Citation
Semiochemically mediated, host-searching behavior of the endoparasitic wasp Cotesia marginiventris (Cresson) (Hymenoptera: braconidae)

Material Information

Title:
Semiochemically mediated, host-searching behavior of the endoparasitic wasp Cotesia marginiventris (Cresson) (Hymenoptera: braconidae)
Creator:
Turlings, Ted C. J., 1959-
Publication Date:
Language:
English
Physical Description:
xiii, 178 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Chemicals ( jstor )
Corn ( jstor )
Female animals ( jstor )
Infestation ( jstor )
Insects ( jstor )
Larvae ( jstor )
Odors ( jstor )
Parasite hosts ( jstor )
Parasitoids ( jstor )
Seedlings ( jstor )
Dissertations, Academic -- Entomology and Nematology -- UF ( lcsh )
Entomology and Nematology thesis Ph. D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 156-177).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Ted C.J. Turlings.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
024294616 ( ALEPH )
23535986 ( OCLC )

Downloads

This item has the following downloads:


Full Text










SEMIOCHEMICALLY MEDIATED, HOST-SEARCHING BEHAVIOR OF THE ENDOPARASITIC WASP Cotesia marginiventris (CRESSON)
(HYMENOPTERA: BRACONIDAE)












BY

TED C. J. TURLINGS












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 1990




UNIVERSITY OF FLORIDA UBRARI,





























To Mams, Paps, Patricia, Emiel, Bix, Caspar, Walman, Dekker, and Sir Patrick.

Ted














ACKNOWLEDGEMENTS


My gratitude goes out to my graduate committee and the entire staff of the Insect Attractants, Behavior, and Basic Biology Research Laboratory, ARS-SEA, USDA, for their continuous support and help throughout the period that the research presented here was conducted. Special thanks go to Barbara Dueben without whom the work would have been much harder and far less enjoyable. In addition, I thank Hans Alborn, Peggy Brennan, Fred Eller, Rob Murphy, Ara Manukian, Tommy Proveaux, Delrea Patrick, Charlie Dillard, Annette Brabham, Bob Doolittle, Bob Heath and Peter Teal for all their patience, valuable advice, and for creating a pleasant working atmosphere. I thank Terri Rossignol for her valuable assistance. For their overseas support, I thank Louise Vet and Ben Kennepohl.

Financial support was granted by the International Research

Division of the Office of International Cooperation and Development and by the Agricultural Research Service, USDA. The research was also partially supported by a Fulbright grant issued by the U.S. Information Agency.

Last, but certainly not least, I would like to express special thanks to Jim Tumlinson and Joe Lewis for not only giving me the opportunity to conduct this research, but also for their patience and motivational support.


iii














PREFACE



Humans primarily use vision and hearing for orientation, while the functions of smell and taste are considered of minor importance. Anthropomorphic thinking may have prevented biologists from realizing how important smell and taste can be for other organisms, particularly in the insect world. During the last few decades, however, interest in how insects find mates, prey, and hosts has stimulated scientists to study the interactions that are mediated by airborne chemicals. As a result, the currently lively field of chemical ecology has revealed that for insect communication and orientation, olfaction is far more important than are the other two major modes, vision and hearing.

Airborne chemical stimuli are used by insects to guide them to

suitable habitats, suitable hosts or prey, to find appropriate mates and to detect danger. By interfering with these chemical interactions, man appears to be able to manipulate insect behavior to his advantage. This is exemplified by several successful attempts to use synthetic versions of insect sex-pheromones (i.e. chemicals emitted by insects to attract mates) to disrupt or trap pest insects. These strategies have been effective and safe alternatives to the use of pesticides to control insects.

Recently, several research groups launched projects to investigate the means by which chemicals can be used to manipulate the behavior of


iv








beneficial insects (e.g. those insects that kill pest insects). It is envisioned that this research will lead to strategies that will increase the effectiveness of these insects as bio-control agents. Preliminary results show that parasitic wasps can be conditioned or stimulated to respond to chemical cues that guide them to hosts (see Chapter I). This is achieved by letting the insects contact hosts or host products in the presence of these odors. This phenomenon of learning by association allows us to condition wasps such that they will utilize certain chemical cues that will lead them to a target pest more effectively. To establish the full potential of this technique, research should focus on the isolation and identification of the essential semiochemicals. Furthermore, the effects of these chemicals on the behavior of entomophagous parasitoids must be studied. When sufficient information is gathered on how different chemicals mediate the various behavioral components that lead to a successful encounter with a host, we may eventually be able to, as Shorey (1977) suggested ",make an insect jump through our hoop."


















v














TABLE OF CONTENTS
PAGE

ACKNOWLEDGEMENTS ............................................... iii

PREFACE ......................................................... iv

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

LIST OF FIGURES ................................................ x

ABSTRACT ........................................ xii

CHAPTERS
I LITERATURE REVIEW AND RESEARCH AIMS .................... 1

Semiochemically Mediated Host Searching Behavior ............ 2
What are Semiochemicals? ........................... . 2
Responses that can be Elicited by Semiochemicals.......... 4
The Mechanisms behind Semiochemical Induced
Responses................................................ 5
Semiochemically Mediated Long-Range Host Searching
Behavior in Parasitic Insects.............................. 8
Searching Behavior in Parasitic Wasps..................... 8
Examples of Host and Host Habitat Location by Olfaction... 11
Mechanisms that Determine the Responses of Parasitic
Insects to Semiochemicals............................. 15
Analysis of Host Searching Behavior....................... 22
Manipulation of Parasitic Wasps for Biological
Control Purposes.......................................... 23
Resource Management...................................... 24
Semiochemical Manipulation............................... 26
Cotesia marginiventris (Cresson) a Prime Candidate
for Augmentative Release Against Lepidopterous pests....... 30
Systematics.............................................. 30
Host and Plant Range..................................... 32
Geographical Range....................................... 36
Biology............ ................................. 36
Host Regulation........................................... 45
Bio-Control.............................................. 46
Research Aims............................................... 49





vi









II INCREASED RESPONSE TO HOST RELATED ODORS AFTER A FORAGING EXPERIENCE: A MATTER OF LEARNING ? ........... 52

Introduction.................. ............................ 52
Materials and Methods...................................... 53
Results.................. ................................. 57
Discussion.................. .............................. 60

III EFFECTS OF FORAGING EXPERIENCES ON ODOR PREFERENCES..... 64

Introduction.................. ............................ 64
Materials and Methods...................................... 65
Procedures and Results..................................... 66
Discussion.................. .............................. 75

IV ANALYSIS OF ORIENTED FLIGHTS TOWARDS A SOURCE OF HOST RELATED ODORS IN A FLIGHT TUNNEL.................. 80

Introduction ................... ........................... 80
Materials and Methods...................................... 81
Results................... ................................ 84
Discussion.................. .............................. 86

V PINPOINTING THE EXACT SOURCE OF VOLATILE ATTRACTANTS
THAT ELICIT ORIENTED FLIGHTS IN Cotesia marginiventris
FEMALES.............. ................................. 89

Introduction.................... .......................... 89
Materials and Methods...................................... 90
Results.................. ................................. 94
Discussion.................. .............................. 99

VI ISOLATION, AND IDENTIFICATION OF ALLELOCHEMICALS THAT ATTRACT Cotesia marginiventris TO THE MICRO-HABITAT OF
ONE OF ITS HOSTS.... .................................. 103

Introduction.................. ............................ 103
Materials and Methods..................................... 104
Results.................................................... 111
Discussion.................. .............................. 119

VII THE ACTIVE ROLE OF PLANTS IN THE PRODUCTION OF THE
VOLATILES THAT GUIDE COTESIA MARGINIVENTRIS FEMALES
TO THEIR HOSTS. ...................................... 125

Introduction.................. ............................ 125
Materials and Methods..................................... 126
Procedures and Results.................................... 127
Discussion.................. .............................. 139




vii









IX SUMMARY AND CONCLUSIONS................................. 144

REFERENCES CITED................................................ 156

BIOGRAPHICAL SKETCH ............................................ 178

















































viii














LIST OF TABLES



Table Page

1-1 Definitions of chemical mediators important in the 3 chemical interaction between organisms.
1-2 Reported host species for C. marginiventris. :_33
2-1 Effect of pre-bioassay experience on response of C. 58 marginiventris females exposed to host related odors in a
four-arm olfactometer.
---------------- -----------------------------6-1 Super Q trapped volatile compounds identified in the 115 atmosphere associated with corn seedlings fed upon by BAW
larvae.
---- ----------------------6-2 Quantative comparison of natural and synthetic blends 118 used for bio-assays
--------------------------------------6-3 Flight tunnel responses of C. marginiventris females with 1120 different experiences to extracts of; 1) volatiles collected from BAW larvae feeding in corn seedlings
(NATURAL), 2) a synthetic mimic of the same volatiles
containing the eleven major components (SYNTHETIC), and
1 3) the extraction solvent, methylene chloride, only
(SOLVENT).
7-1 Amounts of volatiles released by corn seedlings damaged 132 by BAW larvae just after damage (FRESH) compared with
I volatiles released 16 hours after damage (OLD).













ix














LIST OF FIGURES

Figure Page

I I
1-1 : Basic sequence of host-finding activities by females of 10
-.--parasitic insects (after Lewis et al., 1976).
4- -- ----- ---1-2 Schematic depiction of the life-cycle of C. 37 . marginiventris.
------------------------------------------------ ----1-3 i Behavioral ethogram of the host finding and ovipositional 42 sequence of C. marqiniventris females on corn plants
damaged by fall armyworm larvae (after Loke et al,
1983
--4----------------------------------------------------------------------- -----2-1 Experience effect upon response to odors of two plant- , 59
---- host complexes (with vs. without oviposition)
2-2 I Experience effect upon response to odors of two plant- 61 host complexes (CL on cotton vs. FAW on corn).
3-1 * Responses of experienced C. marginiventris females to 3 j 68
----- doses of odors emitted by larvae feeding on leaves.
+- - -- ------------------------3-2 , Effects of experience on the preference of C. , 70 Smarqiniventris females for host-related odors.
---------------------3-3 Responses of C. marginiventris females to the odors of 74 either of two plant-host complexes after a complete
contact experience including ovipositions.
---.--- .---------- ........ .. .------------------------------3-3 Responses of C. marginiventris females to the odors of 76 two plant-host complexes after a complete contact ..... experience with and without ovipositions. 4-1 Kinematic diagram of the flight behavior exhibited by 85 experienced C. marginiventris females in response to
odors emitted by plant-host complex of FAW larvae feeding
.on corn inside a flight tunnel.
5-1 Odor inlet system and insect release funnel for flight 92 tunnel bio-assays.






X








Figure Page

5-2 Flight responses by C. marginiventris females to a 95 complete plant-host complex, compared with flight
� responses to single components of a plant-host complex.
4---4 --- --------------- -------- --------------- -------- ---- ------ ----5-3 , Flight responses by C. marginiventris females during dual j 96
* odor source tests.
4--------------------------------------------- --5-4 Control dual choice flight tunnel tests. 98 4- ------------------------------------------
6-1 | Responses by C. marginiventris females to different doses 112 of extracts of volatiles collected from BAW larvae
feeding on corn seedlings.
4 -- -- s -- - -- - - - - - - - - - -- - - - - - - - - - - -
6-2 A) Profile of volatiles released by a complex of corn 114 seedlings damaged by BAW larvae. B) Profile of volatiles
released by undamaged corn seedlings. C) Structures and
identities of the eleven major compounds, with correspondin peak numbers.
4--------- - --------------------- --------------- -+----6-3 Volatiles released by different components of a complete 117 plant/host complex. A) Complete complex of BAW larvae
feeding on corn seedlings. B) Water-washed corn
seedlings that were damaged by BAW larvae. C) BAW frass
wiped of off the damaged corn seedlings. D) Starved
water-washed BAW larvae.
-----------------------------------------7-1 GC profiles of volatiles released by corn seedlings 129 I subjected to different damage treatments. A) BAW feeding
on seedlings that were damaged overnight by BAW larvae.
B) BAW larvae feeding on fresh seedlings. C)
Artificially damage fresh seedlings. D) Undamaged
seedling s.
7-2 Comparison of volatiles released by corn seedlings with 131 fresh BAW damage and seedlings with older damage.
- - -- --- -- --- -- --- -- --- -- -- - - -- --- -- --- -- -- -- --- -- -- + ----7-3 Comparison of volatiles released by corn seedlings 135 damaged by BAW with volatiles released by seedlings with
various artificial damage treatments. .
----- -----------------------7-4 Responses during dual choice flight tunnel tests by 137 experienced C. marginiventris females to corn seedlings
That underwent the various dama e treatments. .
4-- - ------------- 4 ------r--L''r
7-5 Preferences exhibited by C. marginiventris females for 140 odors released by corn with fresh BAW damage versus odors
, released by corn with old BAW damage.





xi














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

SEMIOCHEMICALLY MEDIATED, HOST-SEARCHING BEHAVIOR
OF THE ENDOPARASITIC WASP Cotesia marginiventris (CRESSON) (HYMENOPTERA: BRACONIDAE)

By

Ted C. J. Turlings

August 1990

Chairman: J. H. Tumlinson
Major Department: Entomology and Nematology

Behavioral and chemical aspects of the long-range, host-searching behavior of Cotesia marginiventris (Cresson), a larval endoparasitoid that attacks many lepidopterous species, were studied. In a four-arm olfactometer, parasitoid females exhibited a striking increase in their response to host-related odors after they previously had contact experience with the source of these odors which was a complex of host larvae feeding on plants. Results indicated that after contacting certain kairomones in the by-products of their hosts, the wasps associate the surrounding odors with the presence of hosts and will subsequently use these odors as cues in their search for additional hosts.

To determine the actual source of the active odors, a complete

plant-host complex was divided into its three main components: 1) corn seedlings damaged by host larvae (beet armyworm), 2) frass produced by


xii








those larvae, and 3) the host larvae. The damaged plants were significantly more attractive to C. marginiventris than frass or larvae alone.

Analysis of collected volatiles revealed the consistent presence of eleven plant released compounds. Four of those were typical green leafy odors and were released in relatively large amounts only if larvae were feeding on the seedlings. The other compounds, six terpenoids and indole, were not released in significant amounts until a few hours after initial feeding by the larvae. Artificially damaged corn seedlings released only minor amounts of the terpenoids and indole unless they were treated immediately after damage with the oral secretions of larvae.

This plant response to herbivore specific damage was exploited by the parasitic wasps. After experiencing the odors during a brief contact with host-infested corn, the wasps were highly attracted to the terpenoid-releasing corn seedlings. Experienced wasps were also attracted to a synthetic blend of the eleven identified compounds.

The flexible response of C. marginiventris to the actively

released plant volatiles is highly adaptive to both the plant and the parasitoid. It can be expected that many similar tritrophic interactions will be unraveled in the future and may be exploited for biological control purposes.









xiii












CHAPTER I

LITERATURE REVIEW AND RESEARCH AIMS


This chapter reviews the literature on semiochemically mediated host searching behavior of parasitoids. First, definitions of the various classes of semiochemicals and their modes of action are addressed. The importance of semiochemicals for long-range host location in parasitoids is discussed, whereby models and examples from the literature are used. Special attention is given to the different mechanisms that have been suggested to determine the responses by host seeking parasitoids to semiochemicals. The mechanisms are also discussed in relation to manipulation by men of parasitoids for biological control purposes. Subsequently, the literature available on the parasitoid Cotesia marginiventris (Cresson) is reviewed and discussed, with the emphasis on the potential of this generalist parasitoid for the use as a control agent. Finally, the research presented in this dissertation on the searching behavior of C. marginiventris is introduced by giving the objectives of the study.












1










Semiochemically Mediated Host Searching Behavior


What are Semiochemicals?

Karlson and Butenandt (1959) introduced the term pheromone and defined it as "substances that are secreted by animals to the outside and cause a specific reaction in a receiving individual of the same species (p. 39)." To also include the interactions between members of different species Law and Regnier (1971) introduced the term semiochemical. This term covers all chemicals that are involved in interactions between individual organisms. Several authors have later attempted to name and define the different types of semiochemicals. Brown et al. (1970) proposed allomone and kairomone as terms for chemicals produced by organisms that elicit reactions in species other than their own. Before and after that several other terms were introduced (Beth, 1932; Frankel, 1959; Chernin, 1970 and Blum, 1974). Nordlund and Lewis (1976) unraveled the tangle that had been created, and Table 1-1 lists the definitions that they gathered and adjusted.

Recently Dicke and Sabelis (1988a) introduced a modified

classification of what they call infochemicals. This classification is based on cost-benefit analysis rather than on the origin of the compounds which was mostly used by the earlier authors. Of course time will tell if new adjustments will have to be made and if other definitions have to be added. But for now the classifications as given by Nordlund and Lewis (1976) and by Dicke and Sabelis (1988a) can both serve as good guides in the world of semiochemicals. Here the classification by Nordlund and Lewis (1976) will be adopted.








3

Table 1-1. Definitions of chemical mediators important in chemical
interactions between organisms.

Semiochemical.-A chemical involved in the chemical interaction between organisms.
1. Pheromone - Substance that is secreted by an animal or a
plant to the outside that causes a specific reaction in a
receiving individual of the same species.
2. Allelochemic - Chemical significant to organisms of a species
different from their source, for reasons other than food as such.
a. Allomone - A substance, produced or acquired by an
organism, which, when it contacts an individual of another
species in the natural context, evokes in the receiver a
behavioral or physiological reaction adaptively favorable to
the emitter but not to the receiver.
b. Kairomone - A substance, produced, acquired by, or
released as a result of the activities of an organism, which,
when it contacts an individual of another species in the natural context, evokes in the receiver a behavioral or
physiological reaction adaptively favorable to the receiver
but not to the emitter.
c. Synomone - A substance, produced or acquired by an
organism, which, when it contacts an individual of another
species in the natural context, evokes in the receiver a
behavioral or physiological reaction adaptively favorable to
both the emitter and the receiver.
d. Apneumone - A substance emitted by a nonliving
material that evokes a behavioral or physiological reaction
adaptively favorable to a receiving organism, but detrimental to an organism, of another species, which may be found in or
on the nonliving material.

After Nordlund and Lewis (1976).








4

Responses that can be Elicited by Semiochemicals

Dethier (1947) and Kennedy (1947) realized that by using only the terms attractant and repellent it was not possible to describe chemicals in terms of their effect on the behavior of insects. Specific reactions to stimuli should get specific names and these names should have a clear definition. The first step in doing so is by analyzing the variety of ways by which aggregation and dispersion can be brought about. According to Dethier et al. (1960) an insect may do one of the following:

1. continue without change of rate of linear progression, rate of
turning, or direction,
2. stop,
3. slow its rate of linear progression,
4. increase its rate of turning,
5. increase its rate of linear progression,
6. decrease its rate of turning,
7. orient toward a source,
8. orient away from a source.

Responses 2 to 6 were grouped together as non-directed responses and called kineses (singular, kinesis). Later Kennedy (1977) added another possible response to this group; the initiation of movement when an insect was previously at rest. Responses 7 and 8 are placed in a second major group, directed responses with reference to the source, taxes (singular, taxis). The listed reactions do not only refer to movement but can also be used for other aspects of behavior such as feeding, mating and oviposition. Having categorized the types of behavior elicited by chemicals, Dethier et al. (1960) were able to list the type of chemicals in terms of what they do.








5

(1) Arrestant (stops and slows)-- a chemical which causes insects to aggregate in contact with it, the mechanism of aggregation being kinetic or having a kinetic component. An arrestant may slow the linear progression of the insects by reducing actual speed of locomotion or by increasing turning rate.

(2) Locomotor stimulant (starts or speeds)-- a chemical which

causes, by a kinetic mechanism, insects to disperse from a region more rapidly than if the area did not contain the chemical. The effect may be to increase the speed of locomotion, to cause the insect to carry out avoiding reactions, or to decrease the rate of turning (Fraenkel and Gunn, 1940).

(3) Attractant (orients toward)-- a chemical which causes insects to make oriented movements towards its source.

(4) Repellent (orients away)-- a chemical which causes insects to make oriented movements away from its source.

(5) Feeding, mating, or ovipositional stimulant (initiates or drives)-- a chemical which elicits feeding, mating or oviposition in insects.

(6) Deterrent (inhibits)-- a chemical which inhibits feeding or

oviposition when present in a place where insects would, in its absence, feed or oviposit.


The Mechanisms behind Semiochemical Induced Responses

The terms attractant, repellent, arrestant, stimulant and

deterrent are useful to describe the effects of semiochemicals, but give us no information on the behavioral mechanisms involved. Several









6

authors have reviewed the orientation to chemical sources (Shorey, 1973; Farkas and Shorey, 1972; Kennedy, 1977; Cardd 1984). Their papers made clear that it is very hard to classify, and even harder to identify, the mechanisms that are used by insects to respond to semiochemicals, even when the classification is limited to a flying orientation to chemical sources. So far, nobody has been able to give a satisfactory classification.

It is generally accepted that the principal mechanism of "long distance" flying orientation to an airborne chemical stimulus in the wind is an optomotor guided, chemically induced, upwind orientation (or anemotaxis) (Card6, 1984). However, orientation mechanisms in different phylogenetic groups will have evolved rather independently, therefore there is good reason to assume that multiple solutions for finding a chemical source will have developed. Knowledge is lacking to recognize and classify all these solutions.

Still, biologists studying insect behavior evoked by

semiochemicals need some kind of classification to work with, and need to be able to communicate with each other. Of the complex classifications made to date, Shorey's (1973) interpretation seems to be the most comprehensive. He splits the behavioral mechanisms used to aggregate at an odor source in three categories.


1) CHEMOTAXIS: the insects steers its body axis in the direction of the chemical source because it can directly sense the gradient
of odor molecules. In chemotaxis two main types are recognized,
ODOMOTROPOTAXIS and CHEMOKLINOTAXIS. In the first the insect uses
two or more receptor organs (e.g. antennae) to measure
concentration differences of the odorant, it will constantly turn towards the highest concentration. In chemoklinotaxis, the insect
swings its body (receptors) from one side to the other. In this way it assesses relative concentrations of the odorant over time








7

and steers towards the side on which it obtained greatest
stimulation.

2) KINESIS REACTIONS: the insect does not directly sense the direction of the chemical source, but it is caused to move at varying rates (ORTHOKINESIS) or to turn at varying frequencies (KLINOKINESIS) depending on the concentration of chemicals to
which it is exposed. The insect only senses concentration
differences and not the direction of the source.

3) Shorey (1973) did not give the third category a name, but it
includes the responses in which the odorant acts as a releaser, sensitizing the insect to some other stimulus and causing it to orient to that stimulus. ANEMOTAXIS is the major member of the third category. It is the orientation by steering into the wind
(upwind) when the appropriate odor is sensed. When the insect loses contact with the odor, it reacts by, for instance, crosswind flight. This increases the likelihood that the insect will
reenter the odorous airstream after which the upwind flight
resumes.

The occurrence of chemotaxis is unlikely because in the field,

currents in airstreams will make it virtually impossible for organisms to sense clear gradients of concentrations. Concentration differences might play a role when an insect gets close to the source, but at respectable distances from the source anemotaxis seems more obvious.

Anemotaxis is, indeed, accepted as the most common mechanism by which insects orient towards or away from an odor stimulus. In some cases, however, it may seem as if an insect orients to air currents, but it might well be that stimulation with semiochemicals results in a visual orientation to an appropriate object in the environment. Future studies of insect behavior and the stimuli evoking it will have to provide the information that will make it possible to give a clearer classification of the orientation mechanisms.

The purpose of this chapter is to review the studies that involve host searching behavior in entomophagous insects evoked by chemical








8

stimuli, therefore further discussion will only include parasitoid behavior evoked by semiochemicals belonging to the group of allelochemics.


Semiochemically Mediated Long-Range Host Searching Behavior in Parasitic Insects Searching Behavior in Parasitic Wasps

We may assume that action is undertaken by an organism because of the possible effects of such an action. A parasitoid follows a plume containing a volatile chemical associated with its host because it might lead the parasitoid to that host. It is generally accepted that such characteristics are adaptive and the result of natural selection.

The major limiting factor for reproduction of solitary parasitoids is the number of hosts a female finds to lay her eggs in. Females that are able to detect and locate suitable hosts more rapidly and from greater distances than other females would be at a reproductive advantage. Therefore, it can be expected that parasitoids have evolved rather "sophisticated" searching strategies. They will have to use chemical and other cues in such a way that the chances of making "mistakes" are limited, thereby optimizing their time allocation.

Vinson (1976) suggested that for many parasitoids searching behavior consists of a series of internally controlled locomotory patterns, each serving to place the wasps in contact with the searching space in which the parasitoid must randomly search for the next cue. Based on investigations of Salt (1935), Laing (1937), and Doutt (1959) he divided the host selection behavior into five steps:








9

1. habitat location

2. host finding

3. host acceptance

4. host suitability

5. host regulation

This division serves to categorize the various behavioral patterns.

Lewis et al. (1976) developed a diagram to describe the basic

sequence of host finding (Figure 1-1). The purpose of this diagram is to distinguish between the behavioral steps that are evoked by different chemical stimuli, thus enhancing the ability to identify key points for manipulation of parasitoids in pest management programs.

Again, the need to be able to communicate with each other makes it necessary for researchers to have models like the ones presented by Vinson (1976) and Lewis et al. (1976). However, the knowledge that has been obtained so far is certainly not enough to make a general diagram or model that describes all the aspects of host finding. The diagram developed by Lewis et al. (1976) may be a good description of the process involved for certain species, but other species might use a different strategy. At several places in the diagram insects might use other steps in their host finding activities that have yet to be observed. In other species, steps of the diagram, like examination of the host or post-oviposition behavior, might not take place at all. Future studies will have to elucidate whether general models can be developed or that the differences between species make it impossible.








10






General Inactivity

T-1 ,T1 (Sl)

Random Movement
T-2 T2 (S2)

Scanning of Habitat

T-3 T3 (S3)
I
Investigation of
Host Trails
Within Habitat
1 s T4 (S4) T8
T-4

Postoviposition Find and Approach Behavior attack cycle to Host


T5 (S5)
Ovipsitin T6 (S6) Exarnination of Host





Figure 1-1. Basic sequence of host-finding activities by females of
parasitic insects: T1 to T8 and T-1 to T-4 = transitions among the indicated behavioral acts. S1-6 = stimuli releasing the indicated
behavioral patterns. S2 = olfactory, visual, and physical cues associated with host plants on other habitats. S3 = primarily
chemical cues from frass, moth scales, and decomposition products
associated with the presence of host insects. S4 = olfactory, visual, auditory, and other chemical or visual cues from host
insect. S5 and S6 = olfactory, tactile, auditory, and/or
combination of these cues from host individual.

Taken from Lewis et al. (1976).










Examples of Host and Host Habitat Location by Olfaction

Host finding can involve long-range and close-range chemoreception (Kennedy, 1977; Weseloh, 1981). Long-range chemoreception is defined as being olfaction, which is the detection of chemicals in air. It is similar to our sense of smell. According to Weseloh (1981) close-range chemoreception involves only the perception of chemicals after direct physical contact with them in solid or liquid form. Obviously, insects also use olfaction when they are close to their target and it seems that close-range chemoreception should include olfaction as well. Much more is known about close-range chemical orientation than about long-range chemical orientation. This is probably because short-range orientation is easier to observe (Weseloh, 1981). Weseloh (1981) lists 19 species of parasitoids for which it has been demonstrated that they locate hosts by long-range chemoreception.

Although the investigation of host location by olfaction has barely begun, it is known that chemical stimuli can be emitted by a variety of sources. It is obvious that stimuli used in close-range orientation have to be directly related to the host. For long-range orientation that is not necessarily true. The stimuli can come from the host insect, but may also come from the habitat in which the host feeds, from associated organisms or from a combination of these factors (Lewis et al., 1981; Vinson, 1984). Reviews by Weseloh (1981), Lewis et al. (1981), Vinson (1976; 1984), van Alphen and Vet (1987), Kainoh (1987), and Eller (1990) report a fair number of examples of olfactorial host location. Some of the most distinct examples will be presented here.








12

Stimuli from the host (i.e. kairomones). Carton (1971; 1974) found that the parasitoid Pimpla instigator could discover its pupal hosts (Pieris brassicae) from a distance even when they were concealed in a tube of wrapped paper. In other experiments with field traps Kennedy (1979) showed that several parasitoid species were attracted to the aggregation pheromone used by their host, the European elm bark beetle. Lewis et al. (1982) and Noldus and van Lenteren (1983) were the first to report that attraction of Trichogramma spp. involves volatile substances released by virgin females of hosts. Further examples of host odors acting as kairomones for Trichogramma spp. are reviewed by Noldus (1989). Host frass is another common source of kairomones that are exploited by parasitic wasps. For example, Microplitis croceipes is attracted to the feces of Heliothis zea larvae (Eller et al., 1988b), and Diadromus pulchellus a parasitoid of the leek moth and the diamondback moth is attracted to volatiles emanating from the larval frass of the latter two (Auger et al., 1989). An extensive review of identified insect kairomones that attract entomophagous insects is presented by Eller (1990).

Stimuli from plants (i.e. synomones). Plants are the most common source of volatile attractants for parasitoids. Thorpe and Caudle (1938) were some of the first to report this. It is normally argued that the parasitoids locate host habitats by tracking the volatiles that are emitted by plants the host may be feeding on (Vinson, 1975). In several cases parasitoids are attracted to host plants, regardless of the presence or absence of hosts. Vinson et al. (1975) observed Cardiochiles nigriceps searching host-free tobacco plants. Nishida








13

(1956) found that Opius fletcheri was attracted to the plant habitat of its host, the melon fly. Macrocentrus grandi, a parasitoid of the European corn borer, was attracted to 15 out of 51 plant species tested (Ding et al., 1989). Ding et al. (1989) found no correlation between host or nonhost status of the plants for the European corn borer. The generalist parasitoid Campoletis sonorensis appears to be different in this respect. The responses to plant odors by females of C. sonorensis in flight tunnels have been studied extensively (Elzen et al., 1983, 1984, 1986, 1987; Baehrecke et al., 1990; McAuslane et al., 1990a, 1990b, 1990c; Williams et al., 1988). This parasitoid is readily attracted to plants that serve as food for its hosts, but the wasp is significantly less attracted to nonfood plants unless it has previously encountered hosts on such a plant (McAuslane et al., 1990b).

It is frequently found that parasitoids are attracted to some species of plants but not to others. Taylor (1932) reported that Heliothis armigera, which feeds on a variety of plants, was attacked by Microbracon brevicornis only when it fed on Antirrhinum. Arthur (1962) found something similar for Itoplectis conquistor, which attacks larvae of the European pine shoot moth on scots pine but not on red pine. More examples of plant odors attracting parasitoids are given in chapter V and the subject was recently reviewed by Nordlund et al. (1988).

The many cases of parasitoids attacking hosts on one plant but not on the other (for review see Vinson, 1981), are usually explained as plant preferences. The preference in some cases might be explained by the fact that on some plants hosts are more suitable for parasitization because they provide the adult parasitoid with nutritional requirements,








14

such as nectar from flowers (Herzog and Funderburk, 1985; Lewis and Nordlund, 1985). In other cases, hosts on the non-preferred plants might contain harmful chemicals obtained by eating from that particular plant. Besides that, specific components in the food might be necessary for the host to produce the right kairomones and these components may not be present in all plant species.

Plants on which the hosts do not feed can also be attractive to

the parasitoid. Associated plants may provide the parasitoids with food and/or shelter. Altieri and Whitcomb (1980) and Altieri et al. (1981) showed that the presence of selected weeds within and around the crop fields greatly affected the abundance and activity of parasitoids in corn crops.

Stimuli from associated organisms other than plants. Examples of host-associated organisms other than plants that are involved in the attraction of parasitoids are relatively rare. Spradbery (1970) demonstrated that Rhyssa persuasoria a parasitoid of siricid woodwasps is attracted to odors produced by fungal symbionts of their hosts. Greany et al. (1977) showed that a rotting fruit fungus, which is often found together with tephritid fruit fly larvae, produces acetaldehyde, an attractant for the parasitoid Biosteres (Opius) longicaudatus. Dicke (1988a) reviews the studies that demonstrate the involvement of microorganisms in the host location of several parasitoids.

Stimuli from a combination of factors. Most studies do not determine the exact source of the stimuli. In some studies it is demonstrated that it is actually a combination of factors. Leptopilina heterotoma, for instance, is attracted to yeast (host food), but shows a








15
much stronger response to yeast patches in which host larvae have been crawling and feeding (Dicke et al., 1984; van Alphen et al., 1984). This indicates the presence of allelochemics produced by a combination of yeast and host material. Studies on the larval endoparasitoid Microplitis croceipes indicate something similar; it responds better to hosts feeding on plants than to hosts only or plants only (Drost et al., 1986; Eller et al., 1988b).


Mechanisms that Determine the Responses of Parasitic Insects to Semiochemicals

A thorough understanding of the mechanisms involved in the host location by parasitoids is necessary before we can start thinking about the manipulation of these insects. Pheromone work has already resulted in a fair amount of information on how insects respond to attractive volatiles. Although possible, there is no reason to believe that basic behavioral responses of parasitoids to chemicals that will lead them to their hosts will differ from the responses by insects to sex pheromones that will lead them to conspecifics of the opposite sex. The strategies behind these responses and the properties of the chemicals that elicit the responses, however, will be quite different.

In the case of sex pheromone attraction, both the receiver and the emitter profit from the communication between them. In the parasitoid host relationship, however, the attraction is advantageous to the parasitoid, but disadvantageous to the host. Therefore, selection will put constant pressure on the hosts to avoid the production and/or release of the semiochemicals. It can be expected that due to these pressures, constant changes in odor production within host populations








16

are taking place. This will force parasitoids, in their turn, to adapt to these changes, requiring plasticity in their responses to semiochemicals. A predator-prey relationship will obviously result in similar coevolutionary pressures. The need for plasticity in a parasitoid's response to attractants becomes even more important when the sources of reliable attractants vary. When the main cues are not coming directly from the host, but from associated organisms, variability can become enormous. For example, a generalist parasitoid that attacks its hosts on a wide variety of plants, may use plant odors to locate its hosts. Different plants will release different odors, therefore the odor blend that contains the most reliable information on the presence of hosts will vary over time and space.

Considering the preceding arguments, it is unlikely that the response of parasitoids to semiochemicals is hardwired and mainly genetically determined. Studies on the host location by parasitic insects have resulted in several theories on the mechanisms that allow these insects to optimally respond to chemical cues. These theories will be reviewed here.

Innate responses. When the wasps respond to odors without ever having had any previous experience with them the responses are often referred to as innate. As Bateson (1984) points out there is quite some controversy over "innateness". It is now generally accepted that behavior is never completely genetically hardwired, and that the interplay between internal and external factors determines behavior. When discussing the responses of parasitoids to certain stimuli, it seems most useful to talk about unlearned versus learned behavior.








17

Unlearned responses involve behavior that develops without the individual experiencing stimuli to which it will respond, or without practice of the motor pattern that it will perform (Bateson, 1984). Whether or not the responsiveness of individual insects to certain odors is totally unlearned appears hard to demonstrate. The fact that the adult insects will respond to odors without ever experiencing them in association with hosts after emergence can easily be mistaken for an unlearned response. One could argue that the inexperienced females have had no chance to associate certain odors with hosts, therefore their responsiveness must be inherited. Although possible, this is not necessarily true. Information obtained during one of the immature stages might well have resulted in their responsiveness (see later). Still, it is unlikely that all odor responses have to be learned. Although most parasitoids respond to host-related odors much better after a contact experience with the host or its by-products, naive insects will always respond to some degree (Vet, 1984, Drost et al., 1986; Hdrard et al., 1988a; Eller, 1990). Recent studies on the searching behavior of the generalist parasitoid Campoletis sonorensis show a high response level for naive females (Baehrecke et al., 1990; McAuslane et al., 1990a, 1990b) not yet observed in other species. Experience will always affect an insects behavior to some degree. Perhaps it is best to distinguish between insects that rely only slightly on experiences and insects that rely strongly on experiences. A similar distinction could be made between the stimuli that the insects respond to. Vet et al. (1990) developed a model that puts the variability of responses by parasitoids to stimuli in a framework which








18

takes the genetic constitution, physiological state, and environmental factors in consideration. It is clear that the complex interplay of internal and external influences is far from completely understood. Some theories on how external factors may determine a parasitoid's responsiveness have been developed. These theories are reviewed in the next paragraphs.

Hopkins' host-selection principle. In 1917 Hopkins introduced the "Hopkins' host-selection principle". It was simply defined as "an insect species which breeds in two or more hosts will prefer to continue to breed in the host to which it has adapted (p. 190)" (Craighead, 1921). This vague definition could be interpreted in several ways, but has always been assumed to say that insects prefer to oviposit on the type of host they fed upon as immatures (Jaenike, 1983). Hopkins did not mention what he thought was causing this phenomenon. It seems that some authors just assume that Hopkins meant that there is a link between larval feeding behavior and preferences shown by the adult for particular oviposition site selection (Corbet, 1985). Looked at it in that way, Jaenike (1983) was right to suggest the following three shortcomings: 1) it fails to rule out genetic effects since the individuals bred from a given host may be the offspring of females that inheritably prefer that host; 2) it fails to distinguish between the effects of larval versus adult exposure to the particular host; and 3) experiments to demonstrate the principle were not carried out on independent samples or over a span of many generations, to rule out random fluctuations in host preference. However, if the definition of








19

Hopkins' principle is interpreted in the broadest sense, these possibilities are included in it as well.

The concept "memory" is regarded as the central feature of

Hopkins' host-selection principle (Jermy et al., 1968; Jaenike, 1982). They see it as an effect on adult behavior of the early chemical environment mediated by a neural change effected in the larva and persisting to the adult stage (Corbet, 1985). In some cases it has been shown that the larval environment does affect the adult responsiveness to odors, but experiences during the adult stage seem to have more influence on the odor preferences of insects (Jaenike, 1983; Vet, 1983).

The chemical legacy hypothesis. The chemical legacy hypothesis as proposed by Corbet (1985), could be seen as an alternative explanation for the Hopkins' host-selection principle. It suggests that "effects of the early environment on the chemosensory responsiveness of a later stage depend not (or not only) on 'memory', but on the direct effects on the later stage itself of a legacy of chemical cues bequeathed to it from the earlier stage. This legacy consists of minute quantities of certain chemicals that persist from one stage to another inside or outside the insects's body" (Corbet, 1985).

So far, no direct evidence for this hypothesis exists. The

obvious thing to do is to change the chemical environment of the larvae and study the effect of such changes, but since the hypothesis suggests that minute quantities of chemicals will be sufficient to influence the adult behavior this might be impossible. Even if it is possible to manipulate the chemical environment of the larvae, it will be hard to distinguish an effect of the chemicals on the sensory system during the








20

larval period from effects of chemicals obtained by the immature and carried over into the adult stage that influence the adult's behavior.

Learning and conditioning. More and more studies demonstrate that recognition of specific semiochemicalsa adult parasitoid females is partially acquired through the association of hosts and/or host products and their odors during a contact experience (e.g. Thorpe and Jones, 1937; Monteith, 1963; Arthur, 1971; Taylor, 1974; Vinson et al., 1977; Sandlan, 1980; Strand and Vinson, 1982; Vet, 1983, 1988; Vet and van Opzeeland, 1984, 1985; Wardle and Borden, 1985; Dmoch et al., 1985; Drost et al., 1986; Lewis and Tumlinson, 1988). For instance, two alysiine species could distinguish between odors of host (Drosophila larvae) infested and uninfested substrates only after experiencing an oviposition on an infested substrate (Vet and van Opzeeland, 1984). Vet (1983) demonstrated conditioning for the parasitoid Leptopilina clavipes. Most females of this parasitoid are attracted to odors of decaying fungi (a potential habitat for its hosts). When females are given oviposition experience with host larvae that are feeding in yeast, their preference changes to yeast odors over odors of decaying fungi. Vet (1983) also found that the effects of this type of conditioning are much stronger than the effects of rearing the parasitoids on larvae feeding on yeast. The latter form of conditioning also increased response to yeast odors but less dramatically, mushroom odors were still preferred. Nevertheless, learning of odors as an immature (known as pre-imaginal conditioning) has been shown to affect the responses by the adult female significantly in some species (reviewed by Vet, 1983). None of these studies, however, consider that an emerging adult female








21

may experience odors that will influence her responses. H&rard et al. (1988b) show that this may be very important. They found that the wasp Microplitis demolitor responds well to host-related odors in a flight tunnel if it is reared on a host that was fed plant material, but not if the host was fed artificial diet. They subsequently demonstrated that the wasps from plant fed hosts required contact with their cocoon after emergence to learn the odors to which they would respond in the flight tunnel. This explains the often observed poor response by laboratory (artificial diet) reared parasitoids to host related odors.

Learning seems to occur frequently in parasitoids (for a review see van Alphen and Vet, 1987) and it will be discussed more extensively in some of the following chapters.


Which of the above mechanisms is more important might depend on the host specificity of the insect. Species that are able to attack several different host species can be expected to depend on experience as an adult. For those species the information obtained as a larva or a pupa might not necessarily lead it to the most abundant, easier to find and most suitable hosts. These species will find the most useful information while searching for hosts.

This does not necessarily mean that specialized insects will rely more on the innate response and the chemical cues encountered as an immature. Specialists that find their hosts in a variety of habitats may use cues that are different for each habitat. Therefore, even though they will always be in search of the same type of host, they will have to alter their responses to find hosts in different habitats.








22

Nonetheless, the array of chemicals a naive specialist female will respond to may be much more limited than that of a generalist. Particularly in those cases where a specialist wasp relies on cues that are very closely associated with its host unlearned responses may be more important than learned responses. It should be clear that the above discussed mechanisms can not be seen separate from each other. Some or all of them will affect a parasitoid's behavior at the same time.


Analysis of Host Searching Behavior

Few studies had been undertaken to elucidate the behavioral

mechanisms underlying host-habitat location. Until recently, none of the available methods seemed to be suitable for such studies. However, the well designed olfactometer studies conducted by Vet and co-workers (Vet, 1984), have resulted in an avalanche of similar studies. Olfactometer studies serve well to show attraction to and preference for odors, but the confined space does not allow an insect to display all aspects of its behavior. To accomplish that, flight tunnel studies appear more suitable (Drost et al., 1986). Initial experiments with flight tunnels failed because the insects tend to disperse upon release. By exposing parasitoids to kairomones before releasing them, however, the tendency to disperse can be overridden (Gross et al. , 1975; Loke and Ashley, 1984c; Drost et al. , 1986).

Drost et al. (1986) and Eller et al. (1988b) performed the first accurate studies of the long-range searching behavior of a parasitoid. They described the characteristics of flight by Microplitis croceipes in








23

response to a host-plant complex and to individual components of such a host-plant complex in a flight tunnel. M. croceipes females showed "sustained flights" (continuous flights that resulted in a landing on the target) towards a host-plant complex and towards damaged plants. A typical "sustained flight" started with casting (alternate left and right movements across the axis of the odor plume without net upwind or downwind movement) close to the release point. This was followed either by anemotaxis (flight track orientation directly towards the odor source) or by zigzagging (flying alternately to the left and the right of the main axis of the odor plume, while gaining or loosing ground) up to ca. 10 cm distance of the target. There, the parasitoid started hovering (hanging still in the air) followed by a darting approach to the target.

An important aspect of their work is that they found that preflight contact with the host-plant complex or with host feces increased responses of the parasitoids. This phenomenon has been observed in several parasitoids and will get special attention in this dissertation. The effect of experience with host and/or host-related products opens possibilities for manipulation of parasitoid behavior for biological control purposes. The current status of this particular research area will be discussed in the next section of this chapter.


Manipulation of Parasitic Wasps for Biological Control Purposes

Agriculture dealing with a constantly increasing demand for its products, is fighting an ever lasting war against pest insects.








24

Conventional chemical pesticides are indispensable in this war, but the use of pesticides has resulted in some basic problems:

1) insects may develop resistance to broad spectrum chemicals,
resulting in the demand for new more powerful and more expensive
pesticides;

2) many beneficial insects and other organisms perish from the use
of pesticides;

3) outbreaks of secondary pests that were previuously not of
economic importance; and

4) environmental contamination, which rightfully receives more and
more public protest.

Insect pest control scientists now agree that the alternatives have to come out of the biological control sector (van den Bosch and Messenger, 1973; Lewis, 1981). Biological control has traditionally been used to describe the regulation of pests with the use of natural enemies. In a much wider view, biological control involves many other naturally derived strategies for controlling pest populations, including the manipulation of natural enemies. This review will focus on manipulative strategies that have been applied, that are under study, and that may be found useful in the future to increase the effectiveness of parasitic insects as controlling agents. The strategies that will be discussed are divided into two main categories: 1) resource management, and 2) semiochemical manipulation of behavior. The field of genetic manipulation is is not discussed here.


Resource Management

The maintenance of parasitoid densities that will be able to

control pests effectively may be realized by supplying extra resources like nectar sources for the adults, hosts during low host densities and








25

maybe even shelter against unfavorable environmental conditions. Several studies have already shown that provision of specific resources results in better control of pest populations (for reviews see Herzog and Funderburk, 1985; Lewis and Nordlund, 1985; Hagen, 1986).

Providing food sources for adult parasitoids. Agricultural monocultures often lack food sources, such as nectar and pollen producing weeds, which under more natural circumstances would provide adult parasitoids with their nutritional requirements. This problem is in some cases amplified by selection for tolerant and resistant crops, such as cotton that lacks extra-floral nectaries (Schuster, 1980; Herzog and Funderburk, 1985).

Drake (1920) was one of the first to suggest the use of what he called a trap crop to attract and feed the tachinid wasp Trichopoda pennies Fabricius. Drake argued that the trap crop, Crotalaria usaramoenis, has floral nectaries that attract the wasp and pods that are attractive to the parasitoid's host, the stink bug Nezara viridula L. Since then several attempts to provide additional food sources conjunction with a crop have had positive effects on retaining and maintaining parasitoids and predators in high enough numbers (Gardner, 1938; Hagen et al., 1971; Doutt and Smith, 1971; Leeper, 1974; Beglyarov and Smetnik, 1976; Matteson et al., 1984).

Providing additional hosts. Smith and DeBach (1953) suggested that artificial infestation of plants with hosts during periods of low host density is a means by which host-parasitoid populations could be synchronized and could result in increased effectiveness of parasitoids in achieving biological control. Knipling and McGuire (1968) had








26

something similar in mind when discussing the effectiveness of Trichogramma species. They, however, suggested a sustained addition of host eggs to the environment irrespective of the actual density of the insect pest. The practicality of these concepts was demonstrated with several field experiments (e.g. Parker et al., 1971; Gross et al., 1984).

No examples exist of active introduction of alternative hosts into an area infested by a pest insect. However, growing certain plants near the target areas that carry alternative hosts has been found to be effective (Doutt and Smith, 1971).


Semiochemical Manipulation of Parasitoid Behavior

Successful reproduction of a parasitoid female depends partially on the proper use of the allelochemics in its macro-habitat. Recent research shows that man can manipulate the parasitoid behavior with chemicals used for host location. Therefore, by applying specific allelochemics at the right time and in the right place it should be possible to manipulate parasitoid behavior to our advantage (Nordlund et al., 1981a; Nordlund et al., 1981b; Lewis, 1981; Lewis and Nordlund, 1985).

Application of allelochemics in the target field. Application of the allelochemics that serve as cues used by parasitoids to locate suitable hosts may result in significantly higher parasitization rates in certain crop systems. The additional allelochemics probably help to retain the parasitoids in the area and may also increase their motivation to search for hosts.








27

Lewis et al. (1972) were the first to demonstrate this effect. A kairomone for Trichogramma evanescens present in the scales of Heliothis zea (Jones et al., 1971) was extracted with hexane. In laboratory as well as in greenhouse and field experiments it was shown that H. zea eggs placed on cotton or pea seedlings that were treated with the extract were parasitized significantly more than eggs on plants that were treated with only hexane. In both the greenhouse and field experiments the level of parasitism and the number of adult parasitoids produced were about twice as high on the leaves treated with kairomone. Even more spectacular results were obtained from field experiments in crimson clover by using tricosane, one of the kairomones identified from moth scales by Jones et al. (1971). Parasitization by naturally occurring Trichogramma spp. was increased from 4% to 15%. The pattern in which the treatment is applied to the field was found to be very important (Lewis et al., 1979). At low host densities constant exposure to allelochemics should be avoided to ensure movement by the parasitoids from one oviposition site to another.

Nordlund et al. (1983) observed an increase in parasitization rate by the parasitoid Telenomus remus, which attacks egg masses of Spodoptera fruqiperda (J.E. Smith), after placing cotton rolls treated with S. fruqiperda pheromone near egg infested cowpea seedlings in a greenhouse. Nordlund et al. (1983) report on how they were able to induce significant parasitization by T. remus in eggs of the non-host Heliothis zea by treating the eggs with S. fruqiperda pheromone.

Some crops are more attractive to certain parasitoids than others crops. Nordlund et al. (1984) found, when testing different cultivars








28

with combinations of tomato, bean, or corn, that parasitization by Trichogramma spp. in cultivars that included tomato was higher than in those without tomato. Parasitization on corn treated with tomato extract was higher than in untreated corn. Parasitization on tomato treated with corn extract was not different from parasitization on untreated tomato. These results indicate that a synomone present in tomato plants "improves" the performance of Trichogramma spp.

Nordlund and Sauls (1981) showed that for the parasitoid

Microplitis croceipes the kairomonal activity of frass from Heliothis zea depends on the plant diet of the host. Female parasitoids responded to extracts of frass from larvae reared on cotton or soybean but not on corn. The lack of response to the frass of corn-fed larvae was due to the absence of some appropriate chemicals.

Besides being useful to attract parasitoids, motivate them to search for hosts and to retain them in a target area, allelochemic attractants may also serve in trapping parasitoids to monitor their establishment and their dispersion.

Application of allelochemics for pre-release stimulation of

parasitoids. The tendency of insects to disperse upon release has often resulted in poor establishment of parasitic insects (Lewis and Nordlund, 1985). Gross et al. (1975) were the first to find that the tendency to disperse can be overcome by exposing parasitoid females to kairomones prior to release. They allowed female Microplitis croceipes to contact Heliothis zea frass before being released close to H. zea larvae feeding on cowpea leaves in a greenhouse. These pre-stimulated females readily








29

parasitized the larva, while control females that were not stimulated before release did not parasitize any larvae.

Gross et al. (1975) also found that contact with the kairomones present in scales of H. zea stimulate host-searching activities in two Trichogramma spp. Stimulated parasitoids had higher rates of parasitism in laboratory experiments than had unstimulated parasitoids. Prerelease stimulation also increased the efficiency of Trichoqramma pretiosum in the field.

Recent research has established that the effect of increased response to host odors after experience is, at least in some cases, caused by associative learning (Vet, 1983; Vet and van Opzeeland, 1984; Lewis and Tumlinson, 1988). In some cases contacting larval by-products causes the females to link the associated or surrounding odors with the possible presence of hosts. Subsequently, the parasitoids will use these odors as cues in their search for more hosts. The first chapters of this dissertation deal with this phenomenon in C. marqiniventris.

Lewis and Tumlinson (1988) showed that Microplitis croceipes uses both a contact kairomone and a volatile kairomone present in the feces of Heliothis zea larvae. They were able to separate the two types of kairomones and found that the parasitoid females associated the surrounding odors with the contact kairomones after they rubbed the feces with their antennae. The parasitoid females could be fooled into flying to the odor of vanilla after they had rubbed the contact kairomone in the presence of vanilla extract.

These findings offer promising possibilities for biological

control. If applied in the right way it should be possible to increase








30
parasitoid responsiveness to host related allelochemics, and also to condition them to respond to those odors that will lead them to the target species. It might even be possible to condition parasitoid females that they will focus on a host species that they normally would not prefer. Agriculture could profit from these findings if we are able to identify and synthesize the allelochemics involved, and apply them as pre-release conditioners of parasitoids used in biological control projects.

In the following chapters the manipulation with semiochemicals

will be explored for the parasitoid Cotesia marginiventris. First, the available literature on this insect will be reviewed.


Cotesia marginiventris (Cresson) a Prime Candidate for
Augmentative Release Against Lepidopterous Insects

The larval endoparasitoid Cotesia marginiventris (Cresson) is one of the most frequently recorded parasitoids in larvae of moths of the family Noctuidae. C. marginiventris attacks many species (Table 1-2), several of which are economically important pests in the USA. Its successful establishment in many different habitats makes it a promising candidate for biological control. Efforts should be undertaken to increase the parasitization efficiency of C. marqiniventris.


Systematics

C. marginiventris (Cresson) was originally described from Cuba

(Muesebeck, 1921) and is considered a native to the West Indies. Over








31
the years this parasitoid has been known under 6 different names (Marsh, 1978; Krombein et al., 1979).


Microgaster marginiventris Cresson, 1865:67.

Apanteles Qrenadensis Ashmead, 1900:278.

Apanteles laphvygmae Ashmead, 1901:38. Nomen nudum.

Apanteles (Protapantelas) harnedi Viereck, 1912:580.

Apanteles marginiventris (Cresson), (Muesebeck, 1921).

Cotesia marginiventris (Cresson), (Mason, 1981).


C. marginiventris is a member of the Braconidae family. This family can be distinguished from related families by a characteristic fore wing venation. Braconids possess only one recurrent vein and usually have a (disco-)cubital vein that separates the discoidal and cubital cells. A third characteristic is the fusion of the second and third abdominal tergites which makes the suture between them inflexible. C. marginiventris is currently classified under the largest and taxonomically difficult subfamily, the Microgastrinae (Matthews, 1974). As most braconids, Microgastrinae are endoparasitoids of immature stages of lepidopterous species. In 1965 Nixon published a correction on the classification which for the first time united the Microgastrinae by the location of the spiracle of tergum I on the lateral membranous margin. That character is still mostly used to distinguish them from other subfamilies,

Mason (1981) introduced the currently used classification of the Microgastrinae into five tribes: Apantelini, Microgastrini, Forniciini,








32

Cotesiini, and Microplitini. C. marginiventris belongs to the Cotesiini. For a quick diagnosis of this tribe, Mason (1981) suggests:

Hypopygium short and evenly sclerotized; ovipositor short, stout basally, and abruptly tapered about mid-length; sheat short with hairs concentrated apically, arising proximally from valvifer. Propodeum with median carina or none; very rarely areolate. Areolet usually open, but if closed hind coxa longer than tergite I and tibial spurs longer than half basitarsus.

For further identification of the adults the keys given by Marsh (1971) or Mason (1981) (to genus) and Muesebeck (1921) (to species) are most useful.


Host and Plant Range

No individual study has been undertaken to determine the actual

host range of C. marginiventris. However, numerous publications present field data for collected caterpillars from which C. marginiventris was reared (Table 1-2). Together with several reports on laboratory reared C. marginiventris an obviously incomplete picture of this parasitoids' host range can be drawn. Table 1-2 gives the species from which C. marginiventris has been reported to emerge. So far, the insect has been reared from three families of Lepidoptera: Noctuidae (14 species), Pluttidae (1 species), and Pyralidae (4 species). The authors have looked mostly at lepidopterous species of economic interest. Undoubtedly C. marginiventris host range will include many more species, possibly several in other lepidopterous families.

Table 1-2 also lists close to 30 different species of plants on

which the hosts for C. marginiventris were found. This relatively large host and plant range includes several serious pest insects and important








33

Table 1-2. Reported host species for C. marQiniventris.

HOST SPECIES i PLANTS COUNTRIES + common name(s) I (ref. #)

Noctuidae:
Heliothis virescens (Fabricius) alfalfa USA = tobacco budworm bicolor lespedeza (9,11,12,16,17, cranesbill 22,27,35,36,37, crimson clover 38,41,42,44,45, cutleaf geranium 46,48,52) deergrass
Ruellia runyonii
sesame
toadflax
tobacco
tomato
velvet-leaf vetch
Heliothis zea (Boddie) alfalfa USA = corn earworm, bollworm, tomato corn (11,12,16,17,22, fruitworm bicolor lespedeza 27,28,29,35,37, cotton 38,41,42,44,45, crimson clover 47,48,50,53,55) cutleaf geranium deergrass
peanuts
potato
Ruellia runyonii
soybean
toadflax
tomato
velvet-leaf
Plathypena scabra (F.) alfalfa USA = green cloverworm corn (7,15,24,26,29, legumes 34,36,47,53,54, soybean 56)
Pseudaletia unipuncta (Haworth) alfalfa USA = armyworm corn (20,36,47,53,54)
Pseudoplusia includens (Walker) alfalfa USA = soybean looper corn (9,18,28,29,50) pigweed
tomato








34

Table 1-2. Continued.
HOST SPECIES i PLANTS I COUNTRIES + common name(s) I (ref. #
Spodoptera fruqiperda (J.E.Smith) alfalfa USA(3,4,5,6,8, = fall armyworm bermuda grass 13,19,30,33,39, cabbage 43,47,53,54,55) corn BRAZIL (2) cotton COLUMBIA (3) peanut LESSER ANTILLES sorghum (2) MEXICO (39)
NICARAGUA (2)
SURINAM (2)
URUGUAY (2,3)
VENEZUELA (2,3)
Spodoptera eridania (Cramer) ? USA = southern armyworm (36)
Spodoptera exempta (Walker) grasses USA-HAWAII = nutgrass armyworm sugar cane (introd.) (8,14,36,40)
Spodoptera includens (Hubner) alfalfa USA = beet armyworm corn (1,17,20,47,50, pigweed 51,53,54) tomato MEXICO
Spodoptera litura (F.) beetroot INDIA cabbage (introd.) castor (21,23) cauliflower
cowpea
knol-kohl
tobacco
Spodoptera mauritia (Boisduval) bermuda grass USA-HAWAII = lawn armyworm (introd.)(14,49)
Spodoptera ornithoqalli (Guende) alfalfa USA = yellowstriped armyworm cotton (38,47,50,55) peanuts
sesame
Spodoptera Draefica (Grote) alfalfa USA = western yellowstriped armyworm (31,32)
Trichoplusia ni (H(bner) Brassica greens USA = cabbage looper cotton (9,10,18,25) weeds








35

Table 1-2 Continued.
HOST SPECIES PLANTS I COUNTRIES + common name(s) (+ ref. #)
Plutell idae:
Plutella xylostella (L.) Brassica greens USA = diamondback mothI (17,25)
Pyralidae:
Achvra rantalis (Guende) alfalfa USA = garden webworm (47)
Diaphania nitidalis (Stoll) squash USA = pickleworm (30)
Herpetogramma bipunctalis (F.) pigweed (in corn USA = southern beet webworm fields) (51)
Hymenia perspectalis (HQbner) ? USA = spotted beet webworm (36)

References:
1. Alvarado-Rodriguez (1987) 29. McCutcheon & Turnipseed (1981)
2. Andrews (1988) 30. McFadden & Creighton (1979)
3. Ashley (review) (1979) 31. Miller (1977)
4. Ashley et al. (1980) 32. Miller & Ehler (1978) 5. Ashley et al. (1982) 33. Mitchell et al. (1984)
6. Ashley et al. (1983) 34. Mueller & Kunnalaca (1979) 7. Barry (1970) 35. Mueller & Phillips (1983)
8. Bianchi et al. (1944) 36. Muesebeck et al. (1951)
9. Boling & Pitre (1970) 37. Neunzig (1969) 10. Boling & Pitre (1971) 38. Pair et al. (1982) 11. Burleigh (1975) 39. Pair et al. (1986) 12. Burleigh & Farmer (1978) 40. Pemberton (1948) 13. Butler (1958) 41. Puterka et al. (1985) 14. Clausen (review) (1978) 42. Roach (1975) 15. Danks et al. (1979) 43. Rohlfs III & Mock (1985) 16. Graham et al. (1972) 44. Shepard & Sterling (1972) 17. Harding (1976a) 45. Smith et al. (1976) 18. Harding (1976b) 46. Snow et al. (1966) 19. Hogg et al. (1982) 47. Soteres et al. (1984) 20. Hotchkin & Kaya (1983) 48. Stadelbacher et al. (1984) 21. Jalali et al. (1987) 49. Tanada & Beardsley (1958) 22. King et al. (1985) 50. Teague et al. (1985) 23. Krishnamoorthy & Mani (1985) 51. Tingle et al. (1978) 24. Kunnalaca & Mueller (1979) 52. Tingle & Mitchell (1982) 25. Latheef & Irwin (1983) 53. Vickery (1925) 26. Lentz & Pedigo (1975) 54. Vickery (1929) 27. Lewis & Brazzel (1968) 55. Wall & Berberet (1975) 28. McCutcheon & Harrison (1987) 56. Whiteside et al. (1967)








36

crop plants. The potential of this parasitoid to control these pests and protect the plants will be discussed later.


Geographical Range

C. marginiventris has been reported from Brazil, Columbia, Cuba, Mexico, Nicaragua, Surinam, Uruguay, the USA (Continental), Venezuela, and the West Indies (Wilson, 1933; Marsh, 1978; Ashley, 1979; Danks et al., 1979; Andrews, 1988). Within the continental United States reports have been mostly from the southern and central states. C. marginiventris has been collected in Alabama, Arizona, Arkansas, California, North and South Carolina, Delaware, Florida, Georgia, Iowa, Kansas, Louisiana, Mississippi, Missouri, Oklahoma, Tennessee, Texas, and Virginia (references in Table 1-2). Furthermore, C. marginiventris has been successfully introduced to Hawaii (Bianchi, 1944; Clausen, 1978) and India (Jalali et al., 1987).


Biology

The life cycle of a C. marginiventris female is depicted in Figure 1-2. The different stages of this cycle are each discussed in more detail in the following paragraphs.

Immature stages. Boling and Pitre (1970) described the immature stages. At oviposition the eggs are cylindrical with rounded ends (0.088 mm long, 0.017 mm wide), with a short, slightly bent, penducle (0.005 mm long) at the caudal end. After 24h the eggs increase in size to 0.041 mm by 0.0211 mm. The eggs float freely in the hemocoel of the host, and hatch 18 to 36 hours after oviposition at 30 *C.








37


















3






5 4












Figure 1-2. Schematic depiction of the life-cycle of C. marginiventris. 1) egg inside the host caterpillar, 2) second instar larva inside the host, 3) final (third) instar larva chews its way out of the host, 4) cocoon in which the parasitoid larva pupates, 5) adult parasitoid emerges from the cocoon, 6) adult female searches for the micro-habitats of potential hosts, 7) female examines host frass, 8) oviposition.








38

Boling and Pitre (1970) found the larval stages mostly in the

posterior part of the host, never in the head. The white caudate first instar larvae were never found attached to the host and measured 0.05 mm to 0.06 mm 24h after egg hatch. It possessed a caudal appendage as described by Clausen (1940). No cannibalism has been observed in this species, but the first instar larvae have a scleroterized head with distinct mandibles and labium. This suggests an ability to attack competitors as has been shown for related species (Allen, 1958). The first stadium lasts for 1.5 to 2 days.

The white vesiculate second instar is more robust, has no visible scleroterized head, and seems less suitable for physical combat. It possesses a very prominent anal vesicle. As is common in parasitic Hymenoptera (Clausen, 1940), the second instar significantly increases in size to 2.5 mm long and 0.59 mm wide, just before the moult to the third instar. The second stadium lasts approximately 2 days.

The moult to the third and final instar takes place just prior to emergence from the host. The third instar larva is 5.5 mm long and 1.0 mm wide, is creamy white at first and turns yellow to brown upon emergence. Color of both the larvae and the pupae appear to depend on the parasitoid's diet.

Parasitoid larvae emerge by biting their way out through the

cuticle of the host larva with their well developed mouth parts. When kept at 30 oC this occurs 6-11 days after egg deposition. The greatest number emerge from Heliothis virescens after 6 days, from Trichoplusia ni and Pseudoplusia includens after 7 days (Boling and Pitre, 1970), and from Plathypena scabra after 8 days (Kunnalaca and Mueller, 1979). At a








39

lower temperature emergence from the host takes longer, 9-15 days for P. scabra at 25 *C. Preliminary observations in our laboratory indicate an average emergence after 7 days from Spodoptera fruqiperda at 25 *C. Emerging parasitoid larvae immediately start spinning a crescent-shaped cocoon on, or in most cases in the vicinity of, the host. The parasitoid first constructs half a shell of the white or yellow cocoon and then crawls inside of it with its posterior end first. Subsequently, the open side is closed with up and down spinning movements followed by cross spinning to reinforce the sides (Boling and Pitre, 1970). To attach the cocoon to a substrate the larva will spin a fine mesh on the bottom when the case is still thin. Then it resumes the spinning of the cocoon. Two hours after it initiated cocoon formation, the larva, barely visible, was still spinning (Boling and Pitre, 1970). On corn and grasses the orientation of the cocoons is almost invariably parallel to the longitudinal axis of the leaf, and is usually placed on the median groove of the leaf (Bianchi, 1944). Approximately 24 hours after formation of the cocoon the parasitoid pupates (at 30 *C) and an adult will emerge from the cocoon 3-5 days later. A wasp exits from a cocoon by gnawing a smooth circular cut near one end, resulting in a conical trap door which is lifted up by the emerging wasp (Bianchi, 1944) (see drawing 5 in Figure 1-2). Data presented by Vickery (1925) emphasize how sensitive developmental time is to temperature. He found that time from oviposition until emergence from the cocoon takes about ten days during summer time, but can take as many as 27 days during cold months.








40
Adults. Developmental times from cocoon to adult takes 4-7 days at 25 *C (Kunnalaca and Mueller,1979). Males emerge on average a day before the females. Longevity for adults is clearly temperature dependent. Kunnalaca and Mueller (1979) found that in the laboratory when provided honey water at 30 OC and 25 �C the wasps would live 5.6 +

2.5 and 9.1 + 4.2 days, respectively. Females live longer than males (Kunnalaca and Mueller, 1979; Jalali, 1987).

Mating. Boling and Pitre (1970) observed that several minutes
after emergence from their cocoon, female wasps are willing to mate. A male approaches a female from the rear, vigorously fanning its wings and tapping her with his antennae. My observations differ from those by Boling and Pitre (1970), who claim that mounting by the male takes only half a second. In our laboratory copulation lasted roughly 5-15 seconds. We also found that once mated, females were not readily willing to mate again. This too contrasts with observations by Boling and Pitre (1970) who observed both sexes to mate freely and many times. Further detailed studies will have to elucidate these contradicting observations. It is possible that females are willing to mate again after they have oviposited in a host. Loke and Ashley (1984a) found that mated females respond more intensely to contact kairomones than unmated females, and they are more ready to search for hosts.

Host finding behavior. Studies on host finding behavior by C.
marginiventris have only been performed at the close-range level (Loke et al., 1983; Loke and Ashley, 1984a, 1984b, 1984c; Dmoch et al., 1985). Loke et al. (1983) described the host-finding behavior of C. marginiventris on corn plants and reported that parasitization is








41
stimulated by plant damage caused by fall armyworm larvae. Females spend significantly more time on leaves damaged by hosts than on artificially damaged leaves or undamaged leaves, which was also reflected in a higher parasitization rate on the host damaged plants. Loke et al. (1983) divided the behavioral sequence of close-range host finding into 12 steps (Figure 1-3). The asterisks in Figure 1-3 represent locations where fixed-action patterns were mediated by stimuli that release specific behavioral patterns. Although this analysis of the different behavioral steps looks very acceptable, Loke et al. (1983) concluded too soon that chemotaxis is involved. However, to draw such a conclusion, it is necessary to demonstrate that contacting certain chemicals isolated from the habitat is sufficient to evoke a specific part of the host-finding behavior. To do this, Loke and Ashley (1984b) isolated several components of the host habitat complex, made extracts out of them, and studied the behavior of female parasitoids when they came in contact with the extracts on filter paper. It appeared that C. marginiventris responded to all of the components of a host habitat with hosts. The strongest responses were elicited by 1) fall armyworm frass

2) moth scales

3) and exuviae
The females were also found to be responsive to damaged leaves, saliva, silk and larval and cuticle material. Loke and Ashley (1984b) found hardly any difference in how C. marginiventris responded to frass produced by larvae fed on different plants. However, they did find that plant material is needed as food for the hosts to produce frass with the








42


(4)
contact
- - damage antennal (5) * ( palpation chemotaxis i


\\ \ ovipositor (6) *
(2) orientation \ probing
I ,.

(1) nonsearching
movement <.. preening <----- contact (7) host


(12) resting
mounting (8) *


(11) ovipositor stinging (9) *
smeering - N oviposition
(10) *






Figure 1-3. Behavioral ethogram of the host finding and ovipositional sequence of C. marQiniventris females on corn plants damaged by fall armyworm larvae. (Solid arrows indicate invariable pathways ; dashed arrows, variable or alternate pathways; asterisks, sign stimuli present).
Taken from Loke et al. (1983).








43

right kairomones. It appeared that the parasitoids were less responsive to the frass of larvae fed on a lab diet. Frass extracts can significantly increase parasitization rates by C. marginiventris. Even the normally not accepted factitious hosts velvetbean caterpillar and wax moth are parasitized if they are sprayed with extracts of fall armyworm frass (Loke and Ashley, 1984c).

Dmoch et al. (1985) confirmed some of the conclusions of Loke and Ashley (1984b), but also demonstrated the importance of learning in the host-finding process for C. marginiventris. This phenomenon of learning forms one of the key elements of the research presented here.

Oviposition. Females are ready to oviposit shortly after emerging from the cocoon, but are more inclined to do so a day or more after emergence (Boling and Pitre, 1970; Loke and Ashley, 1984a). Females on a substrate with host larvae travel quickly over the substrate while touching it vigorously with their antennae. Upon finding a host they will insert one egg with a quick thrust, and subsequently move away from the host. The wasps appear not to be egg limited, and will lay eggs throughout their adult lives (more than 100 eggs under laboratory conditions) (Kunnalaca and Mueller, 1979; Jalali et al., 1987). Miller (1977) concludes that C. marginiventris can lay viable eggs within 24h after emergence from the cocoon. He also found that the female wasp caries 20 + 4 ovarioles per ovary and a total of 210 + 25 ova. Females will lay just one egg during an oviposition attempt but may return to the same host several times to lay more eggs. Dmoch et al. (1985) present evidence showing that experienced females are able to discriminate parasitized and unparasitized hosts. The female wasps









44

oviposit in unparasitized host more readily than in already parasitized hosts. However, Boling and Pitre (1970) found up to 7 eggs in the hemocoel of a single host, although normally only 1 egg was found. They also report a few cases where 2 eggs were dissected from field collected larvae. Normally only one parasitoid will emerge from one host. Kunnalaca and Mueller (1979) found just one example of two wasps emerging from a green cloverworm larvae that had been parasitized more than once in its 3rd-instar. These authors also found that Ist- and 2nd-instar were the preferred larval stages for oviposition by C. marginiventris. Loke and Ashley (1984a) observed the same preference for early instars of fall armyworm, while Jalali et al. (1987) came to the same conclusion when they looked at host age preference for the novel host Spodoptera litura in India. The preferences were solely concluded from parasitoid emergence after choice tests. The observations may therefore have been affected by differential survival rates.

Parasitism by C. marginiventris occurs primarily during daylight hours (Kunnalaca and Mueller, 1979). Wasps were allowed to oviposit in green cloverworm larvae from 6-8 a.m., 6-8 p.m., and 10-12 p.m. They parasitized 49%, 68%, and 3% respectively. This is largely in agreement with findings by Loke and Ashley (1984a), who found that the female wasps responded to contact kairomones throughout the photophase.


Host Requlation

Vinson and Iwantsch (1980) introduced the term "host regulation"

which refers to the many physiological and biochemical changes that take








45

place in a parasitized organism. Considering the wide range of hosts in which C. marginiventris can develop successfully, a sophisticated host regulation process can be expected for this parasitoid.

From our own observations, we can conclude that the parasitoid

dramatically slows down the growth, food utilization and, developmental rate of the host larva. Since the C. marginiventris attacks its host in an early stage this will eventually result in a enormous decrease in plant consumption, something we will consider later when the parasitoids' potential as a bio-control agent is discussed.

So far, Ferkovich et al. (1983) undertook the only serious attempt to study the internal changes in hemolymph proteins of Spodoptera fruqiperda. They detected that as early as 4 hours after parasitization several high-molecular-weight proteins were present in the host hemolymph that showed up much later in unparasitized larvae. Two new protein bands that never appeared in the electrophoretograms of unparasitized host showed up in parasitized hosts after 8 hours and beyond. Clearly, the parasitoid affects the host bio-chemistry the significance, but sources of these changes are far from being determined.

One of the possible causes of changes in the hosts are symbiotic baculovirus particles that are injected with calyx fluid into hosts by many braconid parasitoids (Stolz and Vinson, 1979; Stolz et al., 1981). Styer et al. (1987) discovered a long, filamentous virus (CmFV) replicating in the hypodermal and tracheal matrix cells of larvae parasitized by C. marginiventris. Hamm et al. (1990) describe this non-occluded baculo-like virus and a polydnavirus which both originate









46

from the reproductive tract of the parasitoid and are injected with her calyx fluid into the host. The polydnavirus appear not to replicate inside the host (Styer et al., 1987). As has been demonstrated for other parasitoids (Stolz and Vinson, 1979; Tanaka, 1987), the viruses are thought to suppress the immune responses of the host and effect the host physiology to allow parasitoid development.

Bio-control

Control potential. Several examples exist that indicate that

natural populations of C. marginiventris already prevent pest outbreaks. Whiteside et al. (1967) found that C. marginiventris is one of the most important parasitoids of the green cloverworm in Delaware. Kunnalaca and Mueller (1979) suggest that due to the parasitization activity of C. marginiventris green cloverworm populations seldom reach economic importance in Arkansas. McCutcheon and Turnipseed (1981) concluded that C. marginiventris is the most important parasitoid of lepidopterous species in South Carolina. They collected larvae of the green cloverworm, corn earworm, soybean looper, and velvetbean caterpillar from soybean fields in South Carolina. Of the thirty-three species reared from these larvae, C. marginiventris was the most predominant parasitoid for the first three pest species. C. marginiventris has never been reared from velvetbean caterpillar, but McCutcheon and Turnipseed (1981) reared one parasitoid with a cocoon similar to C. marginiventris that did not develop to maturity. C. marqiniventris is also one of the most common larval parasitoids of two other major pests Heliothis virescens and H. zea (King and Coleman, 1989).









47

C. marginiventris should be particularly effective as a control agent against one of its main hosts the fall armyworm (Gross and Pair, 1986; Lewis and Nordlund, 1980). It is one of the most predominant early attackers of fall armyworm particularly in whorl-stage corn (Ashley, 1979; Hogg et al., 1982; Pair et al., 1986). If present in sufficient numbers when host densities are relatively low, it should be able to prevent pest population buildup. Gross and Pair (1986) characterize C. marginiventris as an r-strategist (i.e. reproductive investments seem to be mostly concentrated on quantity, rather than quality), considering its fast developmental rate, relatively small size, and the females' large egg load, this would be legitimate. As an r-strategist its population buildup should be relatively fast. In general r-strategists should be given priority in introduction and augmentation programs, especially when they are capable of preventing early host population buildup (Pschorn-Walcher, 1977; Ehler and Miller, 1978). For the control of fall armyworm Gross and Pair (1986) suggest the augmentative release of a parasitoid such as C. marginiventris in the host's overwintering areas, thereby preventing dispersing populations from reaching economic levels throughout the geographic range. More specifically, Knipling (1980) proposed releasing 2,000 larval parasitoids per acre in overwintering areas, which should result in effective suppression of fall armyworm.

C. marginiventris has an early effect on its hosts. Ashley et al. (1983) found that of the parasitoids that attack fall armyworm, C. marginiventris allows the least amount of development of the host. Parasitization reduces maximum larval weight by 97 %, and results in the









48

destruction of the host when it reaches the 4th instar (Ashley, 1983). C. marginiventris reduces host development significantly more than the other common fall armyworm parasitoids, Campoletis grioti, Chelonus insularis, and Eiphosoma vitticole. Jalali et al. (1988) demonstrated that Spodoptera litura larvae also consume significantly less food 72hr after parasitization by C. marginiventris than unparasitized larvae. At a later stage the reduction in consumption becomes as high as 97 %. Hence, C. marginiventris not only prevents pest population breakouts, but reduces the destructive potential of the existing population as well.

Costs. Despite its potential to reduce crop damage, C.

marginiventris has not yet been used in augmentative release programs. This is probably mainly because of the great expenses involved in its rearing by conventional means (Greany et al., 1984). Costs could be dramatically reduced with in vitro rearing, but this technique has so far only worked for egg and pupal parasitoids (Thompson, 1986). With the exception of Trichogramma spp., natural enemies of lepidopterous pest can not yet be reared in adequate numbers for augmentative field releases. Studies to develop in vitro rearing for larval parasitoids including C. marqiniventris are currently being conducted (Greany, 1986; Greany et al., 1989). However, even if costs of rearing parasitoids such as C. marginiventris remain high, their potential as control agents should be considered. Well timed releases of carefully manipulated wasps may have significant effects on pest populations, even when they are only released in small numbers. Costs of pesticide applications are









49

and their subsequent detrimental effect to the environment are also very high.

Bio-control experts agree that this wasp has the potential of

becoming a major control agent, particularly if maintenance of effective parasitoid populations can be enhanced (Lewis and Nordlund, 1984). McCutcheon and Turnipseed (1981) state "studies to determine factors which may influence the abundance and parasitic activity of C. marginiventris should enable us to develop more effective management strategies for lepidopterous pests (p. 73)." From Loke and Ashley (1984a, 1984b, 1984c) we learned that within the micro-habitat of its hosts the female wasp uses contact kairomones to track its hosts. No information is yet available on what strategy C. marginiventris relies to locate the (micro)habitats of its hosts. Studies on other parasitoids demonstrate more and more that semiochemicals are the major cues used by female parasitoids to locate host habitats. To elucidate the complete strategy used by C. marginiventris to find its hosts and to determine which semiochemicals are involved, a combined behavioral and chemical study was conducted. The results of that study are presented in this dissertation. The following section states the research aims of that study.


Research Aims

The goal of the research was to obtain information that will

contribute to the development of effective biological control of pest insects by augmentative releases of indigenous parasitic insects. So far, this type of biological control has only been marginally successful









50

because of some reoccurring problems. Most attempts have failed because the wasp disperse upon release and exhibit poor searching efficiencies. To master these problems, more knowledge of the host searching behavior of the parasitoids is required, and techniques need to be developed that will enhance their performances.

The potential of C. marginiventris to become a major control agent has been well established. However, the only information available on the host-searching behavior of C. marginiventris pertains to its closerange searching behavior (i.e. within the micro-habitat of its hosts). The success of a released parasitoid, however, is largely dependent on the efficiency with which it finds the micro-habitats of the target hosts. Long-range host searching behavior is therefore the link in a parasitoid's life-cycle that requires study for possible improvement.

More specifically, techniques need to be developed that will

enable us to manipulate the behavior of C. marginiventris females such that after release in a pest infested area they are retained in that area and will focus their searching on the target pest. As shown earlier, preliminary research indicates that parasitic wasps can be conditioned to respond more intensely to specific host-related cues. In olfactometric studies this possibility needs to be investigated for C. marginiventris. These studies should reveal whether the search efficiency of this parasitoid can be increased by experiencing the wasp with host or host by-products, and whether the preferences for certain volatile cues are determined by such experiences.

Pre-release conditioning appears only practical by artificial

means, therefore it will be necessary to identify not only the chemicals








51

that induce this phenomenon, but also the chemicals that wasps subsequently respond to. In future release programs such information would allow the conditioning of wasps with artificial substrates impregnated with synthetic chemicals. In an ideal situation this technique would not just increase responsiveness of the wasps, but also help the wasps to concentrate on the targeted pest.

The specific aim of the present study was to determine which

chemical cues and behavioral mechanisms enable C. marginiventris females to locate hosts over longer distances. Aspects of the long-range host searching behavior of C. marginiventris were studied with the intent to do the follwing:

1) determine to what degree this behavior is mediated by

volatile semiochemicals associated with hosts,

2) find out how experience with these chemicals affects the

insect's behavior and preferences for specific odors, 3) describe the flight responses exhibited by C.

marginiventris females to host-related odors,

4) establish the exact source of the chemicals that mediate this

behavior,

5) isolate, analyze, and identify the active semiochemicals, and 6) elucidate the different interactions between the three trophic

levels involved (plant, host, and parasitoid) and determine how

these interactions affect successful host location by C.

marqiniventris.














CHAPTER II

INCREASED RESPONSE TO HOST RELATED ODORS AFTER
A FORAGING EXPERIENCE: A MATTER OF LEARNING?


Introduction

In many species of insect parasitoids semiochemicals associated with the hosts and/or the host habitats play a major role in the host location process (Vinson, 1976, 1984). The optimal way for a parasitoid to take advantage of the available semiochemicals may vary in time. For those parasitoid species which can develop in more than one host species, the host species that is predominant at one time may be rare or absent at another. Also, the habitats that the hosts tend to occupy may vary greatly. To optimize host finding, parasitoids should be able to adapt to such changes. This would require the ability to adjust their response to certain semiochemicals.

Indeed, several studies on host habitat and host location have

shown that host searching behavior can be modified by experience. Adult female experience with hosts and host-related substrates have been shown to alter subsequent responses (Thorpe and Jones, 1937; Monteith, 1963; Arthur, 1966, 1971; Taylor, 1974; Vinson et al., 1977; Sandlan, 1980; Strand and Vinson, 1982; Vet, 1983, 1988; Vet and van Opzeeland, 1984, 1985; Wardle and Borden, 1985; Dmoch et al., 1985; Drost et al., 1986; Lewis and Tumlinson, 1988). Learning and conditioning are the terms used most often to describe this effect.


52









53

Recent studies have shown that the flight response of Microplitis croceipes (Cresson) and M. demolitor (Cresson) to host related odors in a flight tunnel, increases significantly after the parasitoids have a brief contact experience with host by-products (Drost et al., 1986; Hdrard et al., 1988a). Oviposition is not required to evoke the increase in response. In fact, increase in response was less if the parasitoids oviposited on larvae that were removed from their habitat. Preliminary observations on the host searching behavior by C. marginiventris indicated that this larval parasitoid of many lepidopteran species also performs better after a similar type of experience.

The exact function and the mechanisms behind the phenomenon of increased responsiveness after experience have yet to be elucidated. One possibility is that perception of certain chemicals in the host by-products sensitizes the females and stimulates them to enter a searching mode, which results in the observed increase in response. It is also possible that during the contact experience the females learn to associate the odors emitted by hosts and their habitat with the presence of hosts. In order to test these two alternative hypotheses, we studied the effects of pre-bioassay contact experience on the response to host related odors by C. marginiventris in a four-arm olfactometer.


Material and Methods

Parasitoids. C. marginiventris (strain Mississippi) were reared on fall armyworm larvae at the USDA-ARS, Insect Biology and Population Management Research Laboratory, Tifton, Georgia, according to the








54

procedure described by Lewis and Burton (1970) for M. croceipes. The pupae were held in 25 X 25 X 25 cm, plexiglass cages, with one side made of fine mesh nylon screen. The parasitoids were allowed to emerge in the cages and were held at 26*C, 50-60% RH and a 15-hr photophase. Only parasitoids that emerged on the same day were kept in the same cage. Males were removed after 2 days, allowing sufficient time for all females to be mated. All experiments were conducted with mated females that were 3 to 4 days old, 7-10 hr after they experienced lights-on.

Hosts. The hosts used in the experiments were second instar

larvae of the fall armyworm (FAW), Spodoptera frugiperda (J.E. Smith), and of the cabbage looper (CL), Trichoplusia ni (Hubner). They were reared according to the method described by King and Leppla (1984). Initially, the larvae were fed on a laboratory-prepared pinto bean diet. Then, for about 18 hr prior to testing, those larvae that were used as part of the odor source were allowed to feed on leaves. Thus, the FAW larvae were fed corn (Zea mays L.) leaves, and the CL larvae were fed cotton (Gossypium hirsutum L.) leaves. The leaves were obtained from plants that were 2-4 weeks old.

Olfactometer. Bioassays were performed in an airflow olfactometer similar to the one described in detail by Vet et al. (1983) with some modifications described by Eller et al. (1988a). It consisted of an exposure chamber connected to four arms through which air flowed into the chamber. The air was pulled out through a center hole in the bottom of the chamber. By balancing the airflows (300 ml/min through each arm), four distinct flow fields were created in the exposure chamber. Each arm was connected to a flowmeter, a water bubbler (to humidify the








55

air), an odor chamber and a catching jar. Materials tested as sources for semiochemicals were placed in the odor chambers. Parasitoids that walked up the arms were captured in the catching jars.

Female parasitoids were introduced through the vertical entry tube after temporarily disconnecting the extraction tube. While walking up the entry tube the test animal was exposed to a mixture of the four flows until it reached the chamber floor, where it moved freely, exploring the different flow fields. One wasp was introduced at a time and its behavior was observed during a 5-min period. If the test animal showed a positive anemotactic response and walked up one of the airflows and out of the exposure chamber it was recorded as a final choice if the female did not return within 15 sec. The temperature in the bioassay room was 28�1*C at all times.

Odor Source. In all experiments, only a single odor source was used in one of the olfactometer arms. The three remaining arms served as controls with only humidified air going through. The test odor source contained five second-instar larvae feeding on young leaves. Either FAW on corn leaves or CL on cotton leaves were used. The larvae were starved for 1 hr before being introduced to the odor chamber with one of the already damaged and contaminated leaves and a fresh seedling. The larvae were allowed to eat for 1.5 hr before wasps were tested in the olfactometer.

Experience. Females were provided with the following types of

experiences prior to testing: 1) no experience with any hosts or host products (INEXP), 2) one oviposition on a second instar CL larva feeding on cotton (CLOVIP), 3) a 20-sec contact experience with a cotton leaf









56

damaged and contaminated by CL larvae (CLDAMAG), 4) one oviposition on a second instar FAW larva feeding on corn (FAWOVIP), and 5) a 20-sec contact experience with a corn leaf damaged and contaminated by FAW larvae (FAWDAMAG). The leaves used in the oviposition and damage experiences were equally damaged by larvae and contaminated with larval by-products to ensure that oviposition was the only difference between these two types of experience.

The females were tested 30-60 sec after they had their experience. Five females of each treatment were tested per day. The two odor sources were alternated between days, as were the arms containing the odor. A total of 30 females were tested for each treatment to each odor source.

Data Recording. The behavior observed in the olfactometer was recorded with an Epson@ Geneva PX-8 portable computer. Response was measured in two ways: 1) the time spent in the quadrant of the introduction chamber containing the odor and, 2) the number of final choices made for the arm through which the odor was entering the exposure chamber.

Statistical analyses. The five treatments were compared with Duncan's new multiple range test after analysis of variance (ANOVA) (Steel and Torrie, 1960). The two response measures were analyzed using: 1) the percent of time spent in the odor field by each individual female and, 2) for each treatment, the percentage of the five females tested daily that made a final choice. The percentages were transformed using the arcsin-square root transformation for statistical analysis.








57

To compare oviposition experience with damage experience, and FAW experience with CL experience, pooled means were analyzed by orthogonal comparison of the sums of squares (Chew, 1986). Significance levels were 0.05 in all tests.




Results

Table 2-1 shows the observed responses for females that had different types of experiences prior to their introduction in the olfactometer. In most cases prior experience was associated with females spending more time in the odor flow and making more final choices for the odor arm. With some exceptions for the time spent in the odor quadrant, an increase in response was significantly less when the females experienced the alternative plant-host complex.

For a better comparison of the different types of experiences, the appropriate means in Table 2-1 were pooled and analyzed by orthogonal comparison of the sum of squares (Chew, 1986), resulting in the following.

Oviposition vs. Damage Experience (Fiqure 2-1). A comparison

between all females that had an oviposition experience and those that only experienced the host-damaged leaves showed that the increase in response did not require contact with a host. For both odor sources, CL on cotton and FAW on corn, no differences were found in the percentage of time spent in the odor quadrant (F=0.0186, F=0.0047) and the percentage of final choices for the odor arm (F=0.79, F=0.05).








58






Table 2-1. Effect of pre-bioassay experience on response of C. marginiventris females exposed to host related odors in a four-arm olfactometer. Response is expressed in 1) percentage of females that made a final choice for the odor (% F.C.), and 2) percentage of the total time spent in the quadrant with the odor flow (% in odor).


Odor
CL feeding on cotton FAW feeding on corn Experience I n % F.C. % in odor % F.C. % in odor INEXP 30 20.0 a 59.4 a 13.3 a 46.9 a CLOVIP 30 73.3 c 85.6 b 43.3 b 69.9 b.c CLDAMAG 30 70.0 c 87.2 b 36.7 b 62.0 a,b FAWOVIP 30 46.7 b 67.1 a 60.0 c 79.2 c,d FAWDAMAG 30 40.0 b 66.7 a 70.0 c 86.8 d Values with the same letter in each column are not significantly different (Duncan's new multiple range test after ANOVA with arcsin-square root transformation of percentages [Steel and Torrie, 19601]).










59

OVIPOSITION vs. DAMAGE EXPERIENCE


' NO 1 OVIPOSITION E2 DAMAGE
EXPERIENCE EXPERIENCE EXPERIENCE
80

z
70


60 //


50


S 40


30
70
a:
0
O 60
0

0 50 U)
o 40

O
30


- 20


10


ODOR SOURCES





Figure 2-1. Experience effect upon response to odors of two plant-host complexes. Shown are the responses of females that 1) had no experience with hosts or hosts products prior to a test, 2) that were allowed to oviposit in one larva (either CL or FAW) feeding on leaves, and 3) that only contacted leaves damaged by CL or FAW larvae prior to a test.









60

CL on Cotton vs. FAW on Corn Experience (Figure 2-2). For each odor source, all females that experienced CL on cotton were compared with those that experienced FAW on corn (oviposition experience and damage experience lumped together). This resulted in highly significant differences for both the time spent in the odor quadrant (F=14.90, F=9.70) and the percent of females that made a final choice for the odor arm (F=25.50, F=12.71). Figure 2-2 shows that the wasps responded much better when exposed to the odors of the plant-host complex with which they have had experience.


Discussion

The response of C. marginiventris females to odors of host larvae feeding on leaves can be increased dramatically by allowing the females to contact host products prior to testing in an olfactometer. Actual encounters with the hosts are not required to evoke this increase in response (Figure 2-1). Apparently, the presence of larval by-products is sufficient to trigger the mechanism behind the increase in response. Furthermore, the increase in response is evoked instantaneously, or at least within 20 seconds. This shows that the process must be a powerful and important modifier of the insects' behavior.

The parasitoids respond best to the odors of the plant-host

complex that they experienced (Figure 2-1). Therefore, the results suggest that we are not merely dealing with a general increase in search motivation (= sensitization), but that associative learning must be involved. It is suspected that the learning process is triggered when a parasitoid contacts a specific kairomone. The parasitoid then links the









61

CL ON COTTON vs. FAW ON CORN


IZI NO EM CL-COTTON 2 FAW-CORN
EXPERIENCE EXPERIENCE EXPERIENCE

90


80 70 60
0



40 30
80

O 70
8
60 U) 50 O 40

-j 30

20 10
CL ON COTTON FAW ON CORN ODOR SOURCES




Figure 2-2. Experience effect upon response to odors of two plant-host complexes. Shown are the responses of females that 1) had no experience with hosts or hosts by-products prior to a test, 2) that experienced CL on cotton (either oviposition or damage), and 3) that experienced FAW on corn (either oviposition or damage).









62

present odors with the possible presence of host larvae. Subsequently, the wasps will use those odors as cues in the search for more hosts. Evidence that supports this theory has been presented for the parasitoid M. croceipes (Lewis and Tumlinson, 1988).

The fact that experience with an alternative host still causes some increase in response (Figure 2-2), suggests that associative learning is not the only process involved, but that sensitization takes place also. Alternatively, however, some of the semiochemicals emitted by one plant-host complex might be the same or similar to those emitted by the other complex. Therefore, learning alone may still explain why experiencing one plant-host complex increases the response to semiochemicals emitted by another plant-host complex.

In the laboratory adult females seem to require a learning experience to make them highly sensitive to host odors. Learning might also take place during or immediately following emergence, when the parasitoid contacts by-products of its own host on the cocoon and on the emergence site. The insects used in the experiments were reared on FAW larvae feeding on artificial diet. The information these parasitoids obtained during emergence might not be adequate to make them respond to the odors to which they were exposed during the experiments. This could explain the relatively poor response by inexperienced females. Females in the field may obtain enough information in an earlier stage to make them more responsive to the appropriate semiochemicals. In fact, Drost et al. (1988) and Hdrard et al. (1988b) showed for M. croceipes and M. demolitor, respectively, that inexperienced adult wasps reared from larvae feeding on cowpea leaves responded better than those reared on









63

larvae feeding on artificial diet, when exposed to odors emitted by larvae feeding on cowpea. The effect of rearing on the response levels is likely to be of importance for C. marginiventris as well. However, considering the host range of this species and the many plant species these hosts might feed upon, modification of response to semiochemicals after experience by an adult female should contribute to the female's host searching efficiency. Therefore, it is likely that adult females in the field continue to increase and adjust their response to semiochemicals each time they contact host by-products.














CHAPTER III

EFFECTS OF FORAGING EXPERIENCES ON ODOR PREFERENCES


Introduction

Females of many insect parasitoids rely on host and host-habitat related chemicals as cues in their search for hosts (for reviews see Vinson, 1976, 1981, 1984; Weseloh, 1981; van Alphen and Vet, 1987). Several studies have demonstrated that the response to these semiochemicals is flexible and can be influenced by learning (see previous chapter). These studies show that experiences with hosts and/or their microhabitats, both by immature and mature stages, may influence an adult parasitoid's response to semiochemicals.

It was shown in the previous chapter that C. marginiventris

(Cresson) shows a significant increase in response to host-related odors after only a brief contact experience with host damaged leaves contaminated with host by-products. After females receive a contact experience with a particular plant-host complex, they respond significantly better when they are exposed to the odors of that plant-host complex than when they are exposed to the odors of an alternative plant-host complex. This suggests that the experience effect is not merely the result of a general increase in response to semiochemicals, but that the insects actually learn to respond to the odors that they encounter during their experience.


64









65

The phenomenon of conditioning through experience has been

suggested as a useful method for biological control programs in which parasitoids could be stimulated to respond to host-related odors prior to their release in a target area (e.g., Lewis and Nordlund, 1985). It would be particularly helpful if the wasps would not just show an increase in their responses to the experienced odors, but would actually prefer these odors over those released by alternative plant-host complexes when given a choice.

This chapter reports on a study in which the effect of experience on odor preferences by C. marginiventris was tested. Shifts in preference in favor of an experienced odor were studied in situations where only the host species was varied and in situations in which only the plant was varied. Thus, we obtained information on the specific roles played by both host larvae and plants in the production and release of semiochemicals essential for host-habitat location.


Materials and Methods

Population of C. marginiventris. Parasitoids of the '85

Mississippi strain were reared on fall armyworm larvae as described in chapter II. All experiments were conducted with 3- to 5-day-old mated females, 6-10 hr into the photophase.

Hosts. The hosts used in the experiments were second instar

larvae of the fall armyworm (FAW), Spodoptera fruqiperda (J.E. Smith), and of the cabbage looper (CL), Trichoplusia ni (Hubner). They were reared according to the procedure described by King and Leppla (1981).








66

Olfactometer. Individual females were exposed to host-related odors and observed in a four-arm olfactometer as described in chapter II. The fraction of an odor-containing flow that actually reached a parasitoid in the olfactometer could be controlled by diverting part of the flow before it entered the arena. The fraction that was split off was then replaced by clean humidified air. The total flow entering the central exposure chamber through each arm was kept at 300 ml/min. Further details are given in the previous chapter.

Odor sources. The odor sources consisted of 5 late second or early third instar larvae feeding on three seedlings. The larvae of either FAW or CL were put on the seedlings of either corn (Zea mavs L.) or cotton (Gossypium hirsutum L.) 1.5 hr prior to the actual bioassays.

Data recording. The behavior of the females in the olfactometer was recorded with the use of an Epson� Geneva PX-8 portable computer. After a female was introduced into the olfactometer the time it spent in each odor quadrant was recorded during a 5-min period. If the parasitoid walked into one of the arms and did not return within 15 sec, this was recorded as a final choice for that arm. The remaining time was added to the time spent in the quadrant of the final choice. For the dual choice tests the odor quadrant in which a female spent the greatest amount of time was recorded as her odor field preference.


Procedures and Results
Dose-response experiments. Before actual odor preferences were tested, the responses to different odor doses were tested for each plant-host complex separately. Thus, dose-response curves could be









67

generated to determine which concentrations of the different plant-host complexes evoke similar response levels. The optimum concentrations were then used in the preference experiments to reduce the possibility that preferences were influenced by concentration differences.

The responses of female parasitoids were observed to the odors of FAW on corn, FAW on cotton, CL on corn, or CL on cotton. During each test the odor of one of these complexes was offered through one of the four flows, while the other flows contained humidified air and served as controls. For each of the four odor sources three concentrations were tested; 25, 50 and 100% of the original odor flow.

Just prior to being tested, a female was placed, for 20 sec, on a plant-host complex like the one used as the odor source. The parasitoids were prevented from actually encountering hosts. This type of experience significantly increases a female's response to hostrelated odors as was shown in the previous chapter II. The female was then introduced into the olfactometer. Odor sources were rotated to the next air flow after 6 females were tested for each of the three concentrations of a particular odor. A total of 24 females were tested for each concentration of all odor sources.

Figure 3-1 shows that the wasps responded in a dose-related manner to all four complexes. Regression analyses of the time spent in the odor quadrant (Figure 3-ib) and of the time it took the females to make a final choice (Figure 3-1c), show a significant increase in responsiveness to the odors with increasing odor dose. The total number of final choices made for the odor arms was also found to be significantly dose related, with the exception of CL on cotton (Figure












S100 68
C, a


90
go
S--+-- FAW + corn D y=0.36x+66.66: r=.72: p<.01 SA80 - - FAW + cotton
S0.37xO 64.58: r=.77: p<.01
-E- CL + corn
/y=0.40x+62.50 r=.71: p<.01
70
70 0 CL + cotton
y=023x+77.09; r=.49: ra

60 __0 25 50 75 100


0 100
.j


S90
0
z
o -+-- FAW + corn
8j 80 0 y=0.19x+278.96: r=.26: p<,05
P- - FAW + cotton
0y=O32x+263.06: r=.24: p<.05
70 -e-- CL + corn
4 y=0.24x+274.63; r=30; p<.05
W O CL + cotton
y=0.66x+218.65: r=.24; p<.05 60 _ _ _ _ _ _,
0 25 50 75 100

0 -+- FAW + corn
UJ 200 y=-1.06x+ 186.00: r=.33; o<.01 C ( -- FAW + cotton I W*\ y=-1.20x+ 196.71: r=.44; p<.01 "\ -- - CL + corn ' \ y=-1.54x+212.73; r=.49; p<.01
-150 CL + cotton





u.



> 50
0 25 50 75 100


ODOR DOSAGE (%)


Figure 3-1 Responses of experienced C. marqiniventris females to 3 doses of odors emitted by larvae feeding on leaves. Responses were measured as; a) Average percentage of the females that made a final choice for the odor arm; b) Average percentage of time that the females spend in the quadrant with the odor; c) Average time it took a female to make a final choice. The drawn lines connect the average values, while the equations for the actual linear regressions are given with each graph.









69

3-la). No significant differences were found between the same doses of the four odor sources. In all of the following preference tests 50% doses were used.

Effect of short contact experience on odor preference. The effect of a short-term contact experience on a female's preference for the odors emitted by different plant-host complexes, was tested in a series of dual choice experiments. A female wasp was allowed to contact for 20 sec, a plant-host complex from which the larvae had been removed. Immediately following this contact experience, the wasp was transferred to the olfactometer. In the olfactometer the wasp was exposed to the odors of two different plant-host complexes introduced through adjacent arms of the olfactometer, the two remaining arms carried humidified air only. One of the odors was from the complex that the female had just encountered, the other odor from a different complex. One combination of two odor sources was tested on a given day, with 8-10 females that experienced one source, and a same number of females that experienced the other source before being introduced into the olfactometer. This was replicated 6 times for each combination.

The results, as summarized in Figure 3-2, show that the

probability that a female chooses the odor of a particular complex is higher if she has had experience with that complex than if she has had experience with the other complex. The T-test (SAS Institute, 1987) was used to make an overall comparison of the responses to the odors of complexes with which females had experience to those with which they had no experience. The differences were highly significant for both the number of final choices (N=6; T=4.17; p=0.009) and the odor field









70





a B Odor A EM Odor B Different hosts on same plant I. A = FAW on corn B = CL on corn
U. A = FAW on cotton 8 = CL on cotton
Same host on different plants IIL A = FAW on corn B = FAW on cotton IV. A = CL on corn B = CL on cotton
Different hosts on different plants V. A = FAW on corn B = CL on cotton VI. A = FAW on cotton 8 = CL on corn

C d
C II III IV V VI 1oo % I 1I III IV V VI 10 W a*140$
i t (12O1 ( 3 ( ) ( ) ( ( 4
J ( 3 12 7
A (odo
Odor A dor
I L

SA A A A AS1g AS AS A AS A A2 1 w (is121 Od or B 112 LL4 L -1 (141 (19) 115) Odor (261 4231 126) (2M t27 8
100 % 100 % A B

ODOR SOURCE CONTACTED PRIOR TO PREFERENCE TEST ODOA SOURCE CONTACTED PRIOA TO PREFERENCE TEST



Figure 3-2. Effects of experience on the preference of C. marginiventris females for host-related odors. a) Diagram of the olfactometer with the odors of two different planthost complexes entering the exposure chamber through adjacent arms. b) List of the six plant-host complexes that were tested. c) Summary of results using the percentage of females that made a final choice for a specific odor arm as the measure of response. Females that did not make a final choice were excluded. d) Summary of results using the percentage of females that spend most of their time in a specific odor field (= odor field preference) as the measure of response. The few females that never entered one of the two odor fields were excluded.
The roman numerals (I-VI) in both c) and d) refer to the combinations listed in b). In c) and d) the bars above the x-axis represent the females that choose odor.A and below the x-axis the females that choose odor B. The actual numbers are shown in parentheses. For each combination 48 to 50 females were tested of both experience types. The asterisks indicate significant differences in total numbers due to experience (chi-square; p < 0.05).









71

preference (N=6; T=4.69; p=0.005). However, for each individual combination the differences in odor preferences were not always significant (Figure 3-2). For the numbers of final choices made for the odors, a significant difference was found only for the combinations FAW on corn versus CL on corn (x2df1=8.624; p=0.003) (I, Figure 3-2c) and CL on corn versus CL on cotton (x2df1=3.907; p=0.048) (IV, Figure 3-2c). Odor field preference was affected significantly for the combinations FAW on corn versus CL on corn (x2df1=6.596; p=0.010) (I, Figure 3-2d) and FAW on corn versus CL on cotton (x2df,=4.318; p=0.038) (V, Figure 32d).

Pooling the combinations with different hosts feeding on the same plants (I and II in Figure 3-2) and the combinations with the same hosts feeding on different plants (III and IV) revealed an overall preference for FAW and for corn odors. The females divided their final choices more or less equally among FAW and CL, (73:62) but the overall odor field preference deviated significantly from a 1:1 ratio (116:76; X2df1=3.802, p=0.051). Plant preference in favor of corn was demonstrated with both the number of final choices (98:45; X2df1=9.778, p=0.002) and the odor field preference (131:68; X2df1=9.285, p=0.002).

Effects of experience were also analyzed with a loglinear model (SAS Institute, 1987) with five dependent variables; experienced host, experienced plant, alternative host, alternative plant, and preference (experienced or alternative odor). The number of females responding on a test day was used as one observation. There were 24 response levels with a total frequency (N) of 415 for the final choices and 585 for the odor field preference. Again the overall effect of experience in favor








72

of the experiences odor was highly significant (Final choices (F.C.): X2df1=12.21, p=0.0005; Odor field preference (O.F.P.): x2df1=13.84, p=0.0002). Significantly more females chose the experienced odor if they had experienced corn than if they had experienced cotton (F.C. X2df1=15.82, p=0.0001, O.F.P.: X2df1=12.55, p=0.0004). No such difference in experience effect was found for the two host species. The general preference for FAW odors was demonstrated by the fact that significantly more females would choose the alternative odor if the alternative host was FAW (F.C.: x2df1=10.32, p=0.0013; O.F.P.: x2df1=19.85, p<0.0001). No such difference was found for the plant species. Note that in those cases where the wasps experienced the source with the least preferred host and plant (i.e., CL on cotton; combinations II, IV, and V) no preference for either odor source was observed.

The combination FAW on corn - CL on cotton was chosen for the

following additional preference experiments. This combination contains all four components and the results obtained for this combination allow room for measurable increases and decreases in the effect of experience.

A more complete experience. Previous experiments were conducted with females that had a 20 sec contact experience without ovipositions just prior to their introduction into the olfactometer. Further experiments were performed to determine whether a longer, more complete experience which included ovipositions would result in a stronger effect upon the odor preference by the parasitoids, and whether this effect of experience would be retained over time.








73

Females were experienced by placing them in a glass container (26 cm in diam., 10 cm high) containing either 70 FAW larvae on 12 corn seedlings or 70 CL larvae on 12 cotton seedlings. The containers were then covered with a plexiglass plate. All females made contact with the plants and frass, and parasitized more than one larva. Females were exposed to the plant-host complex until they left the plants and attempted to leave the container. The exposure time varied from 4 to 11

min.

The persistence of the experience effect over time was tested by giving one group of females their experience in the morning 3-4 hours before being tested in the olfactometer (Group 1); a second group was given their experience just a few minutes prior to the bioassay (Group

2).

Results for the two groups are presented in Figure 3-3. Again, the differences between females that experienced different complexes were slight but consistent. The preference for the odor of FAW on corn was less for the females that experienced CL on cotton. For the females of Group 1 the difference in preference was only found to be significant for the number of final choices (x2df1=5.326; p=0.021). For Group 2 a significant difference between females experienced on a different complex was only found in the odor field preference (x2df1=15; p=0.001).

Group 1 and Group 2 females did not differ from each other in their response to the odors except for females experienced on FAW on corn. Group 2 females with a FAW on corn experience preferred FAW on corn odors significantly more than Group 1 females that had experienced








74





Group 1 Group 2 Group 1 Group 2
100% 129) (26) (38) 100%
IM (29) -x
FAW (23) (19) (26) FAW 0
(22) C
on on a U) L a0 corn corn a) 4

0 L



L (11). CL (13) (10) (14) CL iL on (18) on
cotton cotton -0
0
100% 100%
FAW CL FAW CL FAW CL FAW CL
com cotton corn cotton com cotton corn cotton

ODOR SOURCE CONTACTED PRIOR TO PREFERENCE TEST











Figure 3-3. Responses of C. marginiventris females to the odors of
either of two plant-host complexes after a complete contact experience including ovipositions as indicated in the figures. Group 1 had their experience 3-4 hours prior to a bioassay, Group 2 had their experience just prior to a bioassay. The actual numbers are shown in parentheses.
The asterisks indicate significant preference shifts.








75

the same complex (X2dfl= 5.87; p = 0.015). This difference was not observed in the number of final choices made by the two groups. Experiences with ovipositions versus experiences without ovipositions.

Finally, treatments were tested simultaneously to reveal possible differences between experiences with ovipositions and experiences without ovipositions, which may so far have been hidden by interday variation. Two sets of females were given a complete experience as described above, one hour before the bioassays. The first set encountered larvae and could oviposit freely during the experience. The second set of females, however, was experienced on a complex where the larvae were removed so that only the contaminated and damaged leaves could be contacted.

The results were very similar to those found for the treatments discussed before (Figure 3-4). No differences were found between females that had a total experience including ovipositions and females that only contacted the damaged and contaminated leaves.


Discussion

The dose-response tests revealed that the females' responses increase with an increasing dose of the host-related odors. No significant differences were found in the attractiveness of the four different plant-host complexes when the parasitoids were exposed to the odors in single choice tests. The pooled results of the preference experiments, however, indicate a preference for FAW odors over CL odors and an even stronger preference for corn odors over cotton odors.








76





+Ovip. -Ovip. +Ovip. -Ovip.
100% 100%
(29)
(21) (29) (38) 0 FAW * FAW C
on (15) (16) (20) on La) corn corn a)

0.
-O

S(7)
CL (8) (1 1) (10) CL
- on cotton (20) (20) (23) cotton -o
0

100% 100%
FAW CL FAW CL FAW CL FAW CL
corn cotton corn cotton corn cotton corn cotton

ODOR SOURCE CONTACTED PRIOR TO PREFERENCE TEST










Figure 3-4. Responses of C. marginiventris females to the odors of two
plant-host complexes after a complete contact experience with (+Ovip.)
and without (-Ovip.) ovipositions. The actual numbers are shown in parentheses. The asterisks indicate significant preference shifts.








77

Since FAW seems to be C. marginiventris' most important host and FAW larvae are found predominantly feeding on corn, other grasses and legumes (Ashley, 1986), an initial preference for the odors of FAW larvae and damaged corn is not surprising. Furthermore, CL appears to be a very poor host since initial rearing experiments show minimal emergence from this host (unpublished data; M. R. Strand, personal communication). On the other hand, since all test animals were reared on FAW larvae, the observed preference for FAW may also have been the result of conditioning of the parasitoids as immatures. However, since host larvae are routinely fed artificial diet, the rearing procedure could not account for the corn preference.

The results not only indicated an innate preference for FAW and corn odors, they also showed that the preferences were affected by contact experiences with the plant-host complexes. For all plant-host complex combinations it was found that a particular complex was chosen more often by females that had experienced that complex than by the females that had experienced the alternative complex. Although not always significantly, this experience effect caused a change in preference in each individual combination. The overall effect was found to be highly significant.

The results are in agreement with earlier results of single choice experiments presented in chapter II. The increase in response to hostrelated odors after experience is greatest to the odors emitted by the plant-host complex that the females experienced. The learning process that must be involved is triggered by a brief contact with host byproducts, and does not require actual contact with the hosts. When








78

females were given a longer experience period, including ovipositions, they did not appear to respond differently than females that had a 20 sec experience without ovipositions. The effect of experience on the preference for host-related odors lasted at least several hours, and is therefore likely to be an important factor determining the host searching behavior of these parasitoids in the field.

Significant differences in preference were found when females were offered odor source combinations where only the host species varied as well as in combinations in which only the plant species varied and in those combinations where both the host and the plant varied. We can therefore conclude that females are able to distinguish between different host species and between different plant species. Evidently, both host and plant are somehow involved in the emission of the semiochemicals that evoke a response in the parasitoid females, either by producing the essential volatiles or by affecting the volatiles released by another component of the complex. The parasitoids are therefore likely to respond to more than one compound, the intensity of their response to each compound probably increases when it is encountered in association with a foraging experience. The results suggest that each plant-host complex releases its own blend of semiochemicals that is detected by C. marqiniventris. After exposure to a particular complex, a female will subsequently be attracted to an odor blend that is most similar to the blend she perceived during her experience. Future research will hopefully reveal whether females of C. marginiventris and of other parasitoid species distinguish between








79

variations in specific semiochemical blends or that they are able to differentiate between different compounds altogether.












CHAPTER IV

ANALYSIS OF ORIENTED FLIGHTS TOWARDS A SOURCE
OF HOST-RELATED ODORS IN A FLIGHT TUNNEL Introduction

In the two previous chapters it was shown that females of C.

marginiventris are attracted to volatile cues released by host larvae feeding on plants. For a more detailed study of the long-range host location strategies of C. marQiniventris, a bio-assay needed to be developed that would allow the females to exhibit the behavioral sequence that gets them in the vicinity of a host in a more natural situation. The bio-assay should be designed such that it will be possible to determine which component or components of a host-plant complex is (are) responsible for the release of the active semiochemicals.

Studying the behavior of females flying to an odor source in a flight tunnel appeared to be a possible way to establish the exact source of the semiochemicals. Drost et al. (1986) provided the first detailed flight tunnel study of a parasitoid, Microplitis croceipes (Cresson), responding to host-related odors. The key to the success of the latter study was that pre-flight exposure to contact kairomones did override the tendency of the wasp to disperse upon release. This phenomenon, which was later found to be the result of associative learning (Lewis and Tumlinson, 1988), has facilitated the study of parasitoid host-searching behavior in flight tunnels for several others


80








81

(Elzen et al., 1986, 1987; Eller et al., 1988b; Eller, 1990; Hdrard et al., 1988a, 1988b; Zanen et al., 1989; McAuslane et al., 1990a, 1990b, 1990c).

C. marqiniventris too is able to associate odors with the possible presence of hosts after contacting host by-products (chapters II and III). After such an experience the wasp is significantly more responsive to these odors. In this chapter the flight responses of naive and experienced females to host-related odors are compared. Furthermore, a detailed description of the behavioral sequence exhibited by experienced females is presented in the form of ethograms.


Material and Methods

The parasitoids. The wasps were reared and kept as described in chapter II. Female C. marginiventris were 3 to 5 days old when used in the experiments.

Experience of the wasps. Female wasps that were used in the bioassays were either kept naive (no contact with hosts, host by-products or plants), or allowed to forage on corn leaves that had been fed upon by FAW larvae the night before and were contaminated with host frass. This experience lasted 20 seconds. The wasps made no actual contact with host larvae.

Flight Tunnel. Flight responses of C. marginiventris females were observed in a plexiglass flight tunnel 60 cm x 60 cm in cross-section and 2.4 m long. Two sheets of nylon mosquito netting (10 cm apart) at the open upwind end and one sheet of nylon screen (7 x 7 mesh/cm2) at the downwind end provided near laminar flow. Air was pulled through the








82

tunnel at 0.2 m/sec and was exhausted via a 30 cm diam. flexible pipe with a fan. Four overhead incandescent lights (90 W) were dimmed so that they provided approximately 500 lux inside the tunnel. The room housing the tunnel was maintained at 27.5-29*C and 55-80% RH. A more detailed description of the tunnel is given by Eller et al., (1988b).

The odor source. Twelve late second instar FAW larvae were

allowed to feed on 2 corn seedlings in a plastic petri dish (9 cm diam., 1.2 cm high) overnight. Two hours prior to the actual experiment six of the larvae were starved to ensure that they would be feeding during the experiments. One of the leaves they had been feeding on was removed and used for the above described experience of the parasitoids. Just before the flight tests, the remaining six larvae were removed and replaced by the starved larvae. One extra fresh seedling was added. The petri dish without a cover was then put upwind in the flight tunnel on a stand about 30 cm above the flight tunnel floor and served as the odor source.

Observations. Females were released downwind in the flight tunnel about 1 m distant from the odor source. The insects were released from an open vial on top of a stand 30 cm above the flight tunnel floor. As soon as they reached the rim of the vial, their behavior was closely observed. To compare responses of naive insects with responses of experienced insects only one criterion was used; whether or not an insect would fly all the way to the odor source and land on it. On four different test days fifteen naive and fifteen experienced females were tested. Each female was given 3 trials. If she had not reached the odor source in any of those trials her flight was considered incomplete.








83

The latter experiment was also used to distinguish the different components of the flight behavior. The following behaviors were observed:

Take flight: the female jumps from the releasing vial into the air
stream.

Plume casting: the female shifts from side to side and up and down
in the airflow, facing upwind. The female does not noticeably
move forwards or backwards relative to the airflow. The area of movement is limited to only a few square decimeters in the center
of the tunnel.

Wide casting: this behavior is similar to plume casting, but
differs from it in that the female uses the whole diameter of the
flight tunnel for her movements.

Straight flight: the female flies upwind towards the odor source,
more or less parallel to the airflow.

Downwind loop: while in flight the female suddenly drops back to a position downwind from where she was and continues with one of the
other types of behavior.

Dart: when close to the source (about 4-10 cm), a female flies
quickly forward to the source, this is always followed by a
landing on the source.

Land source: female lands on the source of the odor. This can be
anywhere in or on the petri dish.

Land other: female lands somewhere else than the source. In all
flights observed this was only on the ceiling or one of the sides
of the flight tunnel.


After the different components of the flight behavior were

established, the flights of fifty experienced insects were observed recorded and analyzed. Each insect was allowed three flights. Some escaped and some only flew twice. In total 142 flights were analyzed.

Data recording. The behaviors of the females in the flight tunnel were recorded with the use of an Epson@ Geneva PX-8 portable computer.








84

Results

Naive vs. experienced females. All females took off and flew after spending some time on the rim of the vial. Within the three trials given to them only 20.0 % (12 out of 60) of the naive insects flew all the way to the odor source and landed on it, while 81.6 % (49 out of 60) of the experienced insects landed on the source. The difference in responsiveness between naive and experienced insects is highly significant (x2df1=43.21; p < 0.0001).


Analysis of the flight behavior. After the different types of

behavior were described, the frequency of occurrence of these behaviors was observed and recorded. The results of these recordings are summarized in the kinematic diagrams in figures 4-1. The diagram gives the number of transitions from one type of behavior into another and, the probability one behavior would follow another. Darts were always followed by a landing on the source and were therefore included with the landings in the diagram. The thick arrows indicate the sequence of behavior that usually preceded an actual landing on the source. Such flights would consistently start with casting or straight flight and the wasp would alternate between these two behaviors until she was close to the source. Then the female would dart forward and land on the source.








85
.06
N- 142


2N=22
.22 Down Wind
.47 ..41 Loop


.45
.02.







.041 .02



.04

N=223
Straight .02 W-72 LandLand on Walls





N=70
Land on
Source






Figure 4-1. Kinematic diagram of the flight behavior exhibited by experienced C. marginiventris females in response to odors emitted by plant-host complex of FAW larvae feeding on corn inside a flight tunnel. A total of 142 flights were observed, the probability that particular transitions from one behavior into another occurred are indicated.








86

Discussion

In a four-arm olfactometer C. marginiventris females are

significantly more responsive to host-related odors after they have had a previous contact with the by-products of hosts in association with these odors (chapters II and III). The results of the flight tunnel study clearly show that experience also affects their flight responses. This phenomenon has been studied in detail for the wasp Microplitis croceipes (Drost et al., 1986, 1988; Lewis and Tumlinson, 1988; Eller, 1990), and was found to be the result of associative learning (extensively discussed in chapter I).

The detailed description of the flight behavior to a complete odor source presented here may be useful when the responses to different odor sources are compared. Speculating on the functions of the behaviors, it seems that plume casting is a way to locate the center of the odor plume. Wide casting might be a way to locate the plume itself. Straight flight seems to occur whenever the wasp has found the center of the odor plume. These speculations are supported by the fact that wide casting was usually followed by plume casting and that plume casting was followed by either wide casting or straight flight.

A downwind loop occurred only occasionally and might be the result of losing the odor plume. By dropping back into the airflow the wasp will increase its chances of finding the plume again, because the plume will be wider downwind.

Studies on the host-location flight behaviors of parasitic wasps are rare. M. croceipes is the only wasp for which detailed information on its flight behavior is available (Drost et al., 1986; Eller et al.,








87

1988b). This wasp's oriented flights towards an odor source are very similar to the flights exhibited by C. marginiventris females. Drost et al. (1986) observed the wasps to zig-zag immediately after they took flight. They described this initial zig-zagging as "making sideways excursions mainly in the horizontal plane and perpendicular to the wind direction soon after take off". Eller et al. (1988b) referred to this behavior as casting (as is done here) and reserved the term zig-zagging for side to side flight with upwind movement. Neither of the two papers on M. croceiDes distinguishes between plume casting and wide casting, two behaviors that appear well distinguishable for C. marginiventris. Drost et al. (1986) and Eller et al. (1988b) use the terms straight flight and darting as they are used here, and Drost et al. (1986) also observed downwind loops. M. croceipes females were also observed to hover 5-10 cm in front of the odor source before they would dart and land. This behavior was never observed for C. marginiventris. However, the forward movement of the wasps did slow down when they got close to the source. Hrard et al. (1988a) described the flight of M. demolitor as essentially the same as the flight described for M. croceipes by Drost et al. (1986).

The results indicate that there are some common features to the oriented flights of different parasitic insects, but that generalizations are hard to make. It also appears that the flight responses and the behavioral sequences are strongly dependent on the odor plume shape that is offered to the insects. When different odor source containers were used, wasps exhibited all of the described behaviors, but the frequencies with which they occurred varied with the




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ENY1HIVCI_DKUP0V INGEST_TIME 2015-04-13T19:28:09Z PACKAGE AA00029890_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

SEMIOCHEMICALLY MEDIATED, HOST-SEARCHING BEHAVIOR OF THE ENDOPARASITIC WASP Cotesia marqiniventris (CRESSON) (HYMENOPTERA: BRACONIDAE) BY TED C. J. TURLINGS 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 1990 UNWERSm OF FLORIDA LIBRARIES,

PAGE 2

To Mams, Paps, Patricia, Ernie! , Bix, Caspar, Walman, Dekker, and Sir Patrick. Ted

PAGE 3

ACKNOWLEDGEMENTS My gratitude goes out to my graduate committee and the entire staff of the Insect Attractants, Behavior, and Basic Biology Research Laboratory, ARS-SEA, USDA, for their continuous support and help throughout the period that the research presented here was conducted. Special thanks go to Barbara Dueben without whom the work would have been much harder and far less enjoyable. In addition, I thank Hans Alborn, Peggy Brennan, Fred Eller, Rob Murphy, Ara Manukian, Tommy Proveaux, Delrea Patrick, Charlie Dillard, Annette Brabham, Bob Doolittle, Bob Heath and Peter Teal for all their patience, valuable advice, and for creating a pleasant working atmosphere. I thank Terri Rossignol for her valuable assistance. For their overseas support, I thank Louise Vet and Ben Kennepohl . Financial support was granted by the International Research Division of the Office of International Cooperation and Development and by the Agricultural Research Service, USDA. The research was also partially supported by a Fulbright grant issued by the U.S. Information Agency. Last, but certainly not least, I would like to express special thanks to Jim Tumlinson and Joe Lewis for not only giving me the opportunity to conduct this research, but also for their patience and motivational support. i i i

PAGE 4

PREFACE Humans primarily use vision and hearing for orientation, while the functions of smell and taste are considered of minor importance. Anthropomorphic thinking may have prevented biologists from realizing how important smell and taste can be for other organisms, particularly in the insect world. During the last few decades, however, interest in how insects find mates, prey, and hosts has stimulated scientists to study the interactions that are mediated by airborne chemicals. As a result, the currently lively field of chemical ecology has revealed that for insect communication and orientation, olfaction is far more important than are the other two major modes, vision and hearing. Airborne chemical stimuli are used by insects to guide them to suitable habitats, suitable hosts or prey, to find appropriate mates and to detect danger. By interfering with these chemical interactions, man appears to be able to manipulate insect behavior to his advantage. This is exemplified by several successful attempts to use synthetic versions of insect sex-pheromones (i.e. chemicals emitted by insects to attract mates) to disrupt or trap pest insects. These strategies have been effective and safe alternatives to the use of pesticides to control insects. Recently, several research groups launched projects to investigate the means by which chemicals can be used to manipulate the behavior of

PAGE 5

beneficial insects (e.g. those insects that kill pest insects). It is envisioned that this research will lead to strategies that will increase the effectiveness of these insects as bio-control agents. Preliminary results show that parasitic wasps can be conditioned or stimulated to respond to chemical cues that guide them to hosts (see Chapter I). This is achieved by letting the insects contact hosts or host products in the presence of these odors. This phenomenon of learning by association allows us to condition wasps such that they will utilize certain chemical cues that will lead them to a target pest more effectively. To establish the full potential of this technique, research should focus on the isolation and identification of the essential semiochemicals. Furthermore, the effects of these chemicals on the behavior of entomophagous parasitoids must be studied. When sufficient information is gathered on how different chemicals mediate the various behavioral components that lead to a successful encounter with a host, we may eventually be able to, as Shorey (1977) suggested ",make an insect jump through our hoop."

PAGE 6

TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS iii PREFACE iv LIST OF TABLES ix LIST OF FIGURES x ABSTRACT xii CHAPTERS I LITERATURE REVIEW AND RESEARCH AIMS 1 Semiochemicany Mediated Host Searching Behavior 2 What are Semiochemical s? 2 Responses that can be Elicited by Semiochemical s 4 The Mechanisms behind Semiochemical Induced Responses 5 Semiochemically Mediated Long-Range Host Searching Behavior in Parasitic Insects 8 Searching Behavior in Parasitic Wasps 8 Examples of Host and Host Habitat Location by Olfaction... 11 Mechanisms that Determine the Responses of Parasitic Insects to Semiochemicals 15 Analysis of Host Searching Behavior 22 Manipulation of Parasitic Wasps for Biological Control Purposes 23 Resource Management 24 Semiochemical Manipulation 26 Cotesia marqiniventris (Cresson) a Prime Candidate for Augmentative Release Against Lepidopterous pests 30 Systematics 30 Host and Plant Range 32 Geographical Range 36 Biology 36 Host Regulation 45 Bio-Control 46 Research Aims 49 vi

PAGE 7

II INCREASED RESPONSE TO HOST RELATED ODORS AFTER A FORAGING EXPERIENCE: A MATTER OF LEARNING ? 52 Introduction 52 Materials and Methods 53 Results 57 Discussion 60 III EFFECTS OF FORAGING EXPERIENCES ON ODOR PREFERENCES 64 Introduction 64 Materials and Methods 65 Procedures and Results 66 Discussion 75 IV ANALYSIS OF ORIENTED FLIGHTS TOWARDS A SOURCE OF HOST RELATED ODORS IN A FLIGHT TUNNEL 80 Introduction 80 Materials and Methods 81 Results 84 Discussion 86 V PINPOINTING THE EXACT SOURCE OF VOLATILE ATTRACTANTS THAT ELICIT ORIENTED FLIGHTS IN Cotesia marginiventris FEMALES 89 Introduction 89 Materials and Methods 90 Results 94 Discussion 99 VI ISOLATION, AND IDENTIFICATION OF ALLELOCHEMICALS THAT ATTRACT Cotesia marginiventris TO THE MICRO-HABITAT OF ONE OF ITS HOSTS 103 Introduction 103 Materials and Methods 104 Results Ill Discussion 119 VII THE ACTIVE ROLE OF PLANTS IN THE PRODUCTION OF THE VOLATILES THAT GUIDE COTESIA MARGINIVENTRIS FEMALES TO THEIR HOSTS 125 Introduction 125 Materials and Methods 126 Procedures and Results 127 Discussion 139 vi i

PAGE 8

IX SUMMARY AND CONCLUSIONS 144 REFERENCES CITED 156 BIOGRAPHICAL SKETCH 178 vi i i

PAGE 9

LIST OF TABLES Table Page 1-1 i Definitions of chemical mediators important in the j 3 chemical interaction between organisms. Reported host species for C. marqiniventris. 33 2-1 Effect of pre-bioassay experience on response of C. \ 58 marqiniventris females exposed to host related odors in a ! four-arm olfactometer. ^ 6-1 Super Q trapped volatile compounds identified in the | 115 atmosphere associated with corn seedlings fed upon by BAW j larvae. ^ 6-2 Quantative comparison of natural and synthetic blends j 118 used for bio-assays 6-3 Fliqht tunnel responses of C. marqiniventris females with i 120 different experiences to extracts of; 1) volatiles 1 collected from BAW larvae feeding in corn seedlings j (NATURAL), 2) a synthetic mimic of the same volatiles 1 containing the eleven major components (SYNTHETIC), and | 3) the extraction solvent, methylene chloride, only ! ^ [SOLVENT) . 1 jy-i Amounts of volatiles released by corn seedlings damaged 132 by BAW larvae just after damage (FRESH) compared with volatiles released 16 hours after damage (OLD). j

PAGE 10

Figure LIST OF FIGURES Page 1-1 j Basic sequence of host-finding activities by females of ! parasitic insects (after Lewis et al . , 1976). 10 1 1-2 [ Schematic depiction of the life-cycle of 1 marqiniventris. 37 1 1-3 1 — — — — 1 Behavioral ethogram of the host finding and ovipositional ! sequence of C. marqiniventris females on corn plants 1 damaged by fall armyworm larvae (after Loke et al., i 1983). r42 2-1 Experience effect upon response to odors of two planthost complexes (with vs. without oviposition) . r 59 2-2 Experience effect upon response to odors of two planthost complexes (CL on cotton vs. FAW on corn). 61 3-1 Responses of experienced C. marqiniventris females to 3 uubeb 01 oQors ediitLeu oy larvae leeainy on leaves. 68 3-2 Effects of experience on the preference of marqiniventris females for host-related odors. 70 3-3 Responses of C. marqiniventris females to the odors of either of two plant-host complexes after a complete contact experience including ovijDositions. 74 3-3 L., Responses of C. marqiniventris females to the odors of two plant-host complexes after a complete contact experience with and without ovipositions. 76 Kinematic diagram of the flight behavior exhibited by experienced C. marqiniventris females in response to odors emitted by plant-host complex of FAW larvae feeding on corn inside a flig^ht tunnel. j 85 5-1 j Odor inlet system and insect release funnel for flight 1 tunnel bio-assays. 92 X

PAGE 11

Figure Page 5-2 Flight responses by C. marqiniventris females to a complete plant-host complex, compared with flight ^responses to single components of a plant-host complex. 95 h 5-3 Flight responses bv C. marqiniventris females during dual odor source tests. 96 h 5-4 [^Control dual choice flight tunnel tests. 6-1 Responses bv C. marqiniventris females to different doses of extracts of volatiles collected from BAW larvae feeding on corn seedlinc|s. 112 h 6-2 A) Profile of volatiles released by a complex of corn seedlings damaged by BAW larvae. B) Profile of volatiles released by undamaged corn seedlings. C) Structures and identities of the eleven major compounds, with corresponding peak numbers. 114 1 6-3 Volatiles released by different components of a complete plant/host complex. A) Complete complex of BAW larvae feeding on corn seedlings. B) Water-washed corn seedlings that were damaged by BAW larvae. C) BAW frass wiped of off the damaged corn seedlings. D) Starved water-washed BAW larvae. 117 7-1 GC profiles of volatiles released by corn seedlings subjected to different damage treatments. A) BAW feeding on seedlings that were damaged overnight by BAW larvae. B) BAW larvae feeding on fresh seedlings. C) Artificially damage fresh seedlings. D) Undamaged seedl ings. 129 7-2 Comparison of volatiles released by corn seedlings with fresh BAW damage and seedlings with older damage. 131 7-3 Comparison of volatiles released by corn seedlings damaged by BAW with volatiles released by seedlings with various artificial damage treatments. 135 7-4 Responses during dual choice flight tunnel tests by experienced C. marginiventris females to corn seedlings that underwent the various damage treatments. 137 7-5 Preferences exhibited by C. marqiniventris females for odors released by corn with fresh BAW damage versus odors released by corn with old BAW damage. 140 xi

PAGE 12

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 SEMIOCHEMICALLY MEDIATED, HOST-SEARCHING BEHAVIOR OF THE ENDOPARASITIC WASP Cotesia margini ventri s (CRESSON) (HYMENOPTERA: BRACONIDAE) By Ted C. J. Turlings August 1990 Chairman: J. H. Tumlinson Major Department: Entomology and Nematology Behavioral and chemical aspects of the long-range, host-searching behavior of Cotesia marqiniventris (Cresson), a larval endoparasitoid that attacks many lepidopterous species, were studied. In a four-arm olfactometer, parasitoid females exhibited a striking increase in their response to host-related odors after they previously had contact experience with the source of these odors which was a complex of host larvae feeding on plants. Results indicated that after contacting certain kairomones in the by-products of their hosts, the wasps associate the surrounding odors with the presence of hosts and will subsequently use these odors as cues in their search for additional hosts. To determine the actual source of the active odors, a complete plant-host complex was divided into its three main components: 1) corn seedlings damaged by host larvae (beet armyworm), 2) frass produced by x11

PAGE 13

those larvae, and 3) the host larvae. The damaged plants were significantly more attractive to marqiniventris than frass or larvae alone. Analysis of collected volatiles revealed the consistent presence of eleven plant released compounds. Four of those were typical green leafy odors and were released in relatively large amounts only if larvae were feeding on the seedlings. The other compounds, six terpenoids and indole, were not released in significant amounts until a few hours after initial feeding by the larvae. Artificially damaged corn seedlings released only minor amounts of the terpenoids and indole unless they were treated immediately after damage with the oral secretions of larvae. This plant response to herbivore specific damage was exploited by the parasitic wasps. After experiencing the odors during a brief contact with host-infested corn, the wasps were highly attracted to the terpenoid-releasing corn seedlings. Experienced wasps were also attracted to a synthetic blend of the eleven identified compounds. The flexible response of marqiniventris to the actively released plant volatiles is highly adaptive to both the plant and the parasitoid. It can be expected that many similar tritrophic interactions will be unraveled in the future and may be exploited for biological control purposes. , ; xi i i

PAGE 14

CHAPTER I LITERATURE REVIEW AND RESEARCH AIMS This chapter reviews the literature on semiochemical ly mediated host searching behavior of parasitoids. First, definitions of the various classes of semiochemical s and their modes of action are addressed. The importance of semiochemical s for long-range host location in parasitoids is discussed, whereby models and examples from the literature are used. Special attention is given to the different mechanisms that have been suggested to determine the responses by host seeking parasitoids to semiochemicals. The mechanisms are also discussed in relation to manipulation by men of parasitoids for biological control purposes. Subsequently, the literature available on the parasitoid Cotesia marqiniventris (Cresson) is reviewed and discussed, with the emphasis on the potential of this general ist parasitoid for the use as a control agent. Finally, the research presented in this dissertation on the searching behavior of C^ marqiniventris is introduced by giving the objectives of the study. 1

PAGE 15

2 Semiochemicall V Mediated Host Searching Behavior What are Semiochemicals? Karl son and Butenandt (1959) introduced the term pheromone and defined it as "substances that are secreted by animals to the outside and cause a specific reaction in a receiving individual of the same species (p. 39)." To also include the interactions between members of different species Law and Regnier (1971) introduced the term semiochemical . This term covers all chemicals that are involved in interactions between individual organisms. Several authors have later attempted to name and define the different types of semiochemicals. Brown et £L (1970) proposed allomone and kairomone as terms for chemicals produced by organisms that elicit reactions in species other than their own. Before and after that several other terms were introduced (Beth, 1932; Frankel , 1959; Chernin, 1970 and Blum, 1974). Nordlund and Lewis (1976) unraveled the tangle that had been created, and Table 1-1 lists the definitions that they gathered and adjusted. Recently Dicke and Sabelis (1988a) introduced a modified classification of what they call infochemicals. This classification is based on cost-benefit analysis rather than on the origin of the compounds which was mostly used by the earlier authors. Of course time will tell if new adjustments will have to be made and if other definitions have to be added. But for now the classifications as given by Nordlund and Lewis (1976) and by Dicke and Sabelis (1988a) can both serve as good guides in the world of semiochemicals. Here the classification by Nordlund and Lewis (1976) will be adopted.

PAGE 16

3 Table 1-1. Definitions of chemical mediators important in chemical interactions between organisms. Semiochemical .-A chemical involved in the chemical interaction between organisms. 1. Pheromone Substance that is secreted by an animal or a plant to the outside that causes a specific reaction in a receiving individual of the same species. 2. Allelochemic Chemical significant to organisms of a species different from their source, for reasons other than food as such. a. Allomone A substance, produced or acquired by an organism, which, when it contacts an individual of another species in the natural context, evokes in the receiver a uciiaviurdi ur piiy ^ 1 u 1 uy 1 cd 1 rcdCLiuii dudptivciy Tdvurduic lu the emitter but not to the receiver. b. Kairomone A substance, produced, acquired by, or released as a result of the activities of an organism, which, when it contacts an individual of another species in the natural context, evokes in the receiver a behavioral or physiological reaction adaptively favorable to the receiver but not to the emitter. c. Synomone A substance, produced or acquired by an organism, which, when it contacts an individual of another species in the natural context, evokes in the receiver a behavioral or physiological reaction adaptively favorable to both the emitter and the receiver. d. Apneumone A substance emitted by a nonliving material that evokes a behavioral or physiological reaction adaptively favorable to a receiving organism, but detrimental to an organism, of another species, which may be found in or on the nonliving material. After Nordlund and Lewis (1976).

PAGE 17

4 Responses that can be Elicited by Semiochemical s Dethier (1947) and Kennedy (1947) realized that by using only the terms attractant and repellent it was not possible to describe chemicals in terms of their effect on the behavior of insects. Specific reactions to stimuli should get specific names and these names should have a clear definition. The first step in doing so is by analyzing the variety of ways by which aggregation and dispersion can be brought about. According to Dethier et al_:. (1960) an insect may do one of the following: 1. continue without change of rate of linear progression, rate of turning, or direction, 2. stop, 3. slow its rate of linear progression, 4. increase its rate of turning, 5. increase its rate of linear progression, 6. decrease its rate of turning, 7. orient toward a source, 8. orient away from a source. Responses 2 to 6 were grouped together as non-directed responses and called kineses (singular, kinesis). Later Kennedy (1977) added another possible response to this group; the initiation of movement when an , insect was previously at rest. Responses 7 and 8 are placed in a second major group, directed responses with reference to the source, taxes ^ (singular, taxis). The listed reactions do not only refer to movement but can also be used for other aspects of behavior such as feeding, mating and oviposition. Having categorized the types of behavior elicited by chemicals, Dethier et aK (1960) were able to list the type of chemicals in terms of what they do.

PAGE 18

5 (1) Arrestant (stops and slows)-a chemical which causes insects to aggregate in contact with it, the mechanism of aggregation being kinetic or having a kinetic component. An arrestant may slow the linear progression of the insects by reducing actual speed of locomotion or by increasing turning rate. (2) Locomotor stimulant (starts or speeds)-a chemical which causes, by a kinetic mechanism, insects to disperse from a region more rapidly than if the area did not contain the chemical. The effect may be to increase the speed of locomotion, to cause the insect to carry out avoiding reactions, or to decrease the rate of turning (Fraenkel and Gunn, 1940). (3) Attractant (orients toward)-a chemical which causes insects to make oriented movements towards its source. (4) Repellent (orients away)-a chemical which causes insects to make oriented movements away from its source. (5) Feeding, mating, or ovipositional stimulant (initiates or drives)-a chemical which elicits feeding, mating or oviposition in insects. (6) Deterrent (inhibits)-a chemical which inhibits feeding or oviposition when present in a place where insects would, in its absence, feed or oviposit. The Mechanisms behind Semiochemical Induced Responses The terms attractant, repellent, arrestant, stimulant and deterrent are useful to describe the effects of semiochemicals, but give us no information on the behavioral mechanisms involved. Several

PAGE 19

authors have reviewed the orientation to chemical sources (Shorey, 1973; Farkas and Shorey, 1972; Kennedy, 1977; Carde 1984). Their papers made clear that it is very hard to classify, and even harder to identify, the mechanisms that are used by insects to respond to semiochemical s, even when the classification is limited to a flying orientation to chemical sources. So far, nobody has been able to give a satisfactory classification. It is generally accepted that the principal mechanism of "long distance" flying orientation to an airborne chemical stimulus in the wind is an optomotor guided, chemically induced, upwind orientation (or anemotaxis) (Carde, 1984). However, orientation mechanisms in different phylogenetic groups will have evolved rather independently, therefore there is good reason to assume that multiple solutions for finding a chemical source will have developed. Knowledge is lacking to recognize and classify all these solutions. Still, biologists studying insect behavior evoked by semiochemical s need some kind of classification to work with, and need to be able to communicate with each other. Of the complex classifications made to date, Shorey's (1973) interpretation seems to be the most comprehensive. He splits the behavioral mechanisms used to aggregate at an odor source in three categories. 1) CHEMOTAXIS: the insects steers its body axis in the direction of the chemical source because it can directly sense the gradient of odor molecules. In chemotaxis two main types are recognized, ODOMOTROPOTAXIS and CHEMOKLINOTAXIS. In the first the insect uses two or more receptor organs (e.g. antennae) to measure concentration differences of the odorant, it will constantly turn towards the highest concentration. In chemokl inotaxis, the insect swings its body (receptors) from one side to the other. In this way it assesses relative concentrations of the odorant over time

PAGE 20

7 and steers towards the side on which it obtained greatest stimul ation . 2) KINESIS REACTIONS: the insect does not directly sense the direction of the chemical source, but it is caused to move at varying rates (ORTHOKINESIS) or to turn at varying frequencies (KLINOKINESIS) depending on the concentration of chemicals to which it is exposed. The insect only senses concentration differences and not the direction of the source. 3) Shorey (1973) did not give the third category a name, but it includes the responses in which the odorant acts as a releaser, sensitizing the insect to some other stimulus and causing it to orient to that stimulus. ANEMOTAXIS is the major member of the third category. It is the orientation by steering into the wind (upwind) when the appropriate odor is sensed. When the insect loses contact with the odor, it reacts by, for instance, crosswind flight. This increases the likelihood that the insect will reenter the odorous airstream after which the upwind flight resumes. The occurrence of chemotaxis is unlikely because in the field, currents in airstreams will make it virtually impossible for organisms to sense clear gradients of concentrations. Concentration differences might play a role when an insect gets close to the source, but at respectable distances from the source anemotaxis seems more obvious. Anemotaxis is, indeed, accepted as the most common mechanism by which insects orient towards or away from an odor stimulus. In some cases, however, it may seem as if an insect orients to air currents, but it might well be that stimulation with semiochemicals results in a visual orientation to an appropriate object in the environment. Future studies of insect behavior and the stimuli evoking it will have to provide the information that will make it possible to give a clearer classification of the orientation mechanisms. The purpose of this chapter is to review the studies that involve host searching behavior in entomophagous insects evoked by chemical

PAGE 21

8 stimuli, therefore further discussion will only include parasitoid behavior evoked by semiochemicals belonging to the group of allelochemics. Semiochemicall V Mediated Long-Range Host Searching Behavior in Parasitic Insects Searching Behavior in Parasitic Wasps We may assume that action is undertaken by an organism because of the possible effects of such an action. A parasitoid follows a plume containing a volatile chemical associated with its host because it might lead the parasitoid to that host. It is generally accepted that such characteristics are adaptive and the result of natural selection. The major limiting factor for reproduction of solitary parasitoids is the number of hosts a female finds to lay her eggs in. Females that are able to detect and locate suitable hosts more rapidly and from greater distances than other females would be at a reproductive advantage. Therefore, it can be expected that parasitoids have evolved rather "sophisticated" searching strategies. They will have to use chemical and other cues in such a way that the chances of making "mistakes" are limited, thereby optimizing their time allocation. Vinson (1976) suggested that for many parasitoids searching behavior consists of a series of internally controlled locomotory patterns, each serving to place the wasps in contact with the searching space in which the parasitoid must randomly search for the next cue. Based on investigations of Salt (1935), Laing (1937), and Doutt (1959) he divided the host selection behavior into five steps:

PAGE 22

1. habitat location 2. host finding 3. host acceptance 4. host suitability 5. host regulation This division serves to categorize the various behavioral patterns, Lewis et aL. (1976) developed a diagram to describe the basic sequence of host finding (Figure 1-1). The purpose of this diagram is to distinguish between the behavioral steps that are evoked by different chemical stimuli, thus enhancing the ability to identify key points for manipulation of parasitoids in pest management programs. Again, the need to be able to communicate with each other makes it necessary for researchers to have models like the ones presented by Vinson (1976) and Lewis et aL. (1976). However, the knowledge that has been obtained so far is certainly not enough to make a general diagram or model that describes all the aspects of host finding. The diagram developed by Lewis et aL. (1976) may be a good description of the process involved for certain species, but other species might use a different strategy. At several places in the diagram insects might use other steps in their host finding activities that have yet to be observed. In other species, steps of the diagram, like examination of the host or post-oviposition behavior, might not take place at all. Future studies will have to elucidate whether general models can be developed or that the differences between species make it impossible.

PAGE 23

10 Postoviposition Behavior T8 General Inactivity T1 (S1) Random Movement -9^ T-2 .T2 (S2) Scanning of Habitat I T3 (S3) Investigation of Host Trails Within Habitat T4 (S4) Find and attack cycle ,T6 (S6) Examination of Host Figure 1-1. Basic sequence of host-finding activities by females of parasitic insects: Tl to T8 and T-1 to T-4 = transitions among the indicated behavioral acts. Sl-6 = stimuli releasing the indicated behavioral patterns. S2 = olfactory, visual, and physical cues associated with host plants on other habitats. S3 = primarily chemical cues from frass, moth scales, and decomposition products associated with the presence of host insects. S4 = olfactory, visual, auditory, and other chemical or visual cues from host insect. S5 and S6 = olfactory, tactile, auditory, and/or combination of these cues from host individual. Taken from Lewis et aL. (1976).

PAGE 24

11 Examples of Host and Host Habitat Location by Olfaction Host finding can involve long-range and close-range chemoreception (Kennedy, 1977; Weseloh, 1981). Long-range chemoreception is defined as being olfaction, which is the detection of chemicals in air. It is similar to our sense of smell. According to Weseloh (1981) close-range chemoreception involves only the perception of chemicals after direct physical contact with them in solid or liquid form. Obviously, insects also use olfaction when they are close to their target and it seems that close-range chemoreception should include olfaction as well. Much more is known about close-range chemical orientation than about long-range chemical orientation. This is probably because short-range orientation is easier to observe (Weseloh, 1981). Weseloh (1981) lists 19 species of parasitoids for which it has been demonstrated that they locate hosts by long-range chemoreception. Although the investigation of host location by olfaction has barely begun, it is known that chemical stimuli can be emitted by a variety of sources. It is obvious that stimuli used in close-range orientation have to be directly related to the host. For long-range orientation that is not necessarily true. The stimuli can come from the host insect, but may also come from the habitat in which the host feeds, from associated organisms or from a combination of these factors (Lewis et aK, 1981; Vinson, 1984). Reviews by Weseloh (1981), Lewis et al^ (1981), Vinson (1976; 1984), van Alphen and Vet (1987), Kainoh (1987), and Eller (1990) report a fair number of examples of olfactorial host location. Some of the most distinct examples will be presented here.

PAGE 25

stimuli from the host (i.e. kairomones). Carton {1971; 1974) found that the parasitoid Pimpla instigator could discover its pupal hosts ( Pieris brassicae ) from a distance even when they were concealed in a tube of wrapped paper. In other experiments with field traps Kennedy (1979) showed that several parasitoid species were attracted to the aggregation pheromone used by their host, the European elm bark beetle. Lewis et aK (1982) and Noldus and van Lenteren (1983) were the first to report that attraction of Trichogramma spp. involves volatile substances released by virgin females of hosts. Further examples of host odors acting as kairomones for Trichogramma spp. are reviewed by Noldus (1989). Host frass is another common source of kairomones that are exploited by parasitic wasps. For example, Micropl itis croceipes is attracted to the feces of Heliothis zea larvae (Eller et aL., 1988b), and Diadromus pulche11us a parasitoid of the leek moth and the diamondback moth is attracted to volatiles emanating from the larval frass of the latter two (Auger et aK, 1989). An extensive review of identified insect kairomones that attract entomophagous insects is presented by Eller (1990). Stimuli from plants (i.e. svnomones) . Plants are the most common source of volatile attractants for parasitoids. Thorpe and Caudle (1938) were some of the first to report this. It is normally argued that the parasitoids locate host habitats by tracking the volatiles that are emitted by plants the host may be feeding on (Vinson, 1975). In several cases parasitoids are attracted to host plants, regardless of the presence or absence of hosts. Vinson et al . (1975) observed Cardiochiles niqriceps searching host-free tobacco plants. Nishida

PAGE 26

13 (1956) found that Opius fletcheri was attracted to the plant habitat of its host, the melon fly. Macrocentrus grandi , a parasitoid of the European corn borer, was attracted to 15 out of 51 plant species tested (Ding et al^, 1989). Ding et aL. (1989) found no correlation between host or nonhost status of the plants for the European corn borer. The general ist parasitoid Campoletis sonorensi s appears to be different in this respect. The responses to plant odors by females of C^. sonorensis in flight tunnels have been studied extensively (Elzen et aL., 1983, 1984, 1986, 1987; Baehrecke et aK, 1990; McAuslane et aK, 1990a, 1990b, 1990c; Williams et aL., 1988). This parasitoid is readily attracted to plants that serve as food for its hosts, but the wasp is significantly less attracted to nonfood plants unless it has previously encountered hosts on such a plant (McAuslane et aL.> 1990b). It is frequently found that parasitoids are attracted to some species of plants but not to others. Taylor (1932) reported that Hel iothis armiqera , which feeds on a variety of plants, was attacked by Microbracon brevicornis only when it fed on Antirrhinum . Arthur (1962) found something similar for Itoplectis conquistor , which attacks larvae of the European pine shoot moth on scots pine but not on red pine. More examples of plant odors attracting parasitoids are given in chapter V and the subject was recently reviewed by Nordlund et aL. (1988). The many cases of parasitoids attacking hosts on one plant but not on the other (for review see Vinson, 1981), are usually explained as plant preferences. The preference in some cases might be explained by the fact that on some plants hosts are more suitable for parasitization because they provide the adult parasitoid with nutritional requirements.

PAGE 27

14 such as nectar from flowers (Herzog and Funderburk, 1985; Lewis and Nordlund, 1985). In other cases, hosts on the non-preferred plants might contain harmful chemicals obtained by eating from that particular plant. Besides that, specific components in the food might be necessary for the host to produce the right kairomones and these components may not be present in all plant species. Plants on which the hosts do not feed can also be attractive to the parasitoid. Associated plants may provide the parasitoids with food and/or shelter. Altieri and Whitcomb (1980) and AUieri et aL. (1981) showed that the presence of selected weeds within and around the crop fields greatly affected the abundance and activity of parasitoids in corn crops. Stimuli from associated organisms other than plants . Examples of host-associated organisms other than plants that are involved in the attraction of parasitoids are relatively rare. Spradbery (1970) demonstrated that Rhyssa persuasoria a parasitoid of siricid woodwasps is attracted to odors produced by fungal symbionts of their hosts. Greany et aL. (1977) showed that a rotting fruit fungus, which is often found together with tephritid fruit fly larvae, produces acetaldehyde, an attractant for the parasitoid Biosteres ( Opius ) lonqicaudatus . Dicke (1988a) reviews the studies that demonstrate the involvement of microorganisms in the host location of several parasitoids. Stimuli from a combination of factors . Most studies do not determine the exact source of the stimuli. In some studies it is demonstrated that it is actually a combination of factors. Leptopil ina heterotoma . for instance, is attracted to yeast (host food), but shows a

PAGE 28

much stronger response to yeast patches in which host larvae have been crawling and feeding (Dicke et aL., 1984; van Alphen et aL., 1984). This indicates the presence of al lelochemics produced by a combination of yeast and host material. Studies on the larval endoparasitoid Micropl itis croceipes indicate something similar; it responds better to hosts feeding on plants than to hosts only or plants only (Drost et al . . 1986; Eller et aK, 1988b). Mechanisms that Determine the Responses of Parasitic Insects to Semiochemical s A thorough understanding of the mechanisms involved in the host location by parasitoids is necessary before we can start thinking about the manipulation of these insects. Pheromone work has already resulted in a fair amount of information on how insects respond to attractive volatiles. Although possible, there is no reason to believe that basic behavioral responses of parasitoids to chemicals that will lead them to their hosts will differ from the responses by insects to sex pheromones that will lead them to conspecifics of the opposite sex. The strategies behind these responses and the properties of the chemicals that elicit the responses, however, will be quite different. In the case of sex pheromone attraction, both the receiver and the emitter profit from the communication between them. In the parasitoid host relationship, however, the attraction is advantageous to the parasitoid, but disadvantageous to the host. Therefore, selection will put constant pressure on the hosts to avoid the production and/or release of the semiochemical s. It can be expected that due to these pressures, constant changes in odor production within host populations

PAGE 29

are taking place. This will force parasitoids, in their turn, to adapt to these changes, requiring plasticity in their responses to semiochemicals. A predator-prey relationship will obviously result in similar coevolutionary pressures. The need for plasticity in a parasitoid's response to attractants becomes even more important when the sources of reliable attractants vary. When the main cues are not coming directly from the host, but from associated organisms, variability can become enormous. For example, a general ist parasitoid that attacks its hosts on a wide variety of plants, may use plant odors to locate its hosts. Different plants will release different odors, therefore the odor blend that contains the most reliable information on the presence of hosts will vary over time and space. Considering the preceding arguments, it is unlikely that the response of parasitoids to semiochemicals is hardwired and mainly genetically determined. Studies on the host location by parasitic insects have resulted in several theories on the mechanisms that allow these insects to optimally respond to chemical cues. These theories will be reviewed here. Innate responses. When the wasps respond to odors without ever having had any previous experience with them the responses are often referred to as innate. As Bateson (1984) points out there is quite some controversy over "innateness" . It is now generally accepted that behavior is never completely genetically hardwired, and that the interplay between internal and external factors determines behavior. When discussing the responses of parasitoids to certain stimuli, it seems most useful to talk about unlearned versus learned behavior.

PAGE 30

17 Unlearned responses involve behavior that develops without the individual experiencing stimuli to which it will respond, or without practice of the motor pattern that it will perform (Bateson, 1984). Whether or not the responsiveness of individual insects to certain odors is totally unlearned appears hard to demonstrate. The fact that the adult insects will respond to odors without ever experiencing them in association with hosts after emergence can easily be mistaken for an unlearned response. One could argue that the inexperienced females have had no chance to associate certain odors with hosts, therefore their responsiveness must be inherited. Although possible, this is not necessarily true. Information obtained during one of the immature stages might well have resulted in their responsiveness (see later). Still, it is unlikely that all odor responses have to be learned. Although most parasitoids respond to host-related odors much better after a contact experience with the host or its by-products, naive insects will always respond to some degree (Vet, 1984, Drost et al . , 1986; Herard et aK , 1988a; Eller, 1990). Recent studies on the searching behavior of the general ist parasitoid Campoletis sonorensis show a high response level for naive females (Baehrecke et aK, 1990; McAuslane et iL., 1990a, 1990b) not yet observed in other species. Experience will always affect an insects behavior to some degree. Perhaps it is best to distinguish between insects that rely only slightly on experiences and insects that rely strongly on experiences. A similar distinction could be made between the stimuli that the insects respond to. Vet et (1990) developed a model that puts the variability of responses by parasitoids to stimuli in a framework which

PAGE 31

18 takes the genetic constitution, physiological state, and environmental factors in consideration. It is clear that the complex interplay of internal and external influences is far from completely understood. Some theories on how external factors may determine a parasitoid's responsiveness have been developed. These theories are reviewed in the next paragraphs. Hopkins' host-selection principle. In 1917 Hopkins introduced the "Hopkins' host-selection principle". It was simply defined as "an insect species which breeds in two or more hosts will prefer to continue to breed in the host to which it has adapted (p. 190)" (Craighead, 1921). This vague definition could be interpreted in several ways, but has always been assumed to say that insects prefer to oviposit on the type of host they fed upon as immatures (Jaenike, 1983). Hopkins did not mention what he thought was causing this phenomenon. It seems that some authors just assume that Hopkins meant that there is a link between larval feeding behavior and preferences shown by the adult for particular oviposition site selection (Corbet, 1985). Looked at it in that way, Jaenike (1983) was right to suggest the following three shortcomings: I) it fails to rule out genetic effects since the individuals bred from a given host may be the offspring of females that inheritably prefer that host; 2) it fails to distinguish between the effects of larval versus adult exposure to the particular host; and 3) experiments to demonstrate the principle were not carried out on independent samples or over a span of many generations, to rule out random fluctuations in host preference. However, if the definition of

PAGE 32

19 Hopkins' principle is interpreted in the broadest sense, these possibilities are included in it as well. The concept "memory" is regarded as the central feature of Hopkins' host-selection principle (Jermy et £L , 1968; Jaenike, 1982). They see it as an effect on adult behavior of the early chemical environment mediated by a neural change effected in the larva and persisting to the adult stage (Corbet, 1985). In some cases it has been shown that the larval environment does affect the adult responsiveness to odors, but experiences during the adult stage seem to have more influence on the odor preferences of insects (Jaenike, 1983; Vet, 1983). The chemical legacy hypothesis. The chemical legacy hypothesis as proposed by Corbet (1985), could be seen as an alternative explanation for the Hopkins' host-selection principle. It suggests that "effects of the early environment on the chemosensory responsiveness of a later stage depend not (or not only) on 'memory', but on the direct effects on the later stage itself of a legacy of chemical cues bequeathed to it from the earlier stage. This legacy consists of minute quantities of certain chemicals that persist from one stage to another inside or outside the insects's body" (Corbet, 1985). So far, no direct evidence for this hypothesis exists. The obvious thing to do is to change the chemical environment of the larvae and study the effect of such changes, but since the hypothesis suggests that minute quantities of chemicals will be sufficient to influence the adult behavior this might be impossible. Even if it is possible to manipulate the chemical environment of the larvae, it will be hard to distinguish an effect of the chemicals on the sensory system during the

PAGE 33

20 larval period from effects of chemicals obtained by the immature and carried over into the adult stage that influence the adult's behavior. Learning and conditioning. More and more studies demonstrate that recognition of specific semiochemicalsa adult parasitoid females is partially acquired through the association of hosts and/or host products and their odors during a contact experience (e.g. Thorpe and Jones, 1937; Monteith, 1963; Arthur, 1971; Taylor, 1974; Vinson et aK, 1977; Sandlan, 1980; Strand and Vinson, 1982; Vet, 1983, 1988; Vet and van Opzeeland, 1984, 1985; Wardle and Borden, 1985; Dmoch et aK, 1985; Drost et al^, 1986; Lewis and Tumlinson, 1988). For instance, two alysiine species could distinguish between odors of host ( Drosophila larvae) infested and uninfested substrates only after experiencing an oviposition on an infested substrate (Vet and van Opzeeland, 1984). Vet (1983) demonstrated conditioning for the parasitoid Leptopil ina clavipes . Most females of this parasitoid are attracted to odors of decaying fungi (a potential habitat for its hosts). When females are given oviposition experience with host larvae that are feeding in yeast, their preference changes to yeast odors over odors of decaying fungi. Vet (1983) also found that the effects of this type of conditioning are much stronger than the effects of rearing the parasitoids on larvae feeding on yeast. The latter form of conditioning also increased response to yeast odors but less dramatically, mushroom odors were still preferred. Nevertheless, learning of odors as an immature (known as pre-imaginal conditioning) has been shown to affect the responses by the adult female significantly in some species (reviewed by Vet, 1983). None of these studies, however, consider that an emerging adult female

PAGE 34

21 may experience odors that will influence her responses. Herard et aL. (1988b) show that this may be very important. They found that the wasp Micropl itis demol itor responds well to host-related odors in a flight tunnel if it is reared on a host that was fed plant material, but not if the host was fed artificial diet. They subsequently demonstrated that the wasps from plant fed hosts required contact with their cocoon after emergence to learn the odors to which they would respond in the flight tunnel. This explains the often observed poor response by laboratory (artificial diet) reared parasitoids to host related odors. Learning seems to occur frequently in parasitoids (for a review see van Alphen and Vet, 1987) and it will be discussed more extensively in some of the following chapters. Which of the above mechanisms is more important might depend on the host specificity of the insect. Species that are able to attack several different host species can be expected to depend on experience as an adult. For those species the information obtained as a larva or a pupa might not necessarily lead it to the most abundant, easier to find and most suitable hosts. These species will find the most useful information while searching for hosts. This does not necessarily mean that specialized insects will rely more on the innate response and the chemical cues encountered as an immature. Specialists that find their hosts in a variety of habitats may use cues that are different for each habitat. Therefore, even though they will always be in search of the same type of host, they will have to alter their responses to find hosts in different habitats.

PAGE 35

22 Nonetheless, the array of chemicals a naive specialist female will respond to may be much more limited than that of a general ist. Particularly in those cases where a specialist wasp relies on cues that are very closely associated with its host unlearned responses may be more important than learned responses. It should be clear that the above discussed mechanisms can not be seen separate from each other. Some or all of them will affect a parasitoid's behavior at the same time. Analysis of Host Searching Behavior Few studies had been undertaken to elucidate the behavioral mechanisms underlying host-habitat location. Until recently, none of the available methods seemed to be suitable for such studies. However, the well designed olfactometer studies conducted by Vet and co-workers (Vet, 1984), have resulted in an avalanche of similar studies. Olfactometer studies serve well to show attraction to and preference for odors, but the confined space does not allow an insect to display all aspects of its behavior. To accomplish that, flight tunnel studies appear more suitable (Drost et aK, 1986). Initial experiments with flight tunnels failed because the insects tend to disperse upon release. By exposing parasitoids to kairomones before releasing them, however, the tendency to disperse can be overridden (Gross et jlL , 1975; Loke and Ashley, 1984c; Drost et aK , 1986). Drost et aK (1986) and Eller et aK (1988b) performed the first accurate studies of the long-range searching behavior of a parasitoid. They described the characteristics of flight by Microplitis croceioes in

PAGE 36

23 response to a host-plant complex and to individual components of such a host-plant complex in a flight tunnel. JL. croceipes females showed "sustained flights" (continuous flights that resulted in a landing on the target) towards a host-plant complex and towards damaged plants. A typical "sustained flight" started with casting (alternate left and right movements across the axis of the odor plume without net upwind or downwind movement) close to the release point. This was followed either by anemotaxis (flight track orientation directly towards the odor source) or by zigzagging (flying alternately to the left and the right of the main axis of the odor plume, while gaining or loosing ground) up to ca. 10 cm distance of the target. There, the parasitoid started hovering (hanging still in the air) followed by a darting approach to the target. An important aspect of their work is that they found that preflight contact with the host-plant complex or with host feces increased responses of the parasitoids. This phenomenon has been observed in several parasitoids and will get special attention in this dissertation. The effect of experience with host and/or host-related products opens possibilities for manipulation of parasitoid behavior for biological control purposes. The current status of this particular research area will be discussed in the next section of this chapter. Manipulation of Parasitic Wasps for Biological Control Purposes Agriculture dealing with a constantly increasing demand for its products, is fighting an ever lasting war against pest insects.

PAGE 37

Conventional chemical pesticides are indispensable in this war, but the use of pesticides has resulted in some basic problems: 1) insects may develop resistance to broad spectrum chemicals, resulting in the demand for new more powerful and more expensive pesticides; 2) many beneficial insects and other organisms perish from the use of pesticides; 3) outbreaks of secondary pests that were previuously not of economic importance; and 4) environmental contamination, which rightfully receives more and more public protest. Insect pest control scientists now agree that the alternatives have to come out of the biological control sector (van den Bosch and Messenger, 1973; Lewis, 1981). Biological control has traditionally been used to describe the regulation of pests with the use of natural enemies. In a much wider view, biological control involves many other naturally derived strategies for controlling pest populations, including the manipulation of natural enemies. This review will focus on manipulative strategies that have been applied, that are under study, and that may be found useful in the future to increase the effectiveness of parasitic insects as controlling agents. The strategies that will be discussed are divided into two main categories: 1) resource management, and 2) semiochemical manipulation of behavior. The field of genetic manipulation is is not discussed here. Resource Management The maintenance of parasitoid densities that will be able to control pests effectively may be realized by supplying extra resources like nectar sources for the adults, hosts during low host densities and

PAGE 38

maybe even shelter against unfavorable environmental conditions. Several studies have already shown that provision of specific resources results in better control of pest populations (for reviews see Herzog and Funderburk, 1985; Lewis and Nordlund, 1985; Hagen, 1986). Providing food sources for adult parasitoids. Agricultural monocultures often lack food sources, such as nectar and pollen producing weeds, which under more natural circumstances would provide adult parasitoids with their nutritional requirements. This problem is in some cases amplified by selection for tolerant and resistant crops, such as cotton that lacks extra-floral nectaries (Schuster, 1980; Herzog and Funderburk, 1985). Drake (1920) was one of the first to suggest the use of what he called a trap crop to attract and feed the tachinid wasp Trichopoda pennipes Fabricius. Drake argued that the trap crop, Crotalaria usaramoenis . has floral nectaries that attract the wasp and pods that are attractive to the parasitoid's host, the stink bug Nezara viridula L. Since then several attempts to provide additional food sources conjunction with a crop have had positive effects on retaining and maintaining parasitoids and predators in high enough numbers (Gardner, 1938; Hagen et aK, 1971; Doutt and Smith, 1971; Leeper, 1974; Beglyarov and Smetnik, 1976; Matteson et a]^, 1984). Providing additional hosts. Smith and DeBach (1953) suggested that artificial infestation of plants with hosts during periods of low host density is a means by which host-parasitoid populations could be synchronized and could result in increased effectiveness of parasitoids in achieving biological control. Knipling and McGuire (1968) had

PAGE 39

26 something similar in mind when discussing the effectiveness of Trichoqramma species. They, however, suggested a sustained addition of host eggs to the environment irrespective of the actual density of the insect pest. The practicality of these concepts was demonstrated with several field experiments (e.g. Parker et aL., 1971; Gross et al . , 1984) . No examples exist of active introduction of alternative hosts into an area infested by a pest insect. However, growing certain plants near the target areas that carry alternative hosts has been found to be effective (Doutt and Smith, 1971). Semiochemical Manipulation of Parasitoid Behavior Successful reproduction of a parasitoid female depends partially on the proper use of the allelochemics in its macro-habitat. Recent research shows that man can manipulate the parasitoid behavior with chemicals used for host location. Therefore, by applying specific allelochemics at the right time and in the right place it should be possible to manipulate parasitoid behavior to our advantage (Nordlund et al^, 1981a; Nordlund et aK , 1981b; Lewis, 1981; Lewis and Nordlund, 1985) . Application of allelochemics in the target field. Application of the allelochemics that serve as cues used by parasitoids to locate suitable hosts may result in significantly higher parasitization rates in certain crop systems. The additional allelochemics probably help to retain the parasitoids in the area and may also increase their motivation to search for hosts.

PAGE 40

27 Lewis et aL. (1972) were the first to demonstrate this effect. A kairomone for Trichogramma evanescens present in the scales of Hel iothis zea (Jones et aL., 1971) was extracted with hexane. In laboratory as well as in greenhouse and field experiments it was shown that \L_ zea eggs placed on cotton or pea seedlings that were treated with the extract were parasitized significantly more than eggs on plants that were treated with only hexane. In both the greenhouse and field experiments the level of parasitism and the number of adult parasitoids produced were about twice as high on the leaves treated with kairomone. Even more spectacular results were obtained from field experiments in crimson clover by using tricosane, one of the kairomones identified from moth scales by Jones et aL. (1971). Parasitization by naturally occurring Trichogramma spp. was increased from 4% to 15%. The pattern in which the treatment is applied to the field was found to be very important (Lewis et aL, 1979). At low host densities constant exposure to allelochemics should be avoided to ensure movement by the parasitoids from one oviposition site to another. Nordlund et aL. (1983) observed an increase in parasitization rate by the parasitoid Telenomus remus , which attacks egg masses of SpodoDtera frugiperda (J.E. Smith), after placing cotton rolls treated with frugiperda pheromone near egg infested cowpea seedlings in a greenhouse. Nordlund et aL (1983) report on how they were able to induce significant parasitization by L. remus in eggs of the non-host Hel iothis zea by treating the eggs with frugiperda pheromone. Some crops are more attractive to certain parasitoids than others crops. Nordlund et aL (1984) found, when testing different cultivars

PAGE 41

with combinations of tomato, bean, or corn, that parasitization by Trichoqramma spp. in cultivars that included tomato was higher than in those without tomato. Parasitization on corn treated with tomato extract was higher than in untreated corn. Parasitization on tomato treated with corn extract was not different from parasitization on untreated tomato. These results indicate that a synomone present in tomato plants "improves" the performance of Trichoqramma spp. Nordlund and Sauls (1981) showed that for the parasitoid Micropl itis croceipes the kairomonal activity of frass from Hel iothis zea depends on the plant diet of the host. Female parasitoids responded to extracts of frass from larvae reared on cotton or soybean but not on corn. The lack of response to the frass of corn-fed larvae was due to the absence of some appropriate chemicals. Besides being useful to attract parasitoids, motivate them to search for hosts and to retain them in a target area, allelochemic attractants may also serve in trapping parasitoids to monitor their establishment and their dispersion. Application of al lelochemics for pre-release stimulation of parasitoids. The tendency of insects to disperse upon release has often resulted in poor establishment of parasitic insects (Lewis and Nordlund, 1985). Gross et aK (1975) were the first to find that the tendency to disperse can be overcome by exposing parasitoid females to kairomones prior to release. They allowed female Micropl itis croceipes to contact Heliothis zea frass before being released close to zea larvae feeding on cowpea leaves in a greenhouse. These pre-stimulated females readily

PAGE 42

parasitized the larva, while control females that were not stimulated before release did not parasitize any larvae. Gross et aL. (1975) also found that contact with the kairomones present in scales of iL zea stimulate host-searching activities in two Trichoqramma spp. Stimulated parasitoids had higher rates of parasitism in laboratory experiments than had unstimulated parasitoids. Prerelease stimulation also increased the efficiency of Trichoqramma pretiosum in the field. Recent research has established that the effect of increased response to host odors after experience is, at least in some cases, caused by associative learning (Vet, 1983; Vet and van Opzeeland, 1984; Lewis and Tumlinson, 1988). In some cases contacting larval by-products causes the females to link the associated or surrounding odors with the possible presence of hosts. Subsequently, the parasitoids will use these odors as cues in their search for more hosts. The first chapters of this dissertation deal with this phenomenon in marqiniventris . Lewis and Tumlinson (1988) showed that Microplitis croceipes uses both a contact kairomone and a volatile kairomone present in the feces of Hel iothis zea larvae. They were able to separate the two types of kairomones and found that the parasitoid females associated the surrounding odors with the contact kairomones after they rubbed the feces with their antennae. The parasitoid females could be fooled into flying to the odor of vanilla after they had rubbed the contact kairomone in the presence of vanilla extract. These findings offer promising possibilities for biological control. If applied in the right way it should be possible to increase

PAGE 43

30 parasitoid responsiveness to host related allelochemics, and also to condition them to respond to those odors that will lead them to the target species. It might even be possible to condition parasitoid females that they will focus on a host species that they normally would not prefer. Agriculture could profit from these findings if we are able to identify and synthesize the allelochemics involved, and apply them as pre-release conditioners of parasitoids used in biological control projects. In the following chapters the manipulation with semiochemical s will be explored for the parasitoid Cotesia marqiniventris . First, the available literature on this insect will be reviewed. Cotesia marqiniventris (Cresson) a Prime Candidate for Auqmentative Release Aqainst Lepidopterous Insects The larval endoparasitoid Cotesia marqiniventris (Cresson) is one of the most frequently recorded parasitoids in larvae of moths of the family Noctuidae. C^ marqiniventris attacks many species (Table 1-2), several of which are economically important pests in the USA. Its successful establishment in many different habitats makes it a promising candidate for biological control. Efforts should be undertaken to increase the parasitization efficiency of C^ marqiniventris . Systematics L. marqiniventris (Cresson) was originally described from Cuba (Muesebeck, 1921) and is considered a native to the West Indies. Over

PAGE 44

31 the years this parasitoid has been known under 6 different names (Marsh, 1978; Krombein et al^, 1979). Microqaster marqiniventris Cresson, 1865:67. Apanteles grenadensis Ashmead, 1900:278. Apanteles laphyqmae Ashmead, 1901:38. Nomen nudum. Apanteles ( Protapantel as ) harnedi Viereck, 1912:580. Apanteles marqiniventris (Cresson), (Muesebeck, 1921). Cotesia marqiniventris (Cresson), (Mason, 1981). C^ marqiniventris is a member of the Braconidae family. This family can be distinguished from related families by a characteristic fore wing venation. Braconids possess only one recurrent vein and usually have a (disco-)cubital vein that separates the discoidal and cubital cells. A third characteristic is the fusion of the second and third abdominal tergites which makes the suture between them inflexible. L. marqiniventris is currently classified under the largest and taxonomically difficult subfamily, the Microgastrinae (Matthews, 1974). As most braconids, Microgastrinae are endoparasitoids of immature stages of lepidopterous species. In 1965 Nixon published a correction on the classification which for the first time united the Microgastrinae by the location of the spiracle of tergum I on the lateral membranous margin. That character is still mostly used to distinguish them from other subfamil ies, Mason (1981) introduced the currently used classification of the Microgastrinae into five tribes: Apantelini, Microgastrini , Forniciini,

PAGE 45

32 Cotesiini, and Micropl itini . marqini ventris belongs to the Cotesiini. For a quick diagnosis of this tribe, Mason (1981) suggests: Hypopygium short and evenly sclerotized; ovipositor short, stout basally, and abruptly tapered about mid-length; sheat short with hairs concentrated apically, arising proximally from valvifer. Propodeum with median carina or none; very rarely areolate. Areolet usually open, but if closed hind coxa longer than tergite I and tibial spurs longer than half basitarsus. For further identification of the adults the keys given by Marsh (1971) or Mason (1981) (to genus) and Muesebeck (1921) (to species) are most useful . Host and Plant Range No individual study has been undertaken to determine the actual host range of marqini ventris . However, numerous publications present field data for collected caterpillars from which marqiniventris was reared (Table 1-2). Together with several reports on laboratory reared Cj. marqiniventris an obviously incomplete picture of this parasitoids' host range can be drawn. Table 1-2 gives the species from which marqiniventris has been reported to emerge. So far, the insect has been reared from three families of Lepidoptera: Noctuidae (14 species), Pluttidae (1 species), and Pyralidae (4 species). The authors have looked mostly at lepidopterous species of economic interest. Undoubtedly marqiniventris host range will include many more species, possibly several in other lepidopterous families. Table 1-2 also lists close to 30 different species of plants on which the hosts for marqiniventris were found. This relatively large host and plant range includes several serious pest insects and important

PAGE 46

33 Table 1-2. Reported host species for marqini ventri s . HOST SPECIES PLANTS COUNTRIES + common name(s) (ref. #) Noctuidae: nc 1 lULillb Virc5>CcllJ> ^raUi ICIUb^ d 1 Id 1 la — tuudccu uuuwurin U 1 LU 1 Ur 1 cJ>pt:UcZ.d 11 1? Ifi 17 L.r dilcoU i 1 1 ?? ?7 "^R ^fi "^7 L.I lllloUll LrlUVCi 3Q 41 AO AA AC. ^iii'loa'F novani iim v«UU Icdl Clll lUHl 46,48,52) RiipI 1 i ^ run vnn i i O GIIIIC f narl"f 1 AY tUUdLLU 4" nma t n Vc 1 Vc L 1 cdT Vc LCII Heliothis zea (Boddie) alfalfa USA = corn parwnrni hnllwnrm tnmatn I II cai nut III f 1 inui III) LL/lllaLu mm V^U 1 II Ml 1 ? Ifi 17 9? f ru i tworm hirolnr lp<;npHp7a LyiWIUI ICOLiCVJCZ.U 27 28 29 35 37 rotton 38 41 42 44 45 crimson clover 47 48 50 53 55i cutleaf geranium deergrass peanuts potato Ruellia runvonii soybean toadflax tomato velvet-leaf PlathvDena scabra (F.) alfalfa USA = green cloverworm corn (7,15,24,26,29, legumes 34,36,47,53,54, soybean 56) Pseudaletia uniouncta (Haworth) alfalfa USA = armyworm corn (20,36,47,53,54) Pseudoplusia includens (Walker) alfalfa USA = soybean looper corn (9,18,28,29,50) pigweed tomato

PAGE 47

34 Table 1-2. Continued. HOST SPECIES + common name(s) PLANTS COUNTRIES (ref. #) SDodoptera fruqiperda (J.E.Smith) = fall armyworm alfalfa bermuda grass cabbage corn cotton peanut sorghum USA(3,4,5,6,8, 13,19,30,33,39, 43,47,53,54,55) BRAZIL (2) COLUMBIA (3) LESSER ANTILLES (2) MEXICO (39) NICARAGUA (2) SURINAM (2) URUGUAY (2,3) VENEZUELA (2,3) Soodoptera eridania (Cramer) = southern armyworm USA (36) Soodoptera exempta (Walker) = nutgrass armyworm grasses sugar cane USA-HAWAII (introd. ) (8,14,36,40) SpodoDtera includens (Hubner) = beet armyworm alfalfa corn pigweed tomato USA (1,17,20,47,50, 51,53,54) MCV T rCk MtAlLU Soodoptera 1 itura (F. ) beetroot cabbage castor caul ifl ower cowpea knol -kohl tobacco INDIA (introd. ) (21,23) Soodoptera mauritia (Boisduval) = lawn armyworm bermuda grass USA -HAWAII (introd. )(14, 49) Spodoptera ornithoqall i (Guenee) = yellowstriped armyworm alfalfa cotton peanuts sesame USA (38,47,50,55) SoodoDtera oraefica (Grote) = western yellowstriped armyworm alfalfa USA (31,32) TrichoDlusia ni (Hubner) = cabbage looper Brassica areens cotton weeds USA (9,10,18,25)

PAGE 48

35 Table 1-2 Continued. HOST SPECIES + common name(s) PLANTS COUNTRIES (+ ref. #) Plutellidae: Plutella xvlostella (L.) = diamondback moth Brassica greens USA 1 (17,25) Pyral idae: Achvra rantalis (Guenee) = garden webworm alfalfa USA (47) Diaphania nitidalis (Stoll) = pickleworm squash USA 1 (30) Herpetoqramma bipunctalis (F.) = southern beet webworm pigweed (in corn fields) USA (51) Hvmenia perspectalis (Hiibner) = spotted beet webworm USA (36) 1 References: 1. Alvarado-Rodriguez (1987) 2. Andrews (1988) 3. Ashley (review) (1979) 4. Ashley et al . (1980) 5. Ashley et al . (1982) 6. Ashley et al . (1983) 7. Barry (1970) 8. Bianchi et al . (1944) 9. Boling & Pitre (1970) 10. Boling & Pitre (1971) 11. Burleigh (1975) 12. Burleigh & Farmer (1978) 13. Butler (1958) 14. Clausen (review) (1978) 15. Danks et al . (1979) 16. Graham et al . (1972) 17. Harding (1976a) 18. Harding (1976b) 19. Hogg et al . (1982) 20. Hotchkin & Kaya (1983) 21. Jalali et al . (1987) 22. King et al . (1985) 23. Krishnamoorthy & Mani (1985) 24. Kunnalaca & Mueller (1979) 25. Latheef & Irwin (1983) 26. Lentz & Pedigo (1975) 27. Lewis & Brazzel (1968) 28. McCutcheon & Harrison (1987) 29. McCutcheon & Turnipseed (1981) 30. McFadden & Creighton (1979) 31. Miller (1977) 32. Miller & Ehler (1978) 33. Mitchell et al . (1984) 34. Mueller & Kunnalaca (1979) 35. Mueller & Phillips (1983) 36. Muesebeck et al . (1951) 37. Neunzig (1969) 38. Pair et al . (1982) 39. Pair et al . (1986) 40. Pemberton (1948) 41. Puterka et al . (1985) 42. Roach (1975) 43. Rohlfs III & Mock (1985) 44. Shepard & Sterling (1972) 45. Smith et al . (1976) 46. Snow et al . (1966) 47. Soteres et al . (1984) 48. Stadelbacher et al . (1984) 49. Tanada & Beardsley (1958) 50. Teague et al . (1985) 51. Tingle et al . (1978) 52. Tingle & Mitchell (1982) 53. Vickery (1925) 54. Vickery (1929) 55. Wall & Berberet (1975) 56. Whiteside et al . (1967)

PAGE 49

crop plants. The potential of this parasitoid to control these pests and protect the plants will be discussed later. Geographical Range marginiventris has been reported from Brazil, Columbia, Cuba, Mexico, Nicaragua, Surinam, Uruguay, the USA (Continental), Venezuela, and the West Indies (Wilson, 1933; Marsh, 1978; Ashley, 1979; Danks et al . , 1979; Andrews, 1988). Within the continental United States reports have been mostly from the southern and central states. C. marginiventris has been collected in Alabama, Arizona, Arkansas, California, North and South Carolina, Delaware, Florida, Georgia, Iowa, Kansas, Louisiana, Mississippi, Missouri, Oklahoma, Tennessee, Texas, and Virginia (references in Table 1-2). Furthermore, C. marginiventris has been successfully introduced to Hawaii (Bianchi, 1944; Clausen, 1978) and India (Jalali et aK , 1987). Biology The life cycle of a £_=. marginiventris female is depicted in Figure 1-2. The different stages of this cycle are each discussed in more detail in the following paragraphs. Immature stages . Boling and Pitre (1970) described the immature stages. At oviposition the eggs are cylindrical with rounded ends (0.088 mm long, 0.017 mm wide), with a short, slightly bent, penducle (0.005 mm long) at the caudal end. After 24h the eggs increase in size to 0.041 mm by 0.0211 mm. The eggs float freely in the hemocoel of the host, and hatch 18 to 36 hours after oviposition at 30 °C.

PAGE 50

37 Figure 1-2. Schematic depiction of the life-cycle of marqiniventris . 1) egg inside the host caterpillar, 2) second instar larva inside the host, 3) final (third) instar larva chews its way out of the host, 4) cocoon in which the parasitoid larva pupates, 5) adult parasitoid emerges from the cocoon, 6) adult female searches for the micro-habitats of potential hosts, 7) female examines host frass, 8) oviposition.

PAGE 51

38 Boling and Pitre (1970) found the larval stages mostly in the posterior part of the host, never in the head. The white caudate first instar larvae were never found attached to the host and measured 0.05 mm to 0.06 mm 24h after egg hatch. It possessed a caudal appendage as described by Clausen (1940). No cannibalism has been observed in this species, but the first instar larvae have a scleroterized head with distinct mandibles and labium. This suggests an ability to attack competitors as has been shown for related species (Allen, 1958). The first stadium lasts for 1.5 to 2 days. The white vesiculate second instar is more robust, has no visible scleroterized head, and seems less suitable for physical combat. It possesses a very prominent anal vesicle. As is common in parasitic Hymenoptera (Clausen, 1940), the second instar significantly increases in size to 2.5 mm long and 0.59 mm wide, just before the moult to the third instar. The second stadium lasts approximately 2 days. The moult to the third and final instar takes place just prior to emergence from the host. The third instar larva is 5.5 mm long and 1.0 mm wide, is creamy white at first and turns yellow to brown upon emergence. Color of both the larvae and the pupae appear to depend on the parasitoid's diet. Parasitoid larvae emerge by biting their way out through the cuticle of the host larva with their well developed mouth parts. When kept at 30 °C this occurs 6-11 days after egg deposition. The greatest number emerge from Hel iothis virescens after 6 days, from Trichoplusia ni and Pseudoplusia includens after 7 days (Boling and Pitre, 1970), and from Plathypena scabra after 8 days (Kunnalaca and Mueller, 1979). At a

PAGE 52

39 lower temperature emergence from the host takes longer, 9-15 days for scabra at 25 °C. Preliminary observations in our laboratory indicate an average emergence after 7 days from Spodoptera fruqiperda at 25 °C. Emerging parasitoid larvae immediately start spinning a crescent-shaped cocoon on, or in most cases in the vicinity of, the host. The parasitoid first constructs half a shell of the white or yellow cocoon and then crawls inside of it with its posterior end first. Subsequently, the open side is closed with up and down spinning movements followed by cross spinning to reinforce the sides (Boling and Pitre, 1970). To attach the cocoon to a substrate the larva will spin a fine mesh on the bottom when the case is still thin. Then it resumes the spinning of the cocoon. Two hours after it initiated cocoon formation, the larva, barely visible, was still spinning (Boling and Pitre, 1970). On corn and grasses the orientation of the cocoons is almost invariably parallel to the longitudinal axis of the leaf, and is usually placed on the median groove of the leaf (Bianchi, 1944). Approximately 24 hours after formation of the cocoon the parasitoid pupates (at 30 °C) and an adult will emerge from the cocoon 3-5 days later. A wasp exits from a cocoon by gnawing a smooth circular cut near one end, resulting in a conical trap door which is lifted up by the emerging wasp (Bianchi, 1944) (see drawing 5 in Figure 1-2). Data presented by Vickery (1925) emphasize how sensitive developmental time is to temperature. He found that time from oviposition until emergence from the cocoon takes about ten days during summer time, but can take as many as 27 days during cold months.

PAGE 53

40 Adults . Developmental times from cocoon to adult takes 4-7 days at 25 °C (Kunnalaca and Mueller, 1979) . Males emerge on average a day before the females. Longevity for adults is clearly temperature dependent. Kunnalaca and Mueller (1979) found that in the laboratory when provided honey water at 30 °C and 25 °C the wasps would live 5.6 + 2.5 and 9.1 + 4.2 days, respectively. Females live longer than males (Kunnalaca and Mueller, 1979; Jalali, 1987). Mating . Boling and Pitre (1970) observed that several minutes after emergence from their cocoon, female wasps are willing to mate. A male approaches a female from the rear, vigorously fanning its wings and tapping her with his antennae. My observations differ from those by Boling and Pitre (1970), who claim that mounting by the male takes only half a second. In our laboratory copulation lasted roughly 5-15 seconds. We also found that once mated, females were not readily willing to mate again. This too contrasts with observations by Boling and Pitre (1970) who observed both sexes to mate freely and many times. Further detailed studies will have to elucidate these contradicting observations. It is possible that females are willing to mate again after they have oviposited in a host. Loke and Ashley (1984a) found that mated females respond more intensely to contact kairomones than unmated females, and they are more ready to search for hosts. Host finding behavior. Studies on host finding behavior by marginiventris have only been performed at the close-range level (Loke et aK , 1983; Loke and Ashley, 1984a, 1984b, 1984c; Dmoch et aL, 1985). Loke et aK (1983) described the host-finding behavior of marginiventris on corn plants and reported that parasitization is

PAGE 54

stimulated by plant damage caused by fall armyworm larvae. Females spend significantly more time on leaves damaged by hosts than on artificially damaged leaves or undamaged leaves, which was also reflected in a higher parasitization rate on the host damaged plants. Loke et al_i. (1983) divided the behavioral sequence of close-range host finding into 12 steps (Figure 1-3). The asterisks in Figure 1-3 represent locations where fixed-action patterns were mediated by stimuli that release specific behavioral patterns. Although this analysis of the different behavioral steps looks very acceptable, Loke et aK (1983) concluded too soon that chemotaxis is involved. However, to draw such a conclusion, it is necessary to demonstrate that contacting certain chemicals isolated from the habitat is sufficient to evoke a specific part of the host-finding behavior. To do this, Loke and Ashley (1984b) isolated several components of the host habitat complex, made extracts out of them, and studied the behavior of female parasitoids when they came in contact with the extracts on filter paper. It appeared that L. marqiniventris responded to all of the components of a host habitat with hosts. The strongest responses were elicited by 1) fall armyworm frass 2) moth scales 3) and exuviae The females were also found to be responsive to damaged leaves, saliva, silk and larval and cuticle material. Loke and Ashley (1984b) found hardly any difference in how marqiniventris responded to frass produced by larvae fed on different plants. However, they did find that plant material is needed as food for the hosts to produce frass with the

PAGE 55

(3) (4) contact ^ damage antennal (5) chemotaxis \ palpation ^' ^ \ ' ' ovipositor (6) * (2) orientation \ , / . ^^^^^ (1) nonsearching ^ ^ preening <] contact (7) movement host ^-^'^ ^ ^^^^ A ^/ (12) resting ^ / mounting (8) (11) ovipositor stinging (9) ^ smeering ^^oviposition (10) ^ Figure 1-3. Behavioral ethogram of the host finding and ovipositional sequence of marqini ventri s females on corn plants damaged by fall armyworm larvae. (Solid arrows indicate invariable pathways ; dashed arrows, variable or alternate pathways; asterisks, sign stimuli present). Taken from Loke et aL. (1983).

PAGE 56

r "I. . 43 right kairomones. It appeared that the parasitoids were less responsive to the frass of larvae fed on a lab diet. Frass extracts can significantly increase parasitization rates by marqiniventris . Even the normally not accepted factitious hosts velvetbean caterpillar and wax moth are parasitized if they are sprayed with extracts of fall armyworm frass (Loke and Ashley, 1984c). Dmoch et (1985) confirmed some of the conclusions of Loke and Ashley (1984b), but also demonstrated the importance of learning in the host-finding process for marqiniventris . This phenomenon of learning forms one of the key elements of the research presented here. Oviposition. Females are ready to oviposit shortly after emerging from the cocoon, but are more inclined to do so a day or more after emergence (Boling and Pitre, 1970; Loke and Ashley, 1984a). Females on a substrate with host larvae travel quickly over the substrate while touching it vigorously with their antennae. Upon finding a host they will insert one egg with a quick thrust, and subsequently move away from the host. The wasps appear not to be egg limited, and will lay eggs throughout their adult lives (more than 100 eggs under laboratory conditions) (Kunnalaca and Mueller, 1979; Jalali et al^, 1987). Miller (1977) concludes that marqiniventris can lay viable eggs within 24h after emergence from the cocoon. He also found that the female wasp caries 20+4 ovarioles per ovary and a total of 210 + 25 ova. Females will lay just one egg during an oviposition attempt but may return to the same host several times to lay more eggs. Dmoch et aL. (1985) present evidence showing that experienced females are able to discriminate parasitized and unparasitized hosts. The female wasps

PAGE 57

44 oviposit in unparasitized host more readily than in already parasitized hosts. However, Boling and Pitre (1970) found up to 7 eggs in the hemocoel of a single host, although normally only 1 egg was found. They also report a few cases where 2 eggs were dissected from field collected larvae. Normally only one parasitoid will emerge from one host. Kunnalaca and Mueller (1979) found just one example of two wasps emerging from a green cloverworm larvae that had been parasitized more than once in its 3rd-instar. These authors also found that 1stand 2nd-instar were the preferred larval stages for oviposition by marqiniventris . Loke and Ashley (1984a) observed the same preference for early instars of fall armyworm, while Jalali et aL. (1987) came to the same conclusion when they looked at host age preference for the novel host Spodoptera 1 itura in India. The preferences were solely concluded from parasitoid emergence after choice tests. The observations may therefore have been affected by differential survival rates. Parasitism by L. marqiniventris occurs primarily during daylight hours (Kunnalaca and Mueller, 1979). Wasps were allowed to oviposit in green cloverworm larvae from 6-8 a.m., 6-8 p.m., and 10-12 p.m. They parasitized 49%, 68%, and 3% respectively. This is largely in agreement with findings by Loke and Ashley (1984a), who found that the female wasps responded to contact kairomones throughout the photophase. Host Regulation Vinson and Iwantsch (1980) introduced the term "host regulation" which refers to the many physiological and biochemical changes that take

PAGE 58

45 place in a parasitized organism. Considering the wide range of hosts in which marqiniventris can develop successfully, a sophisticated host regulation process can be expected for this parasitoid. From our own observations, we can conclude that the parasitoid dramatically slows down the growth, food utilization and, developmental rate of the host larva. Since the marqiniventris attacks its host in an early stage this will eventually result in a enormous decrease in plant consumption, something we will consider later when the parasitoids' potential as a bio-control agent is discussed. So far, Ferkovich et aK (1983) undertook the only serious attempt to study the internal changes in hemolymph proteins of Spodoptera fruqiperda . They detected that as early as 4 hours after parasitization several high-molecular-weight proteins were present in the host hemolymph that showed up much later in unparasitized larvae. Two new protein bands that never appeared in the electrophoretograms of unparasitized host showed up in parasitized hosts after 8 hours and beyond. Clearly, the parasitoid affects the host bio-chemistry the significance, but sources of these changes are far from being determined. One of the possible causes of changes in the hosts are symbiotic baculovirus particles that are injected with calyx fluid into hosts by many braconid parasitoids (Stolz and Vinson, 1979; Stolz et aK, 1981). Styer et al^ (1987) discovered a long, filamentous virus (CmFV) replicating in the hypodermal and tracheal matrix cells of larvae parasitized by C^ marqiniventris . Hamm et al^ (1990) describe this non-occluded baculo-like virus and a polydnavirus which both originate

PAGE 59

«6 from the reproductive tract of the parasitoid and are injected with her calyx fluid into the host. The polydnavirus appear not to replicate inside the host (Styer et aL., 1987). As has been demonstrated for other parasitoids (Stolz and Vinson, 1979; Tanaka, 1987), the viruses are thought to suppress the immune responses of the host and effect the host physiology to allow parasitoid development. Bio-control Control potential . Several examples exist that indicate that natural populations of margin iventris already prevent pest outbreaks. Whiteside et al^ (1967) found that marqiniventris is one of the most important parasitoids of the green cloverworm in Delaware. Kunnalaca and Mueller (1979) suggest that due to the parasitization activity of L. marqiniventris green cloverworm populations seldom reach economic importance in Arkansas. McCutcheon and Turnipseed (1981) concluded that marqiniventris is the most important parasitoid of lepidopterous species in South Carolina. They collected larvae of the green cloverworm, corn earworm, soybean looper, and velvetbean caterpillar from soybean fields in South Carolina. Of the thirty-three species reared from these larvae, C^ marqiniventris was the most predominant parasitoid for the first three pest species. C^ marqiniventris has never been reared from velvetbean caterpillar, but McCutcheon and Turnipseed (1981) reared one parasitoid with a cocoon similar to C^ marqiniventris that did not develop to maturity. C^ marqiniventris is also one of the most common larval parasitoids of two other major pests Heliothis virescens and \L zea (King and Coleman, 1989).

PAGE 60

47 marqiniventris should be particularly effective as a control agent against one of its main hosts the fall armyworm (Gross and Pair, 1986; Lewis and Nordlund, 1980). It is one of the most predominant early attackers of fall armyworm particularly in whorl -stage corn (Ashley, 1979; Hogg et aL., 1982; Pair et aK, 1986). If present in sufficient numbers when host densities are relatively low, it should be able to prevent pest population buildup. Gross and Pair (1986) characterize marqiniventris as an r-strategist (i.e. reproductive investments seem to be mostly concentrated on quantity, rather than quality), considering its fast developmental rate, relatively small size, and the females' large egg load, this would be legitimate. As an r-strategist its population buildup should be relatively fast. In general r-strategists should be given priority in introduction and augmentation programs, especially when they are capable of preventing early host population buildup (Pschorn-Walcher, 1977; Ehler and Miller, 1978). For the control of fall armyworm Gross and Pair (1986) suggest the augmentative release of a parasitoid such as marqiniventris in the host's overwintering areas, thereby preventing dispersing populations from reaching economic levels throughout the geographic range. More specifically, Knipling (1980) proposed releasing 2,000 larval parasitoids per acre in overwintering areas, which should result in effective suppression of fall armyworm. marqiniventris has an early effect on its hosts. Ashley et aL. (1983) found that of the parasitoids that attack fall armyworm, marqiniventris allows the least amount of development of the host. Parasitization reduces maximum larval weight by 97 %, and results in the 1 '

PAGE 61

destruction of the host when it reaches the 4th instar (Ashley, 1983). marqiniventris reduces host development significantly more than the other common fall armyworm parasitoids, Campoletis grioti , Chelonus insularis , and Eiphosoma vitticole . Jalali et aL. (1988) demonstrated that Spodoptera 1 itura larvae also consume significantly less food 72hr after parasitization by marqiniventris than unparasitized larvae. At a later stage the reduction in consumption becomes as high as 97 %. Hence, marqiniventris not only prevents pest population breakouts, but reduces the destructive potential of the existing population as well . Costs . Despite its potential to reduce crop damage, C. marqiniventris has not yet been used in augmentative release programs. This is probably mainly because of the great expenses involved in its rearing by conventional means (Greany et aK, 1984). Costs could be dramatically reduced with jn vitro rearing, but this technique has so far only worked for egg and pupal parasitoids (Thompson, 1986). With the exception of Trichoqramma spp., natural enemies of lepidopterous pest can not yet be reared in adequate numbers for augmentative field releases. Studies to develop in vitro rearing for larval parasitoids including C^ marqiniventris are currently being conducted (Greany, 1986; Greany et aK, 1989). However, even if costs of rearing parasitoids such as C^ marqiniventris remain high, their potential as control agents should be considered. Well timed releases of carefully manipulated wasps may have significant effects on pest populations, even when they are only released in small numbers. Costs of pesticide applications are

PAGE 62

49 and their subsequent detrimental effect to the environment are also very high. Bio-control experts agree that this wasp has the potential of becoming a major control agent, particularly if maintenance of effective parasitoid populations can be enhanced (Lewis and Nordlund, 1984). McCutcheon and Turnipseed (1981) state "studies to determine factors which may influence the abundance and parasitic activity of C^. marqiniventris should enable us to develop more effective management strategies for lepidopterous pests (p. 73)." From Loke and Ashley (1984a, 1984b, 1984c) we learned that within the micro-habitat of its hosts the female wasp uses contact kairomones to track its hosts. No information is yet available on what strategy marqiniventris relies to locate the (micro)habitats of its hosts. Studies on other parasitoids demonstrate more and more that semiochemical s are the major cues used by female parasitoids to locate host habitats. To elucidate the complete strategy used by marqiniventris to find its hosts and to determine which semiochemical s are involved, a combined behavioral and chemical study was conducted. The results of that study are presented in this dissertation. The following section states the research aims of that study. Research Aims The goal of the research was to obtain information that will contribute to the development of effective biological control of pest insects by augmentative releases of indigenous parasitic insects. So far, this type of biological control has only been marginally successful

PAGE 63

50 because of some reoccurring problems. Most attempts have failed because the wasp disperse upon release and exhibit poor searching efficiencies. To master these problems, more knowledge of the host searching behavior of the parasitoids is required, and techniques need to be developed that will enhance their performances. The potential of marqiniventris to become a major control agent has been well established. However, the only information available on the host-searching behavior of marqiniventris pertains to its closerange searching behavior (i.e. within the micro-habitat of its hosts). The success of a released parasitoid, however, is largely dependent on the efficiency with which it finds the micro-habitats of the target hosts. Long-range host searching behavior is therefore the link in a parasitoid's life-cycle that requires study for possible improvement. More specifically, techniques need to be developed that will enable us to manipulate the behavior of marqiniventris females such that after release in a pest infested area they are retained in that area and will focus their searching on the target pest. As shown earlier, preliminary research indicates that parasitic wasps can be conditioned to respond more intensely to specific host-related cues. In olfactometric studies this possibility needs to be investigated for marqiniventris . These studies should reveal whether the search efficiency of this parasitoid can be increased by experiencing the wasp with host or host by-products, and whether the preferences for certain volatile cues are determined by such experiences. Pre-release conditioning appears only practical by artificial means, therefore it will be necessary to identify not only the chemicals

PAGE 64

51 that induce this phenomenon, but also the chemicals that wasps subsequently respond to. In future release programs such information would allow the conditioning of wasps with artificial substrates impregnated with synthetic chemicals. In an ideal situation this technique would not just increase responsiveness of the wasps, but also help the wasps to concentrate on the targeted pest. The specific aim of the present study was to determine which chemical cues and behavioral mechanisms enable marqiniventris females to locate hosts over longer distances. Aspects of the long-range host searching behavior of marqiniventris were studied with the intent to do the follwing: 1) determine to what degree this behavior is mediated by volatile semiochemical s associated with hosts, 2) find out how experience with these chemicals affects the insect's behavior and preferences for specific odors, 3) describe the flight responses exhibited by marqiniventris females to host-related odors, 4) establish the exact source of the chemicals that mediate this behavior, 5) isolate, analyze, and identify the active semiochemical s, and 6) elucidate the different interactions between the three trophic levels involved (plant, host, and parasitoid) and determine how these interactions affect successful host location by marqiniventris .

PAGE 65

CHAPTER II INCREASED RESPONSE TO HOST RELATED ODORS AFTER A FORAGING EXPERIENCE: A MATTER OF LEARNING? Introduction In many species of insect parasitoids semiochemicals associated with the hosts and/or the host habitats play a major role in the host location process (Vinson, 1976, 1984). The optimal way for a parasitoid to take advantage of the available semiochemicals may vary in time. For those parasitoid species which can develop in more than one host species, the host species that is predominant at one time may be rare or absent at another. Also, the habitats that the hosts tend to occupy may vary greatly. To optimize host finding, parasitoids should be able to adapt to such changes. This would require the ability to adjust their response to certain semiochemicals. Indeed, several studies on host habitat and host location have shown that host searching behavior can be modified by experience. Adult female experience with hosts and host-related substrates have been shown to alter subsequent responses (Thorpe and Jones, 1937; Monteith, 1963; Arthur, 1966, 1971; Taylor, 1974; Vinson et aK, 1977; Sandlan, 1980; Strand and Vinson, 1982; Vet, 1983, 1988; Vet and van Opzeeland, 1984, 1985; Wardle and Borden, 1985; Dmoch et £L, 1985; Drost et al^, 1986; Lewis and Tumlinson, 1988). Learning and conditioning are the terms used most often to describe this effect. 52

PAGE 66

53 Recent studies have shown that the flight response of Micropi itis croceipes (Cresson) and M. demol itor (Cresson) to host related odors in a flight tunnel, increases significantly after the parasitoids have a brief contact experience with host by-products (Drost et al_^, 1986; Herard et aL., 1988a). Oviposition is not required to evoke the increase in response. In fact, increase in response was less if the parasitoids oviposited on larvae that were removed from their habitat. Preliminary observations on the host searching behavior by marqiniventris indicated that this larval parasitoid of many lepidopteran species also performs better after a similar type of experience. The exact function and the mechanisms behind the phenomenon of increased responsiveness after experience have yet to be elucidated. One possibility is that perception of certain chemicals in the host by-products sensitizes the females and stimulates them to enter a searching mode, which results in the observed increase in response. It is also possible that during the contact experience the females learn to associate the odors emitted by hosts and their habitat with the presence of hosts. In order to test these two alternative hypotheses, we studied the effects of pre-bioassay contact experience on the response to host related odors by marqiniventris in a fourarm olfactometer. Material and Methods Parasitoids . C. marqiniventris (strain Mississippi) were reared on fall armyworm larvae at the USDA-ARS, Insect Biology and Population Management Research Laboratory, Tifton, Georgia, according to the

PAGE 67

54 procedure described by Lewis and Burton (1970) for M. croceipes . The pupae were held in 25 X 25 X 25 cm, plexiglass cages, with one side made of fine mesh nylon screen. The parasitoids were allowed to emerge in the cages and were held at 26°C, 50-60% RH and a 15-hr photophase. Only parasitoids that emerged on the same day were kept in the same cage. Males were removed after 2 days, allowing sufficient time for all females to be mated. All experiments were conducted with mated females that were 3 to 4 days old, 7-10 hr after they experienced lights-on. Hosts . The hosts used in the experiments were second instar larvae of the fall armyworm (FAW), Spodoptera fruqiperda (J.E. Smith), and of the cabbage looper (CL), Trichoplusia ni (Hiibner). They were reared according to the method described by King and Leppla (1984). Initially, the larvae were fed on a laboratory-prepared pinto bean diet. Then, for about 18 hr prior to testing, those larvae that were used as part of the odor source were allowed to feed on leaves. Thus, the FAW larvae were fed corn (Zea mays L.) leaves, and the CL larvae were fed cotton ( Gossypium hirsutum L.) leaves. The leaves were obtained from plants that were 2-4 weeks old. Olfactometer . Bioassays were performed in an airflow olfactometer similar to the one described in detail by Vet et aL. (1983) with some modifications described by Eller et aK (1988a). It consisted of an exposure chamber connected to four arms through which air flowed into the chamber. The air was pulled out through a center hole in the bottom of the chamber. By balancing the airflows (300 ml/min through each arm), four distinct flow fields were created in the exposure chamber. Each arm was connected to a flowmeter, a water bubbler (to humidify the

PAGE 68

55 air), an odor chamber and a catching jar. Materials tested as sources for semiochemicals were placed in the odor chambers. Parasitoids that walked up the arms were captured in the catching jars. Female parasitoids were introduced through the vertical entry tube after temporarily disconnecting the extraction tube. While walking up the entry tube the test animal was exposed to a mixture of the four flows until it reached the chamber floor, where it moved freely, exploring the different flow fields. One wasp was introduced at a time and its behavior was observed during a 5-min period. If the test animal showed a positive anemotactic response and walked up one of the airflows and out of the exposure chamber it was recorded as a final choice if the female did not return within 15 sec. The temperature in the bioassay room was 28±1°C at all times. Odor Source . In all experiments, only a single odor source was used in one of the olfactometer arms. The three remaining arms served as controls with only humidified air going through. The test odor source contained five secondinstar larvae feeding on young leaves. Either FAW on corn leaves or CL on cotton leaves were used. The larvae were starved for 1 hr before being introduced to the odor chamber with one of the already damaged and contaminated leaves and a fresh seedling. The larvae were allowed to eat for 1.5 hr before wasps were tested in the olfactometer. Experience . Females were provided with the following types of experiences prior to testing: 1) no experience with any hosts or host products (INEXP), 2) one oviposition on a second instar CL larva feeding on cotton (CLOVIP), 3) a 20-sec contact experience with a cotton leaf

PAGE 69

56 damaged and contaminated by CL larvae (CLDAMAG), 4) one oviposition on a second instar FAW larva feeding on corn (FAWOVIP), and 5) a 20-sec contact experience with a corn leaf damaged and contaminated by FAW larvae (FAWDAMAG). The leaves used in the oviposition and damage experiences were equally damaged by larvae and contaminated with larval by-products to ensure that oviposition was the only difference between these two types of experience. The females were tested 30-60 sec after they had their experience. Five females of each treatment were tested per day. The two odor sources were alternated between days, as were the arms containing the odor. A total of 30 females were tested for each treatment to each odor source. Data Recording . The behavior observed in the olfactometer was recorded with an Epson® Geneva PX-8 portable computer. Response was measured in two ways: 1) the time spent in the quadrant of the introduction chamber containing the odor and, 2) the number of final choices made for the arm through which the odor was entering the exposure chamber. Statistical analyses . The five treatments were compared with Duncan's new multiple range test after analysis of variance (ANOVA) (Steel and Torrie, 1960). The two response measures were analyzed using: 1) the percent of time spent in the odor field by each individual female and, 2) for each treatment, the percentage of the five females tested daily that made a final choice. The percentages were transformed using the arcsin-square root transformation for statistical analysis.

PAGE 70

57 To compare oviposition experience with damage experience, and FAW experience with CL experience, pooled means were analyzed by orthogonal comparison of the sums of squares (Chew, 1986). Significance levels were 0.05 in all tests. Results Table 2-1 shows the observed responses for females that had « different types of experiences prior to their introduction in the olfactometer. In most cases prior experience was associated with females spending more time in the odor flow and making more final choices for the odor arm. With some exceptions for the time spent in the odor quadrant, an increase in response was significantly less when the females experienced the alternative plant-host complex. For a better comparison of the different types of experiences, the appropriate means in Table 2-1 were pooled and analyzed by orthogonal comparison of the sum of squares (Chew, 1986), resulting in the following. Oviposition vs. Damage Experience (Figure 2-1) . A comparison between all females that had an oviposition experience and those that only experienced the host-damaged leaves showed that the increase in response did not require contact with a host. For both odor sources, CL on cotton and FAW on corn, no differences were found in the percentage of time spent in the odor quadrant (F=0,0186, F=0.0047) and the percentage of final choices for the odor arm (F=0.79, F=0.05).

PAGE 71

Table 2-1. Effect of pre-bioassay experience on response of C. marqiniventris females exposed to host related odors in a four-arm olfactometer. Response is expressed in 1) percentage of females that made a final choice for the odor (% F.C.), and 2) percentage of the total time spent in the quadrant with the odor flow (% in odor). Od or CL feeding on cotton FAW feeding on corn Experience n % F.C. % in odor % F.C. % in odor INEXP 30 20.0 a 59.4 a 13.3 a 46.9 a CLOVIP 30 73.3 c 85.6 b 43.3 b 69.9 b.c CLDAMAG 30 70.0 c 87.2 b 36.7 b 62.0 a,b FAWOVIP 30 46.7 b 67.1 a 60.0 c 79.2 c,d FAWDAMAG 30 40.0 b 66.7 a 70.0 c 86.8 d Values with the same letter in each column are not significantly different (Duncan's new multiple range test after ANOVA with arcsin-square root transformation of percentages [Steel and Torrie, I960]).

PAGE 72

59 OVIPOSITION vs. DAMAGE EXPERIENCE I I NO OVIPOSITION VTTT^ DAMAGE EXPERIENCE EXPERIENCE EXPERIENCE 80 I CL ON COTTON FAW ON CORN ODOR SOURCES Figure 2-1. Experience effect upon response to odors of two plant-host complexes. Shown are the responses of females that 1) had no experience with hosts or hosts products prior to a test, 2) that were allowed to oviposit in one larva (either CL or FAW) feeding on leaves, and 3) that only contacted leaves damaged by CL or FAW larvae prior to a test.

PAGE 73

CL on Cotton vs. FAW on Corn Experience (Figure 2-2). For each odor source, all females that experienced CL on cotton were compared with those that experienced FAW on corn (oviposition experience and damage experience lumped together). This resulted in highly significant differences for both the time spent in the odor quadrant (F=14.90, F=9.70) and the percent of females that made a final choice for the odor arm (F=25.50, F=12.71). Figure 2-2 shows that the wasps responded much better when exposed to the odors of the plant-host complex with which they have had experience. Discussion The response of C. marqini ventri s females to odors of host larvae feeding on leaves can be increased dramatically by allowing the females to contact host products prior to testing in an olfactometer. Actual encounters with the hosts are not required to evoke this increase in response (Figure 2-1). Apparently, the presence of larval by-products is sufficient to trigger the mechanism behind the increase in response. Furthermore, the increase in response is evoked instantaneously, or at least within 20 seconds. This shows that the process must be a powerful and important modifier of the insects' behavior. The parasitoids respond best to the odors of the plant-host complex that they experienced (Figure 2-1). Therefore, the results suggest that we are not merely dealing with a general increase in search motivation (= sensitization), but that associative learning must be involved. It is suspected that the learning process is triggered when a parasitoid contacts a specific kairomone. The parasitoid then links the

PAGE 74

CL ON COTTON vs. FAW ON CORN 61 NO CL-COTTON PZ^ FAW-CORN EXPERIENCE EXPERIENCE EXPERIENCE Z < Q < O DC O Q O UJ 1 Iir o Q o CE o LL (/) ID y o I u < z CL ON COTTON FAW ON CORN ODOR SOURCES Figure 2-2. Experience effect upon response to odors of two plant-host complexes. Shown are the responses of females that 1) had no experience with hosts or hosts by-products prior to a test, 2) that experienced CL on cotton (either oviposition or damage), and 3) that experienced FAW on corn (either oviposition or damage).

PAGE 75

present odors with the possible presence of host larvae. Subsequently, the wasps will use those odors as cues in the search for more hosts. Evidence that supports this theory has been presented for the parasitoid M. croceipes (Lewis and Tumlinson, 1988). The fact that experience with an alternative host still causes some increase in response (Figure 2-2), suggests that associative learning is not the only process involved, but that sensitization takes place also. Alternatively, however, some of the semiochemicals emitted by one plant-host complex might be the same or similar to those emitted by the other complex. Therefore, learning alone may still explain why experiencing one plant-host complex increases the response to semiochemicals emitted by another plant-host complex. i. In the laboratory adult females seem to require a learning experience to make them highly sensitive to host odors. Learning might also take place during or immediately following emergence, when the parasitoid contacts by-products of its own host on the cocoon and on the emergence site. The insects used in the experiments were reared on FAW larvae feeding on artificial diet. The information these parasitoids obtained during emergence might not be adequate to make them respond to the odors to which they were exposed during the experiments. This could explain the relatively poor response by inexperienced females. Females in the field may obtain enough information in an earlier stage to make them more responsive to the appropriate semiochemicals. In fact, Drost et aL. (1988) and Herard et al^ (1988b) showed for M. c roceipes and M. demol itor , respectively, that inexperienced adult wasps reared from larvae feeding on cowpea leaves responded better than those reared on

PAGE 76

larvae feeding on artificial diet, when exposed to odors emitted by larvae feeding on cowpea. The effect of rearing on the response levels is likely to be of importance for C. marginiventris as well. However, considering the host range of this species and the many plant species these hosts might feed upon, modification of response to semiochemicals after experience by an adult female should contribute to the female's host searching efficiency. Therefore, it is likely that adult females in the field continue to increase and adjust their response to semiochemicals each time they contact host by-products.

PAGE 77

CHAPTER III EFFECTS OF FORAGING EXPERIENCES ON ODOR PREFERENCES Introduction Females of many insect parasitoids rely on host and host-habitat related chemicals as cues in their search for hosts (for reviews see Vinson, 1976, 1981, 1984; Weseloh, 1981; van Alphen and Vet, 1987). Several studies have demonstrated that the response to these semiochemicals is flexible and can be influenced by learning (see previous chapter). These studies show that experiences with hosts and/or their microhabitats, both by immature and mature stages, may influence an adult parasitoid's response to semiochemicals. It was shown in the previous chapter that C^ marqiniventris (Cresson) shows a significant increase in response to host-related odors after only a brief contact experience with host damaged leaves contaminated with host by-products. After females receive a contact experience with a particular plant-host complex, they respond significantly better when they are exposed to the odors of that plant-host complex than when they are exposed to the odors of an alternative plant-host complex. This suggests that the experience effect is not merely the result of a general increase in response to semiochemicals, but that the insects actually learn to respond to the odors that they encounter during their experience. 64

PAGE 78

The phenomenon of conditioning through experience has been suggested as a useful method for biological control programs in which parasitoids could be stimulated to respond to host-related odors prior to their release in a target area (e.g., Lewis and Nordlund, 1985). It would be particularly helpful if the wasps would not just show an increase in their responses to the experienced odors, but would actually prefer these odors over those released by alternative plant-host complexes when given a choice. This chapter reports on a study in which the effect of experience on odor preferences by C. marqiniventri s was tested. Shifts in preference in favor of an experienced odor were studied in situations where only the host species was varied and in situations in which only the plant was varied. Thus, we obtained information on the specific roles played by both host larvae and plants in the production and release of semiochemicals essential for host-habitat location. Materials and Methods Population of C. marqiniventris . Parasitoids of the '85 Mississippi strain were reared on fall armyworm larvae as described in chapter II. All experiments were conducted with 3to 5-day-old mated females, 6-10 hr into the photophase. Hosts . The hosts used in the experiments were second instar larvae of the fall armyworm (FAW), Spodoptera fruqiperda (J.E. Smith), and of the cabbage looper (CL), Trichoplusia ni (Hiibner). They were reared according to the procedure described by King and Leppla (1981).

PAGE 79

66 Olfactometer . Individual females were exposed to host-related odors and observed in a four-arm olfactometer as described in chapter II. The fraction of an odor-containing flow that actually reached a parasitoid in the olfactometer could be controlled by diverting part of the flow before it entered the arena. The fraction that was split off was then replaced by clean humidified air. The total flow entering the central exposure chamber through each arm was kept at 300 ml/mi n. Further details are given in the previous chapter. ^ Odor sources . The odor sources consisted of 5 late second or early third instar larvae feeding on three seedlings. The larvae of either FAW or CL were put on the seedlings of either corn (Zea mays L.) or cotton ( Gossypium hirsutum L.) 1.5 hr prior to the actual bioassays. Data recording . The behavior of the females in the olfactometer was recorded with the use of an Epson® Geneva PX-8 portable computer. After a female was introduced into the olfactometer the time it spent in each odor quadrant was recorded during a 5-min period. If the parasitoid walked into one of the arms and did not return within 15 sec, this was recorded as a final choice for that arm. The remaining time was added to the time spent in the quadrant of the final choice. For the dual choice tests the odor quadrant in which a female spent the greatest amount of time was recorded as her odor field preference. Procedures and Results Dose-resDonse experiments. Before actual odor preferences were • tested, the responses to different odor doses were tested for each plant-host complex separately. Thus, dose-response curves could be

PAGE 80

67 generated to determine which concentrations of the different plant-host complexes evoke similar response levels. The optimum concentrations were then used in the preference experiments to reduce the possibility that preferences were influenced by concentration differences. The responses of female parasitoids were observed to the odors of FAW on corn, FAW on cotton, CL on corn, or CL on cotton. During each test the odor of one of these complexes was offered through one of the four flows, while the other flows contained humidified air and served as controls. For each of the four odor sources three concentrations were tested; 25, 50 and 100% of the original odor flow. Just prior to being tested, a female was placed, for 20 sec, on a plant-host complex like the one used as the odor source. The parasitoids were prevented from actually encountering hosts. This type of experience significantly increases a female's response to hostrelated odors as was shown in the previous chapter II. The female was then introduced into the olfactometer. Odor sources were rotated to the next air flow after 6 females were tested for each of the three concentrations of a particular odor. A total of 24 females were tested for each concentration of all odor sources. Figure 3-1 shows that the wasps responded in a dose-related manner to all four complexes. Regression analyses of the time spent in the odor quadrant (Figure 3-lb) and of the time it took the females to make a final choice (Figure 3-lc), show a significant increase in responsiveness to the odors with increasing odor dose. The total number of final choices made for the odor arms was also found to be significantly dose related, with the exception of CL on cotton (Figure

PAGE 81

(0 Ul o 5 z u lU o < GC 111 100 90 BO 70 68 FAW + corn y = 036x + 6666: r = .72: p<.01 A FAW + cotton y = 0.37x + 64.5B: r = .77; p<.01 — 0-CL + com y=0.*0x +62.50; r=.71: p<.01 D CL + cotton y=Oi3x*77.0Q-. r=..49: ns. 60 25 50 75 100 Ul fC O o o 100 90 lU S O < a m 80 70 60 — I — FAW + corn y = 0 19x + 27e.96; r = .26: p<.05 A FAW + cotton y =032x + 263.06; r=.24: p<.05 CL + corn y = 0.24x + 274.63; r = .30: p<.05 CL + cotton y = 0.66x + 21B.65: r = .24: p<.05 25 50 75 100 O u M to Ul Z P lU o 5 X o _l < z 111 o < E u 5 200 r 150 100 50 \ A \ — I — FAW + corn y = -1.06x+ 186.00; r = .33: p<.01 A FAW + cotton y=-1.20x+ 196.71; r = .44; p<.01 —Q-CL + com y = -1.54x*212.73; r = .49; p<.01 D CL + cotton y = -0.59x+ 151.56; r = .24; p<.05 25 50 ODOR DOSAGE <%) 75 100 Figure 3-1 Responses of experienced margin iventris females to 3 doses of odors emitted by larvae feeding on leaves. Responses were measured as; a) Average percentage of the females that made a final choice for the odor arm; b) Average percentage of time that the females spend in the quadrant with the odor; c) Average time it took a female to make a final choice. The drawn lines connect the average values, while the equations for the actual linear regressions are given with each graph.

PAGE 82

69 3-la). No significant differences were found between the same doses of the four odor sources. In all of the following preference tests 50% doses were used. Effect of short contact experience on odor preference. The effect of a short-term contact experience on a female's preference for the odors emitted by different plant-host complexes, was tested in a series of dual choice experiments. A female wasp was allowed to contact for 20 sec, a plant-host complex from which the larvae had been removed. Immediately following this contact experience, the wasp was transferred to the olfactometer. In the olfactometer the wasp was exposed to the odors of two different plant-host complexes introduced through adjacent arms of the olfactometer, the two remaining arms carried humidified air only. One of the odors was from the complex that the female had just encountered, the other odor from a different complex. One combination of two odor sources was tested on a given day, with 8-10 females that experienced one source, and a same number of females that experienced the other source before being introduced into the olfactometer. This was replicated 6 times for each combination. The results, as summarized in Figure 3-2, show that the probability that a female chooses the odor of a particular complex is higher if she has had experience with that complex than if she has had experience with the other complex. The T-test (SAS Institute, 1987) was used to make an overall comparison of the responses to the odors of complexes with which females had experience to those with which they had no experience. The differences were highly significant for both the number of final choices (N=6; T=4.17; p=0.009) and the odor field

PAGE 83

70 Odor A Odor B Different hosts on same plant I A = FAW on corn B = CL on corn ll. A = FAW on cotton B = CL on cotton Same host on different plants „1. A = FAW on corn B = FAW on cotton IV. A = CL on corn B = CL on cotton Different hosts on different plants V. A = FAW on corn B = CL on cotton VI. A = FAW on cotton B = CL on corn Figure 3-2, Effects of experience on the preference of C. marqiniventri s females for host-related odors. a) Diagram of the olfactometer with the odors of two different planthost complexes entering the exposure chamber through adjacent arms. b) List of the six plant-host complexes that were tested. c) Summary of results using the percentage of females that made a final choice for a specific odor arm as the measure of response. Females that did not make a final choice were excluded. d) Summary of results using the percentage of females that spend most of their time in a specific odor field (= odor field preference) as the measure of response. The few females that never entered one of the two odor fields were excluded. The roman numerals (I-VI) in both c) and d) refer to the combinations listed in b). In c) and d) the bars above the x-axis represent the females that choose odor A and below the x-axis the females that choose odor B. The actual numbers are shown in parentheses. For each combination 48 to 50 females were tested of both experience types. The asterisks indicate significant differences in total numbers due to experience (chi-square; p < 0.05).

PAGE 84

preference {N=6; T=4.69; p=0.005). However, for each individual combination the differences in odor preferences were not always significant (Figure 3-2). For the numbers of final choices made for the odors, a significant difference was found only for the combinations FAW on corn versus CL on corn {x^df,=8.624; p=0.003) (I, Figure 3-2c) and CL on corn versus CL on cotton (x^dfi=3.907; p=0.048) (IV, Figure 3-2c). Odor field preference was affected significantly for the combinations FAW on corn versus CL on corn (x^df,=6.596; p=0.010) (I, Figure 3-2d) and FAW on corn versus CL on cotton (x^df,=4.318; p=0.038) (V, Figure 32d). Pooling the combinations with different hosts feeding on the same plants (I and II in Figure 3-2) and the combinations with the same hosts feeding on different plants (III and IV) revealed an overall preference for FAW and for corn odors. The females divided their final choices more or less equally among FAW and CL, (73:62) but the overall odor field preference deviated significantly from a 1:1 ratio (116:76; X dfi=3.802, p=0.051). Plant preference in favor of corn was demonstrated with both the number of final choices (98:45; x^dfi=9.778, p=0.002) and the odor field preference (131:68; x^dfi=9.285, p=0.002). Effects of experience were also analyzed with a loglinear model (SAS Institute, 1987) with five dependent variables; experienced host, experienced plant, alternative host, alternative plant, and preference (experienced or alternative odor). The number of females responding on a test day was used as one observation. There were 24 response levels with a total frequency (N) of 415 for the final choices and 585 for the odor field preference. Again the overall effect of experience in favor

PAGE 85

of the experiences odor was highly significant (Final choices (F.C.): X^dfi=12.21, p=0.0005; Odor field preference (O.F.P.): x^dfi=13.84, p=0.0002). Significantly more females chose the experienced odor if they had experienced corn than if they had experienced cotton (F.C. X^dfi=15.82, p=0.0001, O.F.P.: x2dfi=12.55, p=0.0004). No such difference in experience effect was found for the two host species. The general preference for FAW odors was demonstrated by the fact that significantly more females would choose the alternative odor if the alternative host was FAW (F.C: x^dfi=10.32, p=0.0013; O.F.P.: X^df,=19.85, p<0.0001). No such difference was found for the plant species. Note that in those cases where the wasps experienced the source with the least preferred host and plant (i.e., CL on cotton; combinations II, IV, and V) no preference for either odor source was observed. The combination FAW on corn CL on cotton was chosen for the following additional preference experiments. This combination contains all four components and the results obtained for this combination allow room for measurable increases and decreases in the effect of experience. A more complete experience. Previous experiments were conducted with females that had a 20 sec contact experience without ovipositions just prior to their introduction into the olfactometer. Further experiments were performed to determine whether a longer, more complete experience which included ovipositions would result in a stronger effect upon the odor preference by the parasitoids, and whether this effect of experience would be retained over time.

PAGE 86

73 Females were experienced by placing them in a glass container (26 cm in diam., 10 cm high) containing either 70 FAW larvae on 12 corn seedlings or 70 CL larvae on 12 cotton seedlings. The containers were then covered with a plexiglass plate. All females made contact with the plants and frass, and parasitized more than one larva. Females were exposed to the plant-host complex until they left the plants and attempted to leave the container. The exposure time varied from 4 to 11 min. The persistence of the experience effect over time was tested by giving one group of females their experience in the morning 3-4 hours before being tested in the olfactometer (Group 1); a second group was given their experience just a few minutes prior to the bioassay (Group 2). Results for the two groups are presented in Figure 3-3. Again, the differences between females that experienced different complexes were slight but consistent. The preference for the odor of FAW on corn was less for the females that experienced CL on cotton. For the females of Group 1 the difference in preference was only found to be significant for the number of final choices (x^df,=5.326; p=0.021). For Group 2 a significant difference between females experienced on a different complex was only found in the odor field preference (x^df^=15; p=0.001). Group 1 and Group 2 females did not differ from each other in their response to the odors except for females experienced on FAW on corn. Group 2 females with a FAW on corn experience preferred FAW on corn odors significantly more than Group 1 females that had experienced

PAGE 87

74 (0 O o *o jC o c L Group 1 Group 2 100% rt29) (26) FAW on corn CL on cotton 100% (23) 3) (13) (19) m (3) 1 (10) Group 1 Group 2 (38) (29) 1 (26) (1 1) 1 (22) (2) (14) (18) 100% FAW on corn CL on cotton 100% Q) O c 0 (U i_ a o 0 FAW CL FAW CL FAW CL FAW CL com cotton com cotton com cotton com cotton ODOR SOURCE CONTACTED PRIOR TO PREFERENCE TEST Figure 3-3. Responses of marqiniventris females to the odors of either of two plant-host complexes after a complete contact experience including ovipositions as indicated in the figures. Group 1 had their experience 3-4 hours prior to a bioassay, Group 2 had their experience just prior to a bioassay. The actual numbers are shown in parentheses. The asterisks indicate significant preference shifts.

PAGE 88

75 the same complex {x^clf^= 5.87; p = 0.015). This difference was not observed in the number of final choices made by the two groups. Experiences with ovipositions versus experiences without ovipositions. Finally, treatments were tested simultaneously to reveal possible differences between experiences with ovipositions and experiences without ovipositions, which may so far have been hidden by interday variation. Two sets of females were given a complete experience as described above, one hour before the bioassays. The first set encountered larvae and could oviposit freely during the experience. The second set of females, however, was experienced on a complex where the larvae were removed so that only the contaminated and damaged leaves could be contacted. The results were very similar to those found for the treatments discussed before (Figure 3-4). No differences were found between females that had a total experience including ovipositions and females that only contacted the damaged and contaminated leaves. Discussion The dose-response tests revealed that the females' responses increase with an increasing dose of the host-related odors. No significant differences were found in the attractiveness of the four different plant-host complexes when the parasitoids were exposed to the odors in single choice tests. The pooled results of the preference experiments, however, indicate a preference for FAW odors over CL odors and an even stronger preference for corn odors over cotton odors.

PAGE 89

76 + Ovip. -Ovip 100% 100% 100% FAW CL FAW CL FAW CL FAW CL com cotton com cotton com cotton com cotton ODOR SOURCE CONTACTED PRIOR TO PREFERENCE TEST Figure 3-4. Responses of Ll marqiniventris females to the odors of two plant-host complexes after a complete contact experience with (+Ovip.) and without (-Ovip.) ovipositions. The actual numbers are shown in parentheses. The asterisks indicate significant preference shifts.

PAGE 90

77 Since FAW seems to be L. marqiniventris ' most important host and FAW larvae are found predominantly feeding on corn, other grasses and legumes (Ashley, 1986), an initial preference for the odors of FAW larvae and damaged corn is not surprising. Furthermore, CL appears to be a very poor host since initial rearing experiments show minimal emergence from this host (unpublished data; M. R. Strand, personal communication). On the other hand, since all test animals were reared on FAW larvae, the observed preference for FAW may also have been the result of conditioning of the parasitoids as immatures. However, since host larvae are routinely fed artificial diet, the rearing procedure could not account for the corn preference. The results not only indicated an innate preference for FAW and corn odors, they also showed that the preferences were affected by contact experiences with the plant-host complexes. For all plant-host complex combinations it was found that a particular complex was chosen more often by females that had experienced that complex than by the females that had experienced the alternative complex. Although not always significantly, this experience effect caused a change in preference in each individual combination. The overall effect was found to be highly significant. The results are in agreement with earlier results of single choice experiments presented in chapter II. The increase in response to hostrelated odors after experience is greatest to the odors emitted by the plant-host complex that the females experienced. The learning process that must be involved is triggered by a brief contact with host byproducts, and does not require actual contact with the hosts. When

PAGE 91

females were given a longer experience period, including ovipositions, they did not appear to respond differently than females that had a 20 sec experience without ovipositions. The effect of experience on the preference for host-related odors lasted at least several hours, and is therefore likely to be an important factor determining the host searching behavior of these parasitoids in the field. Significant differences in preference were found when females were offered odor source combinations where only the host species varied as well as in combinations in which only the plant species varied and in those combinations where both the host and the plant varied. We can therefore conclude that females are able to distinguish between different host species and between different plant species. Evidently, both host and plant are somehow involved in the emission of the semiochemicals that evoke a response in the parasitoid females, either by producing the essential volatiles or by affecting the volatiles released by another component of the complex. The parasitoids are therefore likely to respond to more than one compound, the intensity of their response to each compound probably increases when it is encountered in association with a foraging experience. The results suggest that each plant-host complex releases its own blend of semiochemicals that is detected by C. marqiniventris . After exposure to a particular complex, a female will subsequently be attracted to an odor blend that is most similar to the blend she perceived during her experience. Future research will hopefully reveal whether females of marqiniventris and of other parasitoid species distinguish between

PAGE 92

79 variations in specific semiochemical blends or that they are able to differentiate between different compounds altogether.

PAGE 93

CHAPTER IV ANALYSIS OF ORIENTED FLIGHTS TOWARDS A SOURCE OF HOST-RELATED ODORS IN A FLIGHT TUNNEL Introduction In the two previous chapters it was shown that females of C^ marqiniventris are attracted to volatile cues released by host larvae feeding on plants. For a more detailed study of the long-range host location strategies of C^ marqiniventris , a bio-assay needed to be developed that would allow the females to exhibit the behavioral sequence that gets them in the vicinity of a host in a more natural situation. The bio-assay should be designed such that it will be possible to determine which component or components of a host-plant complex is (are) responsible for the release of the active semiochemicals. Studying the behavior of females flying to an odor source in a flight tunnel appeared to be a possible way to establish the exact source of the semiochemicals. Drost et aL. (1986) provided the first detailed flight tunnel study of a parasitoid, Micropl itis croceipes (Cresson), responding to host-related odors. The key to the success of the latter study was that pre-flight exposure to contact kairomones did override the tendency of the wasp to disperse upon release. This phenomenon, which was later found to be the result of associative learning (Lewis and Tumlinson, 1988), has facilitated the study of parasitoid host-searching behavior in flight tunnels for several others 80

PAGE 94

81 (Elzen et aK, 1986, 1987; Eller et aK, 1988b; Eller, 1990; Herard et al 1988a, 1988b; Zanen et aK , 1989; McAuslane et aL., 1990a, 1990b, 1990c). marqiniventris too is able to associate odors with the possible presence of hosts after contacting host by-products (chapters II and III). After such an experience the wasp is significantly more responsive to these odors. In this chapter the flight responses of naive and experienced females to host-related odors are compared. Furthermore, a detailed description of the behavioral sequence exhibited by experienced females is presented in the form of ethograms. Material and Methods The parasitoids. The wasps were reared and kept as described in chapter II. Female marqiniventris were 3 to 5 days old when used in the experiments. Experience of the wasps. Female wasps that were used in the bioassays were either kept naive (no contact with hosts, host by-products or plants), or allowed to forage on corn leaves that had been fed upon by FAW larvae the night before and were contaminated with host frass. This experience lasted 20 seconds. The wasps made no actual contact with host larvae. Fl iqht Tunnel . Flight responses of marqiniventris females were observed in a plexiglass flight tunnel 60 cm x 60 cm in cross-section and 2.4 m long. Two sheets of nylon mosquito netting (10 cm apart) at the open upwind end and one sheet of nylon screen (7x7 mesh/cm^) at the downwind end provided near laminar flow. Air was pulled through the

PAGE 95

tunnel at 0.2 m/sec and was exhausted via a 30 cm diam. flexible pipe with a fan. Four overhead incandescent lights (90 W) were dimmed so that they provided approximately 500 lux inside the tunnel. The room housing the tunnel was maintained at 27.5-29°C and 55-80% RH. A more detailed description of the tunnel is given by Eller et aL., (1988b). The odor source. Twelve late second instar FAW larvae were allowed to feed on 2 corn seedlings in a plastic petri dish (9 cm diam., 1.2 cm high) overnight. Two hours prior to the actual experiment six of the larvae were starved to ensure that they would be feeding during the experiments. One of the leaves they had been feeding on was removed and used for the above described experience of the parasitoids. Just before the flight tests, the remaining six larvae were removed and replaced by the starved larvae. One extra fresh seedling was added. The petri dish without a cover was then put upwind in the flight tunnel on a stand about 30 cm above the flight tunnel floor and served as the odor source. Observations. Females were released downwind in the flight tunnel about 1 m distant from the odor source. The insects were released from an open vial on top of a stand 30 cm above the flight tunnel floor. As soon as they reached the rim of the vial, their behavior was closely observed. To compare responses of naive insects with responses of experienced insects only one criterion was used; whether or not an insect would fly all the way to the odor source and land on it. On four different test days fifteen naive and fifteen experienced females were tested. Each female was given 3 trials. If she had not reached the odor source in any of those trials her flight was considered incomplete.

PAGE 96

83 The latter experiment was also used to distinguish the different components of the flight behavior. The following behaviors were observed: Take flight: the female jumps from the releasing vial into the air stream. Plume casting: the female shifts from side to side and up and down in the airflow, facing upwind. The female does not noticeably move forwards or backwards relative to the airflow. The area of movement is limited to only a few square decimeters in the center of the tunnel . Wide casting: this behavior is similar to plume casting, but differs from it in that the female uses the whole diameter of the flight tunnel for her movements. Straight flight: the female flies upwind towards the odor source, more or less parallel to the airflow. Downwind loop: while in flight the female suddenly drops back to a position downwind from where she was and continues with one of the other types of behavior. Dart: when close to the source (about 4-10 cm), a female flies quickly forward to the source, this is always followed by a landing on the source. Land source: female lands on the source of the odor. This can be anywhere in or on the petri dish. Land other: female lands somewhere else than the source. In all flights observed this was only on the ceiling or one of the sides of the fl ight tunnel . After the different components of the flight behavior were established, the flights of fifty experienced insects were observed recorded and analyzed. Each insect was allowed three flights. Some escaped and some only flew twice. In total 142 flights were analyzed. Data recording. The behaviors of the females in the flight tunnel were recorded with the use of an Epson® Geneva PX-8 portable computer.

PAGE 97

Results Naive vs. experienced females. All females took off and flew after spending some time on the rim of the vial. Within the three trials given to them only 20.0 % (12 out of 60) of the naive insects flew all the way to the odor source and landed on it, while 81.6 % (49 out of 60) of the experienced insects landed on the source. The difference in responsiveness between naive and experienced insects is highly significant (x^dfi=43.21 ; p < 0.0001). Analysis of the flight behavior. After the different types of behavior were described, the frequency of occurrence of these behaviors was observed and recorded. The results of these recordings are summarized in the kinematic diagrams in figures 4-1. The diagram gives the number of transitions from one type of behavior into another and, the probability one behavior would follow another. Darts were always followed by a landing on the source and were therefore included with the landings in the diagram. The thick arrows indicate the sequence of behavior that usually preceded an actual landing on the source. Such flights would consistently start with casting or straight flight and the wasp would alternate between these two behaviors until she was close to the source. Then the female would dart forward and land on the source.

PAGE 98

N=142 85 N=223 Straight Flight N=72 Land on Walls h4=70 Land on Source Figure 4-1. Kinematic diagram of the flight behavior exhibited by experienced marqini ventris females in response to odors emitted by plant-host complex of FAW larvae feeding on corn inside a flight tunnel A total of 142 flights were observed, the probability that particular transitions from one behavior into another occurred are indicated. n

PAGE 99

86 Discussion In a four-arm olfactometer marqiniventris females are significantly more responsive to host-related odors after they have had a previous contact with the by-products of hosts in association with these odors (chapters II and III). The results of the flight tunnel study clearly show that experience also affects their flight responses. This phenomenon has been studied in detail for the wasp Micropi itis croceipes (Drost et al., 1986, 1988; Lewis and Tumlinson, 1988; Eller, 1990), and was found to be the result of associative learning (extensively discussed in chapter I). The detailed description of the flight behavior to a complete odor source presented here may be useful when the responses to different odor sources are compared. Speculating on the functions of the behaviors, it seems that plume casting is a way to locate the center of the odor plume. Wide casting might be a way to locate the plume itself. Straight flight seems to occur whenever the wasp has found the center of the odor plume. These speculations are supported by the fact that wide casting was usually followed by plume casting and that plume casting was followed by either wide casting or straight flight. A downwind loop occurred only occasionally and might be the result of losing the odor plume. By dropping back into the airflow the wasp will increase its chances of finding the plume again, because the plume will be wider downwind. Studies on the host-location flight behaviors of parasitic wasps are rare. \L croceipes is the only wasp for which detailed information on its flight behavior is available (Drost et al., 1986; Eller et al.,

PAGE 100

87 1988b). This wasp's oriented flights towards an odor source are very similar to the flights exhibited by L. marqiniventris females. Drost et al(1986) observed the wasps to zig-zag immediately after they took flight. They described this initial zig-zagging as "making sideways excursions mainly in the horizontal plane and perpendicular to the wind direction soon after take off". Eller et al. (1988b) referred to this behavior as casting (as is done here) and reserved the term zig-zagging for side to side flight with upwind movement. Neither of the two papers on VL croceipes distinguishes between plume casting and wide casting, two behaviors that appear well distinguishable for marqiniventris . Drost et al. (1986) and Eller et al. (1988b) use the terms straight flight and darting as they are used here, and Drost et al. (1986) also observed downwind loops. croceipes females were also observed to hover 5-10 cm in front of the odor source before they would dart and land. This behavior was never observed for marqiniventris . However, the forward movement of the wasps did slow down when they got close to the source. Herard et al. (1988a) described the flight of demol itor as essentially the same as the flight described for croceipes by Drost et al(1986). The results indicate that there are some common features to the oriented flights of different parasitic insects, but that generalizations are hard to make. It also appears that the flight responses and the behavioral sequences are strongly dependent on the odor plume shape that is offered to the insects. When different odor source containers were used, wasps exhibited all of the described behaviors, but the frequencies with which they occurred varied with the

PAGE 101

type of odor plume offered to the wasps (results not presented here). These findings should be kept in mind when experiments are designed to reveal the actual source of the active volatiles. Hence, a device was used for the experiments in the next few chapters that limits the variation in plume shape.

PAGE 102

CHAPTER V PINPOINTING THE EXACT SOURCE OF VOLATILE ATTRACTANTS THAT ELICIT ORIENTED FLIGHTS IN Cotesia margini ventris FEMALES Introduction Numerous studies show that insect parasitoids are attracted to stimuli associated with their hosts (Vinson, 1976, 1981; Weseloh, 1981; van Alphen and Vet, 1986), and the majority of these stimuli appear to be of chemical origin. Volatile chemicals that attract parasitoids over long distances can be emitted by the hosts themselves (e.g., sex and aggregation pheromones) or by host by-products such as feces, silk, and honeydew. These host-derived semiochemical s that serve as cues for parasitoids are termed kairomones (for terminology see Nordlund and Lewis, 1976; Dicke and Sabelis, 1988a; Chapter I). Long-range chemical cues also can be released by plants upon which the host feeds, or by other organisms that are associated with the host or its habitat (Vinson, 1981; Weseloh, 1981; van Alphen and Vet 1986; Kainoh, 1987). Li. marqiniventris is a general ist larval parasitoid that attacks many different lepidopterous species. This parasitoid responds vigorously to contact kairomones present in the by-products of the host such as silk, saliva, and exuviae, but response is strongest to host feces and to the feeding damage caused by the host larvae (Loke and :l. 89

PAGE 103

Ashley, 1984b; Dmoch et aL., 1985). The females are attracted to the odors emanating from a complex of host larvae feeding on plants (previous chapters), however, the exact origin of the attractive volatiles was heretofore unknown. Indirect evidence obtained from experiments on learning suggests that both the plant and the host are involved in the production of the active semiochemical s (Chapter III). In olfactometer tests, females prefer odors previously encountered during a contact experience. This preference is observed when females are given the choice between two odor sources each containing a different host species feeding on the same plant species, but also when the odor source contains two different plant species fed upon by the same host species. Apparently, both the host as well as the plant are involved in the production and/or release of the active volatiles. A flight tunnel study was conducted to reveal the specific roles played by the plant and the host in the long-range attraction of C. marqiniventris . The actual source of the semiochemical s that attract marqiniventris females was determined by testing different components of a complex of beet armyworm (BAW), Spodoptera exiqua (Hiibner) (LepidopterarNoctuidae) , larvae feeding on corn seedlings. Methods and Materials The insects. marqiniventris were reared and maintained as described in chapter II. All experiments were conducted with 3to 5day-old mated females, 6-10 h into the photophase. The hosts used in the experiments were late second or early third instar BAW larvae, reared as previously described by King and Leppla (1981).

PAGE 104

91 Fl jqht Tunnel . Responses of maroiniventris females to volatile chemicals emitted by different odor sources were observed in a plexiglass flight tunnel as described in the previous chapter. Odor sources. Corn seedlings (loana sweet corn) were grown in a greenhouse and cut for immediate use when 10-15 days old. Stems were wrapped in wet cotton wool and 12 seedlings were then placed in glass containers (26 cm in diam., 10 cm high) with 50 second instar BAW larvae. The following day (18-20 h later) various components of the plant-host complex were used as odor sources in the flight tunnel bioassays. To determine the actual source of attractive volatiles the complex was divided into three major components: 1) the host-damaged leaves; 2) frass; and 3) the host larvae. The individual components were obtained by wiping the frass of three host-damaged seedlings with one wet and one dry piece of cotton wool. The seedlings were then water-washed to remove any remaining larval by-products. Fifteen host larvae that had been feeding on these leaves overnight were starved for 2 h to prevent them from defecating during the bioassays, and subsequently water-washed. These three major components of a complete plant-host complex were then used as odor sources in the flight tunnel tests, both in single choice and in dual choice situations. Several recombinations (described in the next section) of the three components were tested in additional dual choice tests to determine what combination of the three components is required for maximum attractiveness. Odor release system. An odor source was placed in a glass cylinder dispenser (Figure 5-la) attached to a 0.5 cm-diam. stainless

PAGE 105

Figure 5-1. A. Odor inlet system. An odor source is placed in a glass chamber attached to a stainless steel 1/4" tube 25 cm above the flight tunnel floor. The wide open end of the glass chamber is covered with fine mesh nylon screen, kept in place with a rubber band. Clean, humidified air is pushed through the stainless steel tube and the glass chamber at a rate of 0.5 ml/mi n. B. Insect release funnel. Female wasps are placed under the glass funnel 80 cm downwind from the odor inlet system. Attraction to light will guide thefn to the open tube on top of the funnel 22 cm above the flight tunnel floor and in the trajectory of an odor plume. When attracted to the odor the wasps will take off from the funnel and fly upwind toward the odor source.

PAGE 106

93 steel tube which entered through the floor at the upwind end of the flight tunnel. The open downwind end of a cylinder was covered with a piece of nylon mosquito netting, held in place with a rubber band (see Figure 5-la). Humidified air was blown via the stainless steel tube through the cylinder at a rate of 500 ml/min to provide a continuous odor plume in the tunnel. For dual choice tests the stainless steel tube was split into a "T" and a glass cylinder was attached to each end, 6 cm apart. With smoke sources it was determined that this system resulted in two distinct spiralling plumes that met 25 cm downwind from the cylinders and overlapped approximately 80% at the wasp release point (80 cm downwind from the cylinder). Bioassavs. A 3to 5-day-old marqiniventris female was given a 20-40 sec contact experience with a complete plant-host complex (BAW larvae on corn seedlings on which they had fed overnight) during which she oviposited once or twice. Experience is known to increase responses to host-related odors significantly (previous chapter). The female wasp was then transferred into a 20 ml vial which was placed under a release funnel (Figure 5-lb) inside the wind tunnel 80 cm downwind from the odor source. The female soon walked up the funnel toward the light and eventually reached the open tube on top of the funnel where she made initial contact with the odor plume. When attracted, the female would walk toward the upwind end of the tube, take off, and fly toward the odor source. Here, we only report on whether or not a female flew all the way toward an odor source and, for the dual choice tests, which of the two sources she would choose.

PAGE 107

94 For the single odor source tests, 8 females were tested to each source on a given day (replicated 5 times). For dual choice tests, 10 females were individually tested to each of three odor source combinations on a given day (replicated 4 or 5 times). Statistics. The results from single source bioassays were analyzed by analysis of variance after angular transformation, followed by Duncan's multiple range test. All results from the dual choice tests were compared using chi-square. In all cases o<0.05 was used to determine significant differences. Results Single choice tests. No significant difference in attractiveness was found between water-washed host-damaged leaves and a complete planthost complex when tested as single odor sources in the flight tunnel (Figure 5-2). Frass and host larvae were attractive as well, but significantly less than the leaves and the plant-host complex. Dual Choice Tests. The above results were confirmed in dual choice tests where in different combinations, two of the three components of a complete plant-host complex were tested parallel to each other in the flight tunnel (Figure 5-3a). The damaged leaves were clearly more attractive than either frass or larvae, and frass was found to be more attractive than larvae. This ranking of attractiveness was also evident in the number of incomplete flights (females that did not reach the odor source). When the choice was between leaves and frass, the number of incomplete flights was the lowest, whereas it was the highest when frass and larvae were tested next to each other.

PAGE 108

95 RESPONSES IN FLIGHT TUNNEL TO DIFFERENT COMPONENTS OF COMPLEX 100 COMPLEX LEAVES FRASS LARVAE ODOUR SOURSES Figure 5-2. Flight responses by marqini ventri s females to a complete plant-host complex, compared with flight responses to single components of a plant-host complex. The bars indicate percentages of completed flights toward a source (n = 40). With each bar, the actual fraction of females flying to the source is shown, while letters indicate significant differences in attractiveness between the odor sources (Duncan's multiple range test after analysis of variance, p<.05).

PAGE 109

96 DUAL CHOICE TESTS IN FLIGHT TUNNEL a INCOMPLETE FLIGHTS 19 13 10 m LEAVES m m * FRAss m 1 1 30 9 ^LEAVES + LARVAE M X12 30 B Ileaves i 1+ FRASSj m LEAVES + LARVAE M 15 28 7 CAGED LARV. + LEAVES 21 |CAGED LARVAEi \ * LEAVES i 23 LARVAE + LEAVES: + FRASS y/my/m larvae ^ % * LEAVES ; 18 1 1 ^^^^^^ ^LEAVES «^ m(y////////////////, 19 ''M. larvae P ' % * LEAVES ^ Figure 5-3. Flight responses by marqiniventris females during dual odor source tests. Two hatched bars show the percentage of females that chose a particular odor in each combination. The open bar to the right of each combination represents the females that did not fly to either of the two odor sources. Given with each bar is the actual number of females that they represent. The asterisks indicate significant preferences for particular odors (x^df^, p<0.005). LEAVES = 3 water-washed corn seedlings that were fed upon by BAW larvae overnight. FRASS = BAW by-products wiped off of the host-damaged corn seedlings. LARVAE = 15 starved and water-washed early third instar BAW larvae. a. Responses to the single components of a complete plant-host complex, tested in combinations of two. b. Responses to host-damaged leaves compared with different recombinations of damaged leaves with either frass or larvae. c. Testing for the importance of larval feeding in attracting marqiniventris females.

PAGE 110

97 Recombinations of odor sources. The host-damaged leaves were recombined with either one of the two other components to determine whether any combination would be more attractive than the damaged leaves alone. It was found that adding back the frass made no difference, but that damaged leaves plus larvae (most of them feeding on the leaves) were significantly more attractive than damaged leaves alone (Figure 53b). Consistent with that, damaged leaves plus larvae were more attractive than damaged leaves plus frass. The Effect of Larval Feeding. Larvae alone were the least attractive, but when recombined with damaged leaves they caused a significant increase in attractiveness. Therefore, it was hypothesized that the additional feeding damage caused by the larvae was required to have this effect. To test the effect of larval feeding, the recombination of damaged leaves plus larvae was compared with a similar odor source in which the larvae were prevented from feeding by keeping them in a brass screen cage (0.5 mm mesh) inside the glass cylinder next to damaged leaves. Indeed, the source with the actively feeding larvae was more attractive than the source with the caged larvae (Figure 5-3c), and, when a source of damaged leaves with caged larvae was compared with just damaged leaves, the odor sources were equally attractive. Attraction to a complete recombination of a plant-host complex (including frass and feeding larvae) was statistically equivalent to an odor source with only feeding larvae on damaged leaves (Figure 5-3c). Controls. The release of the active volatiles may have been affected by taking apart the complete plant-host complex. To determine whether the procedure that we used decreases the attractiveness, a

PAGE 111

98 DUAL CHOICE TESTS IN FLIGHT TUNNEL INCOMPLETE FLIGHTS 19 COMPLETE COMPLEX UNDAMAGED LEAVES UNDAMAGED LEAVES 23 LARVAE + p ^ LEAVES+FRASS^ DAMAGED LEAVES 8 1 1 18 Figure 5-4. Control dual choice tests. The first combination shows the effects of the experimental procedure on the attractiveness of the odor sources. The other two combinations show the relative attractiveness of undamaged corn seedlings. For further details, see the text and the legend with Figure 5-3.

PAGE 112

99 recombined plant-host complex was tested alongside a complete plant-host complex that was not previously taken apart. The first part of Figure 5-4 shows that the recombined complex was not less attractive than an untreated complete complex. In fact, the recombined complex was slightly more attractive, perhaps because it contained starved larvae that ate more vigorously. Finally, undamaged corn seedlings were tested for their attractiveness. Compared to damaged leaves, they were far less attractive. Even when undamaged leaves were tested alongside an odor source that contained only larvae, undamaged leaves appeared only slightly attractive (Figure 5-4). Discussion Host-damaged corn seedlings were found to be the primary source of volatile semiochemical s that attract females of margini ventris . Yet, significant numbers of females were observed to fly to an odor source containing frass only as well as to a source with larvae only (Figure 52). However, when they were tested in choice situations, the relative importance of these two components of a complete plant-host complex in attracting marqiniventris appeared to be minor (Figure 5-3a). Moreover, the amounts of frass used in the bioassays were unrealistically high since in the field most of the frass will drop to the ground. Undamaged plants were far less attractive than plants damaged overnight by BAW larvae (Figure 5-4). Optimal attraction was observed when larvae were put back on the already damaged corn seedlings (Figure

PAGE 113

100 5-3c). This was not clear from the single choice tests which showed no statistical difference between a complete plant-host complex and damaged leaves alone. In dual choice tests, however, a clear difference was found. It appears that choice tests are more suitable to reveal relative importance of different odor sources in attracting an organism. However, dual choice tests alone may not reveal those cases where different steps of an organism's searching behavior are evoked by specific chemicals from different sources. In other words, if two sources are tested together, one may induce initial upwind flight while the other source results in close range attraction. This might not always be recognized unless both single source and choice tests are performed. L. marqiniventris attacks many different host species (at least 19) on a wide range of plants. It could therefore be presumed that this wasp innately responds to compounds that are common to most green plants, such as "greenleafy odors" (Visser et aL., 1979). Learning experiments (chapters II and III) suggest that during foraging experiences the female wasps may learn to respond to odor blends that are specific for a certain plant-host complex. In chapter III it was not only shown that marqiniventris can distinguish between the odors of two different plant species fed upon by the same host species, but also between two different host species feeding on the same plant species. Based on those findings and the data from this study, we suggest that, due to different feeding characteristics, different host species may cause a differential release of plant compounds which can be detected by the parasitoids after experience.

PAGE 114

101 Sources of long-range attractants for parasitoids have only been pinpointed in a few cases, but it appears that attraction to plant volatiles is common. Pine odors have been shown to attract the parasitoids Pimpla ruficoll is Grav. (Thorpe and Caudle, 1938) and Itoplectis conquisitor (Say) (Arthur, 1962). Powell and Zhang Zhi-Li (1983) found that both sexes of two cereal aphid parasitoids are attracted to the odors of aphid free plants, while only females were attracted to aphid odors. Nordlund et aL., (1985) show that female Trichoqramma pretiosum Riley is attracted to tomato plant volatiles. The leaf-miner parasitoid Dapsilarthra rufiventris (Nees) is attracted to ether-soluble plant volatiles that are also attractive to its host (Sugimoto et aL., 1988). These, and other examples (given by Vinson, 1975), of parasitoids being attracted to plant odors are generally seen as long-range location of the macro-habitat of the wasps' hosts. We suggest that for parasitoids like marginiventris . plant cues emitted as a result of larval feeding damage may also guide the wasps into the direct vicinity (micro-habitat) of its hosts. An additional, well documented example is that of the ichneumonid parasitoid Campoletis sonorensis (Cameron) (Elzen, et aL., 1983, 1984, 1986, 1987). Like C^ marginiventris . L. sonorensis attacks many lepidopteran species. It is attracted to plants that are untouched by hosts, but damaged plants are far more attractive than undamaged plants (Elzen, et aL. 1983). In preliminary experiments, C^ marginiventris females were also observed to fly to and land on undamaged plants, but the slightest, even artificial damage, would increase responsiveness and wasps would fly directly to the damaged sites. For the general ist

PAGE 115

102 parasitoids marqiniventris and sonorensis , plants are clearly the main source of volatile attractants that not only guide these wasps into areas that may harbor hosts, but also serve to get them into the direct vicinity of the host. Once a wasp reaches a micro-habitat of its host, it may rely on contact kairomones present in the by-products of the host, such as silk and frass, for the final step of host location. The more specialized parasitoid Micropl itis croceipes , which attacks only Hel iothis spp. is also attracted to the food plants of its hosts (Elzen et al^, 1987; Eller et aK, 1988b; Drost et aK, 1988), but host feces are the key source of odors that attract this wasp to the micro-habitat of its hosts (Eller et aL., 1988b; Elzen et aL., 1987). In general, plant odors appear to play a major role in the host searching process of parasitic insects. The obvious advantage that plants enjoy from attracting parasitoids when under attack by herbivores, makes them a likely source for semiochemicals (synomones). Natural selection should favor plants that attract entomophagous insects, while selection would work to minimize the release of reliable chemical cues by the phytophagous prey or hosts. Several authors have recognized the possible sophisticated tritrophic relationship that may have resulted from these selective pressures (e.g., Vinson 1975, 1981; Vinson et iL, 1987; Price et al.. 1980, 1986; Price 1981; Dicke and Sabelis, 1988b; Dicke et al-, 1990a, 1990c). In the next chapters combined chemical and behavioral studies will explore the complexity of such relationships in greater depth.

PAGE 116

CHAPTER VI ISOLATION AND IDENTIFICATION OF ALLELOCHEMICALS THAT ATTRACT Cotesia marqiniventris (CRESSON) TO THE MICRO-HABITAT OF ONE OF ITS HOSTS Introduction Although many studies have demonstrated parasitoid attraction to volatile semiochemical s, only a few of these chemicals have been isolated and identified (Weseloh, 1981; Vinson, 1981; Kainoh, 1987). Pre-release experience with volatiles that will guide the wasps to larval pests in a target area is likely to increase parasitization rates and thereby their effectiveness as control agents (e.g. Lewis, 1981; Gross, 1981; Nordlund et aK, 1981b; Lewis and Nordlund, 1985). To make this possible, identification and formulation of the essential semiochemical s will be necessary. C^ marqiniventris already frequently controls pests to significant degrees {e.g.. Tingle et £L, 1978; Pair et aL., 1982; McCutcheon and Turnipseed, 1981), but might be more effective after well timed augmentative releases. As shown earlier C^ marqiniventris increases its responses to volatile semiochemicals emitted by a plant-host complex after experiencing certain contact kairomones in association with these volatiles. It has been determined that plants damaged by the host larvae are the main source of the volatile attractants for female marqiniventris (Chapter V). This chapter presents the results of a study in which these volatiles were collected, tested for their attractiveness to the parasitoid, and eventually identified. The 103

PAGE 117

104 activity of the identified compounds was then confirmed by testing synthetic versions in flight tunnel bioassays. Methods and Material The insects . Cotes i a marqiniventris (Cresson) were reared and kept as described in chapter II. All flight tunnel tests were conducted with 3-to 5-day-old mated females 4-8h into the photophase. In all cases late second or early third instar beet armyworm (BAW) f Spodoptera exiqua (Hubner)] larvae were used as hosts in the experiments. They were reared according to the procedure described by King & Leppla (1981). Mass collections of volatiles for bio-assavs and identification. Mass collections of volatiles released by BAW larvae feeding on corn seedlings were made using a push/pull odor collection system similar to the one described by Heath et aL. (1990). Before entering the system air was humidified using a gas dispersion tube and purified with in-line activated charcoal filters (Heath et aK, 1990). The air then entered the first of three cylindric Pyrex® glass components (Southern Scientific, Gainesville, Fla.). This first section contained a glass frit which ensured that a laminar air flow would enter the second (the odor source containing) section. The first section ended in a 110/115 mm male ground-glass joint which was connected to a female counterpart of the second section. This second section was either 35 or 38 cm long (including joints) with an outside diameter of 11 cm. The outlet of the second section was a 50/55 mm male ground-glass joint which fitted the third section; a three port collector base. Two ports were used

PAGE 118

105 simultaneously for collection of the volatiles by attaching stainless steel tubes (0.64 cm OD x 0.5 cm ID) to each port as described by Heath et aL. (1990). The third port was sealed with a glass stopper. On the upwind side a collection trap was connected to each steel tube, while the downwind side was connected with Tygon® tubing to Alborg® flowmeters which were connected to house vacuum. The two collection traps consisted of a 3.7 mm ID, 4 cm long glass tube sealed on one side with a 325-mesh stainless steel frit . Approximately 50 mg of Super Q® adsorbent (80/100 mesh) (Alltech, Deerfield, Illinois) was placed on top of the frit held in place with a small plug of glass wool. Before each collection the traps were rinsed with methylene chloride. In the collection chamber we placed 100 early third instar BAW larvae on 40 ten-day-old greenhouse grown corn seedlings that were cut just before collection and their stems wrapped in wetted cotton. With the odor source and traps in place, a collection was started by balancing house air and house vacuum such that a flow of approximately 300 ml/min was going through each filter and the pressure was only slightly higher inside the system than outside. The pressure could be read with a user-designed glass pressure gauge that was connected to a side arm at the downwind part of the collection chamber. Each collection lasted for 24h. After 10-12h filters were removed and extracted with 300 \i] of pure methylene chloride, then connected back on to the system and extracted once more after 24h. The amount of volatiles collected is expressed in Collection Minute Equivalents (CME). Each collection contains 1440 CME (=24h).

PAGE 119

106 Collections for quantifications. For more exact quantification of the volatiles released by a complete plant-host complex a smaller system was used. The glass collection chambers consisted of two parts. Purified air entered the first part through 2 cm long 1/4" OD inlet, which widened into a section (6 cm long; 3 cm ID) containing a glass frit. The second part, also 3 cm in diam, was 15 cm long with a 2 cm long 1/4" outlet. The second part contained the odor source. Both parts had fitting glass ball joints that were clamped together. One collection trap with 25 mg Super Q® adsorbent was connected to the 1/4" outlet with brass Swagelock® fitting containing Teflon ferules. As odor sources, three corn seedlings that had been fed upon overnight were placed with 15 BAW in the chambers. Volatiles were collected for 1 hr ' (300 ml/min). The traps were rinsed with 200 nl methylene chloride, internal standard was added (1 ^g each of octane and nonyl -acetate in 50 nl methylene chloride). Of each collection sample 1-2 ^^ was analyzed using capillary gas chromatography (GC) (see later). The collection was repeated 6 times. The same procedure was followed to also collect volatiles from undamaged seedlings that were otherwise treated the same as the above seedlings. To determine system impurities, collections with only wetted cotton wool inside the chamber were performed as well. Collections of different components of complete plant-host complex. The latter procedure was also used to collect volatiles of the 3 main components of a complete plant-host complex. A complex of 3 overnight damaged corn seedlings was divided in larvae, frass, and damaged leaves as described in Chapter V. The larvae were starved for 2 hours and water-washed. The frass was wiped off the leaves with 2

PAGE 120

107 pieces of cotton wool (one wet and one dry), and the damaged leaves were water-washed to remove any remaining larval by-products., Volatiles of each of these three components and of a complete plant-host complex were collected for 3 hours simultaneously. Collections were repeated 5 times and GC analyses were used to determine the volatiles present in each of the samples. Chemical analyses. Gas chromatographic (GC) analyses were conducted on a Varian model 3700 GC and a Hewlett-Packard model 5710A GC, both equipped with spl it/spl itless capillary injector systems and flame ionization detectors. Data collection, storage, and subsequent analysis was performed on a Perkin Elmer chromatographic data system. Helium at a linear flow velocity of 19 cm/sec was used as a carrier gas. Most of the initial analyses of the volatile collections were performed on two fused silica capillary columns. They were 50-m x 0.25-mm ID with a 0.25-nm thick film of bonded methyl silicone (007), and 50-m x 0.25-mm ID with 0.25-jim thick film of bonded cyanopropyl methyl silicone (CPS1). Both columns were obtained from Quadrex Corp., New Haven, Connecticut and were run at an initial temperature 50 °C for 3 min, then temperature programmed at 5 °C/min to 180 °C. All injections of 1-3 [il were made in the spl itless mode. Samples were also analyzed by GC-mass spectroscopy (GC-MS) with a Nermag model RIOIO mass spectrometer in both the electron impact and the chemical ionization mode. The methyl silicone and CPS-1 columns used in the previous analyses were used in the GC-MS analyses with helium as carrier gas. Methane and isobutane were used as reagent gases for chemical ionization.

PAGE 121

Vapor phase infrared spectra were obtained from a Nicolet 20SXC GC-FTIR spectrometer. Samples were introduced to the FTIR via the apolar column described above in a Hewlett-Packard model 5890. Several of the compounds, which required [^H]-NMR analysis for full identification, were analyzed with a Nicolet 300 MHz Fourier transform NMR spectrometer interfaced to a Nicolet model 1280 data system. The natural compounds as well as synthetic standards were purified by micropreparati ve GLC on a 30-m x 0.53-mm ID SPB-1 (1.5-^) column (Supelco) in a Hewlett-Packard model 5890 gas chromatograph. To allow injection of large volumes (up to 100 //I) a deactivated column {30-m X 0.53-mm, Quadrex Corporation) preceded the SPB-1 column [see Grob (1982) for details on this technique]. Just before entering the detector the effluent from the column was split, one part going to the FID detector, the other exiting into a collector as originally designed by Brownlee and Silverstein (1968). Split ratio was manipulated such that >95% of the sample entered a glass capillary collection tube (35 cm long, 1.2 mm ID) in the collector (analogous to Murphy et aL., in prep.). Dry-ice in acetone was used to cool the part of the capillary tube furthest away from the GC. Tubes were inserted into the collector just prior to elution of the to-be-collected compound and removed immediately after elution. For the collection of the most volatile compounds approximately 8 mg of super Q was packed between 2 glass wool plugs at the cold end of a collection tube. Temperature programs varied for the collection of different compounds. The collected material was transferred into a NMR tube by rinsing the glass capillary with approximately 25 //I of benzene-D6. The NMR tubes were 5 mm (OD) at the

PAGE 122

109 top, with a 50 X 2 mm (OD) coaxial extension at the bottom (Wilmad Glass Co., Buena, NJ). Data points were collected with a 6 fis pulse (90° tip angle). Where necessary proton decoupling was accomplished by standard decoupling techniques (decoupler power ca. 0.5 watt). The spectral analyses were kindly performed and/or interpreted by A. T. Proveaux, R. E. Murphy, R. R. Heath, and J. H. Tumlinson at the Insects Attractants, Behavior, and Basic Biology Research Laboratory, USDA-ARS, Gainesville, Fla. All spectra of the natural products were compared with those of candidate synthetic compounds. (3E)-4,8dimethyl-l,3,7-nonatriene and (3E,7E)-4,8, 12-trimethyl -1,3,7, 11tridecatetraene were synthesized by Dr. R. E. Doolittle (Insects Attractants, Behavior, and Basic Biology Research Laboratory, USDA-ARS, Gainesville, Fla.) by using the Wittig reaction of geranial and farnasal with methyl enetri phenyl phosphorane (analogous to Maurer et al^, 1986). gtrans -bergamotene was generously provided by Douglas B. Mcllwaine at Brown University, Providence, Rhode Island. All other synthetic standards used in this study were obtained from commercial sources. Flight tunnel bio-assays. The Plexiglas flight tunnel described in chapter IV was used to test the attractiveness of specific samples to females of marginiventris . During each test the following conditions were maintained inside the tunnel; 15 cm/sec airflow; 55-70% RH; 27.529°C temp; and approximately 500 lux lumination. For each experiment a sample dissolved in 200 //I methylene chloride was applied on a strip (1 x 5 cm) of green construction paper. The paper was pinned at the upwind end of the tunnel with an insect pin at a 45° angle from horizontal (highest point upwind), on top of a 1/4" stainless steel tube, centered

PAGE 123

110 30 cm above the flight tunnel floor. Four minutes were allowed for the solvent to evaporate before the first wasps were tested. The wasps were released from a glass release funnel (chapter V) in groups of three (a number that easily allows the observer to recognize each individual throughout each trial) 80 cm downwind from the odor source. Unless p stated otherwise, females were given a 20 sec contact experience with a complex of corn seedlings fed upon by BAW larvae, just prior to their 'introduction into the flight tunnel. This type of experience significantly increases their response to host related odors (previous chapters). Here we report only on the percentage of females that exhibited a complete flight towards each odor source. Dose response tests. The volatiles collected using the above described mass collection procedure were tested at 5 different dosages. Equivalents of 10, 30, 100, 300, and 1000 collection minute equivalents (CME) were diluted with methylene chloride to a total volume of 200 /A per sample. Samples were labeled such that the experimenter did not know the concentration of each sample until after a complete trial. Six females (2 groups of 3) were tested to each concentration on a given day. This was repeated 6 times. Flights to synthetic odor blends. After all major compounds in the complete natural odor blend emitted by BAW larvae feeding on corn were identified, a synthetic blend of these compounds was made. First, all compounds were purified by preparative GLC as described above. After purification the compounds were combined in a blend that closely mimicked a collected natural blend (Table 6-2). Both the synthetic and natural blend were tested for attractiveness to marqiniventris in the

PAGE 124

Ill flight tunnel at a concentration close to the optimal (approximately 250 CME) found in the dose response tests. Five treatments were tested on one day. Females that were experienced on a regular plant-host complex were tested to a natural blend, a synthetic blend and, to solvent (methylene chloride) only. In addition, females that had a pre-flight experience with the synthetic blend were tested to both the natural and the synthetic blend. For the "synthetic" experience a piece of filter paper (Whatman®, quality 1, 5.5 cm) was treated with a synthetic sample (same as used for odor sources) and with 5-7 fecal pellets of BAW fed on corn. Several minutes after the solvent had evaporated wasps were introduced to the paper, which they would travel and antennate vigorously. After 30 sec the females were transferred to the flight tunnel to observe their long-range responses to one of the odor blends. Results Dose-response tests. Female marqiniventris exhibited complete flights to all dosages of the collected volatiles. Responses, however, were clearly dose related, with an optimum response of 60-75% complete flights to a concentration between 200 and 300 CME (Figure 6-1). Perhaps surprisingly, the overall response dropped significantly at the highest dose (1000 CME) tested. The percentages of females flying to the optimal dose were similar to those observed when females were tested to complete plant-host complexes (Chapter V). This is strong evidence that the naturally active compounds were present in the tested extracts.

PAGE 125

112 DOSE RESPONSE CURVE 100 80 y=-64.3 + 118.2X 27.2x^ (n=30; r=0.64; p<0.01) 60 40 20 10 100 DOSE IN CME (log scale) 1000 Figure 6-1. Responses by marqiniventris females to different doses of extracts of volatiles collected from BAW larvae feeding on corn seedlings. Units are collection minute equivalents (=CME), representing the time during which the volatiles were collected from 100 2nd instar BAW larvae feeding on 40 seedlings. Each point in the graph represents the average response (n=6) to a particular dose.

PAGE 126

113 Analyses and identifications of the collected volatiles. The GC profiles of the collected volatiles revealed the consistent occurrence of eleven compounds in significant amounts (Figure 6-2; Table 6-1). None of the compounds were found in the profiles from system blanks (= collections made with only wetted cotton present in the collection chamber), and only trace amounts of some of the volatiles were detected in collections from undamaged leaves (Figure 6-2). Compounds 1-5 and compound 7 were identified by comparing their GC-MS spectra with those obtained from synthetic standards, and reported spectra (Visser et aL., 1979; Stenhagen et aK , 1974). In all cases it was verified that the retention times for the authentic compounds match those of the synthetic candidates on the two different columns. Compounds 1 through 3 were typical greenleaf volatiles; (Z)-3-hexenal , (E)-2-hexenal , and (Z)-3-hexen-l-ol . All three of these have been reported as volatile components of numerous plant species belonging to a variety of plant families (Visser et aK, 1979), with a possible underrepresentation of Z-3-hexenal (see discussion). Compounds 4 and 5, Z-3-hexenyl acetate and linalool, are also volatiles found in the oils of many plants. Indole (compound 7) is common to plants as well (e.g. Tolsten and Bergstrom, 1988). Compounds 6 and 8 through 11 are less common, but have all been reported several times as plant produced chemicals (see discussion). Synthetic candidates for these compounds, when analyzed by GC, and had the same retention times on the two different columns as their natural versions. Moreover, MS, IR and [^H]NMR spectra were found to be identical to the natural products. The [^H]-NMR spectra matched with the spectra reported by others: (3E)-4,8-

PAGE 127

Fig. 6-2 A) Profile of volatiles released by a complex of corn seedlings damaged by BAW larvae. B) Profile of volatiles released by undamaged corn seedlings. C) Structures and identities of the eleven major compounds, with corresponding peak numbers. Gas chromatographic profiles of the volatiles were obtained after analysis on a 50-m 007 methyl silicone column (0.25 mm ID, 0.25 film thickness), after a 2hr collection. ISl and IS2 are the internal standards n-octane and n-nonyl -acetate.

PAGE 128

115 Table 6-1. Super Q trapped volatile compounds identified in the atmosphere associated with corn seedlings fed upon by BAW larvae. Peak" Compound Kovats''^ nq/hr*" Relative% 1 (Z)-3-hexenal 775 2015 (785) 5-21 2 (E)-2-hexenal 834 310 (108) 1-4 3 (Z)-3-hexen-l-ol 845 502 (118) 2-5 4 (Z)-3-hexen-l-yl acetate 991 2058 (286) 12-20 5 1 inalool 1089 427 (40) 2-4 6 (3E)-4,8-dimethyl1110 676 (140) 3-7 1,3,7-nonatriene 7 indole 1266 2333 (617) 9-26 8 0trans -berqamotene 1441 834 (503) 2-9 9 (E)-(P)-farnesene 1451 3153(1867) 8-35 10 (E)-nerol idol 1551 947 (329) 6-10 11 (3E,7E)-4,8,12-trimethyl1569 61 (16) 0-1 1,3,7, 11 -tri -decatetraene ^Peak numbers in figure 6-2. ''Kovats' GLC index (Kovats, 1965) for the 48 m Methyl Silicone capillary column. "^Average amounts released by 3 corn seedlings that were fed upon overnight. During the one hour collections 15 BAW were feeding on the seedlings (n=6). Standard deviations are shown in parentheses.

PAGE 129

116 dimethyl -1, 3, 7-nonatriene and (3E, 7E) -4,8, 12-trimethyl -1,3, 7,11tridecatetraene (Maurer et al^, 1986; Dicke et aL., 1990a), atrans bergamotene (Corey et aK, 1971), (E)-p-farnesene (Bowers et aK, 1972), and (E)-nerolidol (Doskotch et ai^, 1980). Sources of the collected volatiles. Volatiles were collected separately from the three major components of a complete plant-host complex; damaged plants, frass, and larvae. The damaged plants released, by far, most of the volatiles (Figure 6-3). All compounds were collected from the damaged leaves in significant amounts, with the exception of 1-3. Preliminary data, however, indicated that when starved larvae are allowed to feed on the leaves, the three most volatile compounds are released in large amounts. Clearly, the highly volatile compounds are only released in the observed amounts directly as a result of plant damage. No detectable amounts of the eleven volatiles were released by frass or larvae (Figure 6-3). In the collections of the starved larvae, however, an additional peak showed up. It was determined that it was emitted by the oral secretions that the larvae regurgitated while fighting among themselves. The compound was identified as pentadecane by GC-MS. Responses to a synthetic odor blend. A collected natural blend and a synthetic mimic (Table 6-2) were used in further bio-assays. The large differences between the relative percentages given in Tables 6-1 and those in Table 6-2 are due to the different collection procedures that were used. Volatile for Table 6-1 were collected for only 2hr the day after the larvae had started feeding on the corn seedlings, while Table 6-2 represents the massive collections for identification and bio-

PAGE 130

117 IS IS IS IS time Fig. 6-3 Volatiles released by different components of a complete plant/host complex. A) Complete complex of BAW larvae feeding on corn seedlings. B) Water-washed corn seedlings that were damaged by BAW larvae. C) BAW frass wiped of off the damaged corn seedlings. D) Starved water-washed BAW larvae. Peak numbers correspond with those in figure 1. The asterisk marks pentadecane, a volatile emitted by the oral secretions from BAW larvae. L i

PAGE 131

118 Table 6-2. Quantative comparison of natural and synthetic blends used for bio-assays. NATURAL SYNTHETIC Peak^ Compound Relative%^ Relative%^ 1 (Z)-3-hexenal 49.3 48.5 2 {E)-2-hexenal 1.7 1.8 3 (Z)-3-hexen-l-ol 10.4 10.3 4 (Z)-3-hexen-l-yl acetate 15.3 14.9 5 1 inalool 3.2 2.8 6 {3E)-4,8-dimethyl1,3,7-nonatriene 2.5 2.3 7 indole 2.3 2.2 8 a-trans-berqamotene 3.2 3.4 9 (E)-(P)-farnesene 10.0 11.7 10 {E)-nerol idol 1.6 1.6 11 (3E,7E)-4,8,12-trimethyl0.5 0.5 1 ,3,7. 11-tri -decatetraene *Peak numbers in figure 6-2. ''Percentage of total amount in blend (for each test sample 1% = 205 ng).

PAGE 132

119 assays that started out with larvae on fresh leaves. As will be shown in the next chapter several compounds are not released when the larvae just start feeding. Hence, the large differences obtained for some of the major compounds. A majority of the insects would fly to the extracts of volatiles collected from BAW feeding on corn seedlings (Table 6-3). Females that were experienced on the synthetic blend responded equally well to the natural source as females that had a natural experience. However, experience did make a difference for the females that were tested to a synthetic blend of the identified compounds. Significantly fewer wasps that had received a natural experience flew to the synthetic blend than did females that had received a synthetic experience. The responsiveness exhibited by females with a synthetic experience to a synthetic odor blend, did not differ from that of females that were tested to the natural extract. None of the tested females would fly to the solvent alone. Discussion The technique of collecting plant volatiles directly from an airstream passed over the odor source (Heath et aL., 1990) has some major advantages over the often applied harsh methods of extraction or steam distillation. The latter two techniques give no indication of how much of each identified compound is actually released into the environment and also may result in destruction or isomerization of some of the essential chemicals. One of the most extreme deviations from natural release rates is probably the reported representation of (Z)-3-

PAGE 133

120 Table 6-3, Flight tunnel responses of C. marqiniventris females with different experiences to extracts of; 1) volatiles collected from BAW larvae feeding in corn seedlings (NATURAL), 2) a synthetic mimic of the same volatiles containing the eleven major components (SYNTHETIC), and 3) the extraction solvent, methylene chloride, only (SOLVENT). Odor source Experienced Odor n number of flights average % of flights^ NATURAL 30 24 66.7 (16.7) A NATURAL SYNTHETIC 30 21 58.3 (16.0) A NATURAL 30 11 30.6 (11.5) B SYNTHETIC SYNTHETIC 30 19 52.8 (15.0) A SOLVENT NATURAL 30 0 0.0 C ^On six different days five females were tested for each treatment. Analysis of variance was performed on the daily percentages after arcsin-square-root transformation, followed by Duncan's multiple choice test. Different letters indicate significant differences between treatments (P < 0.05).

PAGE 134

121 hexenal in the oils of plants. In our collections (Z) -3-hexenal was one of the major components. Due to its high volatility (Z)-3-hexenal is much harder to collect than the other green leaf volatiles, including (E)-2-hexenal . This is illustrated by our preliminary attempts to recollect the compounds from the GC as described earlier. Initially a (u-)tube immersed in liquid nitrogen was used; collection efficiency was extremely poor for (Z) -3-hexenal , while high for all other volatiles. Only when the collection tubes were used that contained a small amount of Super Q® adsorbent was it possible to collect (Z)-3-hexenal effectively. High amounts of (Z)-3-hexenal are not just characteristic for BAW damage or corn. Preliminary results showed that artificial damage of corn and damage of other plants (i.e. cotton, tomato, and cowpea) resulted in the release of similar relative amounts of (Z)-3hexenal (unpubl. data). Buttery et aK (1987) reported on the fast isomerization of (Z)-3-hexenal (probably into (E) -2-hexenal ) in crushed tomato leaves. Their and our results suggest that damaged green leaves release much more (Z) -3-hexenal than generally reported. When Buttery and Ling (1984) collected the volatiles of corn plants that were cut at the stem, they also found (Z) -3-hexenal , (E)-2hexenal, (Z)-3-hexen-l-ol , (Z)-3-hexen-l-y1 acetate, linalool, and (E)P-farnesene. Thompson et aK (1974) extracted the essential oil from corn and identified 59 compounds, among which were also indole and nerolidol. No previous reports on corn volatiles mention atrans bergamotene, (3E) -4,8-dimethyl -1 ,3, 7-nonatriene, or (3E,7E)-4,8, 12trimethyl-l,3,7,ll-tridecatetraene. The latter two apparently related methylene terpenes were recently reported by Maurer et a]^ (1986) who

PAGE 135

122 found them in the oil of Elettaria cardamomum (cardamom oil). Kaiser (1987) (as ref. by Dicke et aL. 1990a) reported these compounds from night-scented flowers of different plant species that are pollinated by moths. Most interestingly, Dicke et jlL (1990a) found these compounds being released by lima bean leaves that had been subjected to spider mite infestation. Similar to our results these compounds were not released by undamaged plants, nor were they detected from artificially damaged leaves. At least one of the methylene terpenes, (3E)-4,8dimethyl-l,3,7-nonatriene, is involved in the attraction of predatory mites that feed on spider mites (Dicke and Sabelis, 1988b). Several preliminary results suggest that these methylene terpenes are generally released by green plants that have been under attack by insect herbivores, and may therefore elicit behavioral responses in a variety of insect species. Of the other compounds that we identified as being released by BAW damaged corn, (Z)-3-hexen-l-ol , (Z)-3hexen-l-yl acetate, and linalool were also released by spider mite infested lima bean (Dicke et afL, 1990a). The release of these compounds could also be specific for herbivore induced damage. The volatiles and release rates that we report here appear to be specific for corn leaves subjected to caterpillar damage (next chapter), atrans -bergamotene appears to be rare. It has been reported from cotton (Minyard et aK, 1966) and from Buck's horn ( Rhus tvphina ) (Bestmann et al^, 1988). (E)-p-farnesene was the most predominant terpenoid in the collections. It has previously been identified from corn leaves by Buttery and Ling (1984), but they found it in much smaller amounts relative to the other compounds. (E)-p-farnesene has been found to

PAGE 136

123 attract a chalcid wasp (Kamm and Buttery, 1983) and serves as an alarmpheromone for aphids (Bowers et a1 . , 1972; Edwards et a1 . , 1973; Wohlers, 1981). The E isomer of nerolidol is present in the essential oil of several Melaleuca tree species (Jones and Harvey, 1936; Naves, 1960; Doskotch et aL., 1980). Picker et aL. (1976) found it also in the bark of the Australian tree Fl indersia laevicarpa . Doskotch et al_^ (1980) who isolated (E,S)-nerol idol from Melaleaca leucadendron leaves, showed that it functions as a feeding deterrent for gypsy moth larvae. Terpenoids have been suggested and shown to inhibit feeding in several cases (e.g. Doskotch et aL., 1980; Mihal iak et al, 1987; Gunasena, et al . , 1988). It seems that corn seedlings in response to herbivore damage produce the identified terpenoids as well as indole and (Z)-3hexen-l-yl acetate in relatively large amounts. This could be a direct defense against their attackers, but may serve an additional function in attracting natural enemies of these attackers (see next chapter). The parasitoid females responded in a dose-related manner to the collected volatiles (Figure 6-1). The drop in response at the highest dosage tested indicates that for optimal bio-assay results release rates should be carefully balanced. By applying the extracts on filter paper the obtained release rates were obviously far from natural. Before the importance of each individual compound in the attraction of marqiniventris can be determined, exact formulations resulting in release rates similar to a natural situation will have to be obtained. Furthermore, we need to establish which optical isomers of linalool, «trans -bergamotene, and (E)-nerol idol are released by corn seedlings. We

PAGE 137

124 used racemic mixtures of the latter three in our synthetic blend, and this may explain why females with a natural experience did not respond well to the synthetic blend (Figure 6-3). An experience with the blend containing the racemic mixtures, however, increased the responses to a level comparable to the responses to a natural blend. These results again illustrate the significant effects of experience on the responses to airborne semiochemical s in parasitoids. If, in the future, synthetic blends are used to condition parasitoids such that they will perform more effectively when released in host infested areas, the best results are likely to be obtained with the most perfect mimics of naturally released odors. The damaged plants are clearly the main source of the identified compounds (Figure 6-3). This is in agreement with the responses of the parasitoids to the different components of a complete plant-host complex (previous chapter). The wasps are significantly more attracted to the damaged plants than to frass or to larvae. In the last chapter we saw that undamaged plants were far less attractive, here it is shown that when volatiles were collected from undamaged plants only minute amounts of some the compounds were detected (figure 6-1). Striking differences between different types of damage, will be discussed in the next chapter.

PAGE 138

CHAPTER VII THE ACTIVE ROLE OF PLANTS IN THE PRODUCTION OF THE VOLATILES THAT GUIDE Cotesia margini ventri s FEMALES TO THEIR HOSTS Introduction In the two last chapters, the importance of plants in the attraction of C^ margi ni ventri s has clearly been demonstrated. Damage of the plants, however, was required. Many studies have demonstrated that herbivore damage can induce responses in plants that cause major differences in quantities of secondary plant substances, many of which are volatile (for reviews see e.g. Rhoades, 1979, 1985; Kogan and Paxton, 1983; Pimentel, 1988). Rarely are these chemicals found to function as metabolites essential to the development of the plant, but in many cases they are shown to have allomonal activities. Most studies on the significance of herbivore induced production of secondary plant compounds focus on the direct ecological interactions between the plants and the herbivores that feed on them. If these induced chemicals have adverse effects on the organisms that attack the plant the chemicals are often classified as phytoalexins (e.g. Kogan and Paxton, 1983; Ryan, 1983). Only a few authors (Vinson, 1975; Price et aK, 1980; Price, 1981; Vinson et aK, 1987) have suggested a possible mutual istic interaction between herbivore-damaged plants and the third trophic level of insect parasitoids and predators. Recently, such an interaction has been demonstrated between plants, herbivorous spider mites and their 125

PAGE 139

126 natural enemies, predatory mites (Dicke, 1988; Dicke and Sabelis, 1988b; Dicke et iL., 1090a, 1990c). In several elegant experiments it was shown that plants that have been subjected to spider mite infestation produce volatiles that attract predatory mites. Preliminary work on the parasitoid Epidinorcarsis loperi , suggest an active role for the plants in attracting this parasitoid. This wasp is attracted to uninfested leaves of host-infested plants but not to leaves of uninfested plants (Nadel and van Alphen, 1987). To determine if the host location process by marqiniventris involves a similar sophisticated tritrophic interaction the effect of caterpillar damage on the plant needs to be studied more closely. In this chapter a possible response by corn seedlings is studied by comparing the odors emitted by fresh caterpillar (BAW) damage with odors of older damage, and by comparing caterpillar damage with artificial damage. Corn seedlings with different types of damage are subsequently tested for their attractiveness to marqiniventris . The effect of experience on the parasitoid's response were studied as well. Materials and Methods Insects . The insects were reared and maintained as described in Chapter II. Fl iqht tunnel . Bio-assays were conducted with marqiniventris in the flight tunnel as described in Chapter IV, under the conditions described in Chapter VI. Odor sources tested in the flight tunnel contained corn seedlings with various degrees of damage with or without 15 BAW larvae.

PAGE 140

127 Volatile collections . Using the small system (see Chapter VI), volatiles were collected from corn seedlings with various degrees of damage. Filtered, humidified air (300 ml/min) was passed over three seedlings that were placed in each chamber with or without 15 BAW larvae. The entrained volatiles were collected for 2 hrs onto Super Q® filters. Four collections were carried out simultaneously. The collection procedure as well as extraction and analysis techniques were the same as described in Chapter VI. After adding the internal standard, 2 /il of each collection extract was analyzed on the 50-m methyl silicone column. Procedures and Results Significance of long-term larval damage . To establish possible differences in plant odor emission resulting from different types of damage, corn seedlings were treated as follows: 1) OLD DAMAGE-Forty BAW larvae were allowed to feed on 10 corn seedlings overnight, subsequently three arbitrarily chosen seedlings together with fifteen BAW larvae were used for collection of the volatiles; 2) FRESH DAMAGE-Fifteen of the remaining BAW larvae were placed on three undamaged seedlings; 3) ARTIFICIAL DAMAGE-Undamaged seedlings were macerated with a razor blade and quickly placed into a third collection chamber; 4) NO DAMAGE-For the fourth and final collection three undamaged seedlings were carefully placed inside the last collection chamber. GO profiles of the extracts of each collection were used to compare the amounts of the compounds (identified in chapter VI) released by the leaves in each treatment.

PAGE 141

128 Striking differences were observed in the GC profiles of volatiles collected from odor sources that contained corn seedlings with different types of damage (Figure 7-1). Collections from sources with overnight damage (OLD DAMAGE) revealed the familiar (Chapter VI) eleven peaks. The seven larger (less volatile) compounds were missing or only present in minor amounts in the odor blend released by larvae feeding on fresh leaves (FRESH DAMAGE). These profiles were virtually identical to those collected from seedlings that had been artificially damaged (ARTIFICIAL DAMAGE). Finally, it was found that carefully handled undamaged seedlings did not release of the compounds observed in the above collections in measurable amounts. Lag time in plant response? The dramatic difference between "OLD DAMAGE" odor sources and "FRESH DAMAGE" sources was possibly the result of a delayed response by the damaged plants. In other words, the caterpillar damage may have caused the plant to release the seven larger compounds, but this response by the plant was not instantaneous and was therefore not yet visible for the sources with "FRESH DAMAGE". The following experiment was conducted to determine whether caterpillar damage induces a delayed response by the plant resulting in emission of different compounds or different proportions of compounds. Volatiles were collected and analyzed just after, and 16 hours after, corn seedlings were subjected to damage. Corn seedlings were grown in trays (50 X 35-cm, 6-cm deep, approximately 40 seedlings per tray) inside a greenhouse, and used when 8-12 days old. The top and bottom of six cylindrical cardboard micro-pipet containers (138-mm high, 45-mm OD)

PAGE 142

129 IS1 U 2 6 u IS2 9 8 10 JL11 B IS1 i IS2 IS1 U IS2 IS1 IS2 Figure 7-1. GC profiles of volatiles released by corn seedlings subjected to different damage treatments. A) BAW feeding on seedlings that were damaged overnight by BAW larvae. B) BAW larvae feeding on fresh seedlings. C) Artificially damage fresh seedlings. D) Undamaged seedlings. All collections lasted 2 hours. Peak numbers correspond with the compounds given in Figure 6-2. Internal standards were noctane (ISl) and n-nonyl -acetate (IS2), each representing 1000 ng.

PAGE 143

130 were removed. The containers were carefully placed over six different seedlings, and 10 late 2nd instar BAW larvae were dropped inside each of them. The top containers were covered with a piece of nylon screen held in place with a rubber band. After two hours the containers and the larvae were removed. The following day this procedure was repeated with six other seedlings in the same tray. Immediately after the second treatment all treated seedlings were cut at the base, their stems wrapped in wet cotton wool, and placed in volatile collection chambers. Four chambers of the smaller collection system (see Chapter VI) were used each holding three of the seedlings. Two contained seedlings that had been damaged the day before, while the other two contained the seedlings from which larvae had just been removed. Volatiles were collected from the leaves (without larvae) for two hours and were then analyzed by GC. Collection and analytical procedures were the same as described in the previous chapter. Amounts of volatiles were compared for the two treatments. The experiment was repeated four times (n=8). Significant differences were found between the two treatments with equal damage (Figure 7-2; Table 7-1). This time, however, compounds 4-6 and 11 were released in substantial amounts by the leaves with the fresh damage. Their release was not significantly different from what was released by the plants that had been damaged the day before (Table 7-1). This might indicate that the four hour period (2hr damage + 2hr collection) is sufficient for plants to buildup the production and release of these compounds. Still, highly significant differences were found between the two treatments in the amounts of compounds 8-10 (otrans-bergamotene, (E)-p-farnesene, and (E)-nerol idol ) that were

PAGE 144

131 IS FRESH DAMAGE IS IS ii .J.. 4 OLD DAMAGE IS 7 9 8 11 .J..... 10 11 Figure 7-2. Comparison of volatiles released by corn seedlings with fresh BAW damage and seedlings with older damage, FRESH DAMAGE: three seedlings damaged for 2 hours by 15 larvae just prior to collection. OLD DAMAGE: three seedlings damaged for 2 hours by 15 larvae 16 hours prior to collection. No larvae were present on the leaves during the collections.

PAGE 145

132 Table 7-1. Amounts of volatiles released by corn seedlings damaged by BAW larvae just after damaged (FRESH) compared with volatiles released 16 hours after damage (OLD). Numbers represent average amount (ng) collected over a period of 2 hours, standard deviation is given with each number. PEAK # COMPOUND FRESH OLD pvalue^ A •T ^7^ '^-hoYon-l — \/l ar'o+ato \i) o iicAcii 1 jr 1 ai^cUaUc 424 (301) 1 ?R7 (1205) • IOC 5 1 i nal ool 242 (77) 242 (92) .994 6 (3E)-4,8-dimethyl-l,3,7nonatriene 962 (359) 1673 (822) .110 7 indole 159 (115) 156 (267) .980 8 a-trans-berqamotene 14 (13) 115 (72) * .010 9 (E)-p-farnesene 35 (29) 445 (223) * .003 10 (E)-nerol idol 9 (5) 101 (67) * .011 11 (3E,7E)-4,8,ll-trimethyl1,3,7, 11 -tridecatetraene 24 (17) 75 (26) * .010 TOTAL 1868 (722) 4065 (1329) * .019 Pvalues derived from paired T-test for the treatment means (n=8). Asterisks indicate significant differences.

PAGE 146

133 released. As a result, the total amounts of volatiles released by seedlings with older damage were on average more than twice as high as what was released by the seedlings with the fresh damage (Table 7-1). The earliest peaks (1-3) were absent in both collections. As indicated in the previous Chapter (Figure 6-3), these compounds were only released in significant amounts during active damage. Is the plant response specific for larva! damage? In the previous experiment an active response by plants to damage was shown, the question remained whether this response was specific for the type of damage that caterpillars cause, or whether the response could be induced by artificial damage as well. Oral secretions of herbivorous insects have been linked to physiological changes in host plants (Capinera and Roltsch, 1980). It was therefore hypothesized that caterpillar secretions might be involved in the observed plant responses. To answer these questions the following experiment was conducted. During a period of two hours, four groups of three seedlings underwent different treatments. One group was subjected to feeding damage by 10 late 2nd instar BAW larvae. During the same period that these caterpillars were feeding, two other groups were subjected to artificial damage whereby the damage done by the caterpillars was roughly mimicked with micro-scissors and a razor blade. One of the artificially damaged groups of seedlings, and the remaining fourth group of undamaged seedlings were treated with the regurgitated gut content of other corn fed BAW larvae. This was done by grabbing a caterpillar with a pair of forceps, pinching the head region with another pair until regurgitation was induced. On the artificially damaged seedlings the

PAGE 147

regurgitant was immediately rubbed over the sites that had just been damaged. As a control, equal amounts of regurgitant were rubbed over the undamaged seedlings. The following day, approximately 16 hours later, the volatiles were collected separately from each of the treatment groups. The procedure was exactly the same as described for the previous experiment. After analysis by GC, amounts of volatiles released by each group were compared. The comparison between the volatiles released by seedlings damaged by larvae and the volatiles released seedlings with artificial damage, revealed that a larvae-specific factor was involved in the induction (Figure 7-3). As shown earlier, leaves with larval damage released terpenoids in high amounts. Seedlings that only underwent artificial damage released far less. However, those artificially damaged seedlings that were treated with the larval regurgitant released the most predominant compounds in amounts similar to those found for the larval damaged seedlings. The control seedlings, undamaged leaves that were treated with regurgitant, released almost no detectable amounts of volatiles (Figure 7-3). Significance of the plant response to the attraction of the parasitoid. It was hypothesized that, if caterpillar feeding induces plants to actively release volatiles, this plant response might be exploited by the host seeking wasps. To test this experienced female wasps were allowed to fly to seedlings that had undergone the same treatments as described in the previous experiment. The seedlings were offered to the wasps in the flight tunnel in a two-choice experiment.

PAGE 148

135 CATERPILLAR DAMAGE 5 6 i IS 8 1 11 TIME ARTIFICIAL DAMAGE IS JUL 7 JL 11 ARTIFICIAL DAMAGE + REGURGITANT TIME 7 ± IS 8 A....^ aW t. 9 10 11 ..XI TIME NO DAMAGE + REGURGITANT IS 5 6 TIME Figure 7-3. Comparison of volatiles released by corn seedlings damaged by BAW with volatiles released by seedlings with various artificial damage treatments. The day before collections took place, the seedlings were either damaged by BAW caterpillars for 2 hours (CATERPILLAR DAMAGE); artificially damaged with a razor blade during the same period (ARTIFICIAL DAMAGE); artificially damaged and treated with caterpillar regurgitant (ARTIFICIAL DAMAGE + REGURGITANT); left undamaged but treated with regurgitant (NO DAMAGE + REGURGITANT).

PAGE 149

136 Two seedlings of a particular treatment (their stems wrapped in wet cotton wool) were placed upright in a 20-ml vial. Two seedlings of another treatment were placed the same way in a second vial. The two vials were placed 20 cm apart on a stand 18 cm above the flight tunnel floor and approximately 80 cm upwind from the insect release funnel (Chapter IV). The bio-assay procedure was the same as described in Chapters IV and V for experienced insects. Wasps were given experience by allowing them to oviposit in one or two larvae on overnight damaged corn. Four combinations of seedlings were tested: 1) seedlings with just artificial damage versus seedlings with larval damage, 2) seedlings with just artificial damage versus seedlings with artificial damage that were treated with regurgitant, 3) seedlings with larval damage versus seedlings with artificial damage that were treated with regurgitant, and 4) seedlings with just artificial versus undamaged seedlings that were treated with regurgitant. On 5 different days 8 females flew to each combination (n=40) that was presented in the flight tunnel. The positions of the vials holding the seedlings were switched each time four females had been tested. Preference for a treatment in each combination was tested using the chisquare test (a < 0.05). The insects strongly preferred the leaves with larval damage over leaves with just artificial damage (Figure 7-4a). The artificially damaged leaves that were treated with caterpillar regurgitant were clearly preferred over the leaves with only artificial damage (Figure 74b). When given the choice between leaves with larval damage and

PAGE 150

137 INCOMPLETE FLIGHTS ARTIF.-iREG 20l -X22 1 T ARTIF. + REG.i 16 , . ^REG.| •11 Figure 7-4. Responses during dual choice flight tunnel tests by experienced marqiniventris females to corn seedlings that underwent the various treatments described in the legend to Figure 7-3. A day after the leaves were treated, females had the opportunity to choose between the odors released by seedlings with; (A) artificial damage vs. caterpillar damage; (B) artificial damage vs. artificial damage treated with regurgitant; (C) caterpillar damage vs. artificial damage treated with regurgitant; (D) artificial damage vs. no damage treated with regurgitant. On 5 different days 8 females were tested to each combination (n=40). The open bars represent the females that did not fly to the odor sources. Asterisks indicate significant preferences for a particular odor (Chi-square; P < 0.05).

PAGE 151

138 artificially damaged leaves treated with regurgitant the females showed no preference (Figure 7-4c). Fewer females flew to the undamagedl eaves than to the artificially damaged leaves (Figure 7-4d). This is in agreement with the observation that artificial damage alone does result in the release of some of the terpenoids (Figure 7-3), and shows that the regurgitant by itself did not elicit the attraction. Do the parasitoids learn to respond to the plant produced odors? In the previous bio-assay the test insects all experienced a plant-host complex that contained seedlings that were damaged overnight prior to their release in the flight tunnel. Any preferences for leaves that were actively releasing volatiles may have been the result of associative learning during experience. To test this hypothesis females with different experiences flew in a choice test to a plant-host complex with seedlings that had been damaged the day before next to a similar complex with freshly damaged leaves. To obtain the treatments plants were grown in pots in the green house. Six seedlings in one pot were exposed to 30 second instar BAW larvae for 3 hours, after which the larvae were removed. The following day six seedlings in another pot underwent the same treatment. Immediately following the second treatment all seedlings were cut and their stems wrapped in wet cotton. Of each treatment 3 seedlings were placed in two different glass odor source containers as described in Chapter V (Figure 5-1). Together with the plants, 15 BAW larvae were placed in each container and the containers were placed inside the flight tunnel next to each other (Chapter V).

PAGE 152

The remaining seedlings were used to provide the wasps with experience. Females were given either no experience, a complete experience (including oviposition) on leaves that had been damaged the day before, or a complete experience (including oviposition) on freshly damaged leaves. On nine different days 10 females of each experience type were tested to the two sources. The order in which they were tested was alternated on different days and after 5 insects of one type were tested another type was used until all 30 insects were tested. Half way through (after 15 insects) the positions of the odor sources were changed. Again chi -square tests were performed to reveal any significant preferences. Females that had experience with old damage significantly preferred old damage, while females with fresh damage showed no clear preference (Figure 7-5). Only a small fraction of females without experience would fly to the odor sources. Discussion In the previous chapter it was found that all of the identified compounds are released by the caterpillar damaged seedlings and not by the caterpillars themselves nor by their feces or other by-products. In this chapter it is shown that the larger terpenoids, particularly (E)-obergamotene, (E) -p-farnesene, and (E) -nerol idol , were only released in significant amounts by leaves several hours after they had been damaged by caterpillars (Figures 7-1 and 7-2). The fact that the release was not instantaneous, but required an apparent lag time (Figure 7-2), indicates that it was an active response by the plant. The response

PAGE 153

140 DUAL CHOICE TESTS IN FLIGHT TUNNEL Figure 7-5. Preferences exhibited by marginiventris females for odors released by corn with fresh BAW damage versus odors released by corn with old BAW damage. Females were either given no experience (NONE), given experience on corn with fresh damaged (FRESH), or given experience on corn with old damage (OLD). The asterisk indicates a significant preference for the old damage odors (Chi -square; P < 0.05).

PAGE 154

141 could not be induced by artificial damage alone. However, when artificially damaged sites were treated with the regurgitated foregut content of the caterpillars, the seedlings released terpenoids in amounts similar to those released by caterpillar-damaged seedlings (Figure 7-3), Therefore, the observed plant response seemed specifically induced by the feeding of the herbivores tested. It involved not only damage, but required a factor in the regurgitant (most likely in the saliva) of the caterpillars as well. Whether this factor involves enzymes, micro-organisms or something else, remains to be elucidated. Ll marqiniventris females were strongly attracted to the terpenoid releasing seedlings, much more so than to just artificially damaged, or undamaged seedlings (Figure 7-4). The response, however, depended on their previous experiences (Figure 7-5); the wasps responded strongly to the odor of terpenoid releasing leaves after they had had foraging experience on them. This provides evidence for the hypothesis that injury induced release of volatile phytoalexins in plants will not only have a direct allomonal effect on herbivores, but may serve a secondary synomonal function in attracting natural enemies of phytophagous insects. Fast growing plants like corn invest much of their energy in growth and initially little in defense. When under herbivore attack, however, the flexible defense system will allow a rapid induced production of carbon-based defense chemicals (phytoalexins) (Coley et aL., 1985). Terpenoids have been reported as plant defense chemicals (Mabry and Gill, 1979; Doskotch et aK, 1980; Mihaliak et aK, 1987; Gunasena et

PAGE 155

142 a1 . , 1988), and were found to reduce caterpillar feeding and caterpillar development. The terpenoids released by the corn seedlings are likely to serve similar functions. In addition, however, their volatility and high turnover rate should make them reliable indicators of the presence of hosts for parasitoids. The terpenoids are not only reliable cues because they are closely associated with herbivore-damage, but their release is continuous, even during periods when the caterpillars are not feeding. The host larvae of marqiniventris tend to feed in bursts. In preliminary observations this was recorded for Spodoptera fruqiperda , Heliothis zea, Trichoplusia ni* and Pseudoplusia includens . During the non-feeding periods no significant amounts of the typical green leafy volatiles were being released (see Chapter VI). These smaller compounds, however, may also play a role in the attraction of the parasitoid since the wasps are attracted to seedlings with fresh damage (Figure 7-5). In addition, the terpenoids that are closely associated with the damage inflicted by the host will serve as very important cues for experienced wasps, especially during the long periods of time when the caterpillars are not eating. Many herbivores will have developed variable levels of resistance to the plant produced chemicals, making them less effective as direct defense allelochemicals. Attraction of natural enemies of herbivores, however, results in an additional advantage to the plants, thereby perhaps maintaining a selection pressure that favors the production of these chemicals in the observed high quantities. Cost-benefit analyses concerning plant defense strategies should take into account how plants

PAGE 156

can safeguard themselves against severe herbivore injury by attracting predators or parasitoids (Sabelis and de Jong, 1988; Dicke and Sabelis, 1988b; Dicke and Sabelis, 1989). Dicke and co-workers presented the only study in which a comparable interaction was revealed. They (Dicke and Sabelis, 1988b; Dicke and Sabelis, 1989; Dicke et aK , 1990a; 1990c) showed that spider mite infested plants release volatiles that are used by predatory mites to locate their prey {=spider mites). Further examples that indicate similar interactions are rare. However, considering the adaptive advantages to both the entomophage and the plant, these mutual istic pi antinsect interactions may be common in various tri trophic systems. It is likely that the terpenoids and indole are involved in other types of interactions as well. They may, for example, act as oviposition deterrents for females searching for sites to deposit their eggs, or function in the communication between plants (as suggested by Dicke et aj^, 1990c). Further studies will also have to reveal whether the induced reaction is limited to the damaged sites, or whether it is systemic as has been shown in studies on proteinase inhibitors (Green and Ryan, 1972; Sanchez-Serrano et JLLl, 1987; Bowless, 1990), and appears to be the case in the tritrophic mite interaction (Dicke et al . , 1990c). For now, we can say that the induced plant response is specific for herbivore-damage, and as was revealed in the choice tests, may serve as a reliable cue for the parasitoid.

PAGE 157

CHAPTER VIII SUMMARY AND CONCLUSIONS One of the original purposes of this research was to establish the source and identities of the airborne allelochemicals that attract females of C^. marqiniventris into the vicinity of their hosts. As we now know, an overly simple picture of the system was drawn at the onset of the research. It was suspected that factors directly associated with the hosts, such as saliva, silk, feces, or cuticle, emit odors that parasitoids exploit to locate their hosts. Many examples exist of parasitoids responding to by-products of hosts (for reviews see Weseloh, 1981; Eller, 1990). Results from studies by Loke and co-workers indicated the importance of cues associated with host by-products in short range host location by marqiniventris (Loke and Ashley, 1984a, 1984b, 1984c). The first indication that C^ marqiniventris was not merely attracted to specific volatiles that are directly associated with, and that are common to its wide variety of hosts was demonstrated by the results from the first learning experiments presented in Chapter II. In the four-arm olfactometer, it was found that a brief contact experience with a host habitat caused a striking increase in the response to odors released by a plant-host complex. Moreover, the increase in response was significantly higher for odors that were emitted by the plant-host complex with which the parasitoids had had experience as compared with 144

PAGE 158

145 odors of an alternative plant-host complex. This demonstrates the importance of experience, and provides the initial evidence for the hypothesis that associative learning plays a major role in the hostsearching behavior of this parasitoid. These findings also show that we are not dealing with a specific volatile or blend of volatiles that attract the female wasps. The attractive volatiles for the two plant-host complexes vary, since the responses by the wasps to each complex varied. The additional learning experiments presented in Chapter III, demonstrate the flexibility of the parasitoids' response to the host-related airborne semiochemicals. In Chapter III it was also shown that the source of the volatiles can be quite ambiguous. In choice tests a particular plant-host complex was chosen more often by females that had experienced that complex than by females that had experienced the alternative complex. This preference for the experienced odor was shown for combinations where only the host species was varied as well as in combinations where only the plant species was varied. These findings demonstrate that both the host larvae, as well as the plants they feed upon, play a role in the production and/or release of the volatiles that attract C^ marqiniventris. In all experience tests, it was shown that associative learning and the resulting increase in response required only a brief (<20 seconds) contact with a plant-host complex, and that actual encounters with the hosts were not required. Associative learning is clearly a powerful modifier of this insect's behavior, and can be expected to play a major role in its host-searching behavior in the field.

PAGE 159

146 Pinpointing the main source of the active chemicals was accomplished with flight tunnel bio-assays. Experienced females exhibited typical oriented flights to an odor source consisting of host larvae feeding on seedlings (Chapter IV). When similar plant-host complexes of BAW larvae feeding on corn seedlings were divided into their main components (Chapter V), significantly more females of C^ marqiniventris flew to the damaged corn seedlings than to the other two components, frass and BAW larvae. Clearly, plants turned out to be the main source of airborne attractants, but they were not attractive unless damaged by the host. Collection and analysis of the volatiles released by the plant-host complex revealed the consistent presence of eleven volatiles in detectable amounts (Chapter VI). These included the four typical green leafy volatiles (Z)-3-hexenal , (E)-2-hexenal , (Z)-3-hexen-l-ol , and (Z)3-hexen-l-yl acetate. With the exception of indole the remaining compounds were the terpenoids linalool, (3E)-4,8-dimethyl-l,3,7nonatriene, tc) trans -berqamotene. (E)-(P)-farnesene, (E)-nerolidol , and (3E,7E)-4,8,12-trimethyl-l,3,7,ll-tridecatetraene. All of the latter have been isolated from plants before, but some are quite rare. When the volatiles were collected from each of the three main components (damaged leaves, frass, and larvae) separately, it was found that all of the identified volatiles were released by the plants. The two methylene terpenes have only recently been discovered (Maurer et al^, 1986; Dicke et aL., 1990a); additional preliminary collections indicate that they are released from several plants in response to herbivore damage.

PAGE 160

147 Active release of specific volatiles by the plants in response to herbivore damage was demonstrated in Chapter VII. When caterpillars just started feeding on corn seedlings only the four typical green leafy volatiles were released. After several hours, however, the terpenoids as well as indole could be detected in significant amounts. To show that the plant responds specifically to herbivore damage, volatiles released by larval damaged leaves were compared with those released by artificially damaged leaves. Only after artificially damaged seedlings had been treated with oral secretions of BAW larvae, would they release the terpenoids and indole in amounts similar to those released by as seedlings damaged by larvae. Undamaged seedlings treated with oral secretions and artificially damaged seedlings that were not treated with regurgitant released only minor amounts of some of these compounds. The seedlings that underwent the above treatments were tested for their attractiveness to females of marqiniventris . The seedlings that had been damaged by BAW larvae and the seedlings that were treated with oral secretions after artificial damage were far more attractive than seedlings of the other two treatments. It was shown that caterpillar specific feeding damage induced a response in the plants that resulted in the production of certain volatiles in large amounts. This response required damage as well as a factor in the oral secretions of the caterpillars. The parasitoids took advantage of these plant produced chemicals by using them as cues to locate hosts. As in the earlier tests the responses depended strongly on the experiences of the insects. All insects tested had contacted a plant-host complex that contained leaves that had been damaged overnight. Consequently, the

PAGE 161

148 wasps had experienced the terpenoid and indole, which could explain their preference for terpenoid releasing plants. Indeed, in subsequent tests females experienced on the leaves damaged the previous day strongly preferred the odors from a source containing leaves with older damage, whereas females that were experienced on freshly damaged leaves showed no particular preference for either freshly damaged seedlings or seedlings with older damage. Experience also played a role when insects were tested to a synthetic blend that contained the eleven identified compounds (Chapter VI). It was found that the insects respond, in a dose-related manner, to extracts of the collected volatiles when applied on filter paper. A synthetic version of such an odor blend was prepared and tested for attractiveness in the flight tunnel. Attraction of female parasitoids to the synthetic blend was equivalent to the natural blend only if the females had experienced this synthetic blend while they contacted host feces. The synthetic blend was not a perfect mimic of the natural blend, since it contained racemic mixtures of the stereo isomers of linalool, a-trans-bergamotene and (E)-nerol idol . This would explain the observed difference in response, and emphasizes how sensitive the host searching behavior of marqiniventris is to previous experiences. The host searching behavior of marqiniventris is very much influenced by the insect's foraging experiences. marqiniventris attacks at least 19 different host species on more than 30 different plant species (Table 1-2). As shown, plant odors are the main cues used by this parasitoid to locate its hosts. The chemicals produced by the

PAGE 162

149 plants when damaged by herbivores will not only vary for different plants, but also for the different herbivores (host species) that damage the plants. Consequently, the odor blends that are associated with suitable hosts will vary considerably. Being able to learn, allows these wasps to deal with the variations in odor blends that will most reliably guide the wasps to hosts. Unlearned responses will obviously play a role as well, but as shown in Chapters II, IV and VII, inexperienced females respond very poorly. This could be an artifact of the rearing procedure as was demonstrated for Micropl itis demol itor (Herard et al., 1988b; see Chapter I). It is very possible that if the wasps had been reared on plant fed hosts, the "naive" insects would have responded much better. However, this too would have been a result of experience, whereby the wasps would be experienced as immatures or as young adults, learning the odors that are associated with the host they emerged from. Although it may not only occur during the adult stages, it is clear that learning is a powerful modifier of the host searching behavior exhibited by parasitic insects like C. marqiniventris . This is reflected in a highly flexible response to odors that allow these insects to deal with the vast array of odor blends associated with their hosts. Plants clearly profit from the presence of the natural enemies of their herbivorous attackers. For instance, when caterpillars are parasitized by C^ marqiniventris their feeding rate drops significantly and they eventually die. The damage they can do to the plant is dramatically reduced (Ashley, 1983; Jalali, et aK , 1988). Because of

PAGE 163

150 these advantages, plants are likely sources of semiochemicals that attract the natural enemies of their herbivorous attackers. The herbivorous larvae, on the other hand, will have been selected for to be inconspicuous to natural enemies. This is in agreement with the fact that only very minor amounts of volatiles were collected from the larvae themselves or their by-products (Chapter VI; Figure 6-3). It is to the plants' advantage to attract the wasp, but they may not want to give away their presence to herbivores. The fact that the plants do not release any significant amounts of volatiles when they are not under attack (Chapter VII; Figure 7-1) complies with that notion. When under attack by the caterpillars the plants' strategy appears to change and they start releasing relatively large amounts of volatiles that attract C^ marginiventris and most likely other entomophagous insects as well . Independent of the research presented here Dicke and co-workers discovered a very similar tritrophic relationship between spider mites, plants these mites feed on, and predatory mites that feed on spider mites (Dicke and Sabelis, 1988b, 1989; Dicke et aK, 1990a, 1990b; 1990c). Lima bean plants that have been subjected to spider mite damage release volatiles that are not released by undamaged leaves or artificially damaged leaves (Dicke et aK, 1990a). Several of these volatiles were shown to be attractive to the predatory mites (Dicke and Sabelis, 1988b; Dicke et aK , 1990a). Results obtained by Nadel and van Alphen (1987) may indicate a similar interaction between cassava plants infested by cassava mealy bugs that attract the parasitoid Epidinocarsis lopezi . They found that even the uninfested leaves of infested plants

PAGE 164

151 were attractive. The latter leaves were significantly more attractive than the leaves of completely uninfested plants. All in all, evidence for an active involvement of plants in the attraction of natural enemies of the phytophagous insects that feed on these plants is extremely limited. Yet, considering the adaptive advantages to both the plant and the entomophagous insect, such tritrophic interactions can be expected to be common. Speculating on who is in control of the production and/or release of the chemicals, Dicke et aL. (1990c) suggest two possible scenarios. 1) . The phytophagous organism (in their case spider mites) might control the production of the volatiles "to inform conspecifics about local densities, and thus about food quantity and prospects for competition". 2) . The plant controls the production of the volatiles to recruit entomophagous organism (in their case predatory mites) as "bodyguards" against the herbivores. Dicke et aK (1990c) correctly argue that there would be a cost involved for the herbivores by attracting natural enemies with volatiles and discard the first possibility. Their suggestion, however, that therefore the attraction of predatory mites is the primary function of the plant produced compounds would mean that the relationship between entomophagous insects and plants formed the basis of the plant response. I speculate that it is more likely that the production of the terpenoids and other plant-produced volatiles was originally favored by natural selection because of the direct detrimental (toxic) effects of these chemicals on the herbivorous insects and/or invading pathogens, or that these chemicals are byproducts of such defensive chemicals. When a plant in a population

PAGE 165

152 produces certain chemicals this trait will only be maintained and spread through the population if the costs of the production are counter balanced by direct benefits to the plant. It is highly unlikely that when the terpenoid-producing plants showed up, simultaneously entomophagous insects developed the trait of responding to the terpenoids. On the other hand, a toxic effect of the terpenoids on the herbivores may have provided the plants with an immediate profit, which may have resulted in selection pressures that would maintain the trait and allow it to (perhaps rapidly) take over. The latter scenario may be over-simplified, but it conceptualizes the notion that before the third trophic level (entomophagous insects) was involved, the chemical relationship between the plant and its enemies would have been established. Next the parasitoids and predators may have exploited this relationship by taking advantage of the plant produced chemicals, thereby adding further selection pressures that favor continued production. The production of the volatile chemicals then becomes even more advantageous since the chemicals now not only have a direct negative effect on the herbivores, but they also attract the herbivore destroying insects [this is where the second scenario sketched by Dicke et (1990c) comes into play]. These additional selection pressures may have favored the active release of the original defensive chemicals or their by-products into the environment in which they now served secondary functions as attractants to entomophagous organisms and perhaps repellents to herbivores. It is conceivable that many herbivores have developed various levels of resistance and tolerance against the plant produced chemicals.

PAGE 166

153 making them less effective as direct defensive chemicals. For example, enzymatic adaptations in leaf-feeding insects to plant allelochemicals are reported by Brattsten (1988). The consequences of attracting natural enemies of herbivores, however, may maintain selection pressures that favor the continued production of the chemicals in the observed high quantities even if the chemicals no longer work as a direct defense against the herbivores. If this is the case then the original function of these chemicals in the direct defense against the herbivores or pathogens may be hard to show. However, the literature already provides a multitude of examples that show the direct effects of volatile terpenoids on herbivorous insects. For example, nerolidol (one of the compounds identified here) serves as a feeding deterrent to gypsy moth, Lymantria dispar . larvae (Doskotch et aL., 1980). Mihaliak et £L (1987) showed how nitrogen-limited plants accumulate large quantities of leaf volatile terpenes. They suggest that the soybean looper (which also feeds on corn and is a very suitable host for marqiniventris ) is negatively affected by these high terpenoid levels. Larval consumption, growth and survival declined as the leaf volatile terpene content increased. Gossypol a sesquiterpenoid aldehyde present in cotton has been studied by many for its allelochemical effects (Gunasena et al . , 1988). This compound has detrimental effects on several lepidopterous species. Herbivores may not only be adversely affected by the plant produced chemicals, they may also obtain information from these chemicals. The herbivores in their turn may leave and avoid plants that are releasing the chemical "distress" signals, not only to limit the

PAGE 167

154 direct adverse effect of these chemicals on their physiology, but also to avoid detection by their natural enemies. Avoiding plants that release the chemical signals will also reduce competition for food as pointed out by Dicke et aL. (1990c). In short, it seems most likely that originally the production of terpenoids and other secondary substances in plants was favored because of their direct adverse effects on herbivores and pathogens. Evolutionary adaptations in both the herbivores and their natural enemies must have since resulted in more elaborate interactions involving these chemicals, whereby natural enemies exploit the release of these compounds to locate host and prey. Herbivores may also use the chemicals as cues to assess information on the source; the volatiles could inform them of food quality, likelihood of competition, and safety (detectabil ity by natural enemies). Besides providing new information on the complicated interactions between parasitoids, their hosts, and the plants these hosts feed on, the results of this research may point to novel strategies for biological control. The frequently suggested technique of pre-release stimulation of parasitoids (e.g. Lewis, 1981; Gross, 1981; Nordlund et aK, 1981b; Lewis and Nordlund, 1985) appears particularly appropriate to increase the host searching efficiency of marqini ventri s . However, the learning experiments showed that for optimal results of pre-release stimulation it may be necessary to provide the wasps with a blend that mimics the natural odor cues very closely.

PAGE 168

155 The demonstrated active involvement of the plants in the attraction of natural enemies may also be exploited for control purposes. In the past it has been suggested that the breeding of crop cultivars that emit relatively high amounts of synomones may enhance the effectiveness of entomophagous insects (Nordlund et aL., 1981b, 1988). Now that the importance of herbivoreinduced plant responses in the attraction of natural enemies of pest insects has been demonstrated, inducible plant responses should also be considered for cultivar breeding (see also Dicke et iLl, 1990c). Although, investigation of the revealed multi -trophic interactions is in its infancy, it seems likely that more intriguing aspects will be uncovered, which may result in several future applications.

PAGE 169

REFERENCES CITED Allen, W. W., 1958. The biology of Apanteles medicaginis Muesebeck (Hymenoptera: Braconidae) . Hilgardia 27: 515-541. Alphen, J. J. M. van, J. C. van Lenteren, H. W. Nell and H. Eebes, 1984. The response of a polyphagous parasitoid ( Leptopil ina heterotoma (Thomson)) to a kairomone produced by one of its hosts ( Drosophila melanoqaster Meigen). Neth. J. Zool . 34: 215-219. Alphen, J, J. M. van, and L. E. M. Vet, 1987. An evolutionary approach to host finding and selection, in: Insect Parasitoids (Waage, J.K. and Greathead D.J., eds.). Academic Press, London, pp. 23-61. Altieri, M. A., W. J. Lewis, D. A. Nordlund, R. C. Gueldner and J. W. Todd, 1981. Chemical interactions between plants and Trichogramma wasps in Georgia soybean fields. Prot. Ecol . 3: 259-263. Altieri, M. A. and W. H. Whitcomb, 1980. Weed manipulation for insect pest management in corn. Env. Manag. 4: 483-489. Alvarado-Rodriguez, B., 1987. Parasites and disease associated with larvae of beet armyworm Spodoptera exigua (Lepidoptera: Noctuidae), infesting processing tomatoes in Sinaloa, Mexico. Fla. Entomol . 70: 444-449. Andrews, K. L., 1988. Latin american research on Spodoptera fruqiperda (Lepidoptera: Noctuidae). Fla. Entomol. 71: 630-653. Arthur, A. P., 1962. Influence of the host tree on abundance of Itoplectis conquisitor (Say), a polyphagous parasite of the European pine shoot moth Rhvacionia buol iana (Schiff). Can. Entomol. 94: 337-347. Arthur, A. P., 1966. Associative learning in Itoplectis conouisitor (Say) (Hymenoptera: Ichneumonidae) . Can. Entomol. 98: 213-223. Arthur, A. P., 1971. Associative learning by Nemeritis canescens (Hymenoptera: Ichneumonidae). Can. Entomol. 103: 1137-1141. Ashley, T. R., 1976. Computer program for analyzing parasitoid-host or predator-prey relationships. Fla. Entomol. 59: 27-32. Ashley, T. R., 1979. Classification and distribution of fall armyworm parasites. Fla. Entomol. 62: 114-123. 156 ^'

PAGE 170

157 Ashley, T. R., 1983. Growth pattern alterations in fall armyworm, Spodoptera fruqiperda . larvae after parasitization by Apanteles marqiniventris . Campoletis qrioti , Chelonus insularis , and Eiphosoma vitticole . Fla. Entomol . 66: 260-266. Ashley, T. R,, 1986. Geographical distribution and parasitization levels for parasitoids of the fall armyworm, Spodoptera fruqiperda . Fla. Entomol. 69: 516-524. Ashley, T. R., C. S. Barfield, V. H. Waddill and E. R. Mitchell, 1983. Parasitization of fall armyworm larvae on volunteer corn, Bermuda grass, and Paragrass. Fla. Entomol. 66: 267-271. Ashley, T. R., E. R. Mitchell, N. C. Leppla and E. E. Grissell, 1980. Parasites attacking fall armyworm larvae, Spodoptera fruqiperda . in late planted field corn. Fla. Entomol. 63: 136-142. Ashley, T. R., V. H. Waddill, E. R. Mitchell and J. Rye, 1982. Impact of native parasites on the fall armyworm, Spodoptera fruqiperda (Lepidoptera: Noctuidae), in South Florida and release of the exotic parasite, Eiphosoma vitticole (Hymenoptera: Ichnuemonidae) . Environ. Entomol. 11: 833-837. Auger, J., C. Lecomte, J. Paris and E. Thibout, 1989. Identification of leek-moth and diamondback moth frass volatiles that stimulate parasitoid, Diadromus pulchellus . J. Chem. Ecol . IS: I39I-1398. Baehrecke, E. H., S. B. Vinson and H. J. Williams, 1990. Foraging behavior of Campoletis sonorensis in response to Hel iothis virescens and cotton plants. Entomol. Exp. Appl . (in press). Barry, R. M., 1970. Insect parasites of the green cloverworm in Missouri. J. Econ. Entomol. 63: 1963-1965. Bateson, P. P. G., 1984. Genes, evelution, and learning. In: The Biology of Learning (Marler, P. and Terrace, H. S., eds.). SpringerVerlag, Berlin, pp. 325-339. Beglyarov, G. A., A. I. Smetnik, 1976. Seasonal colonization of entomophages in the USSR. In: Biological Control by Augmentation of Natural Enemies (Ridgeway, R. L and Vinson, S.B., eds.). pp. 283-328. Bestmann, H.-J., B. Classen, U. Kobold, 0. Vostrowsky, F. Klingauf, and U. Stein, 1988. Steam volatile constituents from leaves of Rhus tvohina . Phytochemistry 27: 85-90. Beth, A., 1932. Vernachlassigte Hormone. Naturwissenschaften 20: 177-183. Bianchi, F. A., 1944. The recent introduction of armyworm parasites from Texas. Hawaii. Plant. Rec. 48: 203-212.

PAGE 171

158 Blum, M. S., 1974. Deciphering the communicative rosetta stone. Bull. Entomol . Soc. Am. 20: 30-35. Boling, J. C. and H. N. Pitre, 1970. Life history of Apanteles marqiniventris with description of immature stages. J. Kans. Entomol. Soc. 43: 465-470. Boling, J. C. and H. N. Pitre, 1971. Hostal plant preference for oviposition by Trichoplusia ni* and efficiency of Apanteles marqiniventris as a populational regulator of L ni in field-cage tests. J. Econ. Entomol. 64: 411-412. Bosch, R. van den, and P. S. Messenger, 1973. Biological Control. Intext Press, New York. Bowers, W. S., L. R. Nault, R. E. Webb and S. R. Dutky, 1972. Aphid alarm pheromone: Isolation, identification, synthesis. Science 177: 1121-1122. Bowless, D., 1990. Signals in wounded plants. Nature 343: 314-315. Brattsten, L. B., 1988. Enzymic adaptations in leaf-feeding insects to host-plant allelochemicals. J. Chem. Ecol . 14: 1919-1939. Brown, W. L. Jr., T. Eisner and R. H. Whittaker, 1970. Allomones and kairomones: Transpecific chemical messengers. Bioscience 20: 21-22. Brownlee, R. G., and R. M. Silverstein, 1968. A micro-preparative gas chromatograph and a modified carbon skeleton determinator. Anal. Chem. 40: 2077-2079. Burleigh, J. G., 1975. Comparison of Heliothis spp. larval parasitism and Spica ria infection in closed and open canopy cotton varieties. Environ. Entomol. 4: 574-576. Burleigh, J. G. and J. H. Farmer, 1978. Dynamics of Heliothis spp. larval parasitism in Southeast Arkansas. Environ. Entomol. 7: 692-694. Butler, G. D. Jr., 1958. Braconid wasps reared from Lepidopterous larvae in Arizona, 1957. Pan-Pacific Entomol. 34: 221-223. Buttery, R. G. and L. C. Ling, 1984. Corn leaf volatiles: Identification using tenax trapping for possible insect attractants. J. Agric. Food Chem. 32: 1104-1106. Buttery, R. G., R. Teranishi and L. C. Ling, 1987. Fresh tomato aroma volatiles: A quantitative study. J. Agric. Food Chem. 35: 540-544. Capinera, J. L., and W. J. Roltsch, 1980. Response of wheat seedlings to actual and simulated migratory grasshopper defoliation. J. Econ. Entomol. 73: 258-261.

PAGE 172

159 Carton, Y., 1971. Biologie de Pimpla instigator F. (Ichneumonidae: Pimp] ionae) . I. Mode de perception de I'hote. Entomophaga 16: 285-296. Carton, Y., 1974. Biologie de Pimpla instigator (Ichneumonidae: Pimpl ionae). III. Analyse experimental e du processus de reconnaissance de rhote-chrysal ide. Entomol . Exp. Appl . 17: 265-278. Chambers, D. L., J. L. Sharp and T. R. Ashley, 1976. Tethered insect flight: A system for automated data processing of behavioral events. Behav. Res. Methods & Instrum. 8: 352-356. Chernin, E., 1970, Interspecific chemical signals. Bioscience 20: 845. Chew, v., 1986. Multiple comparison procedures: Uses, abuses and alternatives. Paper presented at the 1986 Workshop of the Genetics and Breeding of Southern Forest Trees RIEG. Clausen, C. P., 1940. Entomophagous Insects. McGraw-Hill Book Co., Inc., New York, N. Y. 688 pp. Clausen, C. P. (ed.), 1978. Introduced Parasites and Predators of Arthropod Pests and Weeds: A World Review. USDA, Washington. 545 pp. Coley, P. D., J. P. Bryant, and F. S. Chapin III, 1985. Resource availability and plant antiherbivore defense. Science 230: 895-899. Corbet, S. A., 1985. Insect chemosensory response: A chemical legacy hypothesis. Ecol . Entomol. 10: 143-153. Corey, E. J., D. E. Cane, and L. Libit, 1971. The synthesis of racemic o-transand p trans -bergamotene. J. Amer. Chem. Soc. 93: 7016-7021. Craighead, F. C, 1921. Hopkins host-selection principle as related to certain cerambycis beetles. J. Agric. Res. 64: 189-220. Danks, H. V., R. L. Rabb and P. S. Southern, 1979. Biology of insect parasites of Heliothis larvae in North Carolina. J. Georgia Entomol. Soc. 14: 31-37. Dethier, V. G., 1947. Chemical Insect Attractants and Repellents. Blakiston, Philadelphia. Dethier, V. G., L. B. Browne and C. N. Smith, 1960. The designation of chemicals in terms of the response they elicit from insects. J. Econ. Entomol. 53: 134-136. Dicke, M., 1988a. Microbial allelochemicals affecting the behavior of insects, mites, nematodes and protozoa in different trophic levels. In: Novel Aspect of Insect-Plant Interactions (Barbosa, P. and Letourneau, D., eds.). John Wiley and Sons, New York. pp. 125-163.

PAGE 173

160 Dicke, M., 1988b. Infochemicals in tritrophic interactions. Origin and function in a system consisting of predatory mites, phytophagous mites and their host plants. Ph. D. dissertation. Agricultural University Wageningen, The Netherlands. Dicke, M., T. A. van Beek, M. A. Posthumus, N. Ben Dom, H. van Bokhoven, and ^ de Groot, 1990a. Isolation and identification of volatile kairomone that affects acarine predator-prey interactions. Involvement of host plant in its production. J. Chem. Ecol . 16: 381-396. Dicke, M., J. C. van Lenteren, G. J. F. Boskamp and E. van Dongenvan Leeuwen, 1984. Chemical stimuli in host-habitat location by Leptooil ina heterotoma (Thomson) (Hymenoptera: Eucoilidae), a parasite of Drosophila . J. Chem. Ecol. 10: 695-712. Dicke, M., K. J. van der Maas, J. Takabayashi, and L. E. M. Vet, 1990b. Learning affects response to volatile allelochemicals by predatory mites. In: Proc. 1st Annual Meeting of Section for Applied Entomology of the Dutch Entomological Society (Sommemijer, M. J. and van der Blom, J., eds.) NEV, Amsterdam, pp. 31-36. Dicke, M. and M. W. Sabelis, 1988a. Infochemical terminology: Should it be based on cost-benefit analysis rather than origin of compounds? Funct. Ecol. 2: 131-139. Dicke, M. and M. W. Sabelis, 1988b. How plants obtain predatory mites as bodyguards. Neth. J. Zool . 38: 148-165. Dicke, M. and M. W. Sabelis, 1989. Does it pay plants to advertize for bodyguards? Towards a cost-benefit analysis of induced synomone production. In: Causes and Consequences of Variation in Growth Rate and Productivity of Higher Plants (Lambers H., ed.). Academic Publishing bv.. The Hague, pp. 341-358. Dicke, M., M. W. Sabelis, J. Takabayashi, J. Bruin, and M. A. Posthumus, 1990c. Plant strategies of manipulating predator-prey interactions through allelochemicals: Prospects for application in pest control. J. Chem. Ecol . (in press) . Ding, D., P. D. Swedenborg and R. L. Jones, 1989. Plant odor preferences and learning in Macrocentrus grandii (Hymenoptera: Braconidae), a larval parasitoid of the European corn borer, Ostrinia ""bilalis (Lepidoptera: Pyralidae). J. Kansas Entomol . Soc. 62: 164Dmoch, J., W. J. Lewis, P. B. Martin and D. A. Nordlund, 1985. Role of the host-produced stimuli and learning in host selection behavior of Cotesia (=Apantelesl maroiniventris (Cresson). J. Chem. Ecol. 11: 453-463.

PAGE 174

> 161 Doskotch, R. W., H. -Y. Cheng, T. M. Odell and L. Girard, 1980. Nerolidol: An antifeeding sesquiterpene alcohol for gypsy moth larvae from Melaleuca leucadendron . J. Chem. Ecol . 6: 845-851. Doutt, R. L., 1959. The biology of parasitic Hymenoptera. Annu. Rev. Entomol. 4: 161-182. Doutt, R. L. and R. F. Smith, 1971. The pesticide syndrome-diagnosis and suggested prophylaxis. In: Biological Control. (Huffaker, C. B., ed.). Plenum, New York. pp. 3-15. Drake, C. J. 1920. The southern green stink-bug, Nezara viridula . Q. Bull. St. Plant Board, Florida. 4: 1-75. Drost, Y. C, W. J. Lewis and J. H. Tumlinson, 1988. Beneficial arthropod behavior mediated by airborne semiochemicals. V. Influence of rearing method, host-plant, and adult experience on host-searching behavior of Micropi itis croceipes (Cresson), a larval parasitoid of Heliothis . J. Chem. Ecol. 14: 1607-1616. Drost, Y. C, W. J. Lewis, P. 0. Zanen and M. A. Keller, 1986. Beneficial -insect behavior mediated by semiochemicals. I. Flight behavior and influence of pre-flight handling of Micropi itis croceipes (Cresson). J. Chem. Ecol. 12: 1247-1262. Edwards, L. J., J. B. Siddall, L. L. Dunham, P. Uden and C. J. Kislow, 1973. Trans-p-farnesene alarm pheromone of the green peach aphid, Mvzus persicae Sulzer. Nature 241: 126-127. Ehler, L. E. and J. C. Miller, 1978. Biological control in temporary agroecosystems. Entomophaga 23: 207-212. Eller, F. J., 1990. Foraging behavior of Micropi itis croceipes . a parasitoid of He1 iothis species. Ph. D. dissertation. University of Florida. 221 pp. Filer, F. J., J. H. Tumlinson and W. J. Lewis, 1988a. Beneficial arthropod behavior mediated by airborne semiochemicals. II. Olfactometric studies of the host-location by the parasitoid Microolitis croceipes (Cresson) (Hymenoptera: Braconidae). J. Chem. Ecol. 14: 425434. Filer, F. J., J. H. Tumlinson and W. J. Lewis, 1988b. Beneficial arthropod behavior mediated by airborne semiochemicals: Source of volatiles mediating the host-location flight behavior of Microolitis croceipes (Cresson) (Hymenoptera: Braconidae), a parasitoid of Heliothis zea (Boddie) (Lepidoptera: Noctuidae). Environ. Entomol. 17: 745-753. Elzen, G, W., H. J. Williams, and S. B. Vinson, 1983. Response by the parasitoid Campol itis sonorensis (Hymenoptera: Ichneumonidae) to chemicals (synomones) in plants: Implications for host habitat location. Environ. Entomol. 12: 1872-1876.

PAGE 175

162 Elzen, G. W., H. J, Williams and S. B. Vinson, 1984. Role of diet in host selection of He! iothis virescens by parasitoid Campoletis sonorensis (Hymenoptera: Ichneumonidae) . J. Chem. Ecol . 10: 1535-1541. Elzen, G. W., H. J. Williams and S. B. Vinson, 1986. Wind tunnel flight responses by hymenopterous parasitoid Campoletis sonorensis to cotton cultivars and lines. Entomol . Exp. Appl . 43: 285-289. Elzen, G. W., H. J. Williams, S. B. Vinson and J. E. Powell, 1987. Comparative flight behavior of parasitoids Campoletis sonorensis and Microplitis croceipes . Entomol. Exp. Appl. 45: 175-180. Farkas, S. R. and H. H. Shorey, 1972. Chemical trail -foil owing by flying insects: A mechanism for orientation to a distant odor source. Science 178: 67-68. Ferkovich, S. M., P. D. Greany and C. Dillard, 1983. Changes in hemolymph proteins of the fall armyworm, Spodoptera fruqiperda (J. E. Smith), associated with parasitism by the braconid parasitoid Cotesia marqiniventris (Cresson). J. Insect Physiol. 29: 933-942. Fraenkel, G. S. and D. L. Gunn, 1940. The Orientation of Animals. Clarendon Press, Oxford. Frankel , G., 1959. The raison d'etre of secondary plant substances. Science 129: 1466-1470. Gardner, T. R., 1938. Influences of feeding habits of Tiphia vernal is on the parasitization of the Japanese beetle. J. Econ. Entomol. 31: 204-207. Graham, H. M., N. S. Hernandez, Jr. and J. R. Llanes, 1972. The role of host plants in the dynamics of populations of Hel iothis spp. Environ. Entomol. 1: 4243-4431. Greany, P., 1986. In vitro culture of hymenopterous larval endoparasitoids. J. Insect Physiol. 32: 409-419. Greany, P. D., S. M. Ferkovich and W. R. Clark, 1989. Process towards development of an artificial diet and an in vitro rearing system for Microplitis croceipes . Southwest. Entomol. Suppl . 12: 89-94. Greany, P. D., J. H. Tumlinson, D. L. Chambers and G. M. Boush, 1977. Chemically mediated host finding by Bioster (Opius) lonqicaudatus . a parasitoid of tephritid fruit fly larvae. J. Chem. Ecol. 3: 189-195. Greany, P. D., S. B. Vinson and W. J. Lewis, 1984. Insect parasitoids: Finding new opportunities for biological control. Bioscience 11: 690-696.

PAGE 176

163 Green, T. R., and C. A. Ryan, 1972. Woundinduced proteinase inhibitor in plant leaves: a possible defense mechanism against insects. 175: 776-777. Grob, K. Jr., 1982. Partial solvent trapping in capillary gas chromatography. Description of a solvent effect. J. Cromatogr. 251: 235-248. Gross, H. R., 1981. Employment of kairomones in the management of parasitoids. In: Semiochemicals, Their Role in Pest Control (Nordlund, D. A., Jones, R. L., and Lewis, W. J., eds.). Wiley, New York. pp. 137150. Gross, H. R., Jr., W. J. Lewis, M. Beevers and D. A, Nordlund, 1984. Trichoqramma pretiosum (Hymenoptera: Trichogrammatidae) : Effects of augmented densities and distributions of Hel iothis zea (Lepidoptera: Noctuidae) host eggs and kairomones on field performance. Environ. Entomol. 13: 981-985. Gross, H. R., Jr., W. J. Lewis, R. L. Jones and D. A. Nordlund, 1975, Kairomones and their use for management of entomophagous insects: III. Stimulation of Trichoqramma achaeae, L pretiosum . and Micropl itis croceipes with host seeking stimuli at time of release to improve their efficiency. J. Chem. Ecol . 1: 431-438. Gross, H. R. and S. D. Pair, 1986. The fall armyworm: Status and expectations of biological control with parasitoids and predators. Fla. Entomol. 69: 502-515. Gunasena, G. H., S. B. Vinson, H. J. Williams, and R. D. Stipanovic, 1988. Effects of caryophyllene, caryophyllene oxide, and their interaction with gossypol on the growth and development of Hel iothis virescens (F.) (Lepidoptera: Noctuidae). J. Econ. Entomol. 81: 93-97. Hagen, K. S., 1986. Ecosystem analysis: Plant cultivars (HPR), entomophagous species and food supplements. In: Interactions of Plant Resistance and Parasitoids and Predators of Insects (D. J. Boethel and R. D. Eikenbary, eds.). Ellis Norwood, Chichester, pp. 151-197. Hagen, K. S., E. F. Sawall, R. L. Tassan, 1971. The use of food sprays to increase the effectiveness of entomophagous insects. Tall Timb. Conf. Ecol. Anim. Contr. Habitat Manag. 2: 59-81. Hamm, J. J., E. L. Styer, and W. J. Lewis, 1990. Comparative virogenesis of filamentous virus and polydnavirus in the female reproductive tract of Cotesia margini ventris (Hymenoptera: Braconidae). J. Invert. Path. 55: 357-374. Harding, J. A., 1976a. Heliothis spp.: Parasitism and parasites plus host plants and parasites of the beet armyworm, diamondback moth and two tortricids in the Lower Rio Grande Valley of Texas. Environ Entomol. 5: 669-671.

PAGE 177

164 Harding, J. A., 1976b. Seasonal occurrence, hosts, parasitism and parasites of cabbage and soybean loopers in the lower Rio Grande Valley. Environ. Entomol . 5: 672-674. Heath, R. R., P. L. Landolt, J. H. Tumlinson, and C. 0. Calkins, 1990. A system to collect volatile chemicals emitted by insects. J. Chem. Ecol . (in press) . Herard, F., M. A. Keller, W. J. Lewis and J. H. Tumlinson, 1988a. Beneficial arthropod behavior mediated by airborne semiochemicals. III. Influence of age and experience on flight chamber responses of Microplitis demolitor Wilkinson. J. Chem. Ecol. 14: 1583-1596. Herard, F., M. A. Keller, W. J. Lewis and J. H. Tumlinson, 1988b. Beneficial arthropod behavior mediated by airborne semiochemicals. IV. Influence of host-diet on host-oriented flight chamber responses of Microplitis demolitor Wilkinson. J. Chem. Ecol. 14: 1597-1606. Herzog, D. C. and J. E. Funderbunk, 1985. Plant resistance and cultural practice interactions with biological control. Jn: Biological control in agricultural IPM systems (Hoy, M. A., and Herzog, D. C, eds.). Academic Press, New York. pp. 67-88. Hogg, D. B., R. E. Anderson and H. N. Pitre, 1982. Early-season parasitization of fall armyworm (Lepidoptera: Noctuidae) larvae in Mississippi. Fla. Entomol. 65: 584-585. Hopkins, A. D., 1917. [contribution to discussion]. J. Econ. Entomol. 10: 92-93. Hotchkin, P. G. and H. K. Kaya, 1983. Interactions between two baculoviruses and several insect parasites. Can. Entomol. 115: 841-846. Jackson, D. M., F. C. Tingle and E. R. Mitchell, 1984. Survey of Heliothis spp. larvae found on Florida beggarweed and post-harvest tobacco in Florida. Fla. Entomol. 67: 130-141. Jaenike, J., 1982. Environmental modification of oviposition behavior in Drosophila . Am. Nat. 119: 784-802. Jaenike, J., 1983. Induction of host preference in Drosophila melanoqaster . Oecologia (Berlin) 58: 320-325. Jalali, S. K., S. P. Singh and C. R. Ballal, 1987 . Studies on host age preference and biology of exotic parasite, Cotesia marginiventris (Cresson) (Hymenoptera: Braconidae). Entomon. 12: 59-62. Jalali, S. K., S. P. Singh and C. R. Ballal, 1988. Effect of parasitism by Cotesia marginiventris on consumption and utilization of artificial diet by larvae of Spodoptera litura (Lepidoptera: Noctuidae). Indian J. Agric. Sci. 58: 529-532.

PAGE 178

165 Jermy, T., F. E. Hanson and V. G. Dethier, 1968. Induction of specific food preference in lepidopterous larvae. Entomol . Exp. Appl . 11: 211-230. Jones, R. L., W. J. Lewis, M. C. Bowman, M. Beroza and B. A. Bierl, 1971. Host-seeking stimulant for parasite of corn earworm: Isolation, identification, and synthesis. Science 173: 842-43. Jones, T. G. H. and J. M. Harvey, 1936. Essential oils from the Queensland flora. Part VIII. The identity of melaleucol with nerolidol. Proc. R. Soc. Queensland 47: 92-93. Kainoh, Y., 1987. Behavior regulators of entomophagous insects. In: A Manual of Bio-Pesticides and Growth Regulators (M. Okada, ed.). L.l.C, Tokyo, pp. 254-277. Kaiser, R,, 1987. Night-scented flowers, not only attractive to moths. Abstracts EUCHEM conference 'Semiochemicals in the Plant and Animal Kingdoms'. October 12-16, 1987. Angers, France. Kamm, J. A. and R. G. Buttery, 1983. Response of the alfalfa seed chalcid, Bruchoohaaus roddi . to alfalfa volatiles. Entomol. Exp. Appl. 23: 129-134. Karlson, P. and A. Butenandt, 1959. Pheromones (Ectohormones) in insects. Annu. Rev. Entomol. 4: 39-58. Kennedy, B. H., 1979. The effect of multilure on parasites of the European elm bark beetle, Scolvtus multistriatus . Bull. Entomol. Soc. Am. 25: 116-118. Kennedy, J. S., 1947. The excitant and repellent effects on mosquitoes of sublethal contacts with DDT. Bull. Ent. Res. 37: 593-607. Kennedy, J. S., 1977. Olfactory response to distant and other odor sources. In: Chemical Control of Insect Behavior (Shorey, H. H. and McKelvey, J. J., eds.). Wiley Interscience, New York. pp. 67-91. King, E. G. and R. J. Coleman, 1989. Potential for biological control of Heliothis species. Annu. Rev. Entomol. 34: 53-75. King, E. G. and N. C. Leppla (eds.), 1981. Advances in Insect Rearing. USDA Tech. Bull. King, E. G. and N. C. Leppla, 1984. Advances and Challenges in Insect Rearing. Agric. Res. Serv., USDA, U. S. Government Printing Office, Washington, D.C. King, E. G., J. E. Powell and R. J. Coleman, 1985. A high incidence of parasitism of Heliothis spp. (Lep.: Noctuidae) larvae in cotton in Southeastern Arkansas. Entomophaga 30: 419-426.

PAGE 179

166 Knipling, E. F., 1980. Regional management of the fall armyworm-A realistic approach? Fla. Entomol . 63: 468-480. Knipling, E. F. and J. V. McGuire, 1968. Population models to appraise the limitation and potentialities of Trichogramma in managing host insect populations. U.S. Dept. Agric. Tech. Bull. No. 1387 Krishnamoorthy, A. and M. Mani, 1985. Investigations on Telenomus remus Nixon and Apanteles marqiniventris Cresson against Spodoptera Litura (Fabricius) on cabbage. Entomon. 10: 277-280. Kogan, M., and J, D. Paxton, 1983. Natural inducers of plant resistance to insects. In: Plant Resistance to Insects (P. A. Hedin, ed.). Amer. Chem. Soc, Washington, D. C. pp. 153-171. Krombein, K. V., P. D. Hurd, Jr., D. R. Smith and B. D. Burks, 1979. Catalog of Hymenoptera in America north of Mexico. Vol. 1. Smithsonian Institute Press, Washington, D. C, 1198 pp. Kunnalaca, S. and A. J. Mueller, 1979. A laboratory study of Apanteles marqiniventris. a parasite of green cloverworm. Environ. Entomol. 8: 365-368. Laing, J., 1937. Host-finding by insect parasites. I. Observation on the finding of hosts by Alvsia manducator, Mormoniella vitrioennis and Trichoqramma evanescens . J. Anim. Ecol . 6: 298-317. Latheef, M. A. and R. D. Irwin, 1983. Seasonal abundance and parasitism of lepidopterous larvae on Brassica greens in Virginia, J. Georgia Entomol. Soc. 18: 1154-1168. Law, J. H. and F. E. Regnier, 1971. Pheromones. Annu. Rev. Biochem. 40: 533-548. Leeper, J. R., 1974. Adult feeding behavior of Lixophaqa spenophori a tachinid parasite of the New Guinea sugarcane weevil. Proc. Hawaii. Entomol. Soc. 21: 403-412. Lentz, G. L. and L. P. Pedigo, 1975. Population ecology of parasites of the green cloverworm in Iowa. J. Econ. Entomol. 68: 301-304. Lewis, W. J., 1981. Semiochemicals: Their role with changing approaches to pest control. In: Semiochemicals: Their Role in Pest Control. (Nordlund, D. A., Jones, R. L. and Lewis, W. J., eds.). John Wiley, New York. pp. 3-12. Lewis, W. J., M. Beevers, D. A. Nordlund, H. R. Gross and K. S. Hagen, 1979. Kairomones and their use for management of entomophagous insects. IX. Investigations of various kairomone-treatment patterns for Trichoqramma spp. J. Chem. Ecol. 5: 673-680.

PAGE 180

167 Lewis, W. J. and J. R. Brazzel , 1968. Incidence of parasitic insects on the bollworm in Mississippi. Miss. Agr. Exp. Stn. Bull. 727, 7 pp. Lewis, W. J. and R. L. Burton, 1970. Rearing Microp! itis in the laboratory with Hel iothis Zea as hosts. J. Econ. Entomol . 63: 656-658. Lewis, W. J., R. L. Jones, H. R. Gross, Jr. and D. A. Nordlund, 1976. The role of kairomones and other behavioral chemicals in host finding by parasitic insects. Behav. Biol. 16: 267-289. Lewis, W. J., R. L. Jones, D. A. Nordlund and H. R. Gross Jr., 1975b. Kairomones and their use for management of entomophagous insects: II. Mechanisms causing increase in rate of parasitization by Trichogramma spp. J. Chem. Ecol . 5: 673-680. Lewis, W. J., R. L. Jones, D. A. Nordlund and A. N. Sparks, 1975a. Kairomones and their use for management of entomophagous insects. I. Evaluation for increasing rates of parasitization by Trichogramma spp. in the field. J. Chem. Ecol. 1: 343-347. Lewis, W. J., R. L. Jones and A. N. Sparks, 1972. A host-seeking stimulant for the egg parasite Trichogramma eyanescens: Its source and a demonstration of its laboratory and field activity. Ann. Ent. Soc. Am. 65: 1087-1089. Lewis, W. J. and D. A. Nordlund, 1980. Employment of parasitoids and predators for fall armyworm control. Fla. Entomol. 63: 433-438. Lewis, W. J. and D. A. Nordlund, 1984. Semiochemical s influencing fall armyworm parasitoid behavior: Implications for behavioral manipulation. Fla. Entomol. 67: 343-349. Lewis, W. J. and D. A. Nordlund, 1985. Behavior-modifying chemicals to enhance natural enemy effectiveness. In: Biological Control in Agricultural IPM Systems (Hoy, M.A. and Herzog, D.C., eds.). Academic Press, New York. pp. 89-101. Lewis, W. J., D. A. Nordlund and R. C. Gueldner, 1981. Semiochemical s influencing behavior of entomophages : Roles and strategies for their employment in pest control. Les Mediateurs chimiques. INRA Symp. 7: 225-242. Lewis, W. J., D. A. Nordlund, R. C. Gueldner, P. E. A. Teal and J. H. Tumlinson, 1982. Kairomones and their use for management of entomophagous insects. XIII. Kairomonal activity for Trichogramma spp. of abdominal tips, excretion, and a synthetic sex pheromone blend of Heliothis zea (Boddie) Moths. J. Chem. Ecol. 10: 1323-1331. Lewis, W. J. and J. H. Tumlinson, 1988. Host detection by chemically mediated associative learning in a parasitic wasp. Nature 331: 257-259.

PAGE 181

168 Loke, W. H., 1983. The host-finding behavior of Apanteles marqiniventris (Cresson) in response to kairomones produced by its host, the fall armyworm, Spodoptera fruqiperda (J.E. Smith). Ph. D. dissertation. University of Florida, Gainesville. 115 pp. Loke, W. H. and T. R. Ashley, 1984a. Behavioral and biological responses of Cotesia marqiniventris to kairomones of the fall armyworm, Spodoptera fruqiperda . J. Chem. Ecol . 10: 521-529. Loke, W. H. and T. R. Ashley, 1984b. Sources of fall armyworm, Spodoptera fruqiperda (Lepidoptera: Noctuidae), kairomones eliciting host-finding behavior in Cotesia ( =Apante1es ) marqiniventris (Hymenoptera: Braconidae). J. Chem. Ecol. 10: 1019-1027. Loke, W. H. and T. R. Ashley, 1984c. Potential uses of kairomones for behavioral manipulation of Cotesia marqiniventris (Cresson). J. Chem. Ecol. 10: 1377-1384. Loke, W. H., T. R. Ashley and R. I. Sailer, 1983. Influence of fall armyworm. Spodoptera fruqiperda (Lepidoptera: Noctuidae) larvae and corn plant damage on host finding in Apantheles marqini ventri s (Hymenoptera: Braconidae). Environ. Entomol . 12: 911-915. Mabry, T. J., and J. E. Gill, 1979. Sesquiterpene lactones and other terpenoids. In: Herbivores: Their Interactions with Secondary Plant Metabolites (Rosenthal, G. A., and Janzen, D. H., eds.). Academic Press, New York. pp. 501-537. Marsh, P. M., 1971. Keys to the nearctic genera of the families Braconidae, Aphidiidae, and Hybrizontidae (Hymenoptera). Ann. Entomol. Soc. Am. 64: 841-850. Marsh, P. M., 1978. The braconid parasites (Hymenoptera) of Heliothis species (Lepidoptera: Noctuidae). Proc. Entomol. Soc. Wash. 80: 15-36. Mason, W. R. M., 1981. The polyphyletic nature of Apanteles Foester (Hymenoptera: Braconidae): A phylogeny and reclassification of Microgastrinae. Mem. Entomol. Soc. Can. 115: 1-147. Matteson, P. C, M. A. Altieri, and W. C. Cagne, 1984. Modification of small farmer practices for better pest management. Annu. Rev. Entomol. 19: 383-402. Matthews, R. W., 1974, Biology of Braconidae. Annu. Rev. Entomol. 19: 15-32. Maurer, B., A. Hauser and J. C. Froidevaux, 1986. (E)-4,8-dimethyl1,3,7-nonatriene and (E,E)-4,8,12-trimethyl-l,3,7,ll-tridecatetraene, two unusual hydrocarbons from cardamom, oil. Tetrahedron Lett. 27: 2111-2112.

PAGE 182

McAuslane, H. J., S. B. Vinson and H. J. Williams, 1990a. Influence of host plant on mate location by the parasitoid Campoletis sonorensis (Hymenoptera: Ichneumonidae) . Environ. Entomol . 19: 26-31. McAuslane, H. J., S. B. Vinson and H. J. Williams, 1990b. Influence of adult experience on host micro-habitat location by the general ist parasitoid, Campoletis sonorensis . J. Insect Behavior (in press). McAuslane, H. J., S. B. Vinson and H. J. Williams, 1990c. Effect of host diet on flight behavior of the parasitoid Campoletis sonorensis . Entomol. Exp. Appl . (in press). McCutcheon, G. S. and W. Harrison, 1987. Host range and development of Microplitis rufiventris (Hymenoptera: Braconidae), an imported parasitoid of several lepidopterous pests. Environ. Entomol. 16: 855858. McCutcheon, G. S. and S. G. Turnipseed, 1981. Parasites of Lepidopterous larvae in insect resistant and susceptible soybeans in South Carolina. Environ. Entomol. 10: 69-74. McFadden, T. L. and C. S. Creighton, 1979. Annotated checklist of parasites associated with vegetable insects. J. Georgia Entomol. Soc. 14: 282-285. Mihaliak, C. A., D. Couvet, and D. E. Lincoln, 1987. Inhibition of feeding by a general ist insect due to increased volatile leaf terpenes under nitrate-limiting conditions. J. Chem. Ecol . 13: 2059-2067. Miller, J. C, 1977. Ecological relationships among parasites of Spodoptera praefica . Environ. Entomol. 6: 581-585. Miller, J. C. and L. E. Ehler, 1978. Parasitization of Spodoptera praefica larvae in hay alfalfa. Environ. Entomol. 7: 744-747. Minyard, J. P., J. H. Tumlinson, A. C. Thompson, and P. A. Hedin, 1966. Constituents of the cotton bud. Sesquiterpene hydrocarbons. J. Agr. Food Chem. 14: 332-336. Mitchell, E. R., V. H. Waddill and T. R. Ashley, 1984. Population dynamics of the fall armyworm (Lepidoptera: Noctuidae) and its larval parasites on whole stage corn in pheromonepermeated field environments. Environ. Entomol. 13: 1618-1623. Monteith, L. G., 1963. Habituation and associative learning in Drino bohemica Men. (Diptera: Tachinidae). Can. Ent. 95: 418-426. Mueller, A. J. and S. Kunnalaca, 1979. Parasites of green cloverworm on soybean in Arkansas. Environ. Entomol. 8: 376-379. Mueller, T. F. and J. R. Phillips, 1983. Population dynamics of Heliothis spp. in spring weed hosts in Southeastern Arkansas:

PAGE 183

170 Survivorship and stage-specific parasitism. Environ. Entomol . 12: 18461850. Muesebeck, C. F. W., 1921. A revision of the North American species of ichneumon-flies belonging to the genus Apanteles . Proc. U. S. Nat. Mus. 58: 483-576. Muesebeck, C. F. W., K. V. Krombein and H. K. Townes, 1951. Hymenoptera of America north of Mexico. Synoptic Catalog. U. S. Dept. of Agric. Monograph No. 2 Murphy, R. E., R. R. Heath, and J. G. Dorsey, 1990. The fractionation gap: an optimized coupling of fused silica columns in open tubular gas chromatography, (in prep.) Nadel, H. and J. J. M. van Alphen, 1987. The role of host-plant odours in the attraction of a parasitoid, Epidinocarsis lopezi , to the habitat of its host, the cassava mealybug, Phenacoccus man i hot i . Entomol. Exp. Appl. 45: 181-186. Naves, Y. R., 1960. On the presence of geraniol, nerol , linalool, farnesols, and nerol idols in essential oils. C. R. Hebd. Seances Acad. Sci. C 251: 900-902. Neunzig, H. H., 1969. The biology of the tobacco budworm and the corn earworm in North Carolina with particular reference to tobacco as a host. N. C. Agr. Exp. Stn. Bull. 196, 76 pp. Nishida, T., 1956. An experimental study of the ovipositional behavior of Qpius fletcheri Silvestri, a parasite of the melon fly. Proc. Hawaii. Entomol. Soc. 16: 126-134. Nixon, G. E. J., 1965. A reclassification of the tribe Microgasterini . Bull. Brit. Mus. (Natur. Hist.) Entomol. Suppl . 2, 284 pp. Noldus, L. P. J. J., 1989. Semiochemical s and quality of beneficial arthropods: General considerations with special reference to Trichoqramma spp. J. Appl. Entomol. 108: 425-451. Noldus, L. P. J. J. and J. C. van Lenteren, 1983. Kairomonal effects on searching for eggs of Pieris brassicae . Pieris rapae and Mamestra brassicae of the parasite Trichoqramma evanescens Westwood. Med. Fac. Landbouww. Rijksuniv. Gent. 2: 183-194. Nordlund, D. A., R. 8. Chalfant, and W. J. Lewis, 1984. Arthropod populations, yields, and damage in monocultures and polycultures of corn, beans, and tomatoes. Agric. Ecosystems Environ. 1: 353-367. Nordlund, D. A., R. B. Chalfant, and W. J. Lewis, 1985. Response of Trichoqramma pretiosum females to volatile synomones from tomato plants. J. Entomol. Sci. 20: 372-376.

PAGE 184

171 Nordlund D. A. and W. J. Lewis, 1976. Terminology of chemical releasing stimuli in intraspecific and interspecific interactions. J. Chem. Ecol . 2: 211-220. Nordlund, D. A., W. J. Lewis and M. A. Altieri, 1988. Influences of plant produced allelochemicals on the host and prey selection behavior of entomophagous insects. In: Novel Aspects of Insect-Plant Interactions (Barbosa, P. and Letourneau, D. K., eds.). Wiley, New York. pp. 65-90. Nordlund, D. A., W. J. Lewis and H. R. Gross, 1981b. Elucidation and employment of semiochemicals in the manipulation of entomophagous insects. In: Management of Insect Pests with Semiochemicals (Mitchell, E. R., ed.). Plenum, New York. pp. 463-475. Nordlund, D. A., W. J. Lewis, and R. C. Gueldner, 1983. Kairomones and their use for management of entomophagous insects. XIV. Response of Telonomus remus to abdominal tips of Spodoptera fruqiperda . (Z)-9tetradecene-l-ol acetate and (Z)-9-dodecene-l-ol acetate. J. Chem. Ecol. 6: 695-701. Nordlund, D. A., R. L. Jones and W. J. Lewis (eds.), 1981a. Semiochemicals, Their Role in Pest Control. Plenum, New York. 306 pp. Nordlund, D. A. and C. E. Sauls, 1981. Kairomones and their use for management of entomophagous insects. XI. Effect of host plants on kairomonal activity of frass from Hel iothis zea larvae for the parasitoid Microolitis croceioes . J. Chem. Ecol. 6: 1057-1061. Pair, S. D., M. L. Laster and D. F. Martin, 1982. Parasitoids of Heliothis spp. (Lepidoptera: Noctuidae) larvae in Mississippi associated with sesame interplantings in cotton, 1971-1974: Implications of hosthabitat interaction. Environ. Entomol . 11: 509-512. Pair, S. D., B. R. Wiseman and A. N. Sparks, 1986. Influence of four corn Cultivars on fall armyworm (Lepidoptera: Noctuidae) establishment and parasitization. Fla. Entomol. 69: 566-570. Parker, F. D., F. R. Lawson, and R. E. Pinnell, 1971. Suppression of Pieris raoea utilizing a new control system: Mass release of both the pest and its parasites. J. Econ. Entomol. 64: 721-35. Pemberton, C. E., 1948. The control of the grass armyworm, Laphygma exempta (Walker), in Hawaii by parasites. Hawaii Planters' Rec. 52: 181-200. Picker, K., E. Ritchie and W. C. Taylor, 1976. The chemical constituents of Australian Flindersia species. XXI. An examination of the bark and the leaves of Ll laevicaroa . Aust. J. Chem. 29: 2023-2026. Pimentel, D., 1988. Herbivore population feeding pressure on plant hosts: feedback evolution and host conservation. Oikos 53: 289-302.

PAGE 185

172 Powell, W. and Zhang Zhi-hi, 1983. The reaction of two cereal aphid parasitoids, Aphidius urbekistanicus and ervi to host aphids and their food plants. Physiol. Entomol . 8: 439-443. Price, P. W., 1981. Semiochemical s in evolutionary time. In: Semiochemicals: Their Role in Pest Control (Nordlund, D, A., Jones, R. L. and Lewis, W. J., eds.). John Wiley and Sons, New York. pp. 251-279. Price, P. W., M. Westoby, B. Rice, P. R. Atsatt, R. S. Fritz, J. N. Thompson, and K. Mobley, 1986. Parasite mediationin ecological interactions. Ann. Rev. Ecol . Syst. 17: 487-505. Price, P. W., C. E. Bonton, P. Gross, B. A. McPheron, J. N. Thompson and A. A. E. Weis, 1980. Interactions among three trophic levels: Influence of plant interactions between insect herbivores and natural enemies. Annu. Rev. Ecol. Syst. 11: 41-65. Pschorn-Walcher, H., 1977. Biological control of forest insects. Annu. Rev. Entomol. 22: 1-22. Puterka, G. J., J. E. Slosser and J. R. Price, 1985. Parasites of Heliothis spp. (Lepidoptera: Noctuidae): Parasitism and seasonal occurrence for host crops in the Texas rolling plains. Environ. Entomol. 14: 441-446. Rhoades, D. F., 1979. Evolution of plant chemical defense against herbivores. In: Herbivores: Their Interactions with Secondary Plant Metabolites (Rosenthal, G. A., and Janzen, D. H., eds.). Academic Press, New York. pp. 3-54. Rhoades, D. F., 1985. Offensive-defensive interactions between herbivores and plants: Their relevance in herbivore population dynamics and ecological theory. Am. Nat. 125: 205-238. Roach, S. H., 1975. Heliothis spp.: Larvae and associated parasites and diseases on wild host plants in the Pee Dee area of South Carolina. Environ. Entomol. 4: 725-728. Ryan, C. A., 1983. Insect-induced chemical signals regulating natural plant protection responses. In: Variable Plants and Herbivores in Natural and Managed Systems (Denno, R. F. and McClure, M. S., eds.). Academic Press, New York. pp. 43-60. Sabelis, M. W. and M. C. M. de Jong, 1988. Should all plants recruit bodyguards? Conditions for a polymorphic ESS of synomone production in plants. Oikos 53: 247-252. Salt, G., 1935. Experimental studies in insect parasitism. III. Host selection. Proc. R. Soc. Lond. Ser. B 114: 413-435.

PAGE 186

173 Sanchez-Serrano, J. J., M. Keil, A. O'Connor, J. Schell, and L. Willmitzer, 1987. Wound -induced expression of a potato proteinase inhibitor II gene in transgenic tobacco plants. EMBO J. 6: 303-306. Sand! an, K., 1980. Host location by Coccygomimus turionellae (Hymenoptera: Ichneumonidae) . Entomol . Exp. Appl . 27: 233-245. SAS Institute, Inc., 1987. SAS/STAT Guide for Personal Computers, Version 6 Edition. SAS Institute, Inc., Cary, N. C. 1028 pp. Schuster, M. P., 1980. Insect resistance in cotton. In: Biology and Breeding for Resistance to Arthropods and Pathogens in Agricultural Crops (Harris, M.K., ed.). Texas Agr. Exp. Sta. Misc. Publ . 1451 pp. 101-112. Shepard, M. and W. Sterling, 1972. Incidence of parasitism of He! iothis spp. (Lepidoptera: Noctuidae) in some cotton fields of Texas. Ann. Entomol. Soc. Am. 65: 759-760. Shorey, H. H., 1973. Behavioral responses to insect pheromones. Annu. Rev. Entomol. 18: 349-380. Shorey, H. H., 1977. Interactions of insects with their chemical environment. In: Chemical Control of Insect Behavior (Shorey, H.H. and McKelvey, J.J., eds.). Wiley Interscience, New York. pp. 1-14. Smith, H. S. and P. DeBach, 1953. Artificial infestations of plants with pest insects as an aid in biological control. Proc. 7th Pacific Sci. Congr. 4: 255-259. Smith, J. W., E. G. King and J. V. Bell, 1976. Parasites and pathogens among Heliothis species in the Central Mississippi Delta. Environ. Entomol. 5: 224-226. Snow, J. W., J. J. Hamm and J.R. Brazzel , 1966. Geranium carol ineanum as an early host of Hel iothis zea and tL. virescens (Lepidoptera: Noctuidae) in the Southeastern United States, with special notes on associated parasites. Ann. Entomol. Soc. Am. 59: 506-509. Soteres, K. M., R. C. Berberet and R. W. McNew, 1984. Parasitic insects associated with lepidopterous herbivores on alfalfa in Oklahoma. Environ. Entomol. 13: 787-793. Spradbery, J. P. 1970. Host finding by Rhvssa persuasoria (L.), an ichneumonid parasite of siricid woodwasps. Anim. Behav. 18: 103-114. Stadelbacher, E. A., J. E. Powell and E. G. King, 1984. Parasitism of Hel iothis zea and \L virescens (Lepidoptera: Noctuidae) larvae in wild and cultivated host plants in the Delta of Mississippi. Environ. Entomol. 13: 1167-1172.

PAGE 187

174 Steel R. G. and J. H. Torrie, 1960. Principles and Procedures of Statistics. McGraw-Hill, New York. 481 pp. Stenhagen, E., S. Abrahamsson, and F. W. McLafferty, 1974. Registry of Mass Spectral Data. Wiley, New York. 1257 pp. Stoltz, D. B., P. J. Krell and S. B. Vinson, 1981. Polydisperse viral DNA's in ichneumonid ovaries: A survey. Can. J. Microbiol. 27: 123-130. Stoltz, D. B. and S. B. Vinson, 1979. Viruses and parasitism in insects. Adv. Virus Res. 24: 125-171. Strand, M. R. and S. B. Vinson, 1982. Behavioral response of the parasitoid Cardiochiles niqriceps to a kairomone. Entomol . Exp. Appl . 31: 309-315. Styer, E. L., J. J. Hamm and D. A. Nordlund, 1987. A new virus associated with the parasitoid Cotesia marqiniventris (Hymenoptera: Braconidae): Replication in noctuid host larvae. J. Invert. Path. 50: 302-309. Sugimoto, T., H. Kameoka, S. Kusatani, D. Inui and K. Otsuka, 1988. Foraging for patchil ly-distributed leaf-miners by the parasitoid Dapsilarthra rufiventris (Hymenoptera: Braconidae) V. Plant odour as a cue to long range patch-location. Appl. Entomol. Zool . 23: 135-143. Szeppligeti, G. V., 1904. Hymenoptera, family Braconidae (Wytsmann, P., ed.). Genera Insectorum 22: 1-253. Tanada, Y. and J. W. Beardsley, Jr., 1958. A biological study of the lawn armyworm, Spodoptera mauritia (Boisduval), in Hawaii (Lepidoptera: Phalaenidae) . Proc. Hawaii. Entomol. Soc. 16: 411-436. Tanaka, T., 1987. Morphology and functions of calyx fluid filaments in the reproductive tracts of the endoparasitoid, Microplitis mediator (Hymenoptera: Braconidae). Entomophaga 32: 9-17. Taylor, J. S., 1932. Report on cotton insect and disease investigation. II. Notes on the american bollworm ( Heliothis obsoleta F.) on cotton and its parasite ( Microbracon brevicornis Wesm.). Sci. Bull. Rep. Agric, For. Union. S. Afr. 113. Taylor, R. J., 1974. Role of learning in insect parasitism. Ecol . Monogr. 44: 89-104. Teague, T. G., D. L. Horton, W. C. Yearian and J. R. Philips, 1985. Benomyl inhibition of Cotesia ( =Apanteles ) marqiniventris survival in four Lepidopterous hosts. J. Entomol. Sci. 20: 76-81. Thompson, A. C, P. A. Hedin, and R. C. Gueldner, 1974. Corn bud essential oil. Phytochem. 13: 2029-2032.

PAGE 188

175 Thompson, S. N., 1986. Nutrition and In vivo culture of insect parasitoids. Annu. Rev. Entomol . 31: 197-219. Thorpe, W. H. and H. B. Caudle, 1938. A study of the olfactory responses of insect parasites to the food plant of their host. Parasitology 15: 523-528. Thorpe, W. H. and F. G. W. Jones, 1937. Olfactory conditioning and its relation to the problem of host selection. Proc. Roy. Soc. London, Ser. B 124: 56-81. Tingle, F. C, T. R. Ashley and E. R. Mitchell, 1978. Parasites of Spodoptera exiqua , eridania (Lep.: Noctuidae) and Herpetogramma bipunctal is (Lep.: Pyralidae) collected from Amaranthus hvbridus in field corn. Entomophaga 23: 343-347. Tingle, F. C. and E. R. Mitchell, 1982. Effect of synthetic pheromone on parasitization of Hel iothis virescens (F.) (Lepidoptera: Noctuidae) in tobacco. Environ. Entomol. 11: 913-916. Tollsten, L., and G. Bergstrom, 1988. Headspace volatiles of whole plants and macerated plant parts of Brassica and Sinapis . Phytochem. 27: 4013-4018. Vet, L. E. M., 1983. Host-habitat location through olfactory cues by Leptopilina clavipes (Hartig) (Hym.: Eucoilidae), a parasitoid of fungivorous Drosophila : The influence of conditioning. Neth. J. Zool . 33: 225-248. Vet, L. E. M., 1984. Comparative ecology of hymenopterous parasitoids. Ph. D. dissertation. University of Leiden, The Netherlands. 218 pp. Vet, L. E. M., 1985. Response to kairomones by some alysiine and eucoilid parasitoid species (Hymenoptera) . Neth. J. Zool. 35: 486-496. Vet, L. E. M., 1988. The influence of learning on habitat location and acceptance by parasitoids. In: Proceedings of the third European Workshop on Insect Parasitoids, Les Colloques de I'lNRA. 48: 29-34. Vet, L. E. M., J. C. van Lenteren, M. Heijmans and E. Meelis, 1983. An airflow olfactometer for measuring olfactory responses of hymenopterous parasitoids and other small insects. Physiol. Entomol. 8: 97-106. Vet, L. E. M., W. J. Lewis, D. R. Papaj, and J. C. van Lenteren, 1990. A variable-response model for parasitoid foraging behaviour. J. Insect Beh. (in press). Vet, L. E. M. and K. van Opzeeland, 1984. The influence of conditioning on olfactory microhabitat and host location in Asobara tabida (Nees) and f,-.A. rufescens (Foerster) (Braconidae: Alysiinae), larval parasitoids of Drosophil idae. Oecologia 63: 171-177.

PAGE 189

176 Vet, L. E. M. and K. van Opzeeland, 1985. Olfactory microhabitat selection in Leptopilina heterotoma (Thomson) (Hym.: Eucoilidae), a parasitoid of Drosophil idae. Neth. J. Zool . 35: 497-504. Vickery, R. A., 1925. List of parasitic insects reared from host insects collected in the vicinity of Brownsville, Texas. Proc. Entomol . Soc. Wash. 27: 137-139. Vickery, R. A., 1929. Studies on the fall armyworm in the gulf coast district of Texas. USDA Tech. Bull. 138: 1-63. Viereck, H. L., 1912. Description of one new family, eight new genera, and thirty-three new species of ichneumon-flies. Proc. U. S. Nat. Mus. 44: 555-568. Vinson, S. B., 1975. Biochemical coevolution between parasitoids and their hosts. In: Evolutionary Strategies of Parasitic Insects and Mites (Price, P. W., ed.). Plenum, New York. pp. 14-48. Vinson, S. B., 1976. Host selection by insect parasitoids. Annu. Rev. Entomol. 21: 109-133. Vinson, S. B., 1981. Habitat location. In: Semiochemical s-Their Role in Pest Control (Nordlund, D. A., Jones, R. L. and Lewis, W. J., eds.). John Wiley and Sons, New York. pp. 51-77. Vinson, S. B., 1984. Parasitoid-host relationship. In: Chemical Ecology of Insects (Bell, W. J. and Carde, R. T., eds.). Sinauer Associates Inc., Massachusetts, pp. 111-124. Vinson, S. B., C. S. Barfield and R. D. Henson, 1977. Oviposition behaviour of Bracon mellitor . a parasitoid of the boll weevil ( Anthonomus grandis ). II. Associative learning. Physiol. Entomol. 2: 157-164. Vinson, S. B., G. W. Elzen and H. J. Williams, 1987. The influence of volatile plant allelochemics on the third trophic level (parasitoids) ans their herbivorous hosts. In: Insects-Plants (Labeyrie, V., Fabres, G. and Lachaise, D., eds.). Dr. W. Junk Publishers, Dordrecht, pp. 109114. Vinson, S. B. and G. F. Iwantsch, 1980. Host regulation by insect parasitoids. Q. Rev. Biol. 55: 143-165. Vinson, S. B., R. L. Jones, P. Sonnet, B. A. Beirl and M. Beroza, 1975. Isolation, identification and synthesis of host-seeking stimulants for Cardiochilis nigriceps . a parasitoid of the tobacco budworm. Entomol. Exp. Appl. 18: 443-450.

PAGE 190

Visser, J. H., S. van Straten and H. Maarse, 1979. Isolation and identification of volatiles in the foliage of potato, Solanum tuberosum a host plant of the Colorado beetle, Leptinotarsa deceml ineata . J. Chem. Ecol. 5: 13-25. Wall, R. and Berberet, R. C, 1975. Parasitoids associated with Lepidopterous pests on peanuts; Oklahoma fauna. Environ. Entomol . 4: 877-882. Wardle, A. R. and J. H. Borden, 1985. Age-dependent associative learning by Exeristes roborator (F.) (Hymenoptera: Ichneumonidae) . Can Ent. 117: 605-616. Weseloh, R. M., 1981. Host location by parasitoids. In: Semiochemicals: Their Role in Pest Control (Nordlund, D. A., Jones R. L and Lewis, W. J., eds.). John Wiley, New York. pp. 79-95. Williams, H. J., G. W. Elzen and S. B. Vinson, 1988. Parasitoid host plant interactions, emphasizing cotton ( Gossypium ) . In: Novel Aspects of Insect Plant Interactions (Barbosa, P. and Letourneau, D. K., eds.). John Wiley & Sons, New York. pp. 171-200. Wilson, J. W., 1933. The biology of parasites and predators of Laphyma exiqua Huebner reared during the season of 1932. Fla. Entomol. 17: 1-15. Whiteside, R. C, P. P. Burbutis and L. P. Kelsey, 1967. Insect parasites of the green cloverworm in Delaware. J. Econ. Entomol. 60: 326-328. Wohlers, P., 1981. Effects of the alarm pheromone (E)-p-farnesene on dispersal behaviour of the pea aphid Acyrthosiphon pisum . Entomol. Exp Appl. 29: 117-124. Zanen, P. 0., W. J. Lewis, R. T. Carde, and B. G. Mullinix, 1989. Beneficial arthropod behavior mediated by airborne semiochemicals VI. Flight responses of female Micropl itis croceipes (Cresson), a Braconid endoparasitoid of Hel iothis spp., to varying olfactory stimilus conditions created with a turbulent jet. J. Chem. Ecol. 15: 141-168.

PAGE 191

BIOGRAPHICAL SKETCH Ted C. J. Turlings is the eldest son of Mr. J. L. Turlings and Mrs. M. L. Turl ings-Custers. He was born on July 19, 1959, in Heemstede, The Netherlands. After finishing high school (Mendel College, Haarlem, The Netherlands), Ted attended the University of Leiden in the Netherlands for his undergraduate and master's degrees in biology. His M.Sc. research was conducted under the supervision of Drs. K. Bakker, W. T. F. H. van Strien-van Liempt, and F. D. H. van Batenburg at the Division of Ecology, University of Leiden, The Netherlands. In April 1985 he began his doctoral studies at the Department of Entomology and Nematology of the University of Florida, conducting his research at the Insect Attractants, Behavior, and Basic Biology Research Laboratory (USDA-ARS), Gainesville, Florida. To conduct his doctoral research Ted Turlings was awarded a Fulbright scholarship through the Institute of International Education and the U. S. Information Agency. 178

PAGE 192

:> < I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. James H. Tumlinson, III, Chair Professor of Entomology and Nematology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. W. Joe Lew^S Associate Professor of Entomology and Nematology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. H. Jane Brockmann Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. n uc'i as Associate Professor of Entomology and Nemdtology

PAGE 193

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is full adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 1990 Games L. Nation Professor of Entomology and and Nematology Dean, Graduate School