SELECTION OF SUITABLE ENTOMOPATHOGENIC NEMATODES FOR
BIOLOGICAL CONTROL OF SWEETPOTATO WEEVIL,
CYLAS FORMICARIUS (COLEOPTERA: APIONIDAE)
CATHARINE M. MANNION
A 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
UNIVmRSITY OF FLOrIlp LaIr1S
I express sincere gratitude to my advisor, Dr. R. KJansson, and my
committee, Dr. J. L Capinera, Dr. J. H. Frank, Dr. S. K. O'Hair, and Dr. G. C.
Smart, Jr., for all the help and support they showed me during my stay at the
University of Florida. Additional thanks are extended to Dr. R. McSorley for his
help and advice.
I also extend my gratitude to Nancy Epsky, Linda Mason and Laura
Powers for encouragement, advice, and support. I thank everyone who
assisted in data collection and anyone who provided initial insect and
Most of all, I thank my parents for their encouragement and continued
support. Without them, this would not be possible.
This dissertation was supported by the U.S. Department of Agriculture
under CSRS Special Grant Nos. 88-34135-3564 and 91-34135-6134 managed
by the Caribbean Basin Administrative Group (CBAG).
TABLE OF CONTENTS
1 INTRODUCTION................................................................................................. 1
The Sweet Potato.......................................................................................... 1
The Sweetpotato Weevil...................................................................... 2
Biological Control of the Sweetpotato Weevil............................. ........... 7
Factors that Affect Nematode Infection................................................ 9
Research Objectives.................................................................................. 11
2 COMPARISON OF TEN ENTOMOPATHOGENIC NEMATODES
FOR CONTROL OF SWEETPOTATO WEEVIL....................... 14
Introduction ........................................................................................................ 14
Materials and Methods............................................................................... 16
Results .......................................................................................................... 19
Discussion .................................................................................................... 21
3 INFECTIVITY OF FIVE ENTOMOPATHOGENIC NEMATODES TO
THE SWEETPOTATO WEEVIL IN THREE BIOASSAYS .................. 33
Introduction n ........................................................................................................ 3 3
Materials and Methods............................................................................ 34
Discussion ................................................................................................... 39
4 WITHIN-ROOT MORTALITY OF THE SWEETPOTATO WEEVIL BY
ENTOMOPATHOGENIC NEMATODES ............................................. 58
Introduction ........................................................................................... .. 58
Materials and Methods.............................................................................. ... 60
Discussion .................................................................................................... 65
5 MOVEMENT AND POST-INFECTION EMERGENCE OF
ENTOMOPATHOGENIC NEMATODES ............................................ 74
Introduction ................................................................................................ 74
Materials and Methods....................................... ................... ................76
6 MOVEMENT OF TWO HETERORHABDITID NEMATODES TO
SWEETPOTATO STORAGE ROOTS................................................... 98
Introduction ........................................................................................................ 98
Materials and Methods.............................................................................100
7 CONCLUSION............................................................................................. 120
LIST OF REFERENCES .............................................................................................124
BIOGRAPHICAL SKETCH .......................................................................................136
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
SELECTION OF SUITABLE ENTOMOPATHOGENIC NEMATODES FOR
BIOLOGICAL CONTROL OF SWEETPOTATO WEEVIL, CYLAS FORMICARIUS
Catharine M. Mannion
Chairman: Richard K. Jansson
Major Department: Entomology and Nematology
Entomopathogenic nematodes in the families Steinernematidae and
Heterorhabditidae (Steinernema carpocapsae [Weiser] [=Neoaplectana
carpocapsae, S. feltiae_, Agriotos, All, Breton, Italian, and Mexican strains; S.
feltiae [Filipjev] [=S, bibionis], N-27 strain; S. glaseri (Steiner); S. intermedia
[Poinar]; Heterorhabditis bacteriophora [Poinar] [=H. heliothidis], HP88 and
North Carolina strains; and Heterorhabditis sp. FL2122 [an undescribed
nematode isolated in Florida]) were compared in their ability to infect and kill the
sweetpotato weevil, Cylas formicarius (F.). Initial comparisons (LD5o, host
stage susceptibility, rate of infection, and progeny production) were measured
in a standard Petri plate bioassay. Generally, LDo0s were low (< 10 infective
juveniles per insect). Heterorhabditid nematodes were more pathogenic to
pupae than were steinernematids. Adult weevils were the stage least
susceptible to nematode infection. Larval mortality increased significantly over
a 4 day period for all nematodes with the exception of S. feltiae.
Heterorhabditid nematodes produced significantly more infective juveniles per
cadaver than did steinernematids. Mexican, All, N-27, HP88, and FL2122 were
further compared in their ability to seek and infect (number of invading
nematodes) sweetpotato larvae in 3 experimental arenas (Petri plate, sand, and
soil). Heterorhabditis sp. FL2122 was more infective than all other nematodes
in all three arenas. Nematode infectivity was affected by test arena.
Steinernema carpocapsae strains performed better in the Petri plate arena
compared to the soil or sand arenas while heterorhabditid nematodes and S.
feltiae performed best in the sand arena. The ability of nematodes to seek and
kill weevils within storage roots was compared among the above five
nematodes and S. glaseri. Heterorhabditid nematodes caused higher levels of
mortality within the root than steinernematids. Heterorhabditis sp. FL2122 killed
the most weevils in the shortest period of time. Lastly, emergence and
movement of nematode progeny produced in weevil cadavers within storage
roots were compared among nematodes in sand and soil in the laboratory and
in the field. Also, the distance that nematodes moved from a point source was
measured in a sweet potato field. More heterorhabditid nematodes were
recovered from larval cadavers within infested roots than steinernematids.
Heterorhabditid nematodes are more efficacious than steinernematid
nematodes for controlling the sweetpotato weevil.
The Sweet Potato
Sweet potato, Ipomoea batatas (L.) Lam., is a dicotyledonous plant
belonging to the family Convolvulaceae. The sweet potato originated in or near
northwestern South America around 8,000-6,000 B.C. (Austin 1988). The plant
was first discovered and cultivated around 3,000 B.C. (Austin 1988, 1991,
O'Brien 1972). Sweet potato had spread throughout most of the Neotropics by
2,500 B.C., with the exception of the temperate zones of the New World (Brand
1971). European explorers introduced the crop into Africa and India by the
early 1500s, China by 1594, Taiwan by 1597, and Japan by 1698 (Yen 1974,
1982). Approximately 92% of the world's sweet potato is currently produced in
Asia and the Pacific Islands (Horton 1988b). Most of this production occurs in
Sweet potato ranks seventh among all food crops world-wide, with an
annual production of 115 million metric tons (FAO 1984). It is grown in more
than 100 countries, and is second in importance to the white potato among the
world's root and tuber crops (Horton 1987).
Sweet potato is used as a staple food, vegetable (both fleshy roots,
tender leaves, and petioles), snack food, animal feed, a source of industrial
starch and for fermentation, and for various processed products (Bouwkamp
1985a, Kays 1985, Ln et al. 1985, Sakamoto & Bouwkamp 1985). Sweet
potato is a staple food in some areas of subsistence farming and is a drought-
tolerant crop. Developing countries account for 98% of the world production
(Gregory et al. 1990, Horton1988a).
A survey was conducted to define the constraints to sweet potato
production and use (Horton & Ewell 1991). The top-ranked constraints related
to post-harvest problems: unstable sweet potato supplies and prices, and the
lack of suitable processed products. Their survey also indicated that the
leading production constraints were low soil fertility, drought, and damage from
the sweetpotato weevil, Cylas formicarius (Fabricus). In an earlier survey,
Horton (1989) also found that the most limiting field production problems were
crop losses due to insects followed in decreasing order by plant viruses,
environmental factors, nematodes, available germplasm, and pathogenic fungi
The sweet potato crop is subjected to attack by various insect pests
shortly after planting until the time of harvest and in storage,. These pests may
limit productivity or seriously affect the quality (Ho 1970, Schalk & Jones 1985).
The majority of these pests occur sporadically or in numbers which do not
cause significant damage to the crop. Talekar (1988) listed 270 insect and 17
mite species as pests of sweet potato in the field and in storage around the
world; of these, the weevils C. formicarius, and C. puncticollis (Boheman) are
the most damaging.
The Sweetpotato Weevil
The sweetpotato weevil complex, Cylas spp. and Euscepes postfasciatus
(Fairmaire),varies from region to region, but as a group constitutes the most
important insect threat to sweet potato production on a worldwide basis. Cylas
is diverse in Africa and all known or suspected pest species of Cylas occur in
Africa and/or Madagascar. Only one pest species occurs extensively outside of
Africa/Madagascar, the circumglobally distributed -. formicarius (Wolfe 1991).
The sweetpotato weevil, C. formicarius, attacks nearly all plant parts and
immatures develop successfully in mature stems as well as in storage roots
(Vasquez & Gapasin 1980). Adult weevils feed on the epidermis of vines and
leaves and the external surface of storage roots causing round feeding
punctures. Females lay eggs singly in small cavities excavated in vines or
swollen roots. The eggs are sealed with a fecal plug that preserves moisture,
protects the egg from predacious mites, and also disguises the location of the
oviposition site (Sherman & Tamashiro 1954). Mean lifetime fecundities of
females are between 56 to 256 eggs (Jansson & Hunsberger 1991, Mullen
1981, Sutherland 1986). The developing larvae feed and tunnel throughout the
root or vines causing significant damage. Although 3-5 larval instars have been
reported (Fukuda 1933, Gonzales 1925, Jayaramaiah 1975a, Subramanian
1959), Sherman and Tamashiro (1954) demonstrated three larval instars and
concluded that frequent handling of larvae would interfere with feeding and
increase the number of instars. Pupation occurs in a small chamber prepared
by the final instar. The general adult remains within the pupation chamber or
larval tunnel for a minimum of 4 days (Cockerham et al. 1954). The white adult
gradually matures developing a normal coloration as it tunnels out of the root.
Weevil development, fecundity, and longevity are closely related to
temperature (Mullen 1981). The optimum temperature for development is
between 27 and 300 C, and within this range, the life cycle requires
approximately 33 days. Mean adult female longevity ranges from 76 to 110
days (Mullen 1981, Subramanian 1959, Trehan & Bagal 1957).
The most critical period for weevil attack is late in the season when
storage roots are more available and vulnerable. Storage roots are the
preferred oviposition site (Subramanian 1959). In the field, populations of C.
formicarius increase exponentially at a rate of about one weevil per plant per
day, with most of the increase in population density occurring late in the
growing season (Jansson et al. 1990a). At this time, more weevils are found in
the storage roots than in the vines of most cultivars (Jansson et al. 1990a).
The period between storage root initiation and harvest is critical in
relation to weevil attack. Weevil infestations can become serious if harvest is
delayed (Nawale 1981, Sutherland 1986b). Sutherland (1986b) observed a
linear increase in storage root damage by weevils starting between 80 and 100
days after planting.
Cylas formicarius spends most of its life cycle within the vine or swollen
root, which provides some protection for the developing weevils. For this
reason, the first attempts at chemical control were usually not successful due to
the lack of systemic insecticides (Sutherland 1986a). Sutherland (1986a) listed
59 insecticides, including botanicals of unknown chemical composition, that
were tested against sweetpotato weevil with varying levels of control. Pre-plant
insecticide applications have been used to manage weevils in planting material
(Talekar 1991). This method combined with proper sanitation and other
measures to prevent immigration of weevils from infested plants, has resulted in
satisfactory control of the weevil (Sherman 1951, Sherman & Mitchell 1953,
Sherman & Tamashiro 1954, Wolcott & Perez 1955). Control of the weevil is
difficult with conventional spraying, dusting, or fumigation with currently
available insecticides once weevils are present within the crown or storage
roots (Talekar 1991). Post-plant weevil control from chemical insecticides
requires frequent applications in order to kill adults that might migrate into the
field from other areas. Frequent spraying of insecticides may not be cost-
effective (Talekar 1991).
Cultural practices, such as crop rotation, intercropping, mulching, and
sanitation, were the earliest control measures advocated for reducing damage
by sweetpotato weevils. Talekar (1983) showed that rotations and vine dips of
planting material control weevils if alternate hosts are removed. Although
intercropping with various crop species was shown to reduce weevil infestation,
sweet potato yields were also reduced considerably (Singh et al. 1984). It is
uncertain whether the reduced yield (smaller or fewer roots) contributed to the
lower weevil infestations.
Soil cracks are the major route of weevil access to roots. Prevention of
soil cracking by hilling the area around the plant or irrigating frequently may
help to reduce weevil damage (Sherman & Tamashiro 1954). Mulches
presumably conserve soil moisture and minimize soil cracking as well as
possibly act as a physical barrier between the weevils and the roots. Talekar
(1987b) found plastic film and rice straw mulch reduced weevil infestations as
compared with nonmulched plots. However, Jansson et al. (1987) found that
weevil damage to storage roots of four cultivars of sweet potato did not differ
between plants grown in plots with and without a plastic mulch barrier.
Sanitation practices or clean cultivation may help protect the crop from
insect infestations. Destruction of the crop residue left in the field after harvest is
important because weevils survive in roots and stems (Jansson et al. 1989).
The use of weevil-free sweet potato cuttings is also important because the
weevils lay eggs in the vines in the absence of storage roots or when roots are
inaccessible. Presence of alternate hosts may influence infestations of the
weevil by harboring the weevil between crops and providing an additional food
source for the weevil. Therefore, alternate hosts should be removed.
Cockerham and Deen (1947) first studied the resistance of sweet potato
to Q formicarius in the United States. Since then, several attempts have been
made to develop resistant clones in breeding programs around the world
because of the high level of economic damage caused by these pests. Plant
traits which have been identified as important in weevil resistance include
fleshy root density (Martin 1984), dry matter and starch content (Cockerham &
Deen 1947, Hahn & Leuschner 1981), root depth (Burdeos & Gapasin 1980,
Jayaramaiah 1975), crown hardness (Cockerham & Deen 1947, Jayaramaiah
1975), and fleshy root surface chemistry (Nottingham et al. 1989, Wilson et al.
1988, Wilson et al. 1991). Planting weevil-resistant sweet potato cultivars is a
potential cultural control method; however, cultivars with reliable levels of
resistance to the weevil are not yet available (Talekar 1987a).
The existence of a female-produced sex pheromone in C. formicarius is
well known (Coffelt et al. 1978). Further studies resulted in the isolation,
identification, and synthesis of the active female-produced pheromone (Heath
et al. 1986). After the pheromone was identified, research focused on the
effectiveness of the pheromone at catching male weevils, the development of
an effective trap, and behavioral responses of males to synthetic analogs of the
pheromone (Heath et al. 1991).
