.TS
r^ "
*0@
ao s<
rp ( U r-

os-1
0 o 4 &
c ,i -bJsO

i *t
$ ^3 o
O ,P




O 1
Qua)'-4
.- I. 0)
t%'~~


a0 )0


0 e T .
0 v)
1 D I ,


on
a at
to ai

CN
S ^

c0
soEat


L4IBJpOlj luaOlad


4


I-


I I I


i


I










conidia. In general, conidia derived from ants either from the experiment

with or without soil caused higher mortality than conidia from SDAY

plates. In the first experiment with conidia from the experiment with no

soil, all treatments were significantly different from the control, which

had 8.7% mean mortality. The treatment with 108 conidia from ants/ml,

with 88% ant mortality at 15 days, was significantly higher than all other

treatments, whereas, the 107 ant conidia treatment (69% mortality) was

also higher than all other treatments except the one with 108 plate-derived

conidia/ml.

In the second experiment, conidia derived from ants in soil showed

some superiority in relation to conidia produced on artificial medium

plates. With the exception of the treatment with 106 plate-derived

conidia/ml, all others were significantly different from the control which

had 16% of mean mortality. The 107 plate-derived conidia/ml treatment

was not statistically different from the 106 ant-derived conidia/ml

treatment, but inferior to all other treatments. The last three treatments,
108 plate-derived, and 107 and 108 ant-derived conidia/ml were not

significantly different from each other, with mortalities of 99, 95, and 100%

respectively.










Discussion


Population density in the artificial colonies was directly affected by

the treatments but did not have any effect on the survival of ants. In

treatments with live infected ants, densities changed after fungal mortality

occurred, whereas in treatments with dead infected ants densities were

different from the onset of the experiments. However, in experiments

with live infected ants in soil, the survivals of untreated ants were similar

for all treatments, regardless of a population densities varying from 6 to 13

ants/g soil. Also, in previous experiments (see Chapter 2), the survival of

untreated ants maintained in small groups was apparently not affected by

the low population density of 1 individual/g soil.

Ant survival decreased with an increase in the percentage of

infected ants in experiments without soil but no increase occurred in the

presence of soil, indicating that transmission of the disease within the

colony is minimized by the presence of soil even when ants are confined

within cups. Differences between experiments with and without soil were

not due to difference in infectivity of conidia produced on cadavers under

the two conditions. These conidia were shown to have the same effect on

ants and to be at least as active as plate-grown conidia.

The apparent role of soil in limiting transmission of disease was

further corroborated by the fact that lower ant survival occurred in nests

with soil which received 85% of dead infected ants. Because cadavers were










left on top of the soil, since the small surviving population could not

handle so many cadavers, exposure of the healthy ants to the fungal

conidia was enhanced.

Another fact pointing to limited transmission of the disease in soil

is the presence of survivors in the 100% treatments with soil. Possibly

these ants were not inoculated or they were larvae or pupae when treated

and then molted before the fungus had a chance to penetrate the

hemolymph (Fargues 1972, Vey & Fargues 1977). However, no survivor

was observed in the 100% treatments without soil, indicating that ants that

escaped initial fungal treatment were probably infected later by fungus

grown on cadavers. In field situations where moribund ants can move

away from nest sites prior to death (H11dobler & Wilson 1990) or cadavers

can be removed from the general area of the nest, transmission of disease

from infected ants to healthy ones can be expected to be even less

prevalent.

Effect of the soil in the transmission of disease is due to ant hygienic

behavior which limits dissemination of fungal conidia in at least two

ways. The first one is necrophoresis or corpse removal away from the area

used by the ants, and packing of soil around such cadaver piles as was

observed by Storey (1990). For treatments in experiments described herein

in which necrophoresis was not observed, transmission of the disease

occurred from cadavers to healthy ants.










The second manifestation of hygienic behavior is the reduced

growth of fungus on cadaver piles exposed to live ants. Cadaver piles in

containers with live ants did not have extensive growth and sporulation

of B. bassiana on the surface and results from the estimation of conidia

per ant indicated that ants may somehow manipulate cadaver piles and

prevent some growth of the fungus. Further investigations would be

necessary to determine if this limitation of fungal growth is accomplished

by addition of chemical substances such as venom alkaloid which has

fungistatic action (Storey 1990), by consumption of the fungal growth as

observed by Evans (1989), or other ant behavior.

