(HYMENOPTERA: EULOPHIDAE) by LEIF D. DEYRUP

INTER- AND INTRASPECIFIC INTERACTIONS IN MELITTOBIA DIGITATA AND

LEIF D. DEYRUP

(Under the direction of Robert W. Matthews)

ABSTRACT

This research examined aspects of the nature, roles, and interrelationships between the fundamental wasp behaviors, stinging, chewing/feeding, and aggression using Melittobia as a model. In the field, these tiny gregariously developing wasps attack solitary bees, wasps and associates; in the laboratory, they accept a wide range of alternative hosts. The genus includes 8 North American species; the work focused on M. digitata, a species widely used under the name WOWBug® in science curricular activities.

When individual mated and unmated female wasps were simultaneously offered two flesh fly (Sarcophaga bullata) pupae as potential hosts, virgin females laid eggs on only one of the paired pupae more often than mated females did, but the “unused” second pupa developed to adulthood significantly less often than did controls, suggesting that the female nonetheless had stung it. Because direct visualization of stinging and its effects were not possible with fly pupae due to their enclosing puparia, female Melittobia were placed with mealworm pupae (Tenebrio molitor; Coleoptera) for two days, then removed before they could lay eggs. Exposed mealworms responded to gin-trap-reflex stimulation

significantly less than controls, and those that had melanized sting marks on them were even less likely to move; these findings suggested a paralytic component to the venom.

Highly marked mealworms were significantly more likely to remain wholly or partially in the pupal stage, suggesting a developmental component to the venom.

To confirm this, venom-milking techniques were developed. Milked venom from

M. digitata was injected into mealworms and pupae of an economically important natural host, the leafcutter bee (Megachile rotundata). All responded to injected venom with paralysis and in some cases with developmental delay, indicating that these reactions are a generalized phenomenon and that venom components might be useful for physiological studies or have potential application in pest control.

After mating, Melittobia chew out of their natal host cocoon. Observations showed that a female wasp stings a spot, then other females bite/chew cooperatively at the site. Chewing was experimentally elicited using milked venom and artificial pits.

Using venom and dissected glands as stimulants, similar results were obtained in two other species, M. femorata and M. australica. This apparently cooperative behavior promises insights into possible evolutionary origins of components of eusocial behavior.

The stinging-chewing sequence suggests an evolutionary derivation of the escape chewing behavior of dispersal-ready females from host-feeding. Experiments using mud dauber wasp (Trypoxylon politum) host hemolymph as the critical cue showed it was possible to switch Melittobia females back and forth between the two behaviors, stimulating chewing in a feeding context and feeding in a chewing context.

Whereas chewing and feeding are commonly accepted for female Melittobia, the widely held assumption has been that males do not feed. Experiments showed that males

can and do feed on the hemolymph of another male killed in combat. In addition, males fed hemolymph lived significantly longer than males that were not fed.

Finally, male aggression is widely accepted for male Melittobia, yet M. femorata females suffer body damage when multiple foundresses are placed on a host. Through marking and observation it was determined that this species is territorial and violent. For comparison, tests were done on both M. digitata, and M. australica. Melittobia digitata did not share any territorial or aggressive tendencies, whereas M. australica was territorial but without mutilation and violence.

INDEX WORDS: Aggression, cannibalism, cooperation, developmental delay,

intraspecific competition, paralysis, pheromone, territoriality,

venom

INTER- AND INTRASPECIFIC INTERACTIONS IN MELITTOBIA DIGITATA AND

M. FEMORATA (HYMENOPTERA: EULOPHIDAE)

LEIF D. DEYRUP

B.S., Stetson University, 2001

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

Athens, Georgia

2005

Leif D. Deyrup

INTER- AND INTRASPECIFIC INTERACTIONS IN MELITTOBIA DIGITATA AND

M. FEMORATA (HYMENOPTERA: EULOPHIDAE)

LEIF D. DEYRUP

Major professor: Robert W. Matthews

Committee: Kenneth Ross C. Wayne Berisford

Electronic version approved:

Maureen Grasso Dean of the Graduate School The University of Georgia December 2005

DEDICATION

I would like to dedicate this work to my parents Nancy and Mark Deyrup. I do not think I could have done this with out them, and I thank them far too little for all of their help.

Thanks.

ACKNOWLEDGMENTS

I would like to first thank my major advisor, Robert W. Matthews. His understanding and expertise fostered my graduate career, and will continue to serve as an example to me wherever I go. I would also like to thank Kenneth Ross for stepping in to act as my major advisor as well as his thoughtful advice and conversations. I thank C. Wayne Berisford for all his help on my project, and in defending my course of study early on. I am grateful to David Rivers for his expertise on wasp venom and his help in writing. Thanks are due to all the people who have helped in my lab: Jan Matthews, Jorge González, David

Jenkins, and Christian Silva-Torres. I thank Ray Noblet for his advice and help. I also thank Mike Strand and Judy Willis for their help on my projects and on my initial master’s committee. Thanks to all of the professors whose classes have enriched my research. In addition I would like to thank Doug Seiglaff and Mark Brown for their aid in physiology rich experiments. I have greatly appreciated the help of all the support staff and secretaries. In addition, I would like to thank all my friends and students of the

Department for our long discussions. Finally, I would like to thank my family, my parents Mark and Nancy, and my siblings Stephen and Ingrith. I am sure there are many names which I have left out, so if you are reading this, I probably should be thanking you.

Thank you.

TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS………………..……………………………………………….v

CHAPTER

1 INTRODUCTION AND BRIEF LITERATURE

REVIEW..…………………………………………………...……....…….8

2 HOST PREFERENCE AND UTILIZATION BY MELITTOBIA

DIGITATA (HYMENOPTERA: EULOPHIDAE) IN RELATION TO

MATING STATUS…………...…………………………….………..….19

3 PARALYZATION AND DEVELOPMENTAL DELAY OF A

FACTITIOUS HOST BY MELITTOBIA DIGITATA (HYMENOPTERA:

EULOPHIDAE)…………………………………………………...……..33

4 VENOM FROM THE ECTOPARASITIC WASP MELITTOBIA

DIGITATA DAHMS (HYMENOPTERA: EULOPHIDAE) INDUCES

PARALYSIS AND DEVELOPMENTAL DELAY IN NATURAL AND

FACTITIOUS HOSTS……………………………………………....…..41

vii

5 COOPERATIVE CHEWING IN MELITTOBIA DIGITATA DAHMS, A

GREGARIOUSLY DEVELOPING PARASITOID WASP, IS

STIMULATED BY STRUCTURAL CUES AND A PROBABLE

PHEROMONE IN CRUDE VENOM EXTRACT…….………………...64

6 ESCAPE CHEWING IN RESPONSE TO ALKALINE GLAND AND

VENOM RESERVOIR MARKED SPOTS AND CROSS-

ATTRACTANCY OF MILKED VENOM ACROSS SPECIES GROUPS

IN A GENUS OF PARASITIC WASP, MELITTOBIA

(HYMENOPTERA: EULOPHIDAE).……..………………..………...... 86

7 EXAMINING THE RELATEDNESS OF BEHAVIORS THROUGH

EXPERIMENTATION: SWITCHING ON AND OFF CHEWING AND

FEEDING BEHAVIOR IN A PARASITIC WASP……....………...…...97

8 FEEDING AND SIBLICIDAL CANNIBALISM IN A MALE

PARASITIC WASP (HYMENOPTERA: EULOPHIDAE)………...…109

9 COMPETITION AND AGGRESSION AMONG FEMALE

MELITTOBIA FEMORATA, (HYMENOPTERA: EULOPHIDAE) WITH

COMPARISONS TO M. DIGITATA, AND M. AUSTRALICA.…….…128

viii

10 CONCLUSIONS……………………………………………………….143

CHAPTER 1

INTRODUCTION AND BRIEF LITERATURE REVIEW

The genus Melittobia belongs to the family Eulophidae of the superfamily

Chalcidoidea, which are generally parasitic, but include predatory or, in a few cases, phytophagus wasps (Borror et al. 1992, Schauff et al. 1997). A distinctive feature of

Melittobia is the male’s morphology characterized by reduced wings, reduced eyes, swollen scapes, and a lack of frontofacial sutures (Schauff et al. 1997). The 14 species of this cosmopolitan genus are assigned to four species groups: acasta group, hawaiiensis group, assemi group, and the monospecific clavicornis group. The latter is regarded as the basal member of the genus (Assem et al. 1982, Dahms 1984a).

The main focus of this dissertation is on Melittobia digitata and M. femorata.

Both belong to the acasta group. They occur widely in North America, being sympatric in much of the eastern United States (Dahms 1984a, J. M. González, unpublished).

Melittobia spp. adults are about 1.0-1.5 mm long. In nature they develop as gregarious ectoparasitoids of the pre-pupae of solitary bees and wasps (Dahms 1984b,

Krombein 1967). They appear not to cause any obvious economic damage, although they occasionally infest commercial cultures of leaf-cutter bees (Doroshina 1989, 1990, Farkas and Szalay 1985, 1986, Hobbs 1968, Hobbs and Krunic 1971, Mac Farlane and Donovan

1989). More rarely they can attack social bees (Erickson and Medenwald 1979, Holm

1960, Whitfield and Cameron 1993). As their hosts tend to be useful predators and pollinators, Melittobia are not considered beneficial, but it is unlikely that they seriously depress populations of beneficial insects in natural habitats.

More recently, Melittobia has attracted considerable attention among biologists and science educators, because one species, M. digitata (as the WOWBug®), is used in life science curricula (Matthews et al. 1996, 1997). Because of the violent combat between Melittobia males, this species has been studied in kinship theory models (Griffin and West 2002). Similarly, Hamilton (1967) cited Melittobia as a prime example of how inbred groups tend to evolve extreme female-biased sex ratios.

Melittobia’s generalist host habits are well documented (Dahms 1984b). In addition to their primary hosts Melittobia are also capable of attacking the parasitoids of solitary bees and wasps such as bee flies of the genus Anthrax (Bombyliidae) that commonly attack mud daubers (Sceliphron spp.) (Dahms 1984b). In the laboratory

Melittobia parasitize Sarcophaga bullata (Diptera: Sarcophagidae) and also, less successfully, Tenebrio molitor (Coleoptera: Tenebrionidae) (Deyrup unpublished).

Melittobia dispersal is poorly understood. Clearly, crawling females are able to successfully disperse in habitats where host nests are clumped (e.g., mud dauber wasp nests located under bridges). While Melittobia are capable of flight, some species seem to fly more frequently than others, but none appear to be strong flyers. During laboratory flight trials González (unpublished) found that newly emerged females of M. digitata rarely fly at all, although some flew 40 to 50 cm which could be sufficient to launch a female into a wind column. In a field experiment in Jamaica Freeman (1974) found that the incidence of Melittobia parasitism was correlated to the direction of the Easterly

Trade Winds. In addition to wind, other dispersal solutions such as phoresy have not been investigated, and may aid females dispersing.

After dispersal, a female Melittobia must find a host, a process that is poorly understood. Mated dispersing females tend to be positively phototaxic. This however, may be part of the dispersal phase rather than searching for hosts, since Freeman and

Ittyeipe (1976) found Melittobia sp. more commonly in mud dauber nests that were in the darker areas of the nesting sites. Odor probably plays a part in host location as well.

Freeman and Ittyeipe (1976) also showed that females prefer cells that contain hosts to unoccupied cells. Silva-Torres et al. (2005) demonstrated that females spend more time standing over chambers containing host extract shielded by filter paper than over blank chambers. Host shape may also be important. Cooperband and Vinson (2000) found that round glass dummies the size of host pupae were attractive to M. digitata.

Once a fertilized female Melittobia finds a host, she stings it. Balfour-Browne

(1922) described the first sting as an “exploratory sting”, and raised hosts to adulthood that had only received this initial sting. Once the female stings she backs up to the wound and begins to feed on host fluids. In laboratory observations the amount that a female feeds on a host seems to depend on how long she has gone without food (Deyrup unpublished). After the initial feeding, female M. digitata tend to wander around the container, or sit motionless on the walls and ceiling (Deyrup unpublished). Subsequently, females return to the host to repeatedly sting and feed. Feeding appears to be necessary before oviposition, and presumably provides needed nutrition, but might also be a form of testing the host quality (Dahms 1984b).

After 1-3 days on a host the female starts to lay eggs. Balfour-Browne (1922) reported that stinging at this phase seemed to be lethal, and stung hosts died without developing. He also reported observing what he thought could be liquid moving through the ovipositor on these stings. Dahms (1984b) felt that ovipositor movements could be interpreted in other ways and that more research was needed to determine if females injected a paralyzing or killing venom. This topic is explored in chapters 2-4.

González and Matthews (2002) reported that a single female M. digitata or M. australica Girault could lay as many as 600 eggs on a Trypoxylon politum mud dauber host, but the same species provided a smaller host (Megachile rotundata) laid many fewer eggs. Female Melittobia may use nutritional cues to determine the quantity of eggs that a host can sustain. When artificial host packets containing a limited amount of host hemolymph were offered to M. digitata females, we noticed that they continued to lay eggs long after the food source is overtaxed (Deyrup and Matthews unpublished). This led to a high rate of larval starvation.

Melittobia has two types of females, a long-winged form that disperses after mating, and a short-winged female that develops quickly on a host, but doesn’t disperse, instead laying her eggs on the same host as her mother (Schmieder 1933). This polymorphism allows Melittobia to fully utilize even a very large host, and, theoretically, it allows a foundress female to found several broods on many large hosts. The two morphs differ in many ways behaviorally and physiologically (Schmieder 1933). The short-winged form emerges as an adult with many well developed eggs ready to be laid, whereas the long-winged female emerges with only a couple of developed eggs (Consôli and Vinson 2002). Consôli and Vinson (2004) found that the first females to develop feed

5 on host hemolymph. They attribute feeding on pure hemolymph as the factor responsible for development into the short-wing form. Possibly due to high mortality, we did not get short-wing females developing on artificial host packets made from pure hemolymph, but artificial hosts constructed of full host body tissue homogenates produced both forms with far less larval mortality (Deyrup and Matthews unpublished).

After emergence the females mate with a brother before dispersal (Assem et al.

1982). Hermann et al. (1974) and González et al. (1985) demonstrated that there is a pheromone produced by the male that attracts virgin females. Consôli et al. (2002) found alpha- and beta- trans-bergamotene to be the attractant for M. digitata.

Once mated, the long-winged female Melittobia must find an exit from the cocoon and cell of the host before they can disperse. If no ready exit is available they may start chewing out of the host’s nest. Exit holes have been discovered in many substances (Buckell 1928, Cowley 1961, Hermann 1971, Howard 1892, Maeta and

Yamane 1974, Torchio 1963). Donovan (1976) observed that females of M. hawaiiensis circled together chewing out of clear polystyrene boxes. He concluded that the females were cooperating in their effort to escape. This topic is further explored in chapters 5-7.

Flower or extra floral nectar feeding is not uncommon for parasitic wasps (Quicke

1997), so it is conceivable that the females could feed after having escaped from the host cell. However, it is unknown whether Melittobia feed while searching for a host in nature. In the laboratory, M. digitata and M. femorata will feed readily on a mixture of fructose and water a couple of days after emergence. If the solution is dyed, it can be seen through the wasp’s translucent cuticle to enter the crop, and the color comes out later in the droppings (Deyrup and Matthews unpublished).

Melittobia males are well known for potentially lethal male combat (Abe et al.

2003, Dahms 1984b, Hartley and Matthews 2003). Fighting and courtship activities are energetically costly (González and Deyrup unpublished). Whether a male feeds on a defeated foe, or whether males feed at all, has been subject to controversy, and is investigated in chapter 8.