Historically, the lack of reliable and cost-effective methods to detect the
presence of insect pests at low population densities has been a major
constraint to optimum use of control methods (Knipling 1982). Therefore,
studies were initiated to assess the potential of the female-produced
pheromone of C. formicarius as a monitoring tool for this pest. A monitoring tool
such as this may help growers time management tactics to reduce weevil
densities in seedbeds, in the field, and in storage (Jansson et al. 1991).
The sex pheromone also has potential for suppressing weevil
populations by mass trapping males and perhaps by disrupting mating
(Jansson et al. 1991). Recent studies showed that synthetic sex pheromone of
C. formicarius can disorient adult males and perhaps disrupt mating patterns
(Mason & Jansson 1990, 1991).
Biological Control of the Sweetpotato Weevil
Predators and parasitoids. There are several predators and parasitoids
that attack the sweetpotato weevil; however, these natural enemies seem to be
ineffective at managing weevil populations (Jansson 1991). Three predators
have been reported, two ant species (Cockerham et al. 1954, Castineiras et al.
1982) and a predatory maggot, Drapetis s.s. exilis group (family Empididae)
(Rajamma 1980). Fifteen wasp parasitoids of Cylas spp. have been reported
(Jansson 1991), most of which are not effective at suppressing weevil
populations (Gonzalez 1925, Cockerham 1944, Risbec 1947, Rajamma 1980,
Anonymous 1985, Jansson & Lecrone 1991).
Entomopathogenic fungi. Several fungal pathogens have been reported
to attack Cylas spp. (Jansson 1991). The most predominant fungus isolated
from Cylas spp. has been Beauveria bassiana (Bals.) Vuill. (Jansson 1991).
Epizootics from this pathogen in C. formicarius are rare in the field, although
epizootics have been observed in the laboratory (Jansson 1991). Beauveria
bassiana was shown to have potential for weevil management (Su et al. 1988).
Entomopathogenic Nematodes. One group of organisms with great
potential as biological control agents of sweetpotato weevils is
entomopathogenic nematodes. Nematodes in the families Steinernematidae
and Heterorhabditidae have been shown to be infective against many insects
(Gaugler & Kaya 1990, Poinar 1979). Infection is initiated by a third-stage
juvenile which is morphologically and physiologically adapted to remain in the
environment for a prolonged period without feeding. The alimentary canal is
collapsed and nonfunctional and the mouth and anus are closed (Poinar 1990).
These nematodes harbor symbiotic bacteria, Xenorhabdus spp., in their
alimentary tract. Once contact is made between the host and nematode, the
nematodes enter the host's hemocoel through natural body openings (mouth,
anus or spiracles). Heterorhabditids also possess a dorsal tooth or hook which
may be used to break the outer cuticle of an insect (Bedding & Molyneux 1982)
and allow entry. As soon as the infective juveniles enter the host hemocoel,
nematode development is initiated. The alimentary tract becomes functional
and cells of the bacteria are released through the anus and start to multiply in
the insect's hemocoel. The host ultimately dies from septicemia. The bacteria
are consumed and digested by the developing nematodes. Infective juveniles
of Steinernema develop into amphimictic females or males and those of
Heterorhabditis develop into hermaphroditic females (Poinar 1990). The
second generation, however, consists of amphimictic females and males in both
genera. As resources are depleted within the host, infective juveniles exit the
cadaver to seek new hosts.
Initial studies with an entomopathogenic nematode were conducted in
the 1930s to test S. glaseri (Steiner) as a biological control agent of Japanese
beetle grubs (Popillia japonica Newm.) in soil (Glaser & Farrell 1935). In the
1950s, S. carpocapsae (Weiser) was isolated and used principally as an
above-ground biological insecticide (Kaya 1985). Excellent control has been
obtained against insects in moist, cryptic habitats, but many failures occurred
against foliage-feeding insects (Gaugler 1988). Failures against foliage-
feeding insects were attributed to rapid desiccation and inactivation by sunlight
(Gaugler 1988). Although field trials against soil insects have shown greater
success, results have been inconsistent. (Georgis & Gaugler 1991, Klein 1990).
The sources of variation included application methods, nematode species and
strains, host stage and defense mechanisms, and the biotic and abiotic soil
environments. Application method and timing are crucial to the success of a
control program. Application methods have varied greatly. Nematodes have
been applied as soil drenches, through conventional spray or irrigation
equipment, and with baits with varying results (Georgis 1990). Selection of the
nematode species/strain is also important to success because infectivity differs
among the various nematode species/strains and the susceptibility of hosts
varies among species (Bedding et al. 1983, Belair & Boivin 1985, Capinera et
al. 1988, Fuxa et al. 1988, Georgis & Gaugler 1991, Griffin et al. 1989, Jackson
& Brooks 1989, Klein 1990, Morris et al. 1990). Nematodes can also be
hindered greatly by the environment in which they have been applied.
Therefore, the type of environment in which the nematodes will be applied
should be considered in addition to the application method and nematode
Factors that Affect Nematode Infection
Many factors influence the ability of the nematode to seek out and kill an
insect host. The principal soil factors affecting nematodes are pore size, water,
aeration, temperature, and the chemistry of the soil solution (Wallace 1963). Of
all these factors, soil moisture influences nematode activity the most (Kaya
1990). Soil pH, however, is one factor that probably has little affect on
nematode activity in that the pH of most soils ranges from 4 to 8 which is not
detrimental to the survival or pathogenicity of these nematodes (Kung et al.
Although a small portion of steinernematid and heterorhabditid
populations disperse, there are differences in movement among species/strains
(Kaya 1990). Soil texture and the presence of a host has been shown to
influence nematode movement (Alatorre-Rosas & Kaya 1990, Georgis & Poinar
1983a, 1983b,1983c, Moyle & Kaya 1981, Poinar & Hor 1986, Schroeder &
Beavers 1987). Wallace (1958) indicated that pore spaces of small particle
sizes were too narrow for efficient movement. Nematode movement in heavy
clay soils is reduced with less movement occurring as the percentage of silt and
clay increases in the soil (Georgis & Poinar 1983a, 1983b). Although the
presence of a host in clay soils did not increase dispersal of infective juveniles,
the presence of a host did significantly increase the number of dispersing
infective juveniles in other soil types (Georgis & Poinar 1983a, 1983b, 1983c).
Soil type was also found to affect survival and pathogenicity of S.
carpocapsae and S. glaseri (Kung et al. 1990a). Persistence and pathogenicity
of both species decreased as the proportion of clay increased. Survival for S.
carpocapsae was greatest in sandy loam and was greatest in sand for S.
glaseri. Molyneux and Bedding (1984) also demonstrated reduced
pathogenicity of S. glaseri to larvae of the sheep blowfly in clay loam compared
to sandy soils.
When infective juveniles were placed in the middle of a soil column,
significantly more H. bacteriophora (Georgis & Poinar 1983c) and S.
carpocapsae (Georgis & Poinar 1983b) moved upwards than downwards,
whereas significantly more S. glaseri moved downwards (Georgis & Poinar
1983a). Recent studies indicated that five heterorhabditid infective juveniles
placed on a soil surface were sufficient to infect a high proportion of hosts
buried in soil (Choo et al. 1989). Heterorhabditids and S. glaseri appear to
initiate random movement in soil, and when they are in close proximity to a host,
may use chemical and physical cues for host-finding (Kaya 1990).
The presence of roots may also affect nematode movement by altering
the soil structure, water and the biotic environment. Dense roots were shown to
affect nematode movement adversely and infectivity in sandy soil (Choo et al.
1989) but sparse roots in humus increased nematode (. bacteriophora)
infectivity of the host (Choo & Kaya 1991).
The ability of nematodes to actively move within the environment is
advantageous because it increases the likelihood of encountering a susceptible
host. Some nematodes, however, may employ inactivity as a strategy to
enhance their survival; an inactive nematode reduces energy usage and
attractiveness to predators (e.g., mesostigmatid mites do not attack motionless
infective juveniles) (Ishibashi & Kondo 1986, 1987, Timper & Kaya 1989).
Inactivity in soil has been documented with S. carpocasae, which tends to
remain at the point of application (Moyle & Kaya 1981, Georgis & Poinar 1983).
Major differences in infectivity exist between steinernematid and
heterorhabditid nematode species and strains (Kaya 1990). The reasons for
these differences are unclear and may be related to nematode behavior, quality
of the nematodes, and attraction/adaptation to a given host.
The sweetpotato weevil is the most destructive insect pest of sweet
potato. The weevil is difficult to manage because it spends most of its life cycle
within the storage roots or vines. Entomopathogenic nematodes have great
potential as biological control agents of the sweetpotato weevil, however, little
work has been done in this area. Selecting the most appropriate nematode(s)
for the sweetpotato weevil system and understanding the interaction between
the nematodes and weevils are important to the success of a biological control
The infection process between entomopathogenic nematodes and insect
hosts is not well understood, especially among different nematode species or
strains. There are several laboratory and field bioassays available for
examining this interaction; however, the usefulness and interpretation of these
bioassays is at best confusing. There has been little standardization of
bioassays and results have been variable and inconsistent. Additionally,
laboratory bioassays have not always been indicative of what to expect from
field tests, thereby, questioning the validity of some of these bioassays.
Typically, one type of bioassay is used to evaluate these nematodes which may
also lead to erroneous results. For these reasons, there is a current need to
compare various entomopathogenic nematodes intensively in different
experimental arenas (bioassays) to develop a reliable protocol for selecting
suitable entomopathogenic nematodes for biological control of insect pests. It
has been suggested that a universal bioassay that predicts levels of control
under all conditions will probably never be achieved, but that it is still important
to have a universal bioassay as a first screen to narrow the choice under
particular environmental conditions, for quality control assessment when mass
production is attempted, and for designation of standard preparations (Hominick
& Reid 1990).
The overall objective of this project was to compare several
entomopathogenic nematodes as biological control agents of the sweetpotato
weevil and to define a system of bioassays for selecting nematodes that can be
expanded to other insect hosts. Most of the nematode species/strains used in
these studies were available from a commercial supplier; others were isolated
from soil in Florida. The factors examined included host susceptibility, host
finding, movement, and behavior in relation to the host, C. formicarius.
Each chapter of this dissertation is a separate manuscript which will be
submitted for publication in refereed scientific journals. For this reason, some
repetition occurs across chapters. Chapter 2 compares the suitability of C.
formicarius to ten entomopathogenic nematode species/strains in a standard
Petri plate bioassay. In Chapter 3, the infectivity (number of nematodes capable
of penetrating a host) of five nematodes selected from Chapter 2 is compared in
three different experimental arenas: Petri plates, sand, and soil. These
nematodes are compared further in Chapter 4 by focusing on their ability to
move, seek out, and kill weevils within sweet potato storage roots. The rate of
weevil mortality and the location where infection occurs within the root were
evaluated. The relationship between nematode infection and within-root weevil
density and root size were also investigated. In Chapter 5, post-infection
nematode emergence and movement from cadavers with the storage roots are
evaluated in the laboratory and in the field. Nematode movement from a point
source is also evaluated. In the concluding chapter, Chapter 6, the project is
summarized, gaps in the knowledge base are noted, and the future outlook of
these nematodes in biological control of C. formicarius and other soil insects is
COMPARISON OF TEN ENTOMOPATHOGENIC NEMATODES
FOR CONTROL OF SWEETPOTATO WEEVIL
The sweetpotato weevil, Cyla formicarius (F.), is a serious pest of sweet
potato, pomoea batatas (L.) Lam. (Sutherland 1986, Chalfant et al. 1990).
Weevil larvae and adults develop and feed within the roots and vines causing
damage to the plant. Terpenoids produced in response to weevil feeding make
even slightly damaged roots unpalatable (Akazawa et al. 1960, Uritani et al.
1975, Sato et al. 1981). Marketable yield losses due to weevil infestation have
been reported to be as high as 60-97% (Ho 1970, Mullen 1984, Jansson et al.
Entomopathogenic nematodes in the families Steinernematidae and
Heterorhabditidae have been tested and shown to be infectious against several
weevils, including the sweetpotato weevil (Bedding & Miller 1981, Belair &
Boivin 1985, Georgis & Poinar 1984a, 1984b, Rutherford et al. 1987, Schroeder
1987, Jansson et al. 1990b). These nematodes have great potential as
biological control agents of the sweetpotato weevil because they occur naturally
in soil; seek out well-concealed hosts, such as this weevil; possess high
virulence and reproductive rates; kill their hosts quickly; and attack a wide range
of hosts (Gaugler 1981). Also, mass rearing and subsequent application of
these nematodes is safe and relatively low in cost (Woodring & Kaya 1988).
Although the host range of most entomopathogenic nematodes is very
wide, considerable inter- and intraspecific variation in the infectivity of these
nematodes has been demonstrated (Bedding et al. 1983). Both steinernematid
and heterorhabditid nematodes can be highly infectious to their insect hosts;
however, heterorhabditid nematodes have been shown to be more virulent than
steinernematid nematodes to a variety of insects (Bedding & Miller 1981,
Molyneux et al. 1983, Georgis & Poinar 1984a, 1984b, Rutherford et al. 1987).
In a comparison of the infectivity of several species/strains of entomopathogenic
nematodes for a number of different insect hosts, all species/strains of
nematodes were capable of killing each insect species. However, infectivity by
most of the heterorhabditids was greater than that by steinernematids for four of
the six insects tested at all doses (100 to 104 nematodes per insect) and for the
remaining two at lower doses (Bedding et al. 1983). It also was noted that the
most commonly used nematode, S. carpocapsae, was the least infectious for
two of the insect species and was never among the most infectious.
Molyneux et al. (1983) noted that strains of the same nematode species
can vary widely in infectivity to larvae of the sheep blowfly, Lucilia cuprina
(Wied). LCsos of the eleven species or strains tested ranged from 18 to 5,349
nematodes per larva. Heterorhabditis spp. were able to reproduce in L
cuprina larvae whereas Steinernema spp. were not. Four unidentified isolates
(three Dutch and one Italian) of Heterorhabditis spp. were compared in the
laboratory using Galleria mellonella (L.) larvae as the host (Griffin et al. 1989).
Although all nematodes were able to reach insects through a 56 mm column of
moist sand, there were significant differences in insect mortality due to
nematode infection. Jackson and Brooks (1989) compared four S.
carpocapsae strains (Agriotos, All, Breton, and Mexican) against third-instar
western corn rootworm, Diabrotica virgifera virgifera LeConte. The LCsos
ranged from 32 to 325 nematodes per larva with the Mexican strain being the
most virulent and All strain being the least virulent.
A recent study attributed test failures against Japanese beetle, Popillia
japonica Newman, to unsuitable nematode strains or environmental conditions
after analyzing 380 nematode treatments from 82 field trials (Georgis & Gaugler
1991). An understanding of these basic differences among nematodes may
result in improved selection techniques for biological control agents. Questions
such as how quickly mortality is achieved, which stages) are most susceptible,
and how suitable the host is for reproduction, provide essential information for a
biological control program. Although many factors that affect nematode
performance under field conditions cannot be tested in the laboratory, particular
tests can be used to select the most appropriate nematodes.
In the study reported here, the virulence of ten species/strains of
nematodes to the sweetpotato weevil (LC5o and LCgo values), the susceptibility
of different weevil stages to these nematodes, larval mortality over time, and
reproduction of these nematodes within larval cadavers were compared.
Materials and Methods
Sweetpotato weevils used in the study were obtained from a laboratory
colony that was established from stock received from J. A. Coffelt, USDA, ARS,
Insect Attractants, Behavior, and Basic Biology Research laboratory,
Gainesville, Florida. Weevils were reared at room temperature on 'Jewel' sweet
potato tuberous roots (Coffelt et al. 1987). Late instar weevils and pupae were
dissected from infested roots before experimentation.
The following nematodes were used: Steinernema carpocapsae
(Weiser) (=Neoaplectana carpocapsae. Steinernema feltiae), Agriotos, All,
Breton, Italian, and Mexican strains; Steinernema feltiae (Filipjev)
(=Steinernema bibionis), N-27 strain; Steinernema intermedia (Poinar);
Heterorhabditis bacteriophora (=Heterorhabditis heliothidis) HP88 and North
Carolina strains; and Heterorhabditis sp. FL2122 (an undescribed nematode
isolated in Florida by K. B. Nguyen). All nematodes were cultured in vivo in late
instar G. mellonella (Woodring & Kaya 1988) which were reared at room
temperature on a modified baby cereal diet (Dutky et al. 1962).
All tests were conducted using a modified Petri plate bioassay (Woodring
& Kaya 1988). Ten insects were placed in 9 cm Petri plates lined with a double
layer of Whatman no. 2 filter paper. Nematodes were dispensed to the Petri
plate in 1 ml water. Nematode densities were estimated by counting the
number of infective juveniles in four subsamples. The timing for collection and
examination of weevil cadavers for nematode infection varied to coincide with
the particular test.
LC50 tests. Third-instar weevils were exposed to seven concentrations of
nematode infective juveniles ranging from 0 to approximately 15 infective
juveniles per larva. Each concentration was replicated four times (40
larvae/density). Four trials were conducted resulting in 160 larvae being tested
at each concentration. LC5o and LC9o values (+ 95% fiducial limits) and slope
estimates were determined for each nematode 48 h after inoculation. Mortality
beyond 48 h after inoculation was not determined due to sweetpotato weevil
cannibalism and high natural mortality rates. Cannibalism did not occur in less
than 48 h and infrequently after 48 h.
Pupal and adult mortality. Pupae and adults were exposed to two
densities of infective juveniles (larval LCso and 2x the LC5o). For each
nematode tested, 100 pupae and 100 adults (50 female and 50 male) were
examined per density. Percent mortality was determined 96 h after inoculation.
Larval mortality rate. Two hundred third-instar sweetpotato weevils were
exposed to the LC50 density of infective juveniles for each nematode. The
weevils were divided into four groups of 50 weevils each. Mortality due to
nematode infection was determined at 24, 48, 72 and 96 h after inoculation.
One group was examined at one time period.
Progeny production. Forty weevil larvae were exposed to approximately
1,000 infective juveniles. Three days after exposure to the nematodes, 20
cadavers that appeared to have been killed by the nematodes were randomly
selected for this experiment. The cadavers were placed individually in modified
White traps (White 1927) for nematode emergence. Nematode infective
juveniles emerging from each weevil cadaver were collected and quantified 14
d after inoculation.
Data analysis. Dose response data were analyzed with POLO-PC
(Russell et al. 1977). LCso and LCgo values of weevil larvae, their 95% fiducial
limits, and slope estimates were determined and reported as infective juveniles
per insect. Parameter estimates were compared using the general linear test
(Neter & Wasserman 1974).
Normally distributed data were analyzed by analysis of variance, and
means were separated by Tukey's Studentized Range Test (E<0.05) (Tukey
1953). Non-normally distributed data were square-root transformed and then
subjected to analysis of variance and Tukey's Studentized Range Test. Adult
mortality data were ranked and subjected to analysis of variance, an equivalent
to the Kruskal-Wallis k-sample test (Conover 1980). Step-wise paired t-tests
were performed on the ranked percentage data.
Data from the larval mortality rate test were analyzed by linear regression
analysis. Mean percentages of insect mortality were regressed on time (0-96
h). Homogeneity of slopes of the regression lines were tested by the general
linear test (Neter & Wasserman 1974).
LCAtests. All nematodes tested were capable of killing at least one
stage of the weevil. The LC5o for most of the nematodes was relatively low
(less than 10 infective juveniles per insect). The LCso and LCgo data are
summarized in Table 2-1. The lowest LCso values were obtained by one
heterorhabditid, H. bacteriophora FL2122 strain, and one steinernematid, S.
feltiae (1.9 and 2.9 infective juveniles per larva, respectively). The
heterorhabditid nematodes included the most virulent nematode strain (H.
bacteriophora, FL2122) and the least virulent nematode strain (North Carolina
strain, LC5o=29.7). Virulences of the five S. carpocapsae strains were similar
(LCsos ranging from 6-8 infective juveniles per larva). Although intermediate in
their LCso values, two strains of S. carpocapsae. Breton and Mexican, resulted
in significantly steeper slopes (P < 0.05) than all other nematodes tested.
Pupal and adult mortality. All nematodes tested were pathogenic to
pupae (Fig. 2-1). Heterorhabditid nematodes caused the highest pupal
mortality regardless of the nematode density. Two of the heterorhabditid
strains, HP88 and North Carolina, produced significantly higher pupal mortality
(low rate: F=17.40; df=9,90; P<0.0001; high rate: F=26.78; df=9,90; P<0.0001)
than all other nematodes. Strains of S. carpocapsae caused higher pupal
mortality than those of S. feltiae and S. intermedia at both densities.
With the exception of S. carpocapsae All strain, the nematodes did not
regularly infect adult weevils at either the LCso density or 2X the LC50 density
(< 6% mortality) (Fig. 2-2). Although differences in infectivity of adult weevils
were not significant, the All strain caused 21-25% more mortality at the LC50
density. At the 2X density, All strain caused 61.4 81.8% more adult mortality
than the other nematodes (F=2.79; df=9,90; P<0.0062). Male weevils
accounted for 100 and 96% of the adult mortality of the 1X and 2X LCsos,
Larval mortality rate. There was a significant increase in larval mortality
over time with all nematode strains except S. feltiae (Table 2-2). Although S.
feltiae did not cause a significant increase in larval mortality over time, it had the
highest intercept (0.37). Significant differences were observed among
intercepts; however, there were no explainable trends among species or strains
of nematodes. For example, the intercept for H. bacteriophora, North Carolina
strain, was not significantly different from the strains of S. carpocapsae. The
HP88 strain, another heterorhabditid, differed only from the All strain of S.
carpocapsae. The slope of HP88, however, was significantly different from all
nematodes except the Italian strain of S. carpocapsae. The intercept and slope
for S. feltiae differed from those of all strains of H. bacteriophora, S. intermedia
and most strains of S. carpocapsae.
Mortality of weevil larvae by each nematode was also compared at each
time interval (24, 48, 72 and 96 h) (Fig. 2-3). Larval mortality was generally low
at 24 h with the exception of S. feltiae which caused 42% mortality. Mortality by
S. feltiae was significantly greater than those by all other nematodes except H.
bacteriophora North Carolina strain (26% mortality) (F=6.685; df=7, 32;
P<0.0001). At 48 h, mortality ranged from 14 to 52% with no significant
differences among nematodes. Mortality at 72 h ranged from 38 to 90% with the
highest mortality caused by S. carpocapsae Agriotos strain and H.
bacteriophora. North Carolina and HP88 strains (F=7.89; df=7,32; P<0.0001).
Mortality at 96 h resembled that at 72 h mortality (F=4.96; df=7, 32; E<0.0001).
Progeny production. Heterorhabditid nematode strains, HP88, FL2122
and North Carolina, produced significantly more infective juveniles per cadaver
than all steinernematid nematode strains (F=21.18; df=9, 90; P<0.0001) (Fig. 2-
4). The first generation of heterorhabditid nematodes is hermaphroditic which is
likely to contribute to more progeny production than in steinernematid
nematodes. Although not quantified, most nematodes continued to reproduce
beyond the two week period.
In choosing an entomopathogenic nematode as a biological control
agent, it is important to look at several attributes of the agent in reference to the
prospective host or system. Although many factors are responsible for the level
of infectivity (i.e., attraction, penetration, movement, host defense mechanisms,
the biotic and abiotic environment, etc.), some basic questions may be
answered through laboratory studies such as: (1) can mortality occur? (2) which
stages are most susceptible? and (3) do these nematodes reproduce inside the
Differences among nematodes in their effectiveness against insects have
been demonstrated many times (Morris et al. 1990, Jackson & Brooks 1989,
Bedding et al. 1983). Bedding et al. (1983) demonstrated that no one
nematode was most infectious for all insect hosts, although heterorhabditids
were generally more effective than steinernematids in controlling root weevils.
Georgis and Poinar (1984b) reported that both genera were equally effective
against late-instar Otiorhynchus sulcatus (F.) Often, the favorable results
obtained with heterorhabditid nematodes were attributed to their tendency to
move downward from the point of application (Georgis & Poinar 1984b) and
their strong host-finding ability in soil (Choo et al. 1989). As demonstrated in
the current study, heterorhabditid nematodes may possess additional positive
attributes compared to steinernematid nematodes.
Although the weevil was susceptible to each of the nematode
species/strains tested, there were differences among these nematodes in their
ability to kill and reproduce within a larval cadaver. In general, heterorhabditid
nematodes may be more efficacious against sweetpotato weevil than
steinernematids. This statement, however, was not necessarily reflected in the
LCso data, demonstrating the importance of using more than one type of
bioassay when evaluating nematodes. Two of the three heterorhabditids had
higher LCsos than most of the steinernematids. This LCso test may have been
more useful if it had been continued longer than 48 h post-inoculation to allow
maximum infection. This became apparent in the larval mortality rate test. All
nematodes produced a significant increase in mortality over four days except S.
feltiae. Therefore, most nematodes did not cause maximum mortality until 96 h.
An LCso determined at 48 h (as in this case) will be considerably different from
an LC5o determined at 72 or 96 h. Additionally, point estimates such as LCso,
as determined through probit analysis, represent only one position of the
regression line at a single level of the response. Unless the respective probit
regressions have equal slopes, comparisons cannot be done (Finney 1971).
For example, in a comparison of four strains of S. carpocapsae, Agriotos, All,
Breton and Mexican, against corn rootworm larvae, the regressions of Agriotos
and All strains were clearly unequal but because they converged near the 50%
response level, the LCso estimates were not significantly different (Jackson &
Weevil larval LCsos for HP88 and All were previously reported as 3.4
and 2.6, respectively, by Jansson et al. (1990b). The differences between these
LCsos and the currently reported LCsos (9.3 and 8.4, respectively) can possibly
be accounted for in several ways. First, is time at which the LC50 was
determined, as discussed above. The previous LCsos were determined at 96 h,
two days longer than the current test. Second, the sample size of larvae used to
determine the LCso was considerably greater in the present study (n=1,160)
compared to that sample size used in the previous study (n=327 for All and
n=366 for HP88). Lastly, the nematodes used in the present study were reared
in vivo, whereas those used in the previous study were commercially produced
Heterorhabditid nematodes also may be a more favorable control agent
compared to steinernematid nematodes because weevil pupae were, in
general, more susceptible to heterorhabditids than to steinernematids. Various
stages of the weevil are found within roots at the same time which makes this
attribute especially advantageous. The more stages that are susceptible to the
nematode, the greater chance of a reduction in the insect population density.
Jansson et al. (1990b) noted no differences in susceptibility between weevil life
stages with H. bacteriophora, HP88 strain, and S. carpocapsae. All strain. That
is, pupae and larvae were equally susceptible to these two nematodes. These
differences may also be explained by the difference in sample size as well as
the origin of the nematodes. Although pupal mortality was slightly greater when
the nematode density was increased, the increase in mortality was not
indicative with a doubling of the nematode density.
Adult weevils were not very susceptible to nematode infection. The
exception of adult male susceptibility to S. carpocapsae All strain, was a unique
phenomenon and will need to be investigated further. As with pupal mortality,
doubling the nematode density resulted in little or no increase in adult weevil
Important to the establishment and persistence of these nematodes in the
field is their ability to reproduce in weevil cadavers. Most of the nematodes
were capable of reproducing within weevil larval cadavers; however,
heterorhabditid nematodes were greatly favored over steinernematid
nematodes, at the time that progeny production was recorded. We recognize,
however, that progeny production within larval cadavers extended beyond the
14 d post-infection sample time. The length of time that infective juveniles
emerge from weevil cadavers is not yet known. A recent laboratory study found
that heterorhabditid infective juveniles emerged from sweetpotato weevil larval
cadavers for up to 28 d after inoculation (Jansson, unpublished data).
Heterorhabditis bacteriophora FL2122 strain and an undescribed
Heterorhabditis sp.(Bacardis) isolated from Puerto Rico, both produced
approximately the same number of nematodes as the HP88 and North Carolina
strains in the current study while the FL2122 strain in the current study
produced approximately half of what the FL2122 strain produced in Jansson's
The differences noted among nematode species/strains may be peculiar
to the host or system under investigation. In a comparison of five
entomopathogenic nematodes against white grub larvae, Ligyrus subtropicus
(Blatchley), there were no significant differences in infectivity (Sosa & Beavers
1985). However, Belair and Boivin (1985) found that S. carpocapsae and S.
feltiae caused more mortality of carrot weevil, Ustronotus oregonensis Le
Conte, larvae than did H. bacteriophora. Three European heterorhabditid
isolates differed significantly in their ability to control vine weevil, 0. sulcatus,
larvae in pots, to which the authors attributed some of the variation to
differences in nematode activity in the soil and in penetrating the insects
(Simons & van der Schaaf 1986). Dunphy and Webster (1984) also suggested
behavioral differences to account for differences in virulence of S. feltiae strains.
Griffin et al. (1989) substantiated the hypothesis that the variation between
isolates is at least partly due to differences in the ability of the infective juveniles
to find and/or enter hosts. They also suggested that a possible source of
variation between isolates is in the pathogenicity of the nematode-bacterial
complex following entry into the host.
Therefore, there is a need for better interpretation of laboratory
bioassays and their importance in providing useful information. Basic
comparisons of these nematodes can provide useful information but must be
interpreted carefully and, if possible, correlated to other tests (simulated field or
field tests). Although Petri plate tests are far removed from the reality of
reducing a pest population, meaningful data may still be extracted from these
tests. Establishment of a useful laboratory bioassay system will then aid in the
identification of promising nematode-host combinations.
Table 2-1. Concentration-mortality response of formicarius to entomopathogenic
nematodes at 48 h after exposure in Petri dishes.
Nematodea LC5o (95% FL)b LC9o (95% FL)b Slope + SEC
1.70 0.16 bc
1.86+ 0.21 bc
2.62+ 0.26 a
2.59+ 0.23 a
2.15 1.03 abCd
1.66 0.23 bc
0.73 0.15 d
2.13 0.16 ab
an = 1160 for each dosage-mortality response.
bLCso and LCgo expressed as infective juveniles per insect.
CParameter estimates followed by different letters indicate significant differences
(P=0.05) by the general linear test (Neter & Wasserman 1974)
dNE, not estimable. Index of significance (g) for potency estimation was high
(g, 0.89 for estimating 95% FL). Probit analysis provided a poor fit to the data.
Table 2-2. Parameter estimates, significance levels, and coefficients of
determination (M2) for percentage mortality of C. formicarius larvae in Petri
dish bioassays regressed over time
Nematode Intercepta Slopea F df P r2
0.23 de 0.0050 de
aParameter estimates followed by different letters indicate significant
differences (P<0.05) by the general linear test (Neter & Wasserman, 1974).
bSignificant level of regression; ns, P >0.05; *, P<0.05; **, P<0.001; *,
100 Steinernema dit
so DOSE 2
so b bb
20 od d ed
AGRIO ALL BRET ITAL MEX FELT INTER HP88 NC FL
Fig. 2-1. Mean percentage mortality of C. formicarius pupae 96 h after
inoculation by nematode species/strains in a Petri plate bioassay. Dose
1 = larval LDso rate for each nematode; dose 2 = 2x the larval LD50 rate.
AGRIO = Agriotos strain, ALL = All strain, BRET = Breton strain, ITAL =
Italian strain, MEX = Mexican strain, S. carpocapsae; FELTI = S. feltiae,
N-27 strain; HP88 = HP88 strain, NC = North Carolina strain, K,
bacteriophora; FL = Heterorhabditis sp. FL2122 strain.
I DOust 2 I
b b b
AGRIO ALL BRET ITAL
Fig. 2-2. Mean percentage mortality of C. formicarius adults 96 h after
inoculation by nematode species/strain in a Petri plate bioassay. Dose 1
= larval LDso rate for each nematode; dose 2 = 2x the larval LDso rate.
AGRIO = Agriotos strain, ALL = All strain, BRET = Breton strain, ITAL =
Italian strain, MEX = Mexican strain, ,. carpocapsae; FELTI = S. feltiae, N-
27 strain; HP88 = HP88 strain, NC = North Carolina strain, H.
bacteriophora; FL = Heterorhabditis sp. FL2122 strain.
MEX FELT INTER HP88 NC FL
Fig. 2-3. Relationship between mean percentage mortality of late instar C.
formicarius and nematode species/strains for 24 h periods for 96 h in a Petri
plate bioassay. Nematodes were applied at the larval LDso rate for each
nematode. AGRIO = Agriotos strain, ALL = All strain, BRET = Breton strain,
ITAL = Italian strain, MEX = Mexican strain, S carpocasae; FELTI = S. flta
N-27 strain; HP88 = HP88 strain, NC = North Carolina strain, H.
bacteriophora; FL = Heterorhabditis sp. FL2122 strain.
ALL BRET ITAL MEX FELTIIN
HP88 NC FL
0 I_ c
000 Steiernesma Heterorhabditis
C od d
AGRIO ALL BRET ITAL MEX FELTI INTER HP88 NC FL
Fig. 2-4. Relationship between the number of infective juveniles produced per C
formicarius larval cadaver and nematode species/strain; 2 wk post-inoculation.
AGRIO = Agriotos strain, ALL = All strain, BRET = Breton strain, ITAL = Italian strain,
MEX = Mexican strain, S, carpocapsae; FELTI = S. feltiae, N-27 strain; HP88 = HP88
strain, NC = North Carolina strain, H. bacteriophora; FL = Heterorhabditis sp. FL2122
INFECTIVITY OF FIVE ENTOMOPATHOGENIC NEMATODES TO THE
SWEETPOTATO WEEVIL IN THREE DIFFERENT BIOASSAYS
Nematodes in the families Steinernematidae and Heterorhabditidae are
promising biological control agents of various insect pests (Poinar 1979). The
ability of infective juveniles to penetrate into a host is essential to the infection
process. The infective juveniles penetrate an insect host through natural
openings (mouth, anus, and spiracles) and release a symbiotic bacterium which
ultimately kills the insect from septicemia. Heterorhabditids also possess a
large terminal tooth that may be used to penetrate soft intersegmental areas of
the cuticle (Bedding & Molyneux 1982).
The sweetpotato weevil, CyJl formicarius (F.) (Coleoptera: Apionidae),
is the most destructive insect pest of sweet potato, Ipomoea batata (L.) Lam.
(Schalk & Jones 1985, Sutherland 1986, Jansson & Raman 1991). The weevil
spends most of its life cycle within the swollen roots and vines of the sweet
potato plant thereby making management of this pest difficult (Jansson et al.
Entomopathogenic nematodes have been shown to be pathogenic
toward the sweetpotato weevil in the laboratory and in the field (Jansson et al.
1990b, 1992, Chapter 2). These nematodes are capable of seeking out and
killing the well-concealed hosts within the swollen roots. Although these
nematodes can penetrate, kill and reproduce in sweetpotato weevil larvae and
pupae and to a small extent in adults, a quantitative measure of infection has
not been determined.
The objective of this study was to quantify nematode infection by
counting the number of nematodes establishing in sweetpotato weevil larvae in
three test arenas and to examine the usefulness of various laboratory bioassays
for screening nematodes.
Materials and Methods
Insect and nematode rearing. Sweetpotato weevils were reared in the
laboratory at room temperature on 'Jewel' sweet potato storage roots (Coffelt et
al. 1978). Infested roots were dissected for late instar weevils prior to each
Nematodes included in these studies were: Steinernema carpocapsae
(Weiser) [= Neoaplectana carpocapsae, Steinernema feltiaen All and Mexican
strains; Steinernema feltiae (Filipjev) [= Steinernema bibionis] N-27 strain;
Heterorhabditis bacteriophora Poinar [= Heterorhabditis heliothidis] HP88
strain; and Heterorhabditis sp. FL2122 strain (an undescribed nematode
isolated in Florida by K. B. Nguyen). Nematodes were cultured in vive in late
instar wax moth, Galleria mellonella (L.) (Woodring & Kaya 1988). The wax
moth larvae were purchased from a commercial supplier (JA-DA Bait, Antigo,
Wisconsin) and stored at 50 C until use. Heterorhabditid and steinernematid
nematodes were stored in deionized water at 260 C and 70 C, respectively.
Dilutions of nematode suspensions in deionized water were prepared as
necessary to obtain the appropriate concentration. A mean of four subsamples
(100 ul) from each suspension was used to estimate nematode density.
Infective juveniles were less than 2 wk old before use.
Infectivity bioassays. Infectivity was measured for each of 5 nematode
concentrations in three test arenas for each nematode. The first test arena
consisted of Petri plates (60 x 15 mm) with 2 pieces of filter paper (Fisher P5,
medium grade), moistened with distilled water (1 ml). The second and third
arenas were conducted in plastic vials (30 ml) filled with coarse construction
sand (4.3% clay, 92.7 % sand, 3.0% silt; sieved through No. 10 mesh [2-mm]
sieve) with 5% moisture or Krome very gravelly loam soil (26.8% day, 49.4%
sand, 23.8% silt) with 4 to 5% moisture. The soil was sieved through a No. 10
(2 mm) sieve to remove large rocks. The surface area of the vial was 4.9 cm.
A nematode suspension (100 ul) containing 10, 25, 50, 75 or 100
infective juveniles was added to the bottom of each vial and covered with soil or
sand. A screen (1 mm mesh) was placed on the soil or sand surface to prevent
the larvae from moving throughout the vial. One weevil larva was placed on the
screen in each vial. The vials were capped, inverted, and incubated at 260 C
for 16 d. The same concentrations were added to the moist filter paper in the
Petri plates. One weevil larva was placed in the center of each plate after
nematode application, covered with a lid, and then incubated as previously
Larvae were removed and replaced from all vials and Petri plates at 2 d
intervals. Dead larvae were held for an additional 24 to 48 h to allow
established nematodes to develop to adults to facilitate counting. All dead
larvae were dissected and the numbers of nematodes within each cadaver
Two trials were conducted. Each nematode concentration was replicated
five times per trial per test arena.
Statistical analysis. The cumulative number of infective juveniles
penetrating a larva 16 d after nematodes were applied was regressed on
nematode dose. Parameter estimates were compared using the general linear
test (Neter & Wasserman 1974). Cumulative means and standard errors were
computed for the number of infective juveniles in weevil cadavers on each
sample period (2 d intervals for 16 d). The cumulative mean number of
nematodes within a cadaver 16 d after nematode application for each
nematode-bioassay combination was subjected to analysis of variance and
means were separated by Tukey's Studentized Range Test (Tukey 1953) if a
significant F value (P<0.05) was found.
The number of infective juveniles of Heterorhabditis spp. FL2122 strain
penetrating larvae increased significantly with nematode dose in all three test
arenas in both trials (Table 3-1). Heterorhabditis bacteriophora HP88 strain
closely followed the trend of FL2122 with a significant increase in infective
juveniles penetrating larvae. in all three test arenas in the first trial and in the
sand and soil arenas in the second trial. The results were more variable for the
steinernematid nematodes. Numbers of infective juveniles of Steinernema
feltiae N-27 strain infecting weevils significantly increased with dose in the sand
arena in both trials. Numbers of S. carpocapsae All strain infective juveniles
penetrating larvae increased significantly in the Petri plate arena in both trials
and in the soil arena in the first trial. The slope in the soil arena in the second
trial was significant at P<0.10. The number of infective juveniles of Mexican
strain that penetrated larvae did not increase significantly (P>0.05) with
nematode dose in all arenas in the first trial. The number of infective juveniles
penetrating larvae increased significantly (P<0.05) with nematode dose applied
in all three arenas in the second trial.
The number of infective juveniles penetrating larvae was affected by the
test arena used. Steinemema carpocapsae Mexican strain performed better
(steeper slope = more nematodes penetrating per nematode dose applied) in
the Petri plate arena compared with the soil or sand arenas in both trials (Table
3-1). Mexican performed similarly in the soil and sand arenas. Steinernema
carpocasae All strain also performed better (steeper slope) in the Petri plate
arena compared to the sand arena, but was not significantly different from the
soil arena. The number of infective juveniles of S. feltiae penetrating larvae per
nematode applied was significantly greater in the sand arena than in the Petri
plate arena in both trials. Steinernema carpoasae strains had higher rates of
infection (steeper slope) than S. feltiae in the Petri plate arena in both trials.
The opposite was true in the sand arena, where the rate of infection was greater
for S. feltiae compared to the S. carpocapsae strains. The rate of infection for H.
bacteriophora HP88 strain was similar among all test arenas in the first trial but
was significantly greater in the soil and sand arenas compared to the Petri plate
arena in the second trial. The greatest rate of infection of the FL2122 strain was
in the sand arena with no difference between the sand and soil arenas in the
first trial. In the second trial, the greatest rate of infection by FL2122 strain was
also in the sand arena followed by the Petri plate and the soil arenas,
Significantly more H. bacteriophora FL2122 strain penetrated larvae per
dose than the steinernematid nematodes tested in all arenas in both trials
(Table 3-1). The rate of infection for FL2122 was significantly greater than the
other heterorhabditid nematode, HP88, in the sand arena only in the first trial
and in all three arenas in the second trial.
In a comparison of the rate of infection of the steinernematid nematodes
in the Petri plate arena (Table 3-1), All and Mexican were significantly more
infective (t>1.677; df=48; E<0.05) than S. feltiae in both trials. In the sand
arena, S. feltiae was significantly more infective (t>1.677; df=48; P<0.05) than
All and Mexican in both trials. All strain was significantly more infective
(t>1.677; df=48; P<0.05) than Mexican and S. fealia in the soil arena in the first
trial. In the second trial, there were no differences in rate of infection among the
steinernematid nematodes in the soil arena.
The number of FL2122 infective juveniles penetrating larvae increased
for 6-8 d after nematodes were applied (Fig. 3-1). Numbers of nematodes
infecting weevils tended not to differ between the two lower nematode doses
(10 and 25) nor between the two intermediate doses (50 and 75). The number
of nematodes infecting weevils tended to be similar in all arenas at the lower
doses. Differences in infectivity among test arenas were most evident at the
highest dose (100) and were greatest in the sand arena (both trials) and the
Petri plate arena in the second trial.
The number of HP88 infective juveniles penetrating larvae tended to be
similar in all test arenas (Fig. 3-2). The highest number of infective juveniles
infecting larvae was achieved at the highest dose (100) in all arenas. Fewer
HP88 infective juveniles penetrated larvae than those of FL2122, especially at
the higher doses (50, 75 and 100) and in the sand arena.
Few S. feltiae infective juveniles penetrated larvae regardless of dose
with the exception of the sand arena in the second trial (Fig. 3-3). Infectivity
tended not to differ among doses with the exception of the sand arena in the
second trial. Infectivity by this nematode was lower than by both heterorhabditid
nematodes in all test arenas with the exception of the sand arena in the second
trial when S. feltiae tended to behave similarly to HP88 in the same arena.
The infectivity of both S. carpocapsae strains (Figs. 3-4 and 3-5) was low
in all test arenas compared with those of the heterorhabditid nematodes. Both
of these nematodes appeared to be more infective in the Petri plate arena
compared with the sand and soil arenas. Infectivity tended not to differ among
doses in the soil and sand arenas for these nematodes. Infectivity by these
nematodes was generally lower than by both heterorhabditid nematodes in all
The cumulative mean percentage of nematodes infecting weevils did not
differ among nematodes in the Petri plate arena (Trial 1: E=1.6; df=4, 20;
P=0.21; Trial 2: F=1.4; df=4, 20; E=0.25) (Fig.3-6). The number of nematodes
penetrating weevils ranged from 1 to 26% in the first trial and 7 to 20% in the
second trial in the Petri plate arena. Differences were more apparent in the
sand and soil arenas (Fig. 3-6). Significantly higher percentages of FL2122
nematodes infected larvae than most other nematodes in the sand and soil
arenas in both trials (Trial 1: sand, F=65.9 df=4, 20; P=0.0001; Trial 2: sand,
F=101.6; df=4, 20; P=0.0001; Trial 1:soil, F=36.0; df=4, 20; E=0.0001; Trial 2:
soil, F=14.3; df=4, 20; P=0.0001). The number of nematodes infecting larvae
ranged from 0.2 to 28% and 4 to 20% in sand and soil, respectively, in the first
trial, and from 2 to 53% and 6 to 28% in sand and soil, respectively, in the
Fan and Hominick (1991) found a strong linear relationship between the
mean number of infective juveniles established in G. mellonella larvae and
dose applied in a sand arena similar to that used in this study. They found
Steinernema sp. Nashes strain to be more effective than two Heterorhabditis
spp. (UK and Dutch strains). Slopes for these nematodes were 0.35, 0.30 and
0.23 for Steinernema sp. Nashes strain and the two Heterorhabditis spp. (UK
and Dutch), respectively. Similar dose-response tests were conducted by
different workers in the same research group using different nematodes and
conditions (Fan and Hominick 1991). They also found a high correlation
between the mean number of nematodes that established in the insect hosts
and dose, although protocols varied greatly. It was, therefore, suggested that
the dose-response test is a robust test and that different establishment rates of
nematodes (slope of the line) could be used to compare the efficacy of the
In the current study, however, the correlation between dose and the
number of nematodes that established in sweetpotato weevil larvae was not as
strong (r2< 0.84) as in the dose-response studies mentioned above (r2>0.95).
The correlation between the number of nematodes established and dose was
strongest in the sand arena for both heterorhabditid nematodes and S. feltiae
compared to the Petri plate and soil arenas. The correlation was generally
weak for both S. carpocapsae strains in all three arenas (r2<0.51).
A major difference between Fan and Hominick's (1991) studies and the
current study is in the application methods. In Fan and Hominick's (1991)
bioassay, the vials were shaken to distribute the nematodes throughout the
sand environment. Therefore, those nematodes less capable of movement
and/or host-seeking might be expected to infect larvae to a higher degree
because contact between the nematode and the host is less reliant on
nematode movement and host-seeking. Since several authors have shown that
the majority of S. carpocapsae infective juveniles remain near the point of
placement (Georgis & Poinar 1983a, b, Moyle & Kaya 1981, Schroeder &
Beavers 1987), it is likely that these nematodes would be greatly affected by this
method of application. The ability of nematodes to move and find hosts in the
sand and soil environments was an integral part of the comparisons among
nematodes and test arenas in the present study. Therefore, the nematodes and
larva were placed at opposite ends of the container. Including movement
and/or host-seeking into the bioassay may attribute to the weaker correlation
between nematode establishment and nematode dose.
Another major difference among the studies was the host insect. The
previous studies used G. mellonella and in the current study, Q. formicarius was
used. Galleria mellonella larvae are highly susceptible to infection by
nematodes. In several tests, G. mellonella was found to be the most susceptible
host to these nematodes (Bedding et al. 1983, Molyneux et al 1983, Morris et al.
1990). Thus, the use of G. mellonella may bias results toward increased
infection. These larvae are typically used because they are susceptible and
also because they are easily reared or obtainable. It is questionable whether
comparisons of infectivity among different nematodes against G. mellonella
larvae would be the same or similar to comparisons with a different insect host.
Jansson (unpublished) found that infectivity of H. bacteriophora HP88 strain
and Heterorhabditis sp. FL2122 strain did not differ between G. mellonella and
C. formicarius; however, infectivity of Heterorhabditis sp. Bacardis strain did
differ between the two hosts. Galleria mellonella are useful for standardizing
comparisons of nematodes among different laboratories or for measuring
nematode quality, but probably are not useful for comparing nematode species
or strains for potential control of another target insect.
Using the slope of the regression as an indication of establishment rate
may be a useful tool as long as the protocols are similar. Significant differences
in slopes were found among the three test arenas for all nematodes tested.
Mracek (1982) estimated that 10-20% of a dose of S. carpocapsae applied to
0.3 dm3 of soil was recovered in 20 G. mellonella larvae placed in the soil as
bait. Bednarek and Nowicki (1986) exposed G. mellonella larvae to S. feltiae in
a Petri plate environment and never recovered more that 42% of the dose
applied from the cadavers. In the current study, the recovery of nematodes
varied among nematode species/strains, test arena, and nematode dose. The
highest recovery was obtained with FL2122 in the sand arena (35-46%).
Although determining the number of nematodes that establish in a host
can be informative and allow comparisons among nematodes, it is a very labor
intensive due to the host dissections. It may be too time consuming to justify its
use as an initial screen for testing many nematodes simultaneously.
Petri plate bioassays have been used extensively in the past as a way to
screen the relative infectivities of different species and strains of nematodes. It
was suggested that these bioassays are not recommended because this
approach is far removed from the natural situation (Bedding 1990). Such
bioassays presumably favor nematodes that nictate. The present study
substantiates the above recommendations to some extent in that S.
carpocapsae nematodes, which have been shown to nictate (Ishibashi & Kondo
1990), generally performed better in the Petri plate arena compared with a sand
or soil arena.
It is interesting to note that the results of the present study are generally
corroborated by earlier studies, many of which were conducted in a standard
Petri plate bioassay (Jansson et al. 1990b, Chapter 2). Heterorhabditid
nematodes were more pathogenic than steinernematids to sweetpotato weevil
based on dose-mortality curves, host stage susceptibility and reproductive
potential. As a result of these Petri plate bioassays, FL2122 was found to
perform better than all other nematodes tested (Chapter 2). Although the
present study quantifies infectivity (the number of nematodes establishing within
a host), the end result concurs with our previous findings. Heterorhabditis sp.
FL2122 strain consistently performed better than all other nematodes.
It is clear that a test arena used in a bioassay can affect the outcome of a
study. The infectivity of the S carpocasae strains was generally greater in a
Petri plate environment compared to soil and sand environments, however, the
infectivity of the heterorhabditid nematodes was greatest in a sand environment.
Arguments made against the Petri plate arena could also be made against a
sand environment. For example, sand, in a small container, is also far removed
from a natural situation and one group of nematodes may be selected for over
another. The best arena might be to use a medium (e.g., non-sterile soil) which
more closely resembles the natural situation. This might be especially
important with cryptic insects when nematodes have to move and penetrate
more than one environment. However, bioassays such as these become
inherently more laborious and difficult to handle.
These results demonstrate the importance of selecting a bioassay
carefully, and also that more than one type of bioassay probably should be
considered. Additionally, defining the purpose of the bioassay and determining
the parameters of the bioassay based on the purpose are essential. Petri plate
bioassays (dose-mortality) can be useful, especially for initial screening tests,
because they are relatively easy, inexpensive, and reliable at identifying
potentially effective nematodes. However, additional bioassays should be
conducted so that more definitive conclusions can then be made regarding the
potential of a candidate nematode.
Table 3-1. Parameter estimates, significance levels, and coefficients of determination
(r2) for the cumulative number infective juveniles penetrating C. formicarius larvae in
three bioassay arenas regressed on nematode dose.
Species/Strain arena Intercept Slopea F df pb r2
-1.48 0.07 a BC 23.1 1,23
Species/Strain arena Intercept Slopes F df pb r2
N-27 Petri plate
a Lowercase letters correspond to comparisons among bioassay arenas within
nematodes. Uppercase letters correspond to comparisons among nematodes within a
bioassay arena. Parameter estimates followed by different letters indicate significant
differences by the general linear test (Neter & Wasserman 1974).
b Significance level of regression: ns, P>0.05; *, P<0.05; **, P<0.001; *,
Fig. 3-1. Cumulative mean number of Heterorhabditis sp. FL2122 strain
infective juveniles (+ SEM) established in individual C. formicarius larvae
over time in a petri plate, sand and soil arena.
0 5 10
---- 10 NEMATODES
-- 50 NEMATODES
Fig. 3-2. Cumulative mean number of H. bacteriophora HP88 strain infective
juveniles (+ SEM) established in individual C. formicarius larvae over time in
a petri plate, sand and soil arena.
L o I- glwl-..-a
- 10 NEMATODES
S 50 NEMATODES
--- 75 NEMATODES
Fig. 3-3. Cumulative mean number of S. feltiae N-27 strain infective juveniles
(+ SEM) established in individual C. formicarius larvae over time in a petri
plate, sand and soil arena.
J PETRI PLATE
.1 p 3 U r i I i I I
... p p p p -
-- 25 NEMATODES
---- 75 NEMATODES
Fig. 3-4. Cumulative mean number of S. carpocapsae Mexican strain
infective juveniles (+ SEM) established in individual C. formicarius larvae
over time in a petri plate, sand and soil arena.
0 5 10 15
---- 10 NEMATODES
S 25 NEMATODES
Fig. 3-5. Cumulative mean number of carpocapsae All strain Infective
juveniles ( SEM) established in individual C. formicarius larvae over
time in a petri plate, sand and soil arena.
PETRI PLATE PETRI PLATE
5 10 15
-- 10 NEMATODES
- 25 NEMATODES
- 50 NEMATODES
- 75 NEMATODES
- 100 NEMATODES
Fig. 3-6. Cumulative mean percentage of infective juveniles ( SEM)
established in individual C. formicarius larvae in a petri plate, sand, and
soil arena. Lower case letters correspond to comparisons within Trial 1
and upper case letters correspond to comparisons within Trial 2. Different
letters indicate significant differences among nematodes (P=0.05; Tukey's
studentized range test [Tukey 1953]). Note: figures without letters
indicate no significant differences among treatments.
* TRIAL I
* TRIAL II
i _i R
ALL MEX FELT HP88 FL2122
WITHIN-ROOT MORTALITY OF THE SWEETPOTATO WEEVIL
BY ENTOMOPATHOGENIC NEMATODES
The sweetpotato weevil, Clas formicarius (F.), a serious pest of sweet
potato, spends most of its life cycle within the swollen storage roots and vines of
sweet potato, Ipomoea batatas (L.) Lam. Weevils complete development in
approximately 33 d at 27-30o C (Mullen 1981). Females oviposit eggs singly in
vines or swollen roots. Weevils remain within the root or vines from egg hatch
until adult emergence. The newly emerged adults remain within the larval
tunnel for a short period of time. A minimum of four days is required for adults to
excavate a passageway to the surface of the root or vine and emerge
(Cockerham et al. 1954).
Because of its cryptic life cycle, this pest can be difficult to control with
conventional insecticide applications (Jansson et al. 1990a). The use of
chemical insecticides may be less effective due to the concealed habit of this
pest. Once weevils are present within the crown or swollen roots, control from
spraying, dusting, or fumigation with currently available chemical insecticides is
difficult (Talekar 1991). Control achieved with foliar applications of chemical
insecticides is primarily due to mortality of adult weevils. Frequent applications
of insecticides are necessary for this type of control (Talekar 1991). Because of
problems associated with pesticide residues in food, environmental
contamination, and phytotoxicity, biological approaches to weevil management
would be of value. Additionally, weevil management by chemical insecticides
may be too costly, especially in developing countries, where more than 95% of
all sweet potatoes are grown (Horton 1988).
One biological approach to weevil management is the use of
entomopathogenic nematodes. Nematodes in the families Steinerematidae
and Heterorhabditidae have been shown to be infective against several
weevils, including the sweetpotao weevil (Bedding & Maer 1981., Bei &
Boivin 1985, Georgis & Poinar 1984a, 1984b, Rutherford at al. 1987, Schroeder
1987, Figueroa & Roman 1990, Jansson et al. 1990b, Chaptr 2). These
nematodes have potential for managing this weevil because they are highly
virulent (low LDso values) against sweetpotato larvae and pupae, kill weevils
within 48 h, and are capable of reproducing in weevil cadavers (Chapter 2).
Jansson et al. (1990b) also demonstrated that these nematodes are capable of
seeking out and killing sweetpotato weevil larvae and pupae within the storage
The potential of using entomopathogenic nematodes for control of
sweetpotato weevil relies on many factors including host susceptibility to
nematode infection, contact between the host and nematode, nematode
movement, and nematode persistence. We know that these nematodes can
infect and kill weevil larvae, pupae, and adults efficaciously in Petri plates
(Chapter 2). The ability and efficacy of these nematodes at seeking, finding,
and infecting weevils within a concealed environment, such as the storage
roots, in not well understood. The present studies were conducted to compare
the efficacy of several entomopathogenic nematodes at seeking and killing
sweetpotato weevils within the storage roots.
Materials and Methods
Nematodes used in this study were cultured in vlH in late intar wax
moth, Galleria mellonel L(Woodring & Kaya 1988). The wax moth larvae
were purchased from a commercial supplier (JA-DA Bait, Antigo, Wisconsin)
and stored at 50 C until use. The nematodes included in these studies were:
Steinernema carpocapa (Weiser) [= Neoaplectana a
Steienema teltiae Al and Mexican strains; Steinernam f~lti (Filpjev) [=
Stelnemema bbionisb N-27 strain; Steinernma gasrul (Steine);
Heterorhabditis bacterihora Poinar [= Heterorhabditi helothidia HP88
strain; and Heterorhabdits sp. FL2122 strain (an undescribed nematode
isolated in Florida). Heterorhabditid and steinernematid nematodes were
stored at 26o C and 70 C, respectively. Dilutions of nematode suspensions with
deionized water were prepared as previously described for experimentation
(Chapter 2). Nematode density was estimated from the mean of four
subsamples (100 ul) from each nematode suspension. Infective juveniles were
less than 2 wk old before use.
Sweetpotato weevils were reared at room temperature on 'Jewel' sweet
potato storage roots (Coffelt et al. 1978). Although the roots were not graded for
size, extreme sizes (< 50 g and > 300 g) were not used. All roots were washed
and dipped in 1% Clorox before use. Roots were exposed to 200 adult weevils
(1:1; female:male) in plastic boxes (18x13x9 cm) with screened lids for 3 or 4 d
unless otherwise stated. The weevils then were removed and the roots stored
at 24 + 20 C in plastic boxes (26x19x9 cm) for 2 to 3 wk; the approximate time
needed for weevils to develop to second or third instars and pupae (Gonzalez
1925, Sutherland 1986a). Infested roots then were buried individually in coarse
construction sand (4.3% clay, 92.7 % sand, 3.0% silt; passed through No. 10
mesh [2-mm] sieve) with 5 percent moisture in plastic cups (1 liter). Nematodes
were applied to the sand surface (86.5 cm2) at the equivalent level of 2.5 to 7.5
billion infective juveniles per ha (25 to 74 infective juveniles per cm2).
Weevil denty experiments. Storage roots were exposed to four levels
of adult weevils (20,100,200 and 400; 1:1 sex ratio) in order to determine
weevil density within storage roots exposed to known numbers of weevils in the
laboratory. Forty-eight storage roots (12 roots per weevil density) were held In
empty cups (1 liter) until adults emerged. The numbers of adults emerging from
each root were recorded. This procedure was then repeated using 9 roots for
each level of weevils.
Weevil density was also determined from naturally-infested roots
collected from the field. Six storage roots with severe weevil damage were
collected from a research plot at the Tropical Research and Education Center,
Homestead, FL, and placed in empty cups until adults emerged. The numbers
of weevils that emerged were recorded.
Another 48 storage roots were exposed to the four levels of adult weevils
(12 roots per level) were held for 2 to 3 wk, and then buried In moist sand as
described above. Steinernema carpocasae Mexican strain was applied (25
infective juveniles per cm2) by pouring the suspension (5 ml per cup) over the
sand surface. This density was equivalent to approximately 2.5 billion infective
juveniles per ha. The roots were removed from the sand 3 wk after nematode
application and dissected carefully. The numbers of live and nematode-
infected weevils were recorded within each root. This experiment was repeated
using H. bacteriophora HP88 strain. Roots used in the experiment with HP88
were weighed to estimate fresh root weight.
Field- versus laboratory-infested root experiments. To compare weevil
mortality in laboratory-infested and field-infested storage roots, roots with
external weevil damage were collected from the field and compared to roots
exposed to adult weevils in the laboratory. All roots were buried in moist sand
and exposed to S. carocapsae Mexican strain at two densities, 12 and 35
infective juveniles per cm2 (equivalent rates of 1.25 billon per ha and 3.75
billion per ha, respectively). Roots treated with distilled water of equal volume
served as a control. Eight roots were used for each nematode density per root
type and the control. Roots were dissected at 2 and 3 wk afer applaeloan and
the numbers of live and nematode-infected weevils were recorded.
Weevil mortaty exprimnts. Roots infested wih tI weel In the
laboratory were buried in moist sand and then treated with one of four
nematodes: S. carpocapsae All and Mexican strains, I. bacterdohora HP88
strain, and Heterorhabditis sp. FL2122 strain. Infective juveniles were applied
(25 per cm2) as previously described to each of 12 buried roots per nematode.
Four roots from each nematode treatment were dissected at 1, 2, and 3 wk after
application and numbers of healthy and infected weevils in each stage and their
location within roots were recorded. This experiment was repeated in a second
trial with the above mentioned nematodes plus S. fella N-27 strain and S.
glaser. Six roots for each nematode treatment were dissected at 2 and 3 wk
Data analysis. The means and standard errors were calculated for adult
weevil emergence from laboratory- and field-infested storage roots that were
not buried. All data were transformed to the arcsine of the square root. Data
from roots that were exposed to different levels of adult weevils and then treated
with nematodes in a sand environment were analyzed by least squares
regression analysis (Montgomery 1976). Weevil mortality from nematode
infection was regressed on within-root weevil density and on storage root fresh
weight. Weevil density per root was also regressed on root fresh weight.
Mortality data between laboratory- and field-infested roots were
subjected to unpaired t-tests. Weevil mortality within laboratory-Infested roots
and field-infested roots were subjected to analysis of variance and means were
separated by Tukey's Studentized Range Test (Tukey 1953) If a significant E
value (P<0.05) was found. Total weevil mortality was regressed on time (1 to 3
wk) in the first trial of the mortality experiments. Parameter estimates were
compared using the general linear test (Neter & Wassermn 1974). Weevil
mortality data (by weevil stage and location) were subjected to analysis of
variance and means were separated by Tukey's Studentized Range Test
(Tukey 1953) If a significant E value (P<0.05) was found. Data on the location of
weevil cadavers within the root (inner portion versus outer portion) were
subjected to unpaired t-test within each nematode treatment.
Weevil density experiments. The mean numbers of adults emerging
from roots were 8017, 25833, 50545, and 26150, for roots exposed to 20,
50, 100 and 200 adult weevils respectively, in the first experiment and 8830,
31845, 41158, and 34495, respectively, in the second experiment. The
mean number of weevils collected from field-infested storage roots was 41+11.
No relationship existed between weevil mortality from S. carpocapsae
Mexican strain (E=0.2; df=1, 46; P>0.05) or H. bacteriophora HP88 strain (E=0.1;
df=1, 46; P>0.05) and within-root density of weevils by either. Also, no
relationship existed between weevil density within roots and root fresh weight
(E=1.3; df=l, 46; P>0.05), nor between the level of infection by nematodes and
root fresh weight (E=1.3 df=1, 46; P>0.05).
Field- versus laboratory-infested storage root experiments. Significantly
higher levels of mortality by S. carpocapsae Mexican strain were achieved in
roots infested in the laboratory than in those Infested in the field for both
nematode concentrations (12 Infective juveniles per cm2: t=3.1, df=14, E0.008;
35 infective juveniles per cm2: t=2.8, df=14, E=0.014). Weevil mortality in
laboratory-infested roots (17.6%) did not differ from that in fild-infested roots
(9.6%) in nontreated roots 0=1.519; df=14; P=0.1510).
Higher levels of mortality were achieved in roots treated wMlh rnemodes
than in nontreated roots for both laboratory-infested and eld-infested roots
(laboratory roots: E=6.4; df-2, 21; E=0.007; field roots: E=6.8; df,2, 21; E-0.005)
Weevil mortality experiments. In the first experiment, total weevil
mortality in roots treated with S carocsae All and Mexican strains increased
significantly over time, but did not increase in those treated with heterorhabditid
nematodes (Table 4-1). Slopes did not differ (P>0.05) between HP88 and
FL2122 nor between Mexican and FL2122. The slope for All strain was
significantly steeper than those of all other nematodes.
In the first trial, total weevil mortality among nematodes did not differ
significantly at 1, 2 or 3 wk (wk 1: E=1.3; df=3, 28; E>0.05; wk 2: E=0.4; df=3, 28;
P>0.05; wk 3: E=2.3; df=3, 28; P>0.05) (Table 4-2). In trial 2, weevil mortality did
not differ among nematodes tested on wk 2 (E=2.2; df=5, 42; E>0.05); however,
on wk 3, Heterorhabditis sp. FL2122 strain caused significantly more mortality
than all other nematodes (E=5.3; df=3, 28; E=0.0008) (Table 4-2). Overall,
levels of mortality were lower in trial 2 than in trial 1 for all nematodes tested. It
is unknown why data from the two trials differed.
Larval mortality followed the same trend as total mortality in both trials.
Larval mortality due to nematodes did not differ on any sample period in trial
1 (wk 1: E=0.6; df=3, 28; P>0.05; wk 2: F=0.1; df=3, 28; P>0.05; wk 3: E=2.5; df=3,
28; P>0.05) (Table 4-22). The only significant difference in larval mortality was
on wk 3 in trial 2, when Heterorhabditis sp. FL2122 strain caused 3- to 14-fold
more mortality than al other nematodes tested, but was only algnificanly
different from that caused by S. fetiae (E=2.7; df=5, 42; P-0.0309).
Pupal mortality differed on wk 1 and 2 in trial 1 (wk 1: E=6.7; df=3, 28;
=-0.0015; wk 2: E=3.5; df=3, 28; E=0.0297) (Table 4-2). However, differences
in mortality primarily were due to the lack of pupae in many of the roos. In trial
2, Hterorhabdit sp. FL2122 strain caused sagnificanly moe mortbty than did
all other nematodes at wk 3 (E=4.3; df=5, 42; e=0.0030) (Table 4-2). No pupae
were found in roots treated with S. glaseri.
The numbers of weevil cadavers infected with nematodes in the inner
section of roots did not differ (P>0.05) from that in the outer section, regardless
of nematode treatment or experiment. In the first trial, weevil mortality did not
differ among nematode treatments in the inner section (wk 1: E=0.5; df=3, 28;
1>0.05; wk 2: E=2.2; df=3, 28; P>0.05; wk 3: E=0.1; df=3, 28; E>0.05), nor in the
outer section (wk 1: E=1.5; df=3, 28; P>0.05; wk 2: E=0.5; df=3, 28; P>0.05; wk 3:
E=2.2; df=3, 28; E>0.05). Similar results were obtained in the second trial, on
wk 2, (inner root section: E=1.6; df=5, 42; P>0.05; outer root section: E=2.1; df=5,
42; E>0.05). However, mortality differed among nematode treatments in both
the inner and outer root sections on wk 3 (inner root section: E=3.9; df=5, 42;
E=0.0051; outer root section: E=5.6; df=5, 42; P=0.0005) (Fig. 4-2). Higher
levels of mortality were achieved by Heterorhabditis sp. FL2122 strain than by
all other nematodes.
Laboratory tests conducted with artificially-infested roots can provide
useful information on within-root mortality by entomopathogenic nematodes.
Most of the experiments in this study were conducted using storage roots which
were artificially infested with sweetpotato weevils in the laboratory. Laboratory-
roots appeared more dehydrated than field-infested roots because laboratory-
infested roots were stored for 2-3 wk until use while field-infested roots were
used Immediately after collection from the field. Additionally, roots infested in
the laboratory were stored for an unknown period of time in the store, packing
house and in transit However, these roots were not dehydrated at the time of
purchase. Weevil density was usually greater in labrorWry-Inested roots
compared with field-infested roots; however, there was no relationship between
weevil density within the roots and the level of infection by nematodes. There
was considerable variability in the size (based on weight) of the storage roots
(from the field and the laboratory) used in these experiments. No relationship
was found between weevil density and root size (within the size range tested),
nor between weevil density and infection level by two nematodes, S.
carocapsae Mexican strain and H. bacteriohora HP88 strain.
Although there was significantly more mortality from nematode infection
in laboratory-infested roots than in field-infested roots, the trends for both were
the same. There were considerably higher levels of mortality in storage roots
treated with nematodes than in nontreated roots for both root types. It is unclear
why there was significantly more mortality by nematodes in the laboratory-
infested roots compared with field-infested roots. Nematodes probably enter
roots via the oviposition holes and the laboratory roots might have contained
more oviposition holes than field roots Also, the ovipostion holes in the
laboratory-infested roots were spaced more uniformly on the root surface
compared with patterns on the field-infested roots. Thus, more nematodes
might have entered laboratory-infested roots than field-infested roots.
Laboratory-infested roots might also have produced more C02 than field-
infested roots because of the higher density of weevils within the laboratory-
infested roots. CO2 has been shown to be attractive to infective juveniles
(Gaugler et al. 1980) which may help to account for the higher weevil mortality
within laboratory-infested roots.
All the nematodes tested were capable of entering infested roots and
killing weevils. The heterorhabditid nematodes, especially FL2122 strain,
appeared to gain access to and kill weevils faster than the steinernematid
nematodes. Both heterorhabditid nematodes killed most weevils within 1 wk.
The steinernematids killed fewer weevils in the first wk. The environment in
which these tests were conducted (sand) may bias the results in favor of the
heterorhabditid nematodes because heterorhabditid nematodes have been
shown to be more infectious in a sand environment compared to a soil or Petri
plate environment (Chapter 3). Jansson (1990b) found that H. bacteriophora
HP88 strain killed more weevils than did S. carpocapsae All strain in field-
infested storage roots buried in a Krome very gravelly loam soil for 3 wk in the
laboratory. Also, several field experiments showed that heterorhabditids are
more efficacious than steinernematids at controlling this weevil (Jansson et al.
1990b, 1992, unpublished).
When total weevil mortality was divided by weevil stage, the trends of
larval and pupal mortality were similar to that of the total mortality. Generally,
the heterorhabditid nematodes caused more larval and pupal mortality than
All of the nematodes apparently were able to disperse within roots once
inside because there were no differences in mortality between root sections for
any of the nematodes. Many of these roots were very porous on the inside, due
to dense weevil infestation which may have facilitated nematode movement.
Steinernema glaseri caused more mortality (although not significant)
than other steinernematid nematodes in most of the experiments. This might be
attributed to its ability to move more than other steinernematids (Georgis &
Poinar 1983; Schroeder & Beavers 1987).
Movement and/or host finding abilities are important aspects to the
success of entomopathogenic nematodes. The nematode that was most
efficacious at killing weevils in these studies, FL2122, may also be the most
capable of moving. A recent study showed that FL2122 strain moved more than
12.5 cm vertically in a sand environment (Chapter 6). Seeking hosts is only one
part of the infection process; however, if nematodes cannot find their host,
mortality will not be achieved, regardless of their virulence.
All nematodes tested were capable of entering weevil-infested roots and
killing weevils. As demonstrated in this study as well as others, heterorhabditid
nematodes appear to be more efficacious for management of the sweetpotato
weevil. The encouraging results from these tests demonstrate the potential for
using these as biological control agents of the sweetpotato weevil.
Table 4-1. Parameter estimates, significance levels, and coefficients of
determination (r2) for percentages of within-root mortality of C. formicarius
regressed on time (wk).
Species and strain Intercept Slopes F df P r2
All -0.058 1.970 a 5.22 1,22 0.03 0.19
Mexican -0.100 0.402 b 9.97 1,22 0.005 0.31
HP88 0.664 -0.037 c 0.07 1,22 0.80 0.00
FL2122 0.29 0.205 bc 1.19 1,22 0.29 0.05
aParameter estimates followed by different letters indicate significant
differences by the general linear test (Neter & Wasserman 1974).
Table 4-2. Mean mortality ( SEM) of C. formlcariu by nematode Infection within
the storage root buried in moist sand.
Sample % Mortally
Trial. SpeciesStrain (wk) Larvae Pupae Total
I S. carpocapsae
II S. carlocapsae
Sample % Mortality
Trial. Species/Strain (wk) Larvae Pupae Total
.II .lasri 2 4.0 1.8. 0.1 0.1 2.5 1.5
3 14.3 6.1 0.0 7.9 5.6
HP88 2 12.0 6.6 3.0 2.4 6.2 3.3
3 18.6 9.9 0.6 0.2 2.3+ 0.8
FL2122 2 15.7+ 7.9 7.3 6.1 11.3 6.2
3 45.6+ 16.5 21.6 +11.4 29.8 12.2
S80- FIELD ROOTS b b
a 40- 5
NONTREATED 12 IJs 35 IJs
Fig. 4-1. Mean percentages (and standard error bars) of within-
root mortality of C. formicarius in laboratory- and field-infested
roots 3 wk after application of S. carpocapsae Mexican strain (12
or 35 infective juveniles per cm2). Lower case letters correspond
to comparisons among treatments within laboratory-infested roots
and upper case letters correspond to comparisons among
treatments within field-infested roots. Different letters indicate
significant differences among treatments (P=0.05; Tukey's
studentized range test [Tukey 1953]).
INNER 1/2 RADIUS
b b b
ALL HEX N-27 6LAS. HP88FL2122
Fig. 4-2. Mean percentages (and standard error bars) of within-root
mortality of C. formicarius in the inner 1/2 radius and the outer 1/2 radius
of the root 3 wk after applicationof nematodes. Nematodes tested: S.
carpocapsae All and Mexican strains, S. feltiae N-27 strain, S. glaseri,
Hi bacteriophora HP88 strain and Heterorhabditis. sp. FL2122 strain.
Bars with the same letter are not significantly different (P=0.05; Tukey's
studentized range test [Tukey 1953]).
L b b7
ALL HlEX N-27 6LAS.
MOVEMENT AND POST-INFECTION EMERGENCE OF
Entomopathogenic nematodes in the families Stenernematldae and
Heterorhabditidae have been shown to be infective against many insects
(Gaugler & Kaya 1990). The most consistent efficacious results with these
nematodes have been obtained against insects that inhabit cryptic habitats
(Begley 1990). Cryptic habitats can be characterized as habitats with adequate
moisture (high relative humidity), mild temperatures and no ultraviolet radiation,
all of which enhance nematode survival and infectivity (Begley 1990 and
references therein). The soil environment is the natural reservoir for
entomopathogenic nematodes and 90% of insect pests spend part of their life
cycle in the soil (Klein 1990).
The sweetpotato weevil, Cyla formicarius (F.), is a serious pest of sweet
potato, Ipomoea batatas (L) Lam. (Sutherland 1986, Chalfant et al. 1990,
Jansson & Raman 1991). The weevil spends most of its life cycle within the
storage roots and vines of the sweet potato plant making management of this
pest difficult. Jansson et al. (1990a) showed that approximately 80-90% of the
weevil population was found in plant parts at or below the soil surface. Once
weevils are present within the crown or storage roots, control with currently
available chemical insecticides is poor (Talekar 1991).
Several studies have demonstrated the potential of entomopathogenic
nematodes for managing the sweetpotato weevil (Jansson et al. 1990b, 1992).
In order to control the sweetpotato weevil successfully, the nematodes must
move through the soil environment, locate and penetrate weevil-infested roots,
and locate and infect weevils within the root. In previous studies,
entomopathogenic nematodes were not only shown to be infective against the
sweetpotato weevil in Petri plate bioassays (Chapter 2), but also capable of
killing weevils within roots in laboratory bioassays (Chapter 4) and in the field
(Jansson et al. 1990b). After infection, nematode development and
reproduction occurs inside the insect cadaver. The infective juveniles exit the
cadaver when resources within the cadaver become depleted, seeking new
hosts (Poinar 1979). Nematodes produced in weevil cadavers within storage
roots must not only exit the cadavers but also move within the root in order to
reach the soil environment. It is unknown whether nematodes produced in
cadavers within the storage root remain in the root or exit and return to the soil.
Jansson et al. (1992, unpublished) examined nematode persistence in
the field and were able to recover H. bacteriophora HP88 strain for up to 253
days after application. In additional field studies, they recovered S. fMtia N-27
strain 220 d after application and S. carpcasae All strain was recovered less
than 140 d after application (Jansson, unpublished). Soil moisture, relative
humidity, and soil texture were attributed to the long persistence of the
nematodes. It is unknown, however, whether the nematodes persist for long
periods because of prolonged survivorship under field conditions or through
recycling in weevils or other insect hosts.
The objective of this study was to understand the recycling potential of
these nematodes better by addressing the following questions: (1) Do infective
juveniles produced within cadavers within the storage roots move out of the root
into the soil environment to infect new hosts, and (2) how far can nematodes
move after exiting storage roots? The distance these nematodes moved in the
field from a point source inoculation was also examined.
Materials and Methods
Nematodes were cultured in vivo in late instar Galleria mellonella L (wax
moth larvae) obtained from a commercial supplier (JA-DA Bat, Antigo,
Wisconsin) (Woodring & Kaya 1988). The wax moth larvae were stored at 50
until use. Heterorhabditid and steinernematid infective juveniles were stored in
deionized water at 260 and 70 C, respectively. Infective juveniles were less
than 2 wk old before use. Dilutions of nematode suspensions with deionized
water were prepared and nematode density was estimated as previously
described (Chapter 2).
Sweetpotato weevils were reared in the laboratory at 24 + 20 C (Coffelt et
al. 1978) on 'Jewel' or white-fleshed 'Picadito' sweet potato storage roots
previously described (Chapter 4).
Post-infection nematode emergence in sand. The following nematodes
were used in these experiments: Steinernema carpocapsae (Weiser)
(=Neoaplectana carpocapsae. feiae) All strain; S. feliae (Filipjev) (=S.
bibionis) N-27 strain; Heterorhabditis bacteriophora Poinar (=fH heliothidis)
HP88 strain; and Heterorhabditis sp. FL2122 strain (an undescribed nematode
isolated in Florida).
Two separate experiments were conducted. In the first experiment, 6
weevil-infested 'Picadito' storage roots from the laboratory colony were buried
in coarse construction sand (4.3% clay, 92% sand, 3.0% silt; sieved through No.
10 [2-mm] sieve) with 5% moisture in a plastic box (26x19x9 cm). Nematodes
were applied to the sand surface (494 cm2) in deionized water (100 ml) at
approximately 5 billion infective juveniles per ha (53 infective juveniles per
cm2). Three boxes with 6 weevil-infested roots were treated with each
nematode resulting in 18 roots per nematode. Boxes were covered with
aluminum foil and incubated for 2 wk at 24 20 C. Roots then were removed
from the boxes of sand, washed, and reburied individually in autoclaved sand in
plastic cups (1 liter). At weekly intervals, 10 wax moth larvae were placed on
the sand surface in each cup. After 1 wk, the larvae were removed and
replaced with fresh larvae. The number of larvae killed by nematodes were
recorded. Infection by nematodes was confirmed by disseotng cadavers and
examining them for the presence of nematodes. The experiment was
terminated after 3 wk.
In the second experiment, weevil-infested 'Jewel' sweet potato storage
roots from the laboratory colony were buried individually in coarse construction
sand (as above) with 5% moisture in plastic cups (1 liter). Nematodes were
applied to the sand surface (86.5 cm2) in deionized water (10 ml) at the
equivalent dosage of 7.5 billion infective juveniles per ha (74 infective juveniles
per cm2). A high concentration of nematodes was used to insure that infection
of the weevils would occur. Six infested storage roots were treated with each
nematode. The infested roots were removed from the sand, washed, and then
reburied in autoclaved sand one week after nematodes were applied. Wax
moth larvae were placed in each cup at weekly intervals as described
previously. Wax moth larvae were removed and replaced every 7 d for 6 wk.
Nematode infection was confirmed through dissection as noted above.
Post-infection nematode emergence and movement in soil. The
following nematodes were used in these experiments: S. carpocapsae All
strain, S. glaseri (Steiner), H. bacteriophora HP88 strain, and Heterorhabditis
sp. FL2122 strain. Four weevil-infested 'Jewel' storage roots from the
laboratory colony were buried in each of 4 boxes as above. There were four
roots buried per box and 4 boxes per nematode treatment. Nematodes were
applied to the soil surface (494 cm2) in deionized water (30 ml) at the
approximate dosage of 5 billion infective juveniles per ha (53 infective juveniles
per cm2). Roots were removed from the boxes and then buried individually in
the center of large fiberglass bins (57x42x15 cm) killed with Krome very gravelly
loam soil (26.8% day, 49.4% sand, 23.8% silt) two wk after application. The soil
was tested for the presence of native entomopathogenic nematodes using a
wax moth baiting technique (Bedding & Akhurst 1975) before the experiment.
The bins were covered with a plastic sheet and stored at 24 30 C in the dark.
Bins were watered (150 ml) approximately every 3 wk. Approximately every 3
wk, for 9 weeks, soil samples (20 cc) were removed from 8 locations within each
bin. Samples were removed from the four cardinal points (N, S, E, W); one
sample each at distances of 8 and 15 cm from the buried root in each of four
directions. Soil was added back to each bin at each sample location. Each soil
sample was placed in a plastic cup (280 ml) with 5 wax moth larvae placed at
the bottom of each cup and incubated at 24 30 C for 1 wk. All cadavers were
dissected for the presence of nematodes.
Six weeks after the roots were buried in the bins, 2 weevil-infested roots
were removed from the laboratory colony and buried at opposite ends of each
bin. These roots were removed after 12-14 d, dissected, and the numbers of
live and nematode-infected weevils were recorded. This was repeated three
Post-infection nematode emergence and movement in the field. Vine
cuttings of 'Jewel' sweet potato were hand-planted on 14 June 1991 0.2 m
apart in raised beds with centers 1.9 m apart in a research plot that was
subdivided into 16 subplots; each subplot was 3 rows across by 6.1 m long.
Subplots were surrounded by a 1.5 m strip of sorghum-sudangrass hybrid,
Sorghum color (L) Moench x S. arundinacum (Desv.) Stapf var. sudanese
(Stapf) Hitchc., that was planted by broadcasting seed on 23 July 1991. The
sorghum-sudangrass hybrid was planted to minimize weevil movement which
may potentially spread nematodes between subplots. Subplots were arranged
into a randomized complete block design with 4 replications.
Plants were fertilized (500 kg/ha of 6:12:12 [N: P2Os: KgOD before
planting. No herbicides, fungicides, or insecticides were applied during the
growing season. Plants were drip-irrigated 4 h per d (5.0 Iaerh/h) using a drip
turbo T-tape-irrigation system (model 40) from shortly after planting until the
experiment was terminated.
The nematodes used in this experiment were: S. carpoapsa All strain,
S. fettiae N-27 strain, H. bacteriophora HP88 strain, and Heterorhabditis sp.
FL2122 strain. The methods used for infesting the storage roots with weevils
and exposing these roots to nematodes were identical to those described
previously. In total, 24 weevil-infested roots were exposed to each nematode.
Two weeks after nematodes were applied, the roots were removed from the
boxes of sand and buried in the field in each subplot. Two groups of 3 infested
roots were buried 1.5 m apart, in the center row of each plot. Roots were buried
in block 1 on 16 October, block 2 on 25 October, block 3 on 29 October, and
block 4 on 1 November.
One soil sample was removed from each subplot on 16 September 1991
before the infested roots were buried. Each soil sample was baited with wax
moth larvae as described above to determine abundance of native
entomopathogenic nematodes in the field.
Soil samples (approximately 1,000 ml each) were removed from two
locations, 15 cm from each side of the buried roots. Soil samples were
collected similarly from the second group of buried roots within the same bed.
Subsequent samples were taken adjacent to the previous samples. Soil
samples were placed in plastic cups (1 liter) and 10 wax moth larvae were
placed on the bottom and the middle of each cup. Cups were covered with a
paper towel fastened with a rubber band. Samples then were incubated at 25.5
+ 2.80 C in the dark for 7 d. Cadavers were examined and dissected for
presence of nematodes as described previously. Soil samples were collected
on 19 November, 12 December, and 24 December 1991, and 1 and 28
January, and 18 February 1992.
SMortality of G. mellonella from samples collected from one side of the
buried roots (north) was compared from samples collected from the other side of
the buried roots (south).
Nematode movement from a point source inoculation in the field. Sweet
potato was planted as described above on 5 June 1991. Planting, fertilizer, and
irrigation rates and subplot sizes were identical to those described previously.
Subplots were arranged into a randomized complete block design with 5
The nematode treatments used in this experiment were: S. carpocapsae
All and Mexican strains, S. glaseri, bacteriophora HP88 strain, and
Heterorhabditis sp. FL2122 strain. Nematodes were reared in vivo in wax moth
larvae, G. mellonella, as described above. Nematodes were aerated and
stirred continuously until the necessary concentration was achieved.
Nematodes were applied in a 15 x 30 cm2 band across the center of the middle
row of each subplot, at the approximate rate of 7.5 billion infective juveniles per
ha (74 infective juveniles per cm2). Steinernema glaseri and Heterorhabditis
sp. FL2122 were applied to all replicates on 11 October; S. carpocapsae strains
were applied to replicate 1 on 11 October, replicates 2 and 3 on 23 October,
replicate 4 on 1 November, and replicate 5 on 7 November; and H.
bacteriophora HP88 was applied to replicates 1 and 2 on 23 October, replicate
3 on 1 November, and replicates 4 and 5 on 7 November. The nematodes
were applied as soon as the necessary dose was achieved so that the
nematodes were stored in the laboratory for the least amount of time. Quality of
the nematodes was checked prior to each application date by exposing ten Q.
mellonella larvae to subsamples of the nematodes in a standard Peri plate
bioassay (Woodring & Kaya 1988).
One soil sample was collected from each plot and baited with wax moth
larvae before nematodes were applied to determine population levels of native
entomopathogenic nematodes in the field. Methods were similar to those
Four soil samples (1,000 ml) were removed from the four cardinal points
(N, S, E, W), 30.5 cm from the center of the band of nematodes. An additional 4
samples were taken 61 cm from the center of the band of nematodes between
each cardinal point (NW, NE, SW, SE). Soil samples were collected on 19
December 1991, and 16 January and 10 February 1992. Soil samples were
baited with wax moth larvae and incubated as described above. All dead
larvae were dissected for presence of nematodes as described previously.
On 9 March 1991 additional soil samples were collected from the furrows
and the adjacent 2 rows in each plot to determine if nematodes moved beyond
the previously sampled area. The samples in the furrows were collected 0.9 m
from the center the middle row and samples in the adjacent rows were sampled
1.5 m from the center of the middle row. Two samples, aligned with the samples
collected 30.5 cm from the point source in the middle row, were collected from
each adjacent row and furrow. Soil samples collected from the furrow were not
taken at the same depth as the soil samples from the beds due to compaction.
The soil samples were baited and incubated as described above.
Data analysis. Because the data were not normally distributed, the total
numbers of wax moth larvae killed by nematodes were compared among
treatments by Chi-square analysis in all experiments (Conover 1980). The
expected value for calculating Chi-square was the mean number of S.
mellonella killed by nematodes for the treatments being compared.
Post-infection nematode emergence in sand. Morlaiy of wax moth
larvae significantly differed (Chi2>36.2; df=3; P<0.0001) among nematodes on
all sample dates for both experiments. Higher levels of mortality were achieved
by the heterorhabditid nematodes compared with the steinemematid
nematodes (Fig. 5-1). Steinernema carpcapsae All strain caused more
mortality than S. feMiae N-27 strain.
Post-infection nematode emergence and movement in soil. The soil
used in this experiment appeared to be free of native entomopathogenic
nematodes because there was no mortality of wax moth larvae exposed to the
soil before the experiment.,
The total number of wax moth larvae killed by nematodes differed
significantly among nematode treatments on all sampling dates (Chi2>62.3;
df=3; P<0.0001) (Fig. 5-2). The heterorhabditid nematodes infected more G.
mellonella larvae than the steinemematid nematodes on all sample dates.
Also, S. laseri caused more mortality than S. carpocasa All strain (Fig. 5-2).
Numbers of wax moth larvae killed also differed among nematodes in samples
collected at 8 cm from the buried root (Chi2>58.1; df=3; P<0.0001) and in
samples collected 15 cm from the buried root (Chi2>12.0; df=3; P<0.0007) (Fig.
5-3). On the first sample date, numbers of G. mellonella killed by S. glaseri and
H. bacteriophora HP88 differed between the two sample locations (8 and 15
cm) (S. gmaeri: Chi2=6.2, df=1, P=0.01; HP88: Chi2=18.2, df=1, P=0.0001) (Fig.
5-3). Both nematodes caused more mortality of Q. melonella in samples
collected from 8 cm than in those collected from 15 cm from the buried root. On
the second sample date, differences between sample locations were significant
for the heterorhabditids only (FL2122: Chi2=3.8, df=l, P=0.04; HP88:
Chi2=23.5, df=1, P=0.0001) (Fig. 5-3). Both nematodes killed more 2.
mellonella in the samples collected from 8 cm compared wNh those collected
from 15 cm. Heterorhabditis sp. FL2122 was the only nematode to kil
significantly more G. mellonella (Chi2=4.7, df=1, P=0.03) in the samples
collected from 8 cm compared with those collected from 15 cm on the third
The number of infested roots and percentage infection of weevils within
roots buried in the soil bins are shown in Table 5-1. Little to no weevil mortality
in infested roots was achieved by the first sample date (31 Jan.) with the
exception of FL2122 which caused 3.5% (n=200) mortality in one root. The
number of roots with weevil mortality as well as the percentage infection were
greater on the second and third sample dates (17 and 26 Feb.) compared with
the first sample date for all nematodes except S. carocaae All strain. No
weevil mortality was achieved by All strain on any of the dates. HP88 caused 2-
fold greater weevil mortality than FL2122 on the second and third samples
dates. Steinernema glaseri infected weevils in only one root on both the
second and third sample dates
Post-infection nematode emergence and movement in the field. Mortality
of G. mellonella in soil samples removed before nematodes were applied was
1.8% (n=160), indicating a low population level of native nematodes.
The number of wax moth larvae killed differed among nematodes
(Chi2>20.5, df=3, P<0.0001) on three of the six sample dates (19 Nov., 12 Dec.,
and 28 Jan.) (Fig. 5-4). On these three sample dates, HP88 yielded the highest
mortality. Mortality by the steinemematids was very low on the first two sample
dates (19 Nov. and 12 Dec.) and on the last two sample dates (28 Jan. and 20
Feb.). This may account for the significant differences among nematodes on 19
Nov., 12 Dec., and 28 Jan. On the sixth sample date (20 Feb.), laval mortality
was low for all nematodes tested. Overall, heterorhabdtld caused more
mortality than steinernematids, although mortality was generally low (les than
15%) for all nematodes tested (Fig. 5-4).
Mortality of Q. mellonela did not differ between sample locations
collected north and south of the buried roots for the stelnemematid nematodes
and Heterorhabditis sp. FL2122 (Chi2<2.6, df=1, P>0.11). Heterorhabditis
bacterophora HP88, however, killed more larvae in samples collected south of
the buried roots compared with those collected north of the buried roots
(Chi2=14.3, df=3, P=0.0002).
Nematode movement from a point source inoculation in the field. One
percent of wax moth larvae (n=300) were killed by entomopathogenic
nematodes in soil samples removed before nematodes were applied. This
level of infection by native nematodes concurs with the previous experiment
Total numbers of wax moth larvae killed by nematodes differed
significantly among nematodes on all sample dates (Chi2>23.2, df=4,
P<0.0001) (Fig. 5-5). Heterorhabditis bacteriophora HP88 killed the most
larvae followed by S. glaseri and Heterorhabditis sp. FL2122 (Fig. 5-5).
Steinemema carpocapsae All and Mexican strains caused less than 1%
mortality. Mortality of wax moth larvae in samples collected 30 cm from the
point source differed significantly among nematodes (Chi2221.2, df=4,
P<0.0003) on all sample dates with H. bacteriophora HP88 and S. glaseri
killing the most larvae (Fig. 5-6). Wax moth larval mortality differed among
nematodes (Chi2=34.8, df=4, P=0.0001) in samples collected 60 cm from the
point source on the first sample date only with Heterorhabdts bacteriphora
HP88 killing the most wax moth.
On the first sample date, mortality of -. mellonlla larvae by the
heterorhabditid nematodes and S. glase differed between the two sample
locations (30 and 60 cm) from the point source (HP88: Chl2=9.3, df-1, P=0.002;
FL2122: Chi2=13.3, df-1, P=0.0003; S. glaser: Chi2=23.1, df-1, P-0.0001)
(Fig. 5-6). On the second sample date, mortality of wax moth differed between
sample locations for HP88 and S. glseri only (HP88: Chi.5.3, df=l, P=0.02;
S. glasae: Chi2=6.2, df=1, P=0.01) (Fig. 5-6). Mortality by FL2122 differed
between locations at the 10% level (Chi2=3.0, df=1, P=0.08) (Fig. 5-6). On the
last sample date, results were similar to those on the first sample date (HP88:
Chl2=14.2, df=1, P=0.0002; FL2122: Chi2=10.0, df=1, P=0.002; S. glaser:
Chi2=13.5 df=1, P=0.0002). Percentage mortality was lower in samples
collected at 60 cm compared with those collected at 30 cm for all nematodes on
all sample dates.
Very few (<0.25%) or no G. mellonella were killed (n=400) by any of the
nematodes tested in soil samples collected in the furrows or in the nontreated
beds suggesting little or no movement of nematodes from the point source into
these areas. The few larvae killed by nematodes were from samples collected
from the adjacent beds only, not from those collected in the furrow.
The ability of infective juveniles to return to the soil environment is
important to the overall persistence and recycling of these nematodes. Infective
juveniles that move back into the soil environment potentially can enter
adjacent weevil-infested roots or infect other soil-dwelling insect hosts. It was
demonstrated in these studies that all nematodes tested were capable of exiting
the storage root, but differences existed among the nematodes tested.
The heterorhabditid nematodes were more capable of exiting storage
roots and moving further distances than the S. carpocapsae strains.
Steinernema glaseri behaved more closely to the heterorhabditid nematodes.
Other researchers also have found S. glaseri to move further distances in soil
than S. carpocapsae and H. bacteriophora strains (Schroeder & Beavers 1987).
The high recovery of heterorhabditid nematodes from sand compared to the
other nematodes may also be attributed to their ability to move faster in sand
than in a soil environment (Chapter 3). Additionally, heterorhabditid nematodes
produced more progeny in weevil cadavers than did steinernematids (Chapter
2). These attributes probably contributed to the success of the heterorhabditid
nematodes in the experiments conducted in soil as well.
All nematodes tested were capable of moving 15 cm in soil under
laboratory conditions; however, the heterorhabditid nematodes moved faster.
The heterorhabditid nematodes were also faster at exiting the storage root than
were the steinernematids. This also may be due to the higher cycling of
heterorhabditids and, therefore, more nematodes were likely to exit the root. It
is also possible that some of the nematodes are not new progeny but may have
survived within the roots since the original inoculation. It is likely that recycling
is occurring within the root because of the continued presence of the
Jansson et al. (unpublished) demonstrated persistence of S.
carpocapsae All and S. feltiae N-27 for more than 120 d after application in
Krome very gravelly loam soil. An undescribed heterorhabditid nematode
isolated from Puerto Rico, Heterorhabditis sp. Bacardis, persisted at high levels
for over 230 d after application. In other experiments, Jansson et al. (1992)
recovered H. bacteriophora HP88 for 120 d after application, when the
overhead irrigation ceased and for up to 253 d after application when plants
were drip-irrigated daily. It is not known, however, if the nematodes persisted
because of prolonged survivorship or due to recycling within weevils within
storage roots. Based on the present study, it is likely that nematode recycling in
the root and subsequent emergence from the root contributed to the recovery of
these nematodes in the previous field experiments.
The fact that these nematodes can recycle in infested roots is especially
advantageous. The root environment may provide some protection from the soil
environment while recycling occurs. The storage root may act like a 'container'
which holds the nematodes and releases them slowly into the soil. Additionally,
there is always the potential of other weevils within the same root being infected
which would continue the recycling process.
Once infective juveniles exit the root, they need to move within the soil
environment and enter adjacent roots or locate alternate soil-dwelling insect
hosts. It was shown that the heterorhabditid nematodes were capable of exiting
a storage root, moving more than 15 cm, and killing weevils within another
storage root. Although S. glaseri was capable of moving 15 cm in the soil
environment, it was less effective at killing weevils in adjacent roots. The
reproductive potential of S. glaseri in weevils is not known, but it is likely to be
less than that of heterorhabditids because of the lower nematode: host size ratio
(S. glaseri = 0.14; HP88 = 0.07). Steinernema glaseri is approximately 2-fold
larger than H. bacteriophora HP88 (Poinar 1990). It is likely that fewer S.
glaseri nematodes were produced per cadaver and, therefore, fewer were
available to infect weevils in adjacent roots.
Nematode life cycle would also play a role in that the first generation
heterorhabditid females are hermaphroditic. Males are not required for initial
reproduction. This is not true for the steinernematid nematodes. Both a male
and female are required for initial reproduction. Therefore, there is a potential
for more reproduction from heterorhabditids.
Nematode recovery was generally low (< 15%) when these nematodes
were compared in their ability to move from a point source inoculation within a
sweet potato field. More nematodes were recovered from the two
heterorhabditid nematodes and S. glaseri compared with the S. carpocapsae
strains, again demonstrating the movement capabilities of the heterorhabditid
nematodes and S. glaseri. Others have reported movement of heterorhabditid
nematodes from treated plots into nontreated plots (Shanks & Agudelvo-Silva
1990, Jansson et al. 1990b, 1992). However, the time frame over which this
movement occurred was several months in previous reports compared with
several weeks in the present study. Also in the previous studies, considerably
more nematodes were applied over a larger area compared with the small band
inoculation or infested roots used in the present study.
Although movement by nematode increases the likelihood of contact with
an insect, it is also expensive to the nematode in its energy reserves and
ultimately its survival. Environmental conditions (particle size, moisture, relative
humidity, oxygen, etc.) can also affect nematode activity. The movement of an
active nematode may be reduced by one or more of these factors but the
nematode also may be able to avoid less favorable conditions by moving away.
In spite of the potential energy expenditure and reduced survival, the active
nematodes appear to be the most efficacious. Therefore, heterorhabditid
nematodes, which tend to be more active than steinernematid nematodes,
appear to have more potential as biological control agent of the sweetpotato
Table 5-1. The number of storage roots in which weevils were infected by
nematodes and the mean percentage (+ SEM) of weevils killed within roots.
No. of Roots
Sample with Infected Weevil
Date Nematode Species Strain weevils n Mortality, %
31 Jan. S. carpocapsae All 0 0 0
S. glaseri 0 0 0
H. bacteriophora HP88 0 0 0
Heterorhabditis sp. FL2122 1 200 3.5 (0)
17 Feb. S. carpocapsae All 0 0 0
S. glaseri 1 243 0.4(0)
H. bacteriophora HP88 7 746 20.8 (8.2)
Heterorhabditis sp. FL2122 2 356 11.0 (10.3)
26 Feb. S. carpocapsae All 0 0 0
S. glaseri 1 100 1.1(0)
H. bacteriophora HP88 4 744 11.2 (4.7)
Heterorhabditis sp. FL2122 5 997 5.8 (3.1)
19 NOV. 91
26 NOV. 91
3 DEC. 91
All N-27 HP88
N-27 HP88 FL2122
Fig. 5-1. Mean percentage mortality ( SEM) of G. mellonella larvae at weekly
intervals in sand containing weevil-infested storage roots that were exposed
previously to nematodes.
6 JAN. 92
27 JAN. 92
2 FEB. 92
HP88 FL2122 ALL 6LASERI
Fig. 5-2. Mean percentage mortality ( SEM) of G. mellonella larvae on 3
sampling dates in soil collected from a bin containing a buried weevil-infested
storage root that was exposed previously to nematodes.
6 JAN. 92
27 JAN. 92
2 FEB. 92
Fig. 5-3. Mean percent mortality ( SEM) of G. mellonella larvae in soil
collected from a bin containing a buried weevil-infested storage root that was
exposed previously to nematodes. Samples were collected at two distances
from the infested root on three sample dates. An asterisk above the bar
indicates a significant difference (P<0.05) between the distances sampled for
that particular nematode.