Ant activity in the artificial colonies was normal except for the

abnormal spread of ants in the treatments with a high percentage of

infected ants in cups with no soil. This behavior, which has been

observed before in other laboratory experiments, needs further

investigation but two explanations seem possible. The first possibility is a

behavior change in the sick or injured ants as has been observed in other

fungus-infected ants by Evans (1989) or injured ants by H11dobler &

Wilson (1990). The spread of ants seen in the containers may be the

attempt of the insects to move away from other members of the colony to

avoid transmission of disease. Another possible explanation is that

infected ants avoid clumping due to a higher oxygen consumption.











Higher oxygen consumption was observed (Sussman 1952) on Cecropia

moth infected with Aspergillus flavus.

Apart from the difference in survival of ants in 85% treatments in

nests with soil and a lower general survival, results from the experiments

with dead infected ants were similar to those in which live infected ants

were used. The lower survival may be related to a lack of adequate feeding

prior to initiation of the experiments, since these ants were the only ones

to gain weight during the experiments.

The observed effects of B. bassiana treatments could not be

partitioned between the simple presence of cadavers and the presence of

fungal infection. However, the presence of infected cadavers in

treatments containing low percentages of fungus-killed ants did not

significantly affect survival of ants, suggesting that mortality in treatments

was due to the transmission of the fungal disease. The use of colonies

with noninfected ant cadavers would be necessary to determine the effect

of the presence of different percentage of cadavers in ant colonies.

The loss in mean ant weight during the experiments was

comparable for most treatments, and probably a consequence of the

feeding scheme used. Treatments with large numbers of survivors

sometimes had lower mean weight of ants probably due to lack of

adequate food quantity for all ants. The low final weight of few survivors

in treatments with high percentage of treated ants, however, may be an







78

indication that these ants were underfed as larvae before escaping fungal

infection by molting into pupae and then adults. Lack of adequate care for

brood may be an indirect effect of fungal infection on fire ant colonies.

The evaluation of survival by numbers and total weights of

survivors is important due to changes in mean ant weight during the

course of the experiments and especially changes which differ among

treatments. Although initial population in each treatment cup was

standardized by weight, and only estimates were obtained of the initial

number of ants, evaluation of survival based on both variables provides

more information about the effect of treatments than either variable used

alone.

The results of this study are important for the development of a

strategy for use of B. bassiana against fire ants in the field. Application

methodology will have to rely on distribution of fungal material to reach

as large proportion of the nest population as possible in order to cause-

large mortality and disrupt normal nest activity such as hygienic behavior

and brood care. Disruption of hygienic behavior perhaps with use of

pheromone would also be desirable for inclusion in microbial pesticide

formulations. Reliance on initiation of fungal epizootics in the nest by

infection of small proportion of the population does not seem to be a

viable alternative for control of fire ant colonies with B. bassiana.














CHAPTER 4


GROWTH OF BEAUVERIA BASSIANA IN FIRE ANT NEST SOIL
WITH AMENDMENTS

Introduction

Beauveria bassiana commonly occurs in soil as a saprophyte and

attacks different stages of many insects (McCoy et al. 1988). Its growth and

mass production on artificial medium has been recently reviewed by

Bartlett & Jaronski (1988). Survival of this fungus both under laboratory

and field conditions has been studied in soil (Wojciechowska et al. 1977,

Lingg & Donaldson 1981, Miiller-K6gler & Zimmermann 1986, Gaugler et

al. 1989) as well as on foliage and host cadavers (Daoust & Pereira 1986a,

1986b).

Two types of soil fungistasis have been recognized (Dobbs & Gash

1965): one unstable and microbial in nature, and therefore susceptible to

elimination by sterilization, and another stable and possibly caused by

inorganic iron or calcium carbonate. Ethylene production, possibly by

spore-forming bacteria, has been implicated in the microbial type of

fungistasis of some soils (Dutta & Deb 1984). Fungistatic and fungicidal

action of soil against B. bassiana and other soil fungi have been linked

with several factors. Products of some fungi such as Penicillium urticae











and Aspergillus clavatus have been proven to inhibit B. bassiana very

efficiently (Shields et al. 1981, Majchrowicz et al. 1990). Soil type, humus

content and plant exudates (Sharapov & Kalvish 1984) as well as insect

chemicals (Storey 1990) also have been shown to affect stability of

entomopathogens in soil.

Although saprobic growth may be more important in the spread of

epizootics in the field, stability of fungal propagules in the soil is

important for long term survival of the species in the environment.