Female-female interactions on their hosts in nature or in the laboratory are largely unknown and unstudied. Molumby (1996) found that superparasitism was not uncommon in a population of M. femorata attacking mud daubers in Mississippi. From one to five foundress females (mean 1.83) were reported in host cells. González (unpublished) has often found females of two different Melittobia species in field-collected Trypoxylon politum cells. In the laboratory our typical protocol for establishing new cultures has been to place 5 females on a host. We noticed that M. digitata females appeared to be fully tolerant of conspecifics whereas M. femorata females seemed relatively intolerant. These observations led to further investigation of intraspecific competition in these two species plus M. australica. Findings are reported in chapter 9.

REFERENCES

Abe, J., Y. Kamimura, H. Ito, H. Matsuda, and M. Shimada. 2003. Local mate

competition with lethal male combat: effects of competitive asymmetry and

information availability on a sex ratio game. Journal of Evolutionary Biology 16:

607-613.

Assem, J. van den., H. A. J. in den Bosch, and E. Prooy. 1982. Melittobia courtship

behavior: A comparative study of the evolution of a display. Netherlands Journal

of Zoology 32:427-471.

Balfour-Browne, M. A. 1922. On the life-history of Melittobia acasta Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Borror, D. J., C. A. Triplehorn, N. F. Johnson. 1992. An Introduction to the Study of

Insects. 6th edition. Harcourt College Publishers. Orlando, Fl.

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomologist 5: 14-22.

Consôli, F., and S. B. Vinson. 2002. Clutch size, development and wing morph

differentiation of Melittobia digitata Dahms (Hymenoptera: Eulophidae).

Entomologia Experimentalis et Applicata 102: 135-143.

Consôli, F., and S. B. Vinson. 2004. Wing morph development and reproduction of the

ectoparasitoid Melittobia digitata: nutritional and hormonal effects. Entomologia

Experimentalis et Applicata 112: 47-55.

Cônsoli, F. L., W. J. Williams, S. B. Vinson, R. W. Matthews, and M. F. Cooperband.

2002. Trans-bergamotenes – male pheromone of the ectoparasitoid Melittobia

digitata. Journal of Chemical Ecology 28: 1675- 1689.

Cooperband, M. F., and S. B. Vinson. 2000. Host-acceptance requirements of Melittobia

digitata (Hymenoptera: Eulophidae), a parasitoid of mud dauber wasps.

Biological Control 17: 23-28.

Cowley, D. R. 1961. The associates of Pison spinolae Shuckard (Hymenoptera:

Sphecidae). New Zealand Entomologist 2: 45-46.

Dahms, E. C. 1984a. Revision of the genus Melittobia (Chalcidoidea; Eulophidae) with

the description of seven new species. Memoirs of the Queensland Museum 21:

271-336.

Dahms, E. C. 1984b. A review of the biology of species in the genus Melittobia

Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum 21: 337-360.

Donovan, B. J. 1976. Co-operative material penetration by Melittobia hawaiiensis

(Hymenoptera: Eulophidae) and its adaptive significance. New Zealand

Entomologist 6: 192-193.

Doroshina, L. P. 1989. Adaptive features of Melittobia acasta (Chalcidoidea,

Eulophidae), a parasite of solitary bees. Zoologichesky Zhurnal 68: 60-69.

Doroshina, L. P. 1990. Protection from chalcids and optimization of nesting conditions

for bees (Hymenoptera: Apoidea, Megachilidae). Zoologichesky Zhurnal 69: 55-

59.

Erickson, E. H, and R. Medenwald. 1979. Parasitism of queen honeybee pupae by

Melittobia acasta. Journal of Agricultural Research 18: 73-76.

Farkas, J., and L. Szalay. 1985. Controlling of insect-parasites of alfalfa leaf cutting bee

stock Megachile rotundata F., Hymenoptera, Megachilidae). Apidologie 6: 171-

180.

Farkas, J., and L. Szalay. 1986. Melittobia acasta Walker (Hymenoptera: Eulophidae) the

most dangerous indirect pest of Lucerne seed production. Acta Agronomica

Hungarica 35: 103-106.

Freeman, B. E. 1974. The distribution in Jamaica of the mud wasp Sceliphron assimile

Dahlbom (Sphecidae) and its associates. Caribbean Journal of Science 14: 115-

124.

Freeman, B. E, and K. Ittyeipe. 1976. Field studies on the cumulative response of

Melittobia sp. (hawaiiensis complex) (Eulophidae) to varying host densities.

Journal of Animal Ecology 45: 415-423.

González, J. M., and R. W. Matthews. 2002. Development and sex ratio of Melittobia

australica and M. digitata (Hymenoptera: Eulophidae) on M. rotundata

(Hymenoptera: Megachilidae) and Trypoxylon politum (Hymenoptera:

Sphecidae). The Great Lakes Entomologist 35: 85-91.

González, J. M., R. W. Matthews, and J. R. Matthews. 1985. A sex pheromone in males

of Melittobia australica and M. femorata (Hymenoptera: Eulophidae). Florida

Entomologist 68: 279-286.

Griffin, A. S., and S. A. West. 2002. Kin selection: Fact and fiction. Trends in Ecology

and Evolution 17:15-21.

Hamilton, W. D. 1967. Extraordinary sex ratios. Science 156: 477-488.

Hartley, C. S, and R. W. Matthews. 2003. The effect of body size on male-male combat

in the parasitoid wasp Melittobia digitata (Hymenoptera: Eulophidae). Journal of

Hymenoptera Research 12: 272-277.

Hermann, L. D. 1971. The mating behavior of Melittobia chalybii. MSc. Thesis,

University of Georgia.

Hobbs, G. A. 1968. Controlling insect enemies of the alfalfa leafcutter bee, Megachile

rotundata. Canadian Entomologist 100: 781-784.

Hobbs, G. A. and Krunic M. D. 1971. Comparative behavior of three chalcidoid

(Hymenoptera) parasites of the alfalfa leaf cutter bee, Megachile rotundata, in the

laboratory. Canadian Entomologist 103: 674-685.

Holm, S. N. 1960. Experiments on the domestication of humble bees (Bombus Latr.), in

particular B. lapidarius and B. terrestris L. Royal Veterinary and Agricultural

College of Copenhagen, Yearbook 1960: 1-19.

Howard, L. O. 1892. The habits of Melittobia. Proceedings of the Entomological Society

of Washington 2: 244-249.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington, DC.

Mac Farlane, R. P., and B. J. Donovan. 1989. Melittobia spp. as parasitoids of bumble

and Lucerne leafcutting bees and their control in New Zealand. Proceedings of the

42nd New Zealand Weed and Pest Control Conference 1989: 274-277.

Maeta, Y., and S. Yamane. 1974. Host records and bionomics of Melittobia japonica

Masi (Hymenoptera: Eulophidae). Bulletin of the Tohoku National Agricultural

Experimental Station 47: 115-131.

Matthews, R. W., T. R. Koballa, Jr., L. R. Flage, and E. J. Pyle. 1996. WOWBugs: New

Life for Life Science. Riverview Press, LLC, Athens, GA.

Matthews, R. W., L. R. Flage, and J. R. Matthews. 1997. Insects as teaching tools in

primary and secondary education. Annual Review of Entomology 42: 269-289.

Molumby, A. 1996. The Evolutionary Ecology of a Gregariously Nesting Wasp,

Trypoxylon politum. Ph.D. dissertation. Chicago, Illinois: The University of

Chicago.

Quicke, D. L. J. 1997. Parasitic Wasps. Chapman and Hall, London.

Schauff, M. E., J. LaSalle, and L. D. Coote. 1997. Annotated Keys to the Genera of

Nearctic Chalcidoidea (Hymenoptera). NRC Research Press, Ottawa.

Schmieder, R. G. 1933. The polymorphic forms of Melittobia chalyibii Ashmead and the

determining factors involved in their production (Hymenoptera: Chalcidoidea,

Eulophidae). Biological Bulletin 65: 338-352.

Silva-Torres, C. S. A., R. W. Matthews, J. R. Ruberson, and W. J. Lewis. 2005. Role of

chemical cues and natal rearing effect on host recognition by the parasitic wasp

Melittobia digitata. Entomological Science 8: 355-362.

Torchio, P. F. 1963. A chalcid wasp parasite of the alfalfa leaf cutting bee. Utah Farm

Home Science 24: 70-71.

Whitfield, J. B., and S. A. Cameron. 1993. Comparative notes on hymenopteran

parasitoids in bumble bee and honey bee colonies (Hymenoptera: Apidae) reared

adjacently. Entomological News 104: 240-248

CHAPTER 2

HOST PREFERENCE AND UTILIZATION BY MELITTOBIA DIGITATA

(HYMENOPTERA: EULOPHIDAE) IN RELATION TO MATING STATUS

Deyrup, L. D., and R. W. Matthews. Journal of Entomological Science 38: 682-687. Reprinted here with permission of publisher.

ABSTRACT

Host preferences of virgin and mated females of Melittobia digitata Dahms were compared in the laboratory using pupae of the flesh fly Neobellieria (= Sarcophaga) bullata (Parker) as hosts. When simultaneously offered two hosts, virgin females used only one of the hosts more often than mated females did. However, the unused second host developed to adulthood significantly less often than did controls, suggesting that the female stung and paralyzed it. Because virgin females lay only a few eggs that always develop into males which utilize very little of the host resource, this behavior seems adaptive in that potential hosts remain available, but developmentally arrested, for later full exploitation by the same female (now mated by her offspring). An additional implication of these results is that females can discriminate one flesh fly host from the other, and choose to avoid oviposition on both.

KEY WORDS: parasitism, parasitoid, host development, host recognition, venom, clutch size

INTRODUCTION

Melittobia is a genus of small, cosmopolitan, gregarious ectoparasitoids of prepupae and pupae of many insect species ranging across different orders. Melittobia commonly parasitize solitary wasps and bees in nature (Maeta and Yamane 1974,

Edwards and Pengelly 1966, Krombein 1967), and several species can be found parasitizing mud dauber wasps (Hymenoptera: Sphecidae) (Dahms 1984).

Insects can be valuable tools in science education (Matthews et al. 1997). Being easy to rear, handle and differentiate between the sexes, M. digitata Dahms has emerged as a particularly user-friendly classroom insect for exploring various concepts in biology such as life history strategies, orientation behavior, and population dynamics (Matthews et al. 1996). However, there is a paucity of basic knowledge about many aspects of the biology of this species.

Melittobia digitata has the ability to lay many eggs on a single host. Thus, a female could potentially lay so many eggs that all host food reserves would be used before her offspring fully developed, resulting in high brood mortality. However, it has been demonstrated that the number of eggs oviposited by some other species of

Melittobia correlates with the amount of resources the host provides (Ittyeipe and

Freeman, Unpubl.). Our observations (Unpubl.) indicate that M. digitata also does not overburden a host with offspring.

A normal clutch of Melittobia offspring produced by a single female on one host includes a disproportionate number of females as compared to males. For example, the offspring of unidentified species of Melittobia were shown to be, on average, about 95%

15 females (Ittyeipe and Freeman, Unpubl.), and M. chalybii Ashmead has been shown to produce 97% females (Schmieder 1938). Melittobia digitata also shares this skewed sex ratio favoring females (Dahms 1984). The disproportionate sex ratio seems to violate

Fisher’s principle, i.e., sex ratios should equilibrate over time (Hamilton 1967). However, in species where inbreeding is frequent, as occurs in Melittobia, high numbers of females to males are common (Hamilton 1967).

As in other Hymenoptera, M. digitata females are able to control offspring sex; unfertilized eggs are produced by withholding stored sperm, and such eggs always produce males. This mechanism allows some reproduction (yielding males only) by females that have failed to find mates (Hobbs and Krunic 1971), but its principal importance may be to allow resource allocation appropriate to the sex of the offspring.

Among species that practice sibling mating, as is the case for M. digitata, resources can be devoted disproportionately to females, thus increasing the growth and spread of the family group, which is acting as a population.

The behavior of mated versus virgin M. digitata can be observed in the laboratory. Virgin M. digitata are easily obtained by isolating female pupae. Despite the ease of this procedure, little has been published about behavioral differences between virgin and mated females, and practically nothing has been done to investigate host use by virgin females. It is known that virgins initially oviposit only a few male eggs, and that the females survive long enough for their offspring to become reproductively mature.

The virgins are then inseminated by their sons and proceed to oviposit a larger female- dominated brood on the same host (Dahms 1984, Balfour-Browne 1922).

Balfour-Browne (1922) reported a “paralyzation sting” by Melittobia females, but that observation could be interpreted in other ways (Dahms 1984). A non-feeding insertion of the stinger observed in M. chalybii was interpreted by Buckell (1928) as a paralyzing sting. It has been shown that hosts fed on by adult female Melittobia still matured and even emerged as long as eggs were not laid (Dahms 1984). It is unclear which of these attributes M. digitata might possess.

In attempting to duplicate production of a high proportion of male brood as reported by Whiting and Blouch (1948), we noticed that when virgin M. digitata females were given two hosts, males rarely developed from both hosts. Previously, Whiting

(1947) noted that when unmated female Melittobia were placed with 132 blow flies, none emerged. Although they were identified only as Melittobia, based on a photo included in the paper some could have been M. digitata (Gonzalez pers. comm.). Therefore, the objectives of the study reported here were to determine: (1) whether a virgin female is able to prevent potential hosts from completing their normal development, and (2) whether a virgin female, when presented with two hosts, will use only one for male production.

MATERIALS AND METHODS

Two initial cultures were founded by placing a mated female M. digitata with three flesh fly, Neobellieria (=Sarcophaga) bullata Parker, puparia obtained from

Carolina Biological Supply Co., Burlington, NC, into 1-dram glass shell vials stoppered with cotton to allow gas exchange. Cultures were maintained in an incubator at 26 °C,

17 and when the first females emerged, host puparia in each culture were opened with forceps to expose the remaining uneclosed pupae of M. digitata. These were sorted into two groups, one with males and females and the other with females only based on the presence of eyes (males lack compound eyes).

One day after the M. digitata adults eclosed, equal numbers of females from each initial culture were combined to form one group of virgins and another group of mated females (in M. digitata mating normally occurs soon after eclosion when males are present). These groups were used to set up the experimental cultures using the same procedures and conditions as for the initial cultures, except that only two randomly chosen host fly puparia were provided in each culture vial. From prior observations, we expected that some individuals of the virgin female group would die without producing offspring. Therefore, more replicates were set up in the virgin group. In all, 100 mated female cultures, 127 virgin female cultures, and 25 controls (hosts without females) were simultaneously established.

Thirteen days later, by which time any M. digitata progeny would be beginning to pupate or unparasitized fly puparia would have eclosed as adults, all host puparia were opened. For the virgin group we recorded number of hosts with developing M. digitata, number of males and developmental stage of each, and number of unemerged adult flesh flies. The mated group was scored for number of hosts containing M. digitata progeny and the developmental stages of the offspring on each host. For the controls we recorded the number of emerged adult flesh flies.

1. Prevention of host development. For virgin female cultures in which only one of the two hosts was used (no matter how many males were laid on that one host) we

18 counted the number of unused flesh flies that matured to adulthood. Using a Chi-square test (Zar 1974), we compared this number to the expected value derived from the control group.

2. A Chi-square test (Zar 1974) was used to determine whether the virgin female group restricted oviposition to a single host. Observed results were determined by assessing how many of the hosts had offspring at different developmental stages (adult, pupae, or larvae), and the number of hosts used in each culture (one or both). Cases in which multiple offspring were at the same developmental stage were omitted because those might have resulted from a single oviposition session. The expected value was deduced by taking the number of mated females that used both hosts (to take into account the previously observed proclivity of mated females to use just one host), multiplied by

0.5, the probability that if only two eggs were randomly oviposited that they would be on the same host.

Periodic observations of the virgin female cultures showed that during the first several days of the experiment these females would leave the host pupae and be found crawling on the vial walls. After 13 days most females were observed resting on one of the host pupae. This validates introduction of the probability factor, because if the female remained on the initially chosen host for the entire 13 days, then no opportunity for making choices would exist. Because females did leave the hosts periodically, a host choice had to be made each time the female returned. If she were able to recognize previously visited hosts (an ability possessed by other species of parasitic wasps), then she could bias her choice.

RESULTS AND DISCUSSION

In 85 of the 127 virgin female cultures male offspring were produced on only one of the two available hosts. In these, significantly fewer flesh fly adults eclosed compared to the control group (Table 2.1). Although stinging and host feeding were not directly observed, the simplest explanation is that these unused and uneclosed hosts were stung by the virgin females, and the sting inhibited further development. However, the chemical nature of M. digitata venom is unknown. In addition to venom, other compounds affecting development potentially could be injected during stinging. In

Dendrocerus carpenteri Curtis (Hymenoptera: Megaspilidae), an aphid parasite, juvenile hormone injected at the time of attack has been shown to stop development (Holler et al.

1993).

In 50 of 85 cases where males at distinctly different developmental stages were found, virgin M. digitata oviposited on only one of the two hosts (Table 2.2). This suggests that although they typically did not remain on their original host (particularly during the first week), the female wasps were significantly more likely to preferentially return to the same host to lay their next egg.

Because M. digitata lays its eggs through the puparial “shell” of the flesh fly, previously laid eggs provide no visual cues, yet the female can clearly tell which host has received them. Several explanations are possible, though not mutually exclusive. For example, the mother might chemically mark the host. This occurs in other Hymenoptera such as D. carpenteri (Holler et al. 1993). Alternatively, she might obtain and remember the physical dimensions of the host on which she previously laid her eggs during her

20 initial inspections of the available host resources. An additional possibility is that the female may simply repeat the assessment process by which she made her initial egg host choice. Many species of parasitoid wasps are capable of assessing host suitability

(Vinson and Iwantsch 1980); they might be expected to consistently make the same decision again.

It seems likely that a mated female Melittobia is capable of laying eggs continuously until she exhausts either hosts or sperm. If a female fails to be mated or exhausts her sperm but still has hosts available, she might be expected to select one host, lay a few eggs, then wait quietly nearby until her sons mature, mate with one of them, and commence egg laying again, as was shown by Balfour-Browne (1922).

The skewed sex ratio in Melittobia seems adaptive for both virgin and inseminated females. For a mated female, producing either just males or more males than necessary for insemination of her female brood is not adaptive, because males are not equipped for dispersal and the vast majority of dispersing females that might find them would likely have already mated (like most parasitoid wasps, female Melittobia mate only once (Assem et al. 1982)). In contrast, a virgin female that produces just a few males ultimately ensures that she will become mated (to a son) and thereby be able to produce a predominantly female brood often on the same host used to produce her clutch of sons. When more than one host is available, a virgin is well served by the host utilization behavior we observed. By preventing both hosts from developing past the stage where they would be suitable for her offspring, then using only one for male production, a virgin female can “save” the other for her full, predominantly female brood.

ACKNOWLEDGMENTS

We thank Mark Deyrup, Janice Matthews, and Jorge Gonzalez for their insightful ideas and reviews of earlier drafts of this paper.

REFERENCES

Assem, J. van den, H. A. J. den Bosch, and E. Prooy. 1982. Melittobia courtship

behaviour: a comparative study of the evolution of a display. Netherl. J. Zool. 32:

427-471.

Balfour- Browne, M. A. 1922. On the life-history of Melittobia acasta, Walker, a chalcid

parasite of bees and wasps. Parasitol. 14: 349-369

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead.

(Chalcidoidea: Elachertidae) Pan-Pacif. Entomol. 5: 14-22.

Dahms E. C. 1984. Review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Mem. Queensland Mus. 21: 337-360

Edwards, C.J. and D.H. Pengelly. 1966. Melittobia chalybii Ashmead (Hymenoptera:

Eulophidae) parasitizing Bombus fervidus Fabricius (Hymenoptera: Apidae).

Proc. Entomol. Soc. Washington 96: 98-99.

Hamilton, W. D. 1967. Extraordinary sex ratios. Science 156: 477-488

Hobbs, G.A. and M.D. Krunic. 1971. Comparative behavior of three chalcidoid

(Hymenoptera) parasites of the alfalfa leafcutter bee, Megachile rotundata, in the

laboratory. Can. Entomol. 103(5): 674-685.

Holler, C., H. Bargen, S.B. Vinson, D. Witt. 1993. Evidence for the external use of

juvenile hormone for host marking and regulation in a parasitic wasp,

Dendrocerus carpenteri. J. Insect Physiol. 40(4): 317-322.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington DC. 570 pp.

Maeta, Y. and S. Yamane. 1974. Host records and bionomics of Melittobia japonica

Masi, (Hymenoptera: Eulophidae). Bull. Tohoku Nat. Agric. Exp. Station 47:

115-131.

Matthews, R. W., T. R. Koballa, L. R. Flage and E. J. Pyle. 1996. WOWBugs: New Life

for Life Science. Riverview Press, LLC, Athens, GA. 318 pp.

Matthews, R., W. L. R. Flage, and J. R. Matthews. 1997. Insects as teaching tools in

primary and secondary education. Ann. Rev. Entomol. 42: 269-289.

Schmieder, R. G. 1938. The sex ratio in Melittobia chalybii Ashmead, gametogenesis and

cleavage in females and in haploid males (Hymenoptera: Chalcidoidea) Biol.

Bull. 74: 256-266

Vinson, S.B. and G.F. Iwantsch. 1980. Host suitability for insect parasitoids. Ann. Rev.

Entomol. 25: 397-419

Whiting P.W. 1947. Some experiments with Melittobia and other wasps. J. Hered. 38:

11-20

Whiting, P.W. and B. M. Blouch. 1948. The genetic block to free oviposition in the

chalcidoid wasp Melittobia sp. -C. Biol. Bull. 95: 243-244.

Zar, J.H. 1974. Biostatisitical Analysis. Prentice-Hall, Englewood Cliffs, NJ. 620 pp.

Table 2.1. Number of flesh flies that emerged as adults in the virgin group compared to controls with no female present. The data include only those cultures where only one of the two hosts was used.

Group Emerged No Emergence

Virgins 5 80

Expected 78.2 6.8

Controls 46 4 p<0.001, 2=856.496, 1 df.

Table 2.2. Cases in the virgin female group where one or two hosts were used compared to the expected values. Observed is the number of virgin cultures having males of distinctly different developmental stages. The expected numbers come from the observed total (71) multiplied by 0.5, the probability of a random choice multiplied by the percentage of mated females who used both hosts (0.94) for “2 Hosts Used” and (1.06) for “1 Host Used”.

1 Host Used 2 Hosts Used

Observed 50 21

Expected 37.6 33.4 p=0.003, 2=8.693, 1 df.

CHAPTER 3

PARALYZATION AND DEVELOPMENTAL DELAY OF A FACTITIOUS HOST BY

MELITTOBIA DIGITATA (HYMENOPTERA: EULOPHIDAE)

Deyrup, L. D., M. Deyrup, and R. W. Matthews. Journal of Entomological Science 38: 703-705. Reprinted here with permission of publisher.

Ectoparasites face at least two potential obstacles. Since ectoparasitoids lay their eggs on the outside of the host, their eggs are in danger of being dislodged if the host moves. Some ectoparasitoids deal with this problem by injecting a paralytic venom (Piek and Spanjer 1986. pg. 161-307 in Venoms of the Hymenoptera: Biochemical,

Pharmacological and Behavioral Aspects. Ed. T. Piek. Academic Press, London).

Continued host development poses another danger, potentially limiting the amount of time that the host is available for the parasitic larvae. To cope with this problem, ectoparasitoids have evolved a variety of mechanisms that cause developmental stasis or delay in host development, such as the injection of specialized venom proteins, some of which influence host ecdysteroid production (Marris et al. 2001. Physiol. Entomol. 26:

229-238)

Melittobia digitata Dahms (Hymenoptera: Eulophidae) is a ectoparasitoid that usually attacks the prepupae of solitary and, sometimes, social bees and wasps (Maeta and Yamane 1974. Bull. Tohoku Nat. Agric. Exp. Sta. 47: 115-131, Edwards and

Pengelly 1966. Proc. Entomol. Soc. Wash. 96: 98-99). It is also a facultative hyperparasitoid, attacking both Hymenoptera and Diptera, such as bee flies of the genus

Anthrax, that attack its habitual hosts (Krombein 1967. Trap-nesting Wasps and Bees:

Life Histories, Nests, and Associates. Smithsonian Press, Washington, DC).

The objective of this study was to determine whether Melittobia digitata is able to paralyze or halt development of a host. An impediment to any investigation of the possible paralytic effects of the venom of M. digitata is that its most common host, the mud dauber Trypoxylon politum (Say), is already virtually immobile in the prepupal and pupal stages. Alternate, easily accessible and mobile hosts, useful for a bioassay for

28 paralytic effects, are pupae of the yellow mealworm, Tenebrio molitor L. Preliminary tests showed that M. digitata would readily sting mealworm pupae.

Two hundred four newly-molted T. molitor pupae from a laboratory culture maintained at the University of Georgia Department of Entomology were placed individually in 1-dram glass vials (47 x 15 mm) with cotton stoppers. Five female M. digitata from a culture maintained in the laboratory were added to each of 102 vials containing mealworm pupae; the remaining 102 vials of pupae were kept as controls. All were maintained at ambient room temperatures (20 to 22oC).

After 48h M. digitata were removed, and all mealworm pupae were immediately tested for movement, using the gin-trap reflex (Hinton 1946. Trans. R Ent. Soc. Lond. 97:

473-96). This reflex occurs when a pupa, whose abdomen in a relaxed state is somewhat decurved in relation to the thorax, suddenly straightens the abdomen, causing a set of sharp edges on the anterior and posterior margins of the medial tergites to come together.

These structures are thought to help defend the pupa from small insects and mites

(Crowson 1981. The Biology of the Coleoptera. Academic Press, London). To test for movement or mobility, each pupa was held individually by the thorax, with the thumb and forefinger of the investigator clasping its dorsal and ventral surfaces. With the forefinger and thumb of the other hand, the investigator gently squeezed the sides of the pupal head which invariably elicits abdominal twitching in unparasitized pupae. A second investigator, who was unaware of which larvae had been exposed to the wasps, performed a second test intended to elicit movement, in which the pupae were held as before and the abdominal tergites stroked, first with a small paint brush, then with a finger tip. In addition to mobility, we also recorded the number of pupae that

29 metamorphosed into adults, and the number of visible dark spots on the pupae that appear at sting sites as noted by Malyshev (1968. Genesis of the Hymenoptera and the

Phases of their Evolution. Edited Translation by O. W. Richards and B. Uvarov. Richard

Clay [The Chaucer Press], Ltd., Bungay) using M. acasta (Walker).

The mean number of dark spots on pupae classified as immobile was 5.16

(SD=3.41); the mean number of spots on pupae classified as mobile was 0.06 (SD=0.24).

A Mann-Whitney test showed that these means were significantly different (U = 113.00, df =1, P < 0.001) (Statistica 6.0. StatSoft, Inc. Tulsa).

Estimations of mobility taken by the two investigators agreed 97.55% of the time.

In the five cases in which only one investigator classified a pupa as mobile, the pupa was considered to be mobile for the purposes of this study. Mealworm pupae exposed to M. digitata were significantly less likely to move, with 68 out of 102 classified as immobile,

2 compared to 3 out of 102 in the control group ( = 91.27, df =1, P < 0.001) (Statistica

6.0. StatSoft, Inc. Tulsa) (Table 3.1). A group of the immobile pupae observed 2d later under a dissecting microscope (40X) revealed internal movement of fluid that could be seen through the translucent pupal integument.

Significantly fewer mealworms developed to adulthood in the experimental group

( = 59.46, df =1, P < 0.001) (Statistica 6.0. StatSoft, Inc. Tulsa) (Table 3.2). Of the 68 mealworms in the experimental group that were classified as immobile, only one eclosed as an adult. Statistically this was compared with those that eclosed in the unparalyzed population of the meal worms exposed to M. digitata using a McNemar test (T = 18.27, df =1, P < 0.001) (Statistica 6.0. StatSoft, Inc. Tulsa).

Our results confirm that M. digitata stings the host, as evidenced by the significantly greater number of dark spots on mealworm pupae that did not respond to stimuli, and that the substance injected by M. digitata affects both mobility and development of its host. The data support the hypothesis that the sting of M. digitata causes paralysis in mealworm pupae. It seems probable that the paralysis is caused by the injected venom, rather than physical trauma resulting from insertion of the sting. To test this hypothesis, it would be necessary to obtain and inject pure venom from the wasps, with corresponding injection of insect saline in controls. The reduction in mobility in mealworm pupae exposed to M. digitata could be due to death rather than to paralysis, but the observed internal fluid movement suggests that these hosts were simply paralyzed. Moreover, the immobilized pupae did not immediately decay or desiccate, but remained in an arrested state of development for an extended period without metamorphosing into adults.

Host paralysis following parasitization by chalcidoids has been reported in the aphelinid Aphelinus jucundus Gahan and the eulophids Sympiesis viridula (Thomson)

(Parker and Smith 1933. Ann. Entomol. Soc. Amer. 26: 21-39) and Dahlbominus fuscipennis (Zettersted) (Clausen 1940. Entomophagous Insects. McGraw-Hill, New

York). Coudron et al. (2000. Ann. Entomol. Soc. Amer. 93: 890-897) also arrested developmental activity in hosts injected with an aqueous extract from the venom gland of the eulophid Necremnus breviramulus Gahan.

Interruption of T. molitor pupal development by M. digitata sting components is strongly supported (Table 3.2). Although 20% of the controls failed to develop, it is possible that handling or other factors were detrimental to development. Because one of

31 the 68 immobilized pupae developed into an adult, the possibility of a dosage-sensitive factor or factors that affect both paralysis and development exists. The ability to halt or delay development has been shown in at least two other species of ectoparasitic eulophids(Marris et al. 2001. Physiol. Entomol. 26: 229-238., Coudron and Brandt 1996.

Toxicon 34: 1431-1441).

The M. digitata-mealworm interaction could also provide a useful classroom demonstration of how minuscule amounts of substances injected by a tiny wasp into a relatively enormous host can have dramatic effects on the host. Although the mealworm pupa is a factitious host, probably similar effects occur in normal hosts. Though less easily observed, these venom effects are no doubt of considerable adaptive significance in the parasitoid-host relationship.

Table 3.1. Tenebrio molitor pupal movement after 48h exposure to five female M. digitata

Experimental Treatment Movement Evident No Movement

Female M. digitata present 34 68

Controls (female M. digitata 99 3 absent)

2 p<0.001, =91.27.

Table 3.2. Tenebrio molitor pupal development after 48h exposure to five female M. digitata

Experimental Treatment Developed to adult Did Not Develop

Female M. digitata present 26 76

Controls (female M. digitata 81 21 absent)

2 p<0.001, =59.46.