Beauveria bassiana can grow as a saprophyte in the soil under favorable

conditions but germination of conidia is inhibited by soil extracts and

conidia can stay dormant until a host is present (Clerk 1969). This

microorganism-associated fungistasis can be eliminated by filtration or

heat sterilization of soil extracts or by addition of nutrients to the soil.

Growth of B. bassiana in nonsterile soil was not observed by Walstad et

al. (1970), but in sterile soil and on the nonsterile body wall of an insect

host the conidia were able to germinate. When a host cadaver was

present, B. bassiana has been observed growing through nonsterile soil

from this inoculum source and eventually infecting other insects in the

vicinity (Gottwald & Tedders 1984).

The addition of nutrient amendments to the soil may in some cases

allow microorganisms to overcome the soil fungistasis. The addition of

chitin decreased the fungistasis against B. bassiana, M. anisopliae,










Paecilomyces farinosus, and P. fumoso-roseus of a dark gray forest soil

(Sharapov & Kalvish 1984). Chitin and other soil amendments also

caused an increase in number of B. bassiana conidia in sterile soil but a

decrease in fungal material in nonsterile soil. This decrease was not

caused by competition for nutrients by other microorganisms (Lingg &

Donaldson 1981). Nematode populations were also reduced by the

addition of chitin into soil due to an increase in fungal population which

attacks nematode eggs (Mian et al. 1982, Godoy et al. 1983).

Because fire ants live in nests with large numbers of individuals

and fungus-infected cadavers are removed to burial piles (Storey 1990),

there is good possibility for B. bassiana growth in soil after a portion of

the ant population is killed by the fungus. The present studies were

conducted to evaluate the capacity of B. bassiana to grow, spread, and

sporulate in soil from fire ant nests, with or without nutrient

amendments. Amendments included fire ant carcasses, chitin and rice,

and B. bassiana was inoculated on soil either by mycelial particles,

conidial suspension or infected cadavers.


Materials and Methods


Fungus

Fungal isolation and culture were as previously described (see

Chapter 2, page 19). Dry mycelium used was produced in SDBY







82
(Sabouraud Dextrose Broth + 5% Yeast Extract), agitated at high speed (400

RPM), and harvested on the third day after medium inoculation.

Harvesting of mycelium was accomplished by vacuum-filtering the

medium out using a buchner filter fitted with a filter paper disc. A

mycelial mat of c. 1.5 mm thickness was formed and cut into squares of

approximately 9 mm2. The mycelial particles were then air-dried

overnight in a vertical-flow hood at c. 220C and maintained in the

refrigerator until used for inoculation of soil.


Amendments

Four amendments were chosen to be used in the soil: fire ant

carcasses, Sodium dodecyl sulfate (SDS) treated fire ants (Boucias et al.

1989), chitin, and rice. Ant carcasses were used as a soil amendment

because fire ants are present in the nest soil and fungal growth may occur

on these carcasses when B. bassiana is applied in the control of S. invicta

nests (Stimac et al. 1989). The SDS-clean fire ant carcasses were used in

an effort to eliminate other nutrients or antagonists present in the

carcasses but the cuticular material was preserved. The chitin (Kodak)

used was a practical grade material obtained by alkali-treatment of several

crustacea shells, and served as a comparison to the ant-derived chitin.

Rice, which is used as a medium in the large-scale production of the

fungus (Alves 1986, Alves & Pereira 1989), can be incorporated in the final

formulation of a microbial insecticide.










Ant carcasses used in the experiments were obtained by freeze

killing of ants from colonies collected in the field but maintained in the

laboratory. In order to grind ant material to particles that would pass

through a 60-mesh sieve (0.25 mm opening), 30g of ants were blended at

high speed in 250 ml of deionized water. The ant powder was then filtered

and freeze dried. Ant carcasses that were treated with SDS were first

blended with 95% ethanol, air dried in a vertical flow hood, and then

boiled in 10% SDS in deionized water for 30 min. The final ant powder

was washed with deionized water until no residue of the SDS was left and

then freeze dried. Rice and chitin were ground in a blender and sifted, and

only particles passing through a 60-mesh sieve were used.


Preparation of Soil Plates and Inoculation

The soil used in all the experiments described here was obtained by

excavating fire ant nests in the field. The ants were flooded out of the soil,

and the soil was allowed to air-dry in shade and then sifted to eliminate

plant materials and large rocks. Further sifting through a 60-mesh sieve

eliminated most residual ant cadavers and other larger organic material.
Each of the amendments was added to constitute 1% (w/w) of dry

soil. Part of the soil was then sterilized in an autoclave, and part left
nonsterile. Sterile deionized water (27%, v/w) was mixed into the soil and
15 g of soil was added to 60-mm-diameter plastic Petri dishes. The surface

of the soil was smoothed with a metal spatula and plates were maintained










in a refrigerator at 40C until ready to be used. Six plates of each soil
amendment/sterilization combination were prepared for each

experiment. Also plates of SDAY and water/agar were prepared with all
inoculum types to verify their viability. Soil plates were maintained in
humid chambers in incubators at 252 OC.

Three types of inoculum were used in separate experiments: dry

mycelial particles, conidial suspensions, and B. bassiana-infected ant

abdomens. Dry mycelial particles (4 mm2 of surface area, and 1 mm

thickness) were placed on the center of the plate and leveled with the soil.

Conidial inoculum (1.5 x 109 conidia/ml suspension) was prepared with

sterile deionized water, and 5 pl were pipeted in to a small hole made in

the center of the soil plate, each plate receiving approximately 7.5 x 106

conidia. Conidia in suspension were c. 99% viable as determined by

plating of conidia on SDAY plates and microscopic observation of

germinated and ungerminated conidia after 20-24 h. To obtain ant

inoculum, large worker fire ants were infected with the same conidial

suspension as described above and dead ants were collected daily, washed

in 95% ethanol and individually placed in wells of a microtiter plate

maintained in the refrigerator to avoid further development of the

infecting fungus. At the time of inoculation, the abdomens were severed

from the rest of the body, placed in the center of the soil plate and leveled

with the soil surface. The remaining carcasses were maintained in a

humid chamber to allow sporulation of B. bassiana on the ant bodies.












Observations, Measurements and Analysis of Results

Fungal growth was observed and/or measured on the third, fifth,

seventh, tenth, and fifteenth day after inoculation. Two diameter

measurements were made, averaged and rounded up to the next

milimeter. Observations were made on the sporulation and general

growth of B. bassiana colonies as well as on the presence and growth of

other contaminant microorganisms.

On the fifteenth day, either the center part or all of the soil in each

plate was transferred into a plastic centrifuge tube (50 ml), and 35-50 ml of

sterile deionized water was added. The tubes were shaken for a minimum

of 30 min at 400 RPM in order to suspend the fungal material contained in

the soil sample. Dilutions of the soil extract (0.1 ml/plate) were plated on

oatmeal-Dodine agar medium (Beilharz et al. 1982) as modified by Steve

Krueger (pers. comm.) The plates were incubated in the dark for 7 days

after which the number of B. bassiana colonies were counted. After

incubation, the number of colony forming units (CFU's) of B. bassiana

per plate of fungal colony was calculated.

ANOVA and Fisher's Protected Least Significant Difference (PLSD)

were used to determine differences in log of CFU's produced in each of the

treatments, as well as average diameter of the colonies on the soil surface

at 3, 5, 7, 10 and 15 days post-inoculation.









Results


Mycelial Inoculum

Soon after addition of mycelial particles to the soil, they absorbed

water and swelled. By the third day on the soil new fungal growth was

observed on all particles regardless of the treatment, except for one of the

six soil plates with control nonsterile soil. This soil plate never showed

any signs of new fungal growth and was eliminated from the experiment

on the assumption that some unknown factor had caused the death of the

mycelial particle. Sporulation of the fungus was observed in all

treatments by day 5.

Mycelial growth in all particles placed on nonsterile soil was very

uniform among the different amendment treatments and diameter of the

fungal colonies was around 3 mm throughout the experiment, with no

significant difference in diameter of colonies being detected among the

nonsterile soil treatments (Figure 4-1). In nonsterile soil, contaminants

were observed growing on the soil surface but no antagonistic fungus was

observed killing B. bassiana growth on top of mycelial particles.

In general mycelial particles on sterile soil showed more new

growth than particles on nonsterile soil starting only 3 days after

inoculation. In ant and rice amendment treatments, the fungus

continued to grow during the experiment (Figure 4-1) and by day 5 two of

the soil plates on sterile soil with ant amendment had mycelial growth







87




r.o

--o- t v
I / .




\' "0 g'^

li o.



S: 0



M a



*1*
*r a








I
I a o *U
1 1-
I a I
x a& "






a U\ l










(m0) Anoio3 jo jaIIWBia