CHAPTER 4

VENOM FROM THE ECTOPARASITIC WASP MELITTOBIA DIGITATA DAHMS

(HYMENOPTERA: EULOPHIDAE) INDUCES PARALYSIS AND

DEVELOPMENTAL DELAY IN NATURAL AND FACTITIOUS HOSTS

Deyrup, L. D., D. B. Rivers, and R. W. Matthews. To be Submitted to Annals of the Entomological Society of America.

ABSTRACT

Melittobia digitata is a small gregarious eulophid. Although there is a large amount of information on the life history of this wasp, very little is known about its venom. We tested milked venom to determine if it would paralyze and/or cause developmental delay. Prepupae of Megachile rotundata (alfalfa leafcutter bee), an economically important natural host species, were injected with either venom or saline solution without venom. We found that the venom acted as a paralyzer (P<0.001, df 1), and that there was a general decrease in time to paralysis as dose was increased (r= -0.70,

P<0.01). We also injected pupae of mealworms (Tenebrio molitor), a factitious host, and obtained similar results, with a highly significant difference between venom and controls

(P<0.001, df 1) and a decrease in time to paralysis with an increase in dose(r= -0.40,

P=0.01). In addition we found that injected venom induced variable degrees of developmental delay depending on dosage. This work settles a controversy in the literature about whether developmental delay and/or paralysis can be caused by the venom alone. In addition, it is a first step in venom characterization and points to the potential value of further work aimed at chemically characterizing the compounds comprising M. digitata venom and their modes of action.

KEY WORDS: Developmental delay, Melittobia digitata, paralysis, venom

INTRODUCTION

Melittobia digitata Dahms (Hymenoptera: Eulophidae) is a cosmopolitan ectoparasitoid that parasitizes pupae of solitary bees and wasps (Edwards and Pengelly

1966, Maeta and Yamane 1974). The wasp also can function as a hyperparasite, for example when attacking diapausing larvae of the bee fly, Anthrax sp., which in turn is a parasite of mud dauber wasps (Krombein 1967). As an idiobiont, M. digitata displays many behaviors typical of this life history strategy: Its larvae typically develop externally on concealed hosts with limited mobility (pupae), and adult females appear to inject substances (presumably venom) during oviposition that inhibit continued growth and development of parasitized hosts (Dahms 1984, Deyrup et al. 2003, Quick 1997).

Immobilization of the host appears to be essential to prevent dislodging of parasitoid eggs if the host moves or molts. Likewise, the adult female faces risk of injury during oviposition or being interrupted if the host is still mobile (Piek and Spanjer 1986, Rivers et al., 1999).

Like many ectoparasitic species after host location, the adult female inserts her ovipositor into an acceptable host and feeds on hemolymph exuding from the puncture

(Balfour-Browne 1922, Buckell 1928, Dahms 1984). This may be performed repeatedly during the first 24-48 hours after finding a host, presumably as a means to acquire the resources needed for egg maturation (Flanders 1943). Prior to laying eggs but after host feeding, the wasp inserts the ovipositor into the host again, and according to Balfour-

Browne (1922), a fluid presumed to be venom moves down the ovipositor and into the host during the later stings. Hosts envenomated in this fashion, fail to progress in

37 development, regardless of whether feeding parasitoids are present or not (Balfour-

Browne 1922). Though these observations are consistent with other investigations using

M. digitata (Deyrup et al. 2003, Deyrup and Matthews 2003b), and with the effects of envenomation by many species of ectoparasitoids (Coudron and Puttler 1988, Doury et al. 1995, Quicke 1997, Rivers 2004, Rivers et al. 1993), the use of venom during parasitization by M. digitata has been disputed (Dahms 1984). The issue as to whether this wasp uses maternal secretions to halt host development remains unresolved.

The aim of this study was to examine the response of two host species to venom from M. digitata. Specifically, we tested whether venom from M. digitata was capable of eliciting paralysis and/or development delay in natural [alfalfa leafcutter bee, Megachile rotundata (F.) (Hymenoptera: Megachilidae)] and factitious [yellow mealworms,

Tenebrio molitor L. (Coleoptera: Tenebrionidae)] hosts.

METHODS

Insect rearing

Melittobia digitata were cultured at the University of Georgia (Department of

Entomology) using pupae of the mud dauber, Trypoxylon politum Say, as hosts. Cultures were reared in continuous darkness at 25ºC. Adult females (3 days-old ±1 d after emergence at 25oC) used for milking were from cultures that have been maintained in the laboratory for many generations.

Leafcutter bees (Megachile rotundata) were purchased from Pioneer (Napa, Id.) as prepupae, and were kept refrigerated (12ºC) until used for venom injections. The bees

38 were allowed to warm to room temperature and tested positive for movement before injection. Pupae of the yellow mealworm, T. molitor, were taken from laboratory cultures maintained at room temperature on a diet of bran, rolled oats, wheat flour and occasional pieces of apple. Pupae used for injection were 2-4 days old.

Venom collection

Venom from M. digitata was collected by the manual milking method described by Deyrup and Matthews (2003a). Individual females were placed dorsal side down on sticky tape and pressure was applied using a dissecting needle to crush the head. Each wasp then ejected venom that solidified as a long thread. That thread was collected on a dry number 2 insect pin, and added to saline [10 mM sodium phosphate, 0.9% (w/v)

NaCl, 15% (w/v) sucrose, 1mM EDTA, pH 8.0] before injection. The venom collected from one female was termed 1 female equivalent dose (FED).

Venom injections

Prepupae of leaf cutter bees were injected using a 5 µl Hamilton syringe with either saline or venom (0-5 FED) in 2 µl saline. Injections were accomplished by inserting the syringe tip into the lateral surface of the abdomen. The bees were then maintained at 25°C in clear polystyrene boxes (50 mm x 25 mm x 18 mm) while observations were made to record the onset of paralysis, developmental arrest, or death.

Mealworm pupae were injected in a similar manner. Venom and saline injections

(2 µl volume) were accomplished by inserting the syringe needle between the 2nd and 3rd

39 or between the 3rd and 4th abdominal segments near the gin-traps. After injection, pupae were placed in clear plastic boxes at 25°C.

Paralysis and development

Paralysis was examined in leaf cutter bee prepupae by brushing their mandible hairs lightly with a fine camel’s hair brush. Prepupae were considered paralyzed if no movement resulted from stimulating the mandibular hairs. Paralysis was tested in each bee every hour for the first 12 hours following injections, and then every two hours for an additional 6 hours. The developmental fate of all individuals was followed until either adult eclosion or death.

Paralysis was monitored in pupae of T. molitor using the gin-trap reflex (Hinton

1946). To test for movement in saline- or venom-injected pupae, a calibrated glass tool was applied to each wing pad and the gin-trap reflex action recorded. The testing tool was constructed from a 50-µl glass capillary tube that was inserted into a 100-µl capillary tube that had been cut 25 mm from the top. The tool was calibrated to 4 g of pressure by applying weight to the 50 µl tube while held vertically on an Ohaus® (Navigator™) analytical balance. When tested on 50 untreated mealworm pupae, a gin-trap reflexive action was produced 100% of the time over a period of 24 hours using this technique.

Injected mealworms were tested on each wing pad once every hour for the first 12 hours after treatment, and then the reflex response monitored every 2 hours for an additional 6 hours.

A separate group of 20 mealworm pupae was injected with serial dilutions of one batch of collected venom (10, 5, 2.5, 1.5, and 0 FED), so that 4 individuals received each

40 dose. These pupae were allowed to develop at 25oC and were monitored every 24 hours for 4 weeks.

Statistical analyses

Gin-trap reflex and mandibular stimulation responses were analyzed by Chi-

Square analysis and regression analysis (Statistica 6.0).

RESULTS

Artificial envenomation effects on the leafcutter bee

Some saline-injected individuals (0 FED) scored as paralyzed on the movement test in the 18h. This was a normal cessation of movement during development. All injected individuals were scored as paralyzed within the 18h. Despite the handicap of cessation in movement in the controls, the results were highly significant for venom injected versus saline injected individuals chi-square test (P< 0.001, 2 = 28.8, 1df). Since we were able to use several different doses, we could determine if there was a negative correlation between time to paralysis and dose in FED. Overall, the time to paralysis decreased with an increase in dose (Including controls r= -0.70, P<0.01, F1,28=27.33;

Excluding controls r= -0.47, P=0.02, F1,22=6.32). Time until paralysis was much longer for saline-injected individuals average time in hours of 12.17±1.99% (n=6), in contrast to the venom-injected group which had an average time of 4.04±0.39 (n=24). Hours until paralyzed under differing venom doses within the envenomated group were 2 FED

3.67±0.37%, n=9; 2.5 FED 6.6±0.68%, n=5; 3 FED 5.5±0.5%, n=2; 5 FED 2.5±0.33%, n=8 (X + SEM, number of individuals).

Artificial envenomation effects on the mealworm

At 24 h post-injection we noticed that some of the mealworms had become more translucent, but were still clearly alive. It was possible to see through the cuticle to the dorsal vessel which could be observed beating. Upon dissection, we noticed that the abdominal muscles had degenerated and sloughed from the cuticle leaving the translucent appearance. Paralysis tests showed no effect on the saline or 0 FED treated individuals, but in venom injected individuals the percent of individuals paralyzed after 18 hours increased as the dose was increased. These results were significantly different using a chi square test of venom or saline alone injected treatments (P< 0.001, 2 = 29.95, 1df). The average time to paralysis of individuals paralyzed varied from 4-18h (N=36). The specific time to paralysis, in general, decreased (r= -0.40, P=0.01, F1,34=6.63) as the dose increased: 1.5 FED 12.33±3.18%, n=2; 2 FED 18±N/A, n=1; 2.5 FED 10.6±0.62%, n=10; 3 FED 10.5±1.5%, n=2; 5 FED 8.9±0.71%, n=20 (X + SEM, n).

Development was more frequently delayed as the FED was increased (Fig. 4.1).

At 10 FED the pupae did not develop into adults. As dose decreased, partial adult development occurred (Fig. 4.2). Finally as the FED was decreased to 1.25, individuals started to develop normally. One individual died prematurely in the 1.25 FED group, and was excluded from the analysis since it was likely due to natural causes or experimenter error.

DISCUSSION

Because their host range is not limited to hymenopterans females of M. digitata behave as generalists in terms of host selection. In fact, in the laboratory these parasitoids will readily accept a wide range of insects as hosts, including pupae of

Sarcophaga bullata (Diptera: Sarcophagidae). Females will even accept and lay eggs on

T. molitor, however, offspring seldom complete development (Deyrup unpublished).

Such a wide range of hosts necessitates venom with a broad spectrum of activity for subduing and/or manipulating all hosts encountered. Indeed, observations from natural

(Deyrup et al. 2003, Deyrup and Matthews 2003a) and our artificial venom injections revealed that both natural and factitious hosts were susceptible, albeit not equally, to venom from M. digitata. All susceptible insects responded to natural envenomation by becoming paralyzed, and developmental progression of paralyzed hosts was dependent on venom concentration. The fact that milked venom evoked the same host responses as natural envenomation demonstrates that venom alone elicits the previously reported disruption of host development (Balfour-Browne 1922), and it seems unlikely that other maternal secretions are involved.

Three mechanisms are known by which parasitoid wasps may prevent host molting and subsequent development normally induced by ecdysteroid peaks. The first is the wasp’s secretion, by the mother or larva, of juvenile hormone (JH), a lipid hormone, into the hemolymph of the host (Beckage and Riddiford 1982, Shopf et al. 1996).

Alternatively, such secretions may act to reduce the amount of juvenile hormone esterase, the enzyme that breaks down JH (Balgopal et al. 1995, Dover et al. 1994). These have

43 been shown to prevent ecdysteroid peaks (Beckage 1985). A second method for preventing an ecdysteroid peak is through the release of specialized cells, teratocytes, produced by the wasp embryo (Dover et al. 1994, Strand and Wong 1991). The third, found in many koinobionts, involves the disruption of molting by polydnaviruses that are injected along with the egg or eggs (Chelliah and Johnes 1990, Cusson et al. 2000, Johnes et al. 1985). The venom of such parasitoids has also been shown to abet the effects caused by the polydnavirus (Balgopal et al. 1995, Grossniklaus-Burgin et al. 1998).

Polydnavirus from the braconid wasp Ascogaster quadridentata does not survive in the host Cydia pomonella without injection of venom (Johnes et al. 1990). A study conducted by Stoltz et al. (1988) suggested that the venom helped in the uncoating of the polydnavirus. Venom, therefore, may have several roles in host molt regulation. These mechanisms may not be mutually exclusive, and one or more of them could be used by different kinds of parasitoid wasps.

In idiobionts the venom is thought to be the primary factor responsible for arrest of host development. An example is Euplectrus comstockii, a chalcidoid wasp that attacks several lepidopteran hosts (Coudron and Brandt 1996). Most of the data that has been collected on venom as the cause of low ecdysteroid levels has been from idiobionts.

Several studies show a direct link between venom in idiobionts and the prevention of host molting. By injecting Eulophus pennicornis venom into dishes containing isolated prothoracic glands, it was shown that the venom stopped production of ecdysteroids

(Marris et al. 2001). Coudron and Brandt (1996) characterized the venom protein in E. comstockii responsible for developmental delays in their hosts as a (66kDa) protein with two subunits of 33kDa or less.

How do venoms or other factors prevent molting? The mechanisms behind the absence of an ecdysteroid peak are physiological, and there have been three explanations proposed. One way of halting ecdysteroid production is to destroy or disable the gland that produces it, the prothoracic gland (PTG) (Dover et al. 1988). Another way is to disrupt the production pathway of ecdysteroids, and the third is to metabolize them after production (Dover et al. 1988). Taking into account the wide array of methods by which parasitoids prevent molting, it is not surprising that PTG destruction and pathway disruption are both used by different species of wasps in suppression of an ecdysteroid peak.

The injection of venom had a paralyzing effect in both species that we injected.

This was expected to be the case if the mealworms in the experiment in Deyrup et al.

(2003) were being paralyzed by the venom of the wasp. Host paralysis after parasitism has been reported in many eulophids including Eulophus viridulus, Aphelinus jucundus, and Microplectron fuscipennis (Clausen 1940, Parker and Smith 1933). Although the paralysis is probably due to the mother’s envenomation, it is hard to be confident without venom injection tests, in which the venom is removed from the mother wasp and injected by the researcher directly into the host. While looking for developmental arrests Coudron et al. (2000) discovered arrestment of activity when injecting an aqueous extract from the venom gland of the eulophid Necremnus breviramulus. This corresponds to the results of our experiment, and adds support to the generally accepted notion that it is the venom that causes the paralysis in the host.

Injecting paralyzing venom can have several benefits for wasps. In predatory wasps it can immediately immobilize the host for a secondary sting necessary for other

45 effects, as in Ampulex compressa, with its primary sting for paralysis and a second sting causing long-term sluggishness (Piek et al. 1984) It can be useful for immediate paralysis in situations in which the host has powerful jaws, as is the case for the tarantula-hunting wasps (Steiner 1986). A paralyzing sting can also be useful for making the prey suitable for feeding by larvae. A paralyzed spider cannot run or attack the slow-moving and poorly defended young of a mud dauber.

The time required for paralysis varies. A paralyzing effect can be instantaneous, or require 15 minutes or more to take effect as in the venom from Philanthus triangulum, a sphecid wasp that preys on honey bees (Piek et al. 1982). Similarly, whether the venom has a reversible effect or not also differs from wasp to wasp (Ferber et al. 1999, Fouad et al. 1994, Gnatzy 2001, Piek 1990, Piek et al. 1982, Steiner 1986,). The time it takes a host to recover from envenomation can be related to dose size or just the chemistry of the venom injected.

In contrast to the relatively extensive knowledge of aculeate wasp venom biochemistry, very little has been reported on paralyzing venom proteins in parasitic wasps. One exception is the ichneumonoid Bracon hebetor. Visser et al. (1983) characterized two proteins of 43.7 and 56.7 kDa that were responsible for the paralytic effects found in hosts. These proteins tend to be very labile and are sensitive to heat, freezing, and long term storage (Piek et al. 1982, Visser et al. 1983). Among aculeate wasps there is a wide array of sizes and structures of paralyzing venoms. They can have low weights as in the polyamines that are produced by some sphecids (Piek and Spanjer

1986, Schmidt 1982) that block postsynaptic cation channels or interfere with the reuptake of glutamate (Piek 1990). Alternatively, they can be as large as 56kDa (Piek and

Spanjer 1986). Many venom proteins lie somewhere in between. Acylpolyamine toxins have been found to be non-competitive antagonists of glutamate and nicotinic receptors

(Piek 1990). Several types of kinins including bradykinins found in Scoliidae and Thr6-

BK found in Pompilidae and Scoliidae (Konno et al. 2002) are implicated in the paralysis of the prey. In Campsomeris sexmaculata and Megascolia flavifrons kinins act by blocking the synaptic nicotinic transmissions (Piek 1990, Piek et al. 1987a, Piek et al.

1987b), and pompilidotoxins slow or block sodium channels (Kinoshita et al 2001).

In summary, the overall effect of paralytic venom is to affect the motor neurons involved in movement through ion channel blocking, receptor blocking, inhibition of neurotransmitter reuptake, neurotransmitter overload, or neurotransmitter depletion.

Because there are so many and varied wasp parasitoids, and because so few have had their venom studied, there are probably many more types of parasitoid venom proteins left to be discovered, possibly with unique methods of paralyzation. This adds to the importance of following up this research with studies into the composition and mode of action of M. digitata venom. Such research could also have applications in pest control.

Some venom effects are caused by proteins (Ferber et al. 1999, Kinoshita et al. 2001,

Konno et al. 2002, Piek 1990, Piek and Spanjer 1986), and proteins may be more useful for agricultural pest control than viruses. At the very least, venom technology might be useful in connection with the viruses.

ACKNOWLEDGEMENTS

This work was supported in part by a grant from the National Science Foundation to R.

W. Matthews.

REFERENCES

Balfour-Browne, M. A. 1922. On the life-history of Melittobia acasta, Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Balgopal, M. M., B. A. Dover, W. G. Goodman, and M. R. Strand. 1995. Parasitism by

Microplitis demolitor induces alteration in the juvenile hormone titers and

juvenile hormone esterase activity of its host, Pseudoplusia includens. Journal of

Insect Physiology 42: 337-345.

Beckage, N. E. 1985. Endocrine interaction between endoparasitic insects and their hosts.

Annual Review of Entomology 30: 371-413.

Beckage, N. E., and L. M. Riddiford. 1982. Effects of parasitism by Apanteles

congregatus on the endocrine physiology of the tobacco hornworm Manduca

sexta. General Comparative Endocrinology 47: 308-322.

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomology 5: 14-22.

Chelliah, J., and D. Johnes. 1990. Biochemical and immunological studies of proteins

from polydnavirus Chelonus sp. near curvimaculatus. Journal of General

Virology 71: 2353-2359.

Clausen, C. P. 1940. Entomophagous Insects. McGraw-Hill. New York.

Coudron, T. A., and S. L. Brandt. 1996. Characteristics of a developmental arrestant in

the venom of the ectoparasitoid wasp Euplectrus comstockii. Toxicon 34: 1431-

1441.

Coudron, T. A., M. M. K. Wright, B. Puttler, S. L. Brandt, and W. C. Rice. 2000. Effect

of the ectoparasite Necremnus breviramulus (Hymenoptera: Eulophidae) and its

venom on natural and factitious hosts. Annals of the Entomological Society of

America 93: 890-897.

Coudron, T. A. and B. Puttler. 1988. Response of natural and factitious hosts to the

ectoparasite Euplectrus plathypenae (Hymenoptera, Eulophidae). Annals of the

Entomological Society of America. 81: 931-937.

Cusson, M., M. Laforge, D. Miller, C. Cloutier, and D. Stoltz. 2000. Functional

significance of parasitism-induced suppression of juvenile hormone esterase

activity in developmentally delayed Choristoneura fumiferana larvae. General

and Comparative Endocrinology 117: 343-354.

Dahms, E. C. 1984. A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum. 21: 337-360.

Deyrup, L. D., M. Deyrup, and R. W. Matthews. 2003. Paralyzation and developmental

delay of a factitious host by Melittobia digitata (Hymenoptera: Eulophidae)

Journal of Entomological Science 38: 703-705.

Deyrup, L. D., and R. W. Matthews. 2003a. A simple technique for milking the venom of

a small parasitic wasp, Melittobia digitata (Hymenoptera: Eulophidae). Toxicon

42: 217-218.

Deyrup L. D., and R. W. Matthews. 2003b. Host preference and utilization by Melittobia

digitata (Hymenoptera: Eulophidae) in relation to mating status. Journal of

Entomological Science 38: 682-687.

Dover, B. A., D. H. Davies, and S. B. Vinson. 1988. Degeneration of last instar Heliothis

virescens prothoracic glands by Campoletis sonorensis polydnavirus. Journal of

Invertebrate Pathology 51: 80-91.

Dover, B. A., A. Menon, R. C. Brown, and M. R. Strand. 1994. Suppression of juvenile

hormone esterase in Heliothis virescens by Microplitis demolitor. Journal of

Insect Physiology 41: 809-817.

Doury, G., D. Rojasrousse, and G. Periquet. 1995. Ability of Eupelmus-orientalis

ectoparasitoid larvae to develop on an unparalyzed host in the absence of female

stinging behavior. Journal of Insect Physiology. 41: 287-296.

Edwards, C.J. and D. H. Pengelly. 1966. Melittobia chalybii Ashmead (Hymenoptera:

Eulophidae) Parasitizing Bombus fervidus Fabricius (Hymenoptera: Apidae).

Proceedings of the Entomological Society of Washington 96: 98-99.

Ferber, M., C. Consoulas, and W. Gnatzy. 1999. Digger wasp vs. cricket: Immediate

actions of the predator’s paralytic venom on the CNS of the prey. Journal of

Neurobiology 38: 323-337.

Flanders, S. E. 1943. The role of mating in the reproduction of parasitic Hymenoptera.

Journal of Economic Entomology 36: 802-803.

Fouad, K., F. Libersat, and W. Rathmayer. 1994. The venom of the cockroach-hunting

wasp Ampulex compressa changes motor thresholds: A novel tool for studying the

neural control of arousal. Zoology 98: 23-24.

Gnatzy, W. 2001. Digger wasp vs. cricket: (Neuro-) biology of a predator-prey-

interaction. Zoology – analysis of complex systems 103: 125-139.

Grossniklaus-Burgin, C., R. Pfister-Wilhelm, V. Meyer, K. Treiblmayr, and B. Lanzrein.

1998. Physiological and endocrine changes associated with polydnavirus/venom

in the parasitoid-host system Chelonus inanitus-Spodoptera littoralis. Journal of

Insect Physiology 44: 305-321.

Hinton, H. 1946. The gin traps of some beetle pupae: a protective device that appears to

be unknown. Transactions of the Royal Entomological Society of London 97:

473-96

Jones, D., S. Sreekrishna, M. Iwaya, J. N. Yang, and M. Eberely. 1985. Comparison of

viral ultrastructure and DNA banding patterns from the reproductive tracts of

eastern and western hemisphere Chelonus spp. (Braconidae: Hymenoptera).

Journal of Invertebrate Pathology 47: 105-115.

Jones, D., T. Taylor, R. Farkas, J. Chelliah, B. Haene, J. Brown, and D. Reed-Larsen.

1990. Intercession of parasitic wasps (Cheloninae) in host developmental and

biochemical pathways. IN Advances in Invertebrate Reproduction 5 (eds. Hoshi

M, and Yamashita O.) Elsevier Science Publishers B.V. 157-162.

Kinoshita, E., E. Maejima, K. Yamaoka, K. Konno, N. Kawai, E. Shimizu, H. Nakayama,

and I. Seyama. 2001. Novel wasp toxin discriminates between neuronal and

cardiac sodium channels. Molecular Pharmacology 59: 1457-1463.

Konno, K., M. S. Palma, I. Y. Hitara, M. A. Juliano, L. Juliano, and T. Yasuhara. 2002.

Identification of bradykinins in solitary wasp venoms. Toxicon 40: 309-312.

Krombein, K.V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington D.C.

Marris, G.C., R. J. Weaver, J. Bell, and J. P. Edwards. 2001. Venom from the

ectoparasitoid wasp Eulophus pennicornis disrupts host ecdysteroid production by

regulation of host prothoracic gland activity. Physiological Entomology 26: 229-

238.

Maeta, Y., and S. Yamane. 1974. Host records and bionomics of Melittobia japonica

Masi (Hymenoptera: Eulophidae). Bulletin Tohoku National Agricultural

Experiment Station 47: 115-131.

Parker, H. L., and H. D. Smith. 1933. Eulophus viridulus a parasite of Pyrausta nubilalis

Hubn. Annals of the Entomological Society of America 26: 21-39.

Piek, T. 1990. Neurotoxins from venoms of the Hymenoptera – 25 years of research in

Amsterdam. Comparative biochemistry and physiology C-Pharmacology

Toxicology and Endocrinology 96: 223-233.

Piek, T., W. Spanjer, R. D. Veldsema-Currie, T. Van Groen, N. De Haan, and P. Mantel.

1982. Effect of venom of the digger wasp Philanthus triangulum F. on the sixth

abdominal ganglion of the cockroach. Comparative Biochemistry and Physiology

C-Pharmacology toxicology and endocrinology. 71: 159-164.

Piek, T., J. H. Visser, and R. L. Veenendaal. 1984. Change in behaviour of the cockroach,

Periplaneta americana, after being stung by the wasp Ampulex compressa.

Entomologia Experimentalis et Applicata. 35: 192-203.

Piek, T., and W. Spanjer. 1986. Chemistry and pharmacology of solitary wasp venoms

In: Venoms of the Hymenoptera: Biochemical, Pharmacological and Behavioural

Aspects. (ed. Piek T.) Academic Press Inc. London.

Piek, T., B. Hue, L. Mony, T. Nakajima, M. Pelhate, and T. Yasuhara. 1987a. Block of

synaptic transmission in insect CNS by toxins from the venom of the wasp

Megascolia flavifrons Fab. Comparative Biochemistry and Physiology C-

Pharmacology Toxicology and Endocrinology 87: 287-295.

Piek, T., B. Hue, M. Pelhate, and L. Mony. 1987b. The venom of the wasp Campsomeris

sexmaculata F. blocks synaptic transmission in insects CNS. Comparative

Biochemistry and Physiology C-Pharmacology Toxicology and Endocrinology

87: 283-286.

Quick, D. L. J. 1997. Parasitic Wasps. Chapman and Hall. London.

Rivers, D. B. 2004. Evaluation of host responses to envenomation as a means to assess

ectoparasitic pteromalid wasp’s potential for controlling manure-breeding flies.

Biological Control. 30: 181-192.

Rivers, D. B., M. Genco, and R. A. Sanchez. 1999. In vitro analysis of venom from the

wasp Nasonia vitripennis: Susceptibility of different cell lines and venom-induces

changes in plasma membrane permeability. In Vitro Cellular and Developmental

Biology-Animal. 35: 102-120.

Rivers, D. B., W. F. Hink, and D. L. Denlinger. 1993. Toxicity of the venom from

Nasonia-vitripennis (Hymenoptera, Pteromalidae) toward fly hosts, nontarget

insects, different developmental stages, and cultured insect cells. Toxicon. 31:

755-765.

Schmidt, J. O. 1982. Biochemistry of insect venoms. Annual Review of Entomology 27:

339-368.

Schopf, A., C. Nussbaumer, H. Rembold, and B. D. Hammock. 1996. Influence of the

braconid Glyptapanteles liparidis on the juvenile hormone titer of its larval host,

the gypsy moth Lymantria dispar. Archives of Insect Biochemistry and

Physiology 31: 337-351.

Steiner, A. L. 1986. Stinging behaviour of solitary wasps. In: Venoms of the

Hymenoptera Biochemical, Pharmacological and Behavioural Aspects. (ed. Piek

T.) Academic Press Inc. London.

Stoltz, D. B., D. Guzo, E. R. Belland, C. J. Lucarotti, E. A. MacKinnon. 1988. Venom

promotes uncoating in vitro and persistence in vivo of DNA from a braconid

polydnavirus. Journal of General Virology 69: 903-907.

Strand, M. R., and E. A. Wong. 1991. The growth and role of Microplitis demolitor

teratocytes in parasitism by Pseudoplusia includens. Journal of Insect Physiology

37: 503-515.

Visser, B. J., W. T. Labruyere, W. Spanjer, and T. Piek T. 1983. Characterization of 2

paralyzing protein toxins (A-MTX and B-MTX), isolated from a homogenate of

the wasp Microbracon hebetor (Say). Comparative Biochemistry and Physiology

B- Biochemistry and Molecular Biology 75: 523-530.

FIGURE LEGENDS

Fig. 4.1. Mealworm development following injection with differing amounts of venom.

(N =19 [1 prepupae died prematurely in 1.25 FED treatment].) ( = Fully Developed

Adults, = Incompletely Developed Adults, = Undeveloped Pupae)

Fig. 4.2. Mealworm beetles in different states of development. From left to right: undeveloped pupae (two pupae injected with 10 FED), incompletely developed adults

(two pupae injected with 2.5 FED), and fully developed adults (two pupae injected with 0

FED). Note the adult head and legs of the middle group while the wing buds and abdomen retain the pupal condition. This is not a normal progression of imaginal development.

100% 90%

- s

l 80% a

u 70% d i v

i 60% d n

I 50%

o 40%

n 30% e c r

e 20% P 10% 0% 10 5 2.5 1.25 none Dose (Female Equivalents)

Figure 4.1

Figure 4.2

CHAPTER 5

COOPERATIVE CHEWING IN MELITTOBIA DIGITATA DAHMS, A

GREGARIOUSLY DEVELOPING PARASITOID WASP, IS STIMULATED BY

STRUCTURAL CUES AND A PROBABLE PHEROMONE IN CRUDE VENOM

EXTRACT

Deyrup, L. D., R. W. Matthews, and J. M. González. Journal of Insect Behavior 18: 293- 304. Reprinted here with permission of publisher.

ABSTRACT

To disperse after mating, female Melittobia digitata Dahms (Hymenoptera:

Eulophidae) tunnel through the walls of their host’s nest. We found that M. digitata use chemical and structural cues to identify locations for chewing exit holes. An experimental combination of milked venom and artificial pits on the inner surface of rearing containers elicited a stronger response than either stimulus alone. We suggest that the venom- associated cue may be a pheromone that facilitates mutual attraction, aggregation, and focused chewing, and that these behaviors may have arisen from behaviors associated with the initial stages of host attack. This apparently cooperative behavior promises insights into the possible evolutionary origins of components of eusocial behavior.

KEY WORDS: Chemical communication; aggregation; Eulophidae; sociality

INTRODUCTION

The study of insect social behavior and its evolution traditionally has centered on eusocial species. Although this has been productive, we may be missing part of the story if we fail to include non-social and presocial species (Wilson, 1971; Choe and Crespi,

1997). Melittobia digitata Dahms (Hymenoptera: Eulophidae) are small parasitic wasps that gregariously develop in a wide range of hosts, but most commonly on solitary bees and wasps, and especially mud-dauber wasps (Hymenoptera: Sphecidae) (e.g., Krombein,

1967; Maeta and Yamane, 1974; González and Matthews, 2002). Although Melittobia are not known to have reproductive division of labor, they share some traits with social insects, including overlapping adult generations (Dahms, 1984) and the ability to produce variable adult phenotypes depending on conditions (Schmieder, 1933; Freeman and

Ittyeipe, 1982; Dahms, 1984; Consôli and Vinson, 2002, 2004). These conditions led

Malyshev (1968) to regard Melittobia as ‘primitively social’.

Upon finding a host, a female Melittobia lays hundreds of eggs (Balfour-Browne,

1922; Buckell, 1928; Dahms, 1984; González et al., 1999). When the offspring emerge as adults three or more weeks later, they find themselves in total darkness, enclosed in a tough host cocoon surrounded by the thick walls of a mud nest, hardened resins, or similar barriers. The remarkable ability of various Melittobia species to escape through both natural and artificial rearing structures and materials has been noted repeatedly for over 100 years (e.g., Howard, 1892; Buckell, 1928; Cowley, 1961; Torchio, 1963;

Hermann, 1971; Maeta and Yamane, 1974; González and Terán, 2001).

Working with M. hawaiiensis, Donovan (1976) provides the only published observation of females cooperatively chewing exit holes in polystyrene boxes; he does not discuss possible cues eliciting the behavior, but says “it seems possible that the plastic may have been slightly softened by secretion from the mouthparts.”

Rearing M. digitata in polystyrene boxes in our laboratory, we have routinely observed groups of females clustered on the interior, typically on the box top or side; at these locations, one female in the cluster will bite at the plastic for a while, enlarging a central depression (“chew pit”), then step aside to be succeeded by another in the retinue

(Fig. 5.1). Under magnification, the resultant mandibular scratches (“chew marks”) are clearly evident.

Upon close examination, we noticed that each chew pit usually contained a clearly defined sting trace, evident as a tiny line in the plastic (Fig. 5.1). We have also directly observed females stinging the inside of the tough walls of our boxes (Fig. 5.2); the process typically requires several minutes.

As in many other parasitoid wasps, female Melittobia inject venom when attacking a host. For a related study, we were developing techniques to extract Melittobia venom (Deyrup and Matthews, 2003). While we were extracting venom from one female, sometimes another female would repeatedly antennate the exuding venom. The combination of sting marks in the chew pits, associated chewing by multiple females at the same location, and evident interest in freshly milked venom led to the hypothesis that the crude milked venom could be – or contain – a pheromone that elicits chewing. This study was undertaken to test this hypothesis and to examine other possible stimuli for chewing behavior.

METHODS

Insect Rearing

Female M. digitata used in this study were derived from laboratory stock originally established from field-collected nests of Trypoxylon politum Say collected in

Georgia, Kentucky, and Illinois. Each of our cultures was established by placing three mated females with a mature T. politum prepupa. (Approximately three weeks later, this usually yields 500-600 female M. digitata that mate shortly after emergence.) Five to seven days after female emergence, cultures were chilled in a refrigerator for 15 minutes for easier handling. Groups of 250-300 females were removed and transferred into 45 new transparent, tightly lidded, polystyrene holding boxes measuring 50 mm x 25 mm x

18 mm (Carolina Biological Supply Co., Cat. No. ER-14-4584).

Laboratory Experiments

For each experiment, we established box-lid arenas, using a lid-sized paper template with punched holes to assure uniformity of circles drawn on the lid exterior with a fine-point permanent marker. Treatments were applied to the inner side of the transparent lid within the marked circles. These treated lids were used to replace the holding lids over the previously isolated female groups, and the boxes were placed in complete darkness for 12 (expts. 1, 2) or 8 (expt. 3) hours. Then the lids were removed, and the presence or absence of chew marks in each circle was recorded; these marks were distinctive and easily recognized, as there are no other similar marks in the plastic.

Experiment 1: Chemical cues. To investigate whether chemical cues might elicit chewing at a sting mark, we simultaneously compared wasp responses to milked venom, a squashed head, a female gaster, and a blank control within the inner lid areas corresponding to exterior circles marked with a template with four punched circles, 5 mm diameter, situated 3 mm from each of the two sides of each corner. Fifteen replicates were prepared, and relative treatment and control locations were randomized.

Crude venom extract was obtained from unused females from the same source cultures, following published procedures (Deyrup and Matthews, 2003). Collected onto the shaft of a clean number 2 insect pin, venom from two successively milked females was combined to form a single dose that was applied to the treated area within 15 minutes of collection.

Head crushes were made by mashing a single live female head with the tip of an insect pin within the interior of the transparent lid at a location corresponding to the exterior-marked circle. These were included to test the possibility that an oral secretion left by a chewing female in a chew pit might stimulate others to chew at that location.

Each gaster rub was produced by three successive rubs of the ventral abdomen and hind tarsi on inside of the lid within a circle, using a living wasp held in forceps by its thorax. This tested the possibility that some other odor(s) from the stinging wasp’s abdomen might elicit chewing.

After 12 hours in complete darkness, the lids were removed, presence or absence of chew marks on each circle was recorded, and results were analyzed using a Cochran

Test (Statistica 6.0).

Experiment 2. To examine whether chewing could be elicited in response to a physical stimulus, we simulated natural chewed pits by forcibly pushing the tip of an insect pin into the surface of the plastic to create a visible depression. Using the template method described for Expt. 1, ten equidistantly spaced 5 mm diameter circles were drawn in two parallel rows on the outside of each of 15 lids. By coin toss with the restriction of equal numbers of each condition, each circle was designated either to receive one pin indentation (artificial pit) or to remain pit-free.

After each prepared lid was placed over a group of 250-300 females for 12 hours in darkness, lids were removed to record presence or absence of chewing in each circle.

To avoid potential bias due to pseudoreplication, an average chew score for each lid replicate was calculated, giving a value between 1 and 0. (For example, if chew marks were found in 2 of 5 simulated pit treatments, a score of 0.4 was assigned for the pit treatment for that lid.; if one of the 5 blank circles on the same lid contained chew marks, this resulted in a score of 0.2 for the control group.) To compare within replicates, the scores for the treatment and blank control circles were analyzed using a Wilcoxon

Matched Pairs Test (Statistica 6.0).

Experiment 3. To test for the possibility of an interaction between chemical and physical stimuli, we prepared four-circle lid arenas as for Expt. 1, and compared the response of the wasps to four simultaneously presented treatments: a venom dose alone, an artificial pit alone, a venom dose in an artificial pit, and a blank. After lids were exposed to 250-300 females for 8 hours in the dark, the presence or absence of chewing in each of the circles was recorded. Data were analyzed using a Cochran Test (Statistica

6.0).

RESULTS

Experiment 1. Assessment of the number of circles with evident chew marks after venom, head squash, gaster rub, and blank treatments (Table 5.1) showed that some component of the venom extract was important in eliciting chewing, whereas the other tested sources of chemical cues tested were not. Pair-wise comparisons were highly significantly different (venom versus head squash, P=0.002; venom versus gaster rub,

P=0.002).

Experiment 2. Comparison of the average number of chewed circles with artificial pits to the average for the no-pit circles in each replicate (Table 5.2) revealed that M. digitata females chewed significantly more in the areas with a pit in them (P=

0.002).

Experiment 3. Pair-wise comparisons of the treatments were all significantly different (venom and pit versus venom, P<0.001; venom and pit versus pit alone, P<0.05; and venom and pit versus blank, P<0.001), indicating that the combination of both chemical and physical cues elicits a stronger chewing response by females than either cue presented alone (Table 5.3).

DISCUSSION

The data clearly show that a chemical contained in M. digitata crude venom, or more conservatively, the product of venom milking, elicits focused female chewing.

Emerged female wasps are jointly responsive to some odor component, and their individual attraction to such spots results in aggregation around the spot.

Venom-associated chemical signals are not unusual in eusocial Hymenoptera.

For example, attractants and alarm pheromones have been reported in the venoms of several species of Vespula yellowjackets (Landolt et al., 1995; Weston et al., 1997), Apis mellifera L. honey bees (Butler and Free, 1952; Lensky and Cassier, 1995), and

Solenopsis fire ants (Van der Meer et al., 1980, Cruz-Lopez et al., 2001). Interestingly, among parasitic wasps, Holler et al. (1994) have shown that the megaspilid Dendrocerus carpenteri Curtis uses juvenile hormone in its venom as a pheromone to mark hosts.

In this study we found that a simulated chew pit (a pin indentation) would stimulate chewing behavior by M. digitata females that encounter it (Table 5.2).

Structural cues that serve as stimuli for different behaviors also are widespread in the insects, as they are throughout the animal kingdom. For example, the species-specific shape and architecture of large and complex termite mounds result from additive effects of generations of termites responding to structural cues left by their predecessors rather than to direct communication among nest-mates (Wilson, 1971).

The existence of these focused chewing behaviors raises the question of their origin. Among the ectoparasitic wasps, stinging and feeding are widespread behaviors that usually occur after rather than before dispersal, as part of the process of subduing a host and nurturing egg development (Quicke, 1997). When a female Melittobia encounters a host, she inserts her ovipositor, withdraws it, backs up while drumming with her antennae, and returns to feed upon the protein-rich fluids exuding from the punctured host (Dahms, 1984). In M. digitata, host stinging and host feeding appear to have been

66 modified or reorganized to fit a different context – escape from the barriers provided by the natal host – and have elaborated into a complex behavioral suite involving mutual attraction, aggregation, and apparent cooperation. Support for this recontextualization argument comes from our observations that a lid-stinging female subsequently antennates the sting site, applies her mouthparts directly to the spot, and sometimes consumes the residual fluid (see Fig. 5.2).

Chemicals present in the venom may have become associated with feeding and may have come to serve as a marker for feeding sites. If an emerging ancestral female by chance stung a nest wall in the process of attempting to disperse, other females that associated venom chemicals with host feeding may have been attracted. Chewing associated with the process of attempting to feed might open an exit. If these behavioral tendencies had a genetic component, the increased likelihood of escape would provide a selective advantage. Furthermore, as Donovan (1976) noted for M. hawaiiensis, “cooperative boring activity between female parasites would reduce to a minimum the effort required from each female to escape, and the adaptive significance of this behaviour would appear to be obvious.”

Such chemically facilitated escape behavior likely would be reinforced by any accompanied tendency to recognize and exploit physical cues such as pits, cracks, or irregularities commonly present in a host’s nest wall. Our results show that the combination of a physical and chemical cue is a stronger stimulus for eliciting chewing than either cue alone (Table 5.3).

While speculative, such a scenario seems plausible. In Melittobia, there is the additional factor of kinship due to sib-mating; wasps emerging from a host’s cell are

67 likely to be highly related (Hamilton, 1967). This could further facilitate selection and maintenance of focused communal chewing. Genes of any mutant selfish individual that declined to chew in favor of simply exploiting the chewers might initially increase in frequency but would eventually disappear because emerging clutches carrying the mutant gene would fail to escape the host nest. Indeed, in field-collected mud dauber nests we occasionally have found Melittobia-parasitized cells containing an entire clutch of dead adults entombed (unpublished observations).

Definitions of cooperation are many (e.g., Dugatkin, 2002, and references therein), but the focus of this paper is on cues for the communal chewing, rather than quantitative support for defining female chewing circles (such as that shown in Fig. 5.1) as examples of fully developed cooperation. Whether the female Melittobia aggregations that arise in response to chemical and physical cues go beyond simple facultative mutual attraction to a common stimulus remains to be decided by further study of such matters as the extent of interaction among group members and the duration of their association.

It is worth noting that over the past five years, we have examined hundreds of empty mud dauber cocoons from which Melittobia has emerged; not one has had more than one escape hole. Whether such focus provides circumstantial evidence for active cooperation between Melittobia females, simply a strong mutual attraction to a single chew site, or merely opportunistic escape deserves further study. Seemingly similar cooperation in exit-hole chewing occurs the agaonid fig wasps (Ramírez, 1970) another group of chalcidoids in which offspring clutches are highly related (Hamilton, 1979).

Interestingly, among fig wasps, males are the sex that cooperatively chews exit tunnels from the fig syconium.

Cooperative behaviors among closely related individuals are one of the hallmarks of eusocial behavior in insects (Wilson, 1971). Like many other social behaviors, their origins can be difficult to study in highly derived form. Perhaps in the behaviors of species such as Melittobia and fig wasps we can glimpse antecedents that with continuing studies we may come to better understand.

ACKNOWLEDGMENTS

We thank Mark Deyrup, Janice Matthews, and Christian Silva-Torres for their input into this study, including discussions and helpful reviews of various drafts.

Comments from two anonymous reviewers greatly improved this paper. This work was supported by NSF Grant 0088021, R.W. Matthews, principal investigator.

REFERENCES

Balfour-Browne, M. A. (1922). On the life-history of Melittobia acasta Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Buckell, E. R. (1928). Notes on the life-history and habits of Melittobia chalybii

Ashmead (Chalcidoidea: Elachertidae). Pan-Pacif. Entomol. 5: 14-22.

Butler, C. G., and Free, J. B. (1952). The behaviour of worker honeybees at the hive

entrance. Behaviour 4: 262-292.

Choe, J. C., and Crespi, B. J. (1997). Introduction. In Choe, J. C., and Crespi, B. J. (eds.)

The Evolution of Social Behavior in Insects and Arachnids. Cambridge University

Press, New York, pp. 1-7.

Consôli, F., and Vinson, S. B. (2002). Clutch size, development and wing morph

differentiation of Melittobia digitata Dahms (Hymenoptera: Eulophidae).

Entomol. Exper. Applic. 102: 135-143.

Consôli, F., and Vinson, S. B. (2004). Wing morph development and reproduction of the

ectoparasitoid Melittobia digitata: nutritional and hormonal effects. Entomol.

Exper. Applic. 112: 47-55.

Cowley, D. R. (1961). The associates of Pison spinolae Shuckard (Hymenoptera:

Sphecidae). N. Z. Entomol. 2: 45-46.

Cruz-Lopez, L., Rojas, J. C., De la Cruz-Cordero, R., and Morgan, E. D. (2001).

Behavioral and chemical analysis of venom gland secretion of queens of the ant

Solenopsis geminata. J. Chem. Ecol. 27: 2437-2445.

Dahms, E. C. (1984). A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Mem. Queensland Mus. 21: 337-360.

Deyrup, L. D., and Matthews, R. W. (2003). A simple technique for milking the venom

of a small parasitic wasp, Melittobia digitata (Hymenoptera: Eulophidae). Toxicon

42: 217-218.

Donovan, B. J. (1976). Co-operative material penetration by Melittobia hawaiiensis

(Hymenoptera: Eulophidae) and its adaptive significance. N. Z. Entomol. 6: 192-

193.

Dugatkin, L. A. (2002). Animal cooperation among unrelated individuals.

Naturwissenshaften 89: 533-541.

Freeman, B. E., and Ittyeipe, K. (1982). Morph determination in Melittobia, a eulophid

wasp. Ecol. Entomol. 7: 355-363.

González, J. M., and Matthews, R. W. (2002). Development and sex ratio of Melittobia

australica and M. digitata (Hymenoptera: Eulophidae) on Megachile rotundata

(Hymenoptera: Megachilidae) and Trypoxylon politum (Hymenoptera:

Sphecidae). Great Lakes Entomol. 35: 85-91.

González, J. M., and Terán, J. B. (2001). Dispersión, búsqueda y acceso al hospedador

por Melittobia acasta (Hymenoptera: Eulophidae). Bol. Cent. Inv. Biol.

(Maracaibo) 35: 52-64.

González, J. M., Terán, J. B., and Badaraco, M. T. (1999). Efecto de la densidad

poblacional sobre la sobrevivencia en Melittobia acasta (Hymenoptera:

Eulophidae). Bol. Cent. Inv. Biol. (Maracaibo) 33: 197-210.

Hamilton, W. D. (1967). Extraordinary sex ratios. Science 156: 477-488.

Hamilton, W. D. (1979). Wingless and fighting males in fig wasps and other insects. In

Blum, M.S., and Blum, N. A. (eds.), Sexual Selection and Reproductive

Competition in Insects. Academic Press, New York, pp. 167-220.

Hermann, L. D. (1971). The mating behavior of Melittobia chalybii. MSc. Thesis,

University of Georgia, Athens.

Holler, C., Bargen, H., Vinson, S. B., and Witt, D. (1994). Evidence for the external use

of juvenile hormone for host marking and regulation in a parasitic wasp,

Dendrocerus carpenteri. J. Insect Physiol. 40: 317-322.

Howard, L. O. (1892). The habits of Melittobia. Proc. Entomol. Soc. Washington 2: 244-

249.

Krombein, K. V. (1967). Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington, DC.

Landolt, P. J., Heath, R. R., Reed, H. C., and Manning, K. (1995). Pheromonal mediation

of alarm in the eastern yellowjacket (Hymenoptera, Vespidae). Fla. Entomol. 17:

101-108.

Lensky, Y., and Cassier, P. (1995). The alarm pheromones of queen and worker honey

bees. Bee World 76: 119-129.

Maeta, Y., and Yamane, S. (1974). Host records and bionomics of Melittobia japonica

Masi (Hymenoptera: Eulophidae). Bull. Tohoku Natl. Agric. Exp. Sta. 47: 115-

131.

Malyshev, S. I. (1968). Genesis of the Hymenoptera and the Phases of their Evolution.

Methuen & Co., London.

Quicke, D. L. J. (1997). Parasitic Wasps. Chapman and Hall, London.

Ramírez, W. B. (1970). Taxonomic and biological studies of neotropical fig wasps

(Hymenoptera: Agaonidae). Univ. Kansas Sci. Bull. 69:1-44

Schmieder, R. G. (1933). The polymorphic forms of Melittobia chalyibii Ashmead and

the determining factors involved in their production (Hymenoptera: Chalcidoidea,

Eulophidae). Biol. Bull. 65: 338-352.

Torchio, P. F. (1963). A chalcid wasp parasite of the alfalfa leaf cutting bee. Utah Farm

Home Sci. 24: 70-71.

Van der Meer, R. K., Glancey, B. M., Lofgren, C. S., Glover, A., Tumlinson, J. H., and

Rocca, J. (1980). The poison sac of the red imported fire ant queens; source of a

pheromone attractant (Hymenoptera: Formicidae). Ann. Entomol. Soc. Am. 73:

609-612.

Weston, R. J., Woolhouse, A. D., Spurr, E. B., Harris, R. J., and Suckling, D. M. (1997).

Spiroacetals and other venom constituents as potential wasp attractants. J. Chem.

Ecol. 23: 553-558.

Wilson, E. O. (1971). The Insect Societies. Harvard University Press, Cambridge, MA.

Table 5.1. Possible Chemical Cues: Effectsa of treatment on M. digitata Chewing Within

Treated Lid Circles (N = 15)

Treatment Number of replicates

with chew marks

Milked venom 10

Head squash 0

Gaster rub 0

Blank 0 aQ = 30.000; df = 3; P<0.0001.

Table 5.2. Possible Physical Cues: Effectsa of A Simulated Pit on Presence of M. digitata

Chewing Within Treated Lid Circles (N = 15)

Treatment Mean ± standard deviation for all replicates

Simulated chew pit 0.400 ± 0.302

No pit (control) 0.000 ± 0.000

aZ = 3.0594; df = 1; P = 0.002.

Table 5.3. Effectsa of Chemical and Physical Cues, Alone and In Combination, on M. digitata Chewing Within Treated Lid Circles (N = 15)

Treatment Number of replicates with chew marks

Milked venom 3

Artificial chew pit 4

Milked venom + artificial chew pit 15

Blank (control) 0

aQ = 29.76923; df = 3; P<0.001.

FIGURE LEGENDS

Fig. 5.1. A group of female M. digitata clustered about a potential exit hole being chewed by one female (at bottom of the group). Females take turns chewing at the spot to deepen the hole. The abundant small spots on the plastic lid are defecation specks. Female body length is ca. 1.3 mm. Inset: lateral view of a typical chewed spot in the plastic lid, showing the sting trace (arrow).

Fig. 5.2. Behavioral sequence of a female M. digitata stinging the interior surface of a polystyrene rearing box. A) Initial probe with ovipositor tip; B) ovipositor (arrow) inserted into the plastic; C) female applying mouthparts to the sting site.

Fig. 5.1

Fig. 5.2

CHAPTER 6

ESCAPE CHEWING IN RESPONSE TO ALKALINE GLAND AND VENOM

RESERVOIR MARKED SPOTS AND CROSS-ATTRACTANCY OF MILKED

VENOM ACROSS SPECIES GROUPS IN A GENUS OF PARASITIC WASP,

MELITTOBIA (HYMENOPTERA: EULOPHIDAE)

Deyrup, L. D., and R. W. Matthews. To be submitted to Journal of Hymenoptera

Research

ABSTRACT

Melittobia are a genus of small idiobionts in the family Eulophidae. Following emergence, adult females form circles in which they cooperate to chew escape holes. Dry milked venom has been shown to elicit chewing in M. digitata. Here we investigated whether a related species (M. femorata) chewed in response to compounds in its dissected venom reservoir and alkaline-gland., and if venom milked from a member of another species group (M. australica) would elicit chewing in M. digitata. Melittobia femorata chewed significantly more at gland and reservoir-marked spots than at controls (P

<0.0001, Q =17.0000, 1 df). To examine the venom’s effect across species we marked spots with M. australica milked venom, and introduced female M. digitata wasps. M. australica-marked spots elicited chewing similar to that elicited by that of M. digitata- marked spots (P<0.3173, Q =1.0, 1 df), and response to either’s venom was significantly different from blank controls (P <0.0047, Q =8.0, 1 df) and (P <0.0143, Q =6.0, 1 df) respectively. Possible reasons for the lack of specificity in a pheromone for chewing are discussed.

KEY WORDS: Cooperation, escape, pheromone, venom, Melittobia femorata, Melittobia australica, Melittobia digitata

INTRODUCTION

Melittobia are a genus of small gregarious parasitic wasps present on every continent except Antarctica (Balfour-Browne 1922, Buckell 1928, Dahms 1984b). They are commonly found attacking mud dauber prepupae and their associates (Matthews

1997), but may attack a wide range of solitary bees and wasps (Balfour-Browne 1922,

Krombein 1967).

When attacking a mud dauber wasp they face the daunting task of escaping from the thick walled mud nest after emerging as adults, yet females do not have noticeably well developed mandibles. Donovan (1976) observed M. hawaiiensis females circled around another female chewing a pit, and speculated that they cooperated in chewing out.

Such cooperative chewing has been observed in several Melittobia species (Deyrup unpublished).

Deyrup et al. (2005) reported that chewed pits invariably had associated sting marks and showed that a putative pheromone in the milked venom of M. digitata elicited chewing from other females. Because similar chew pits made by other species of

Melittobia also typically show sting marks in their centers (Deyrup unpublished), we decided to investigate whether extracted venom components would elicit chewing in a closely related species, M. femorata (Dahms 1984a).

Such chewing, if demonstrated could be in response to a normal constituent of venom, or a rapidly changing molecule. Regardless, it is difficult to envision selection pressure sufficient to cause evolutionary divergence in such a cue, since there appear to be no negative effects of cooperative escape chewing, even among unrelated females.

However, it is first necessary to verify whether cross species attractancy occurs to the venom of different species. Thus, the second objective of this study was to test M. digitata’s response to venom from M. australica, a species belonging to a different species group (Dahms 1984a).

METHODS

General methods follow those described in Deyrup et al. (2005). Melittobia australica responded to the venom milking procedures described in Deyrup and

Matthews (2003), yielding adequate amounts of venom for the experiment. However, M. femorata does not respond to the venom milking technique. Therefore, as an alternative we dissected the lower reproductive tract of the female in insect saline. While there are many possible pheromone sources in the female reproductive system, the two most likely are the alkaline gland and venom reservoir. These were separated from the ovipositor and combined for use in the experiment. Since milked venom used in previous work could possibly be a combination of the fluids contained in both organs we decided to combine them for this experiment.

As described in Deyrup et al. (2005), 20 plastic box lids were prepared for the first set of experimental treatments by making four pin indentations, one in each corner of the inner side. We then smeared the combined alkaline gland and venom reservoin dissected from a single female of M. femorata into one pin indentation and repeated this using a fresh female applied to the pit on the opposite corner. The other two pits served as controls for chewing stimulated by the pit alone as in Deyrup et al. (2005). Treated lids

83 were then placed on 20 boxes of 250-300 1-3 day old mated M. femorata females and left for 12 hours in complete darkness at 25°C, after which they were examined for evidence of chewing at each of the four pits.

To determine if M. digitata would be stimulated to chew by M. australica venom, we set up a second series of boxes. Three corner circles were drawn on the lids as in

Deyrup et al. (2005) and randomly assigned one of three treatments. One circle received

1 FED (female equivalent dose) of milked M. australica venom. In another circle a clean pin rub served as a negative control, and the third circle was a positive control of 1 FED milked venom from a sister M. digitata. Fifteen of these lids were prepared, and placed on boxes of 250-300 females as before. Boxes were then placed in absolute darkness at

25°C, and scored for signs of chewing 12 hours later.

Cochran Q tests were used to analyze chewing frequencies (Statistica 6.0). This test was chosen because the treatments were paired, and the results were scored as chewing presence or absence (1 or 0 respectively).

RESULTS

The experimental group containing smeared M. femorata venom reservoir and alkaline gland contents elicited chewing from M. femorata in at least one of the two treated pits in 19 of the 20 replicates. In contrast, both control pits were chewed on only two occasions out of 20 (Table 6.1). These differences were highly significant (P

<0.0001, Q =17.0000, 1 df).

For the experiment to examine if chewing was elicited in M. digitata by milked venom from M. australica chewing occurred in 9 of the 15 replicates (Table 6.2). The overall Cochran test was significant (P <0.0031, Q =11.5556, 2 df). Therefore using

Fisher’s test for multiple analyses, we ran pairwise Cochran tests that revealed a significant difference between the blank and M. australica venom (P <0.0047, Q =8.0, 1 df) and the blank and M. digitata venom (P <0.0143, Q =6.0, 1 df). There was no significant difference between chewing at the positive control, M. digitata venom, and M. australica venom (P <0.3173, Q =1.0, 1 df).

DISCUSSION

The M. femorata chewing results in response to dissected M. femorata reproductive tract organs suggest that M. femorata has a pheromone in its venom that stimulates chewing at a particular spot (Table 6.1). This adds support to the idea that chewing in response to venom components evolved before the speciation event that resulted in M. digitata and M. femorata.

The positive results for the cross-attractancy effect on M. digitata females (Table

6.2) might seem surprising since the two venom source species belong to different species groups (Dahms 1984a). Curiously, we have been unable to elicit chewing in response to milked venom in M. australica, although, testing has not been exhaustive

(Deyrup unpublished). However, there is little reason to expect that such a pheromone, if there is a pheromone for chewing in M. australica, would not be conserved, since a mutation could leave carriers trapped in the host’s cell. Even if there is no such

85 pheromone present in venom for chewing in M. australica, the chemical that stimulates chewing for M. digitata could be one that is stable and under selection for another purpose (e.g., perhaps containing a constituent causing developmental delay in the host

[Deyrup et al. 2003]). Components of other pheromones apparently have been conserved in Melittobia. Matthews et al. (1985) found that females of M. digitata, M. femorata, and

M. australica were attracted to non-conspecific as well as conspecific males in choice tests.

The data for the M. femorata have an additional implication. They suggest that the source of the pheromone for chewing is in either the alkaline gland, venom reservoir, or both. It is possible that the chewing is only elicited when two constituents are combined from the different glands. The milked venom in M. digitata (Deyrup et al. 2005) is probably mostly composed of fluids from the two sources, and it is probable that the chewing is stimulated by similar factors in the three species. Further work in which each source is presented alone and together is needed to clarify this matter.

ACKNOWLEDGEMENTS

This work was supported in part by a grant from the National Science Foundation to R. W. Matthews.

REFERENCES

Balfour-Browne, M. A. 1922. On the life-history of Melittobia acasta Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomologist 5: 14-22.

Dahms E.C. 1984a. Revision of the genus Melittobia (Chalcidoidea :Eulophidae) with

the description of seven new species. Memoirs of the Queensland Museum 21:

271-336.

Dahms E.C. 1984b. A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum 21: 337-360.

Deyrup, L. D., and R. W. Matthews. 2003. A simple technique for milking the venom of

a small parasitic wasp, Melittobia digitata (Hymenoptera: Eulophidae). Toxicon

42: 217-218.

Deyrup, L. D., M. Deyrup, and R. W. Matthews. 2003. Paralyzation and developmental

delay of a factitious host by Melittobia digitata (Hymenopter: Eulophidae)

Journal of Entomological Science 38: 703-705.

Deyrup, L. D., R. W. Matthews, and J. M. González, J. M. 2005. Cooperative chewing in

a gregariously developing parasitoid wasp, Melittobia digitata Dahms, is

stimulated by structural cues and a pheromone in crude venom extract. Journal of

Insect Behavior 18: 293-304.

Donovan, B. J. 1976. Co-operative material penetration by Melittobia hawaiiensis

(Hymenoptera: Eulophidae) and its adaptive significance. New Zealand

Entomologist 6: 192-193.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington, DC.

Matthews, R. W. 1997. Teaching ecological interactions with mud dauber nests. The

American Biology Teacher 59: 152-158.

Matthews, R. W., J. Yukawa, and J. M. González. 1985. Sex pheromones in Melittobia

parasitic wasps: Female response to conspecific and congeneric males of 3

species. Journal of Ethology 3: 59-62.

Table 6.1. The number of replicates after 12 hours exposure having chewing by M. femorata females in both of the treatment pits that received venom components dissected from M. femorata females or both of the control pits.

Treatment Chewed Replicates

Venom/Alkaline 19 20

Control 2 20

P < 0.0001, Q = 17.00000, df 1

Table 6.2. The number of replicates after 12 hours exposure having chewing by M. digitata females in the treatment circles that contained venom milked from either M. digitata or M. australica. Different letters in the “Significance” column indicate significant differences using a Cochran Q test p< 0.05, df = 1 (Statistica 6.0).

Treatment Chewed Replicates Significance

M. australica Venom 9 15 a

M. digitata Venom 7 15 a

Control 1 15 b

CHAPTER 7

EXAMINING THE RELATEDNESS OF BEHAVIORS THROUGH

EXPERIMENTATION: SWITCHING ON AND OFF CHEWING AND FEEDING

BEHAVIOR IN A PARASITIC WASP

Deyrup, L. D., and R. W. Matthews. To be submitted to Journal of Ethology

ABSTRACT

Melittobia digitata (Hymenoptera: Eulophidae) are small gregarious parasitoids known for their cooperative escape behavior. The first part of this escape chewing behavior has been compared to the first part of their feeding behavior. We sought to experimentally test whether these initial sequences were interchangeable by manipulating the behaviors using a single cue, hemolymph, offered in different contexts. By such manipulation we were successful in turning off chewing and turning on feeding in a chewing situation, and turning on chewing and turning off feeding in a feeding context.

We discuss the implications of this finding for the origin and evolution of these behaviors, and how experimental tests of key cues can be helpful for understanding the evolution of behavioral patterns.

KEY WORDS: Cooperation, evolution of behavior, feeding, hemolymph, Melittobia digitata

INTRODUCTION

Melittobia is a gregariously developing ectoparasitic wasp belonging to the family

Eulophidae (Dahms 1984a). Melittobia are most commonly found attacking mud dauber wasps, but they also attack a variety of solitary bees and wasps, and their associates

(Edwards and Pengelly 1966, Jayasingh and Freeman 1980, Krombein 1967). In addition,

Melittobia can be easily reared in the laboratory on pupae of the flesh fly Sarcophaga

(Neobellieria) bullata (e.g. Silva-Torres and Matthews 2003, Deyrup and Matthews

2003a). Once a female finds a host she paralyzes it and oviposits. Depending on host nesting behavior, her offspring could emerge as adults completely entombed in a thick- walled nest (Balfour-Browne 1922, Deyrup et al. 2003, Deyrup et al. 2005). In order to escape and disperse, the adult females first must chew their way out of the host chamber.

The behavior of chewing out of a host cocoon or nest has been noted repeatedly over the 100+ years of Melittobia research (e.g., Buckell 1928, Cowley 1961, Deyrup et al. 2005, Donovan 1976, Howard 1892, Maeta and Yamane 1974, Torchio 1963).

Donovan (1976) first noticed that females form circles for cooperative chewing through thick substances.

Deyrup et al. (2005) demonstrated that cooperating females initiate chewing at a spot that has been marked by stinging, and discussed the possible behavioral link between chewing an escape exit and the similar sequence of stinging and then feeding on a host prepupa. Deyrup et al. (2005) however, only pointed out the apparent similarities between the early feeding and the early chewing behaviors. To support a theory that one behavioral function is derived from another one must experimentally investigate the

93 presumed cues and/or stimuli that elicit the behavior. If two similar complex behavioral sequences differ in function and are triggered by divergent stimuli, the simplest explanation is that one behavior is derived from the other rather than appearing de novo.

Our goal in this study was to test whether initial sequences of the two behavior schemes that appear similar to researchers (Fig. 7.1), are interchangeable to the wasps. To accomplish this we sought to test whether both behaviors can be switched to the apparent related behavior with the presence or absence of the same cue. After eliminating several possible cues in pilot tests, hemolymph was selected as the key cue for experimental manipulation, because following host stinging, hemolymph can be obtained from the wound and is always consumed by the wasp, whereas during escape chewing, no hemolymph would be encountered in response to stinging.

METHODS

Experiment 1: Turning on feeding in a chewing scheme

To test whether the presence of hemolymph would turn off the chewing and turn on the feeding behaviors, we used the lids of the clear plastic boxes that dispersing wasps tunnel into in the laboratory (Deyrup et al. 2005) (Carolina Biological Supply Co., Cat.

No. ER-14-4584). We took 80 new lids and drilled a 5mm diameter hole through each lid in the approximate center of the lid. Then we sealed the underside of the hole (on the inner surface of the lid) with plastic wrap (Reynolds®) held on with clear tape

(Highland™ Invisible tape, 3M). This created a small cup or well in the lid. Then using a glass pipette we filled this cup with either hemolymph bled from a diapausing prepupa of

Trypoxylon politum or water, depending on the treatment. Finally, we sealed the top of the small well with clear packaging tape (Scotch® Packaging and Mailing Tape, 3M).

Then immediately before using the lids for the experiment we milked venom from two female M. digitata as in Deyrup and Matthews (2003b), and applied it to the inside of the lid directly over the well containing either the hemolymph or water. Venom has been previously shown to elicit chewing at treated spots by emerging M. digitata females

(Deyrup et al. 2005).

The M. digitata were reared from laboratory cultures maintained at the University of Georgia Department of Entomology. Eighty cultures were established by placing single females in our clear plastic boxes each with three flesh fly (Sarcophaga bullata) hosts. Two days after the adult offspring had begun to emerge, we replaced the lids of their boxes with freshly prepared experimental lids. In this way 40 of the cultures had lids with a small well of hemolymph in them, and 40 had lids with a small well of water. We then placed these cultures in complete darkness for 24 hours. After 24 hours we examined the lids, and recorded whether the well had been chewed into, or left alone.

Experiment 2: Turning on chewing in a feeding context

For this experiment we made two types of artificial hosts. One was a small packet

(approximately the size of a mud dauber host) of Parafilm® filled with hemolymph from the host prepupa (Trypoxylon politum) or water. These were then placed in a box with five inexperienced, adult, mated females. The end result was that there were 20 boxes with water packets, and 20 with hemolymph packets. As before, boxes were placed in

95 complete darkness for 24 hours and then the artificial hosts were examined for evidence of chewing.

RESULTS AND DISCUSSION

In experiment 1, 14 of the 40 water replicates had chewing, and there was no chewing evident in any of the 40 hemolymph treatments. This difference was significant using a chi-square test (P <0.0001, 2 =16.970, 1 df) (Statistica 6.0). These results indicate that the chewing behavior is not elicited when females encounter hemolymph.

As evidence that feeding behavior was turned on, several of the spots marked that were hemolymph-filled had received groups of eggs laid after 2 to 4 days although most wells had become dried or fed dry during the extended time period. Host feeding is necessary for egg maturation and subsequent oviposition in Melittobia (Dahms 1984b, Doutt 1959).

In experiment 2, 16 out of 20 water packets were chewed into. None of the packets made with hemolymph had chewing. This difference was significant using a chi- square test (P <0.0001, 2 =16.970, 1 df) (Statistica 6.0). Since females only chewed on water-containing packets, it suggests that chewing behavior is turned on if hemolymph is not encountered. As evidence of feeding behavior being turned on, eggs were laid on all of the 20 hemolymph packets after 2 to 4 days, but none of the water packets.

These results suggest that the behavioral schemes for escape and feeding are related. The presence of a single stimulus cue (venom) appears to serve to switch on the behavior for both feeding and tunneling. Post-initial behavioral sequences are then dependent on further cues, including the presence or absence of hemolymph. It is

96 significant that the presence of hemolymph was sufficient for both turning the behavior off as well as on, since switching to feeding might be expected to be an automatic response to encountering food. However, it is unlikely that the lack of the same stimulus would switch feeding to chewing.

It has been established that most instinctive behaviors evolve from preexisting behaviors (Wilson 1975). While our experimental focus was escape chewing, in M. digitata, an alternative possible evolutionary scenario might involve the initial access to a host. Since at least some hosts, such as T. politum, are encapsulated in rigid cocoons but do not fully fill their cocoons, there are sites on the outside of the cocoon where no host would be encountered by stinging. If upon discovering this host a female Melittobia inserted her sting into such a void, then when backing up to feed she would not encounter hemolymph. However, if that female switched to chewing/tunneling behavior, she would eventually enter the cocoon, find the host, and be able to successfully reproduce. Once such a behavior had reached fixation in a population, it would seem a simple step for access chewing to be recontexualized into escape chewing. Chewing into a host cocoon or puparia by female Melittobia has been confirmed in several circumstances (Dahms

1984b). We have noticed that females also often sting before attempting to chew through a barrier to gain access to a host (Deyrup unpublished).

ACKNOWLEDGEMENTS

This work was supported in part by a grant from the National Science Foundation to R. W. Matthews.

REFERENCES

Balfour-Browne, M. A. 1922. On the life-history of Melittobia acasta Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomologist 5: 14-22.

Cowley, D. R. 1961. The associates of Pison spinolae Shuckard (Hymenoptera:

Sphecidae). New Zealand Entomologist 2: 45-46.

Dahms, E.C. 1984a. Revision of the genus Melittobia (Chalcidoidea; Eulophidae) with

the description of seven new species. Memoirs of the Queensland Museum 21:

271 – 336.

Dahms, E. C. 1984b. A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum 21: 337-360.

Deyrup, L.D., and R.W. Matthews. 2003a. Host preference and utilization by Melittobia

digitata (Hymenoptera: Eulophidae) in relation to mating status. Journal of

Entomological Science 38: 512 – 517.

Deyrup, L. D., and R. W. Matthews. 2003b. A simple technique for milking the venom of

a small parasitic wasp, Melittobia digitata (Hymenoptera: Eulophidae). Toxicon

42: 217-218.

Deyrup, L.D., Deyrup, M. and R.W. Matthews. 2003. Paralyzation and developmental

delay of a factitious host by Melittobia digitata (Hymenoptera: Eulophidae).

Journal of Entomological Science 38: 533 – 535.

Deyrup L.D., R.W. Matthews, and J. M. Gonzalez. 2005. Cooperative chewing in a

gregariously developing parasitoid wasp, Melittobia digitata Dahms, is stimulated

by structural cues and a pheromone in crude venom extract. Journal of Insect

Behavior 18: 293-304.

Donovan, B. J. 1976. Co-operative material penetration by Melittobia hawaiiensis

(Hymenoptera: Eulophidae) and its adaptive significance. New Zealand

Entomologist 6: 192-193.

Doutt, R.L. 1959. The biology of parasitic Hymenoptera. Annual Review of Entomology

4: 131-182.

Edwards, C. J. and D. H. Pengelly. 1966. Melittobia chalybii Ashmead (Hymenoptera:

Eulophidae) parasitizing Bombus fervidus Fabricius (Hymenoptera: Apidae).

Proceedings of the Entomological Society of Washington 96: 98-99.

Howard, L. O. 1892. The habits of Melittobia. Proceedings of the Entomological Society

of Washington 2: 244-249.

Jayasingh, D.B., and B.E. Freeman. 1980. The comparative population dynamics of eight

solitary bees and wasps (Aculeata; Apocrita; Hymenoptera) trap-nested in

Jamaica. Biotropica 12(3): 214 – 219.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington, DC.

Maeta, Y., and S. Yamane. 1974. Host records and bionomics of Melittobia japonica

Masi (Hymenoptera: Eulophidae). Bulletin of the Tohoku National Agricultural

Experimental Station 47: 115-131.

Silva-Torres, C. S. A., and R. W. Matthews. 2003. Development of Melittobia australica

Girault and M. digitata Dahms (Parker) (Hymenoptera: Eulophidae) parasitizing

Neobellieria bullata (Parker) (Diptera: Sarcophagidae) puparia. Neotropical

Entomologist 34: 641-655.

Torchio, P. F. 1963. A chalcid wasp parasite of the alfalfa leaf cutting bee. Utah Farm

Home Science 24: 70-71.

Wilson, E. O. 1975. Sociobiology: The New Synthesis. Belknap Press of Harvard

University Press, Cambridge, Mass.

100

FIGURE LEGEND

Fig 7.1. Observed similarities between the behavioral schemes in the feeding and escape chewing behaviors of Melittobia digitata, and the relationship to the questions addressed in this study. “Insert ovipositor” equals stinging, a behavior seen in both the feeding and escape contexts.

101

Fig. 7.1

102

CHAPTER 8

FEEDING AND SIBLICIDAL CANNIBALISM IN A MALE PARASITIC WASP

(HYMENOPTERA: EULOPHIDAE)

Deyrup, L. D., R. W. Matthews, and M. Deyrup. Submitted to Florida Entomologist.

103

ABSTRACT

Melittobia digitata Dahms is a small parasitic wasp known for its lethal male combat but subject to controversy regarding the existence of male feeding in general and cannibalistic feeding in particular. Here we report our observations supporting filicidal cannibalism. To test the ability of a male’s capability to feed we smeared sugary dye on the wasps’ mouthparts and observed it passing through the digestive system to produce colored feces, confirming that males have a complete digestive tract. To document siblicidal feeding we injected other males with water-soluble dye, and paired them with undyed males. Undyed winners that appeared to feed on dyed losers were monitored; dye was evident in their feces. Finally, to determine if males benefit from feeding, we compared the longevity of fed and unfed males; fed males lived significantly longer than non-fed males (Mann-Whitney U test = 81.5, N1 = 26, N2 = 26, P<0.001). We discuss possible reasons for the comparative rarity of siblicidal cannibalism and its fitness implications.

KEY WORDS: Melittobia, kin selection, uneven sex ratios, male combat

104

INTRODUCTION

Melittobia (Hymenoptera: Eulophidae) are small, gregarious parasitoids of solitary wasps and bees, and assorted associates (Edwards & Pengelly 1966, Krombein

1967, Maeta & Yamane 1974). These parasitoids have intrigued biologists (e.g. Hamilton

1967) because of their unusual and highly inbred reproductive strategy. Melittobia digitata Dahms is also used in educational curricula under the name WOWBug®

(Matthews et al. 1996, 1997).

Upon finding a suitable host, the female Melittobia stings it, feeds on hemolymph exuding from the sting wound, and then lays several hundred eggs, of which over 90% develop into females (Buckell 1928, Schmieder 1938, Dahms 1984). Females mate once with a brother, then cooperate to chew an exit hole and disperse (Deyrup et al. 2005); their brothers remain behind to die within their natal host’s cocoon (Dahms 1984).

While males’ lives may be circumscribed within their natal cocoon, they are nonetheless action-filled. Males of most Melittobia species are highly pugnacious and frequently engage in fatal fights with their brothers (e.g. Graham-Smith 1919, Malyshev

1968, Hamilton 1979, Hartley & Matthews 2003, Abe et al. 2003); attacks on male pupae are also documented (Hermann 1971, Abe et al. 2005) and, because Melittobia are protandrous, male fighting can begin before the first females emerge. Occasionally, however, males also will attack females presented to them (Balfour-Brown 1922,

Hermann 1971, Matthews 1975, Dahms 1984).

There has been continued speculation as to whether – in addition to the obvious advantage of dispatching potential rivals – such attacks might provide an opportunity for

105 males to feed. However, several biologists have gone on record as doubting that

Melittobia males feed at all. For example, while Dahms (1984) observed attacks, he found no evidence of feeding and pointed out that a male’s gaster grows increasingly thinner until he dies. Abe et al. (2005) categorically state that males of M. australica

Girault do not feed. Balfour-Browne (1922) noted chewing attacks, but considered them to be an artifact of experimental conditions. Others disagree, reporting that males sometimes continue to chew on a defeated male sibling (Graham-Smith 1919, Matthews

1975) or on an attacked female (Hermann 1971) for relatively extended periods of time.

If they were to ingest nutrients during this behavior, such cannibalism might provide a competitive advantage (Matthews 1975), enabling a male to live longer or produce more sperm.

Combat between males of Melittobia digitata is particularly intense. We noticed that M. digitata males in our laboratory cultures sometimes spent an extended period of time with their mandibles immersed in the tissues and hemolymph of a defeated male

(Fig. 8.1). In one instance, a male killed an emerging male by biting through the emerging male’s head capsule, and then inserted his mandibles deeper into the head capsule. The victor’s palpi were highly active, with motions resembling those of feeding females. As we watched, the abdomen of the male began to swell slightly, as if hemolymph were filling the crop. This observation lent support to a hypothesis that M. digitata males sometimes feed on a defeated male, and encouraged our experimental approach to male feeding with three objectives. The first was to determine whether male

M. digitata have a working digestive tract. The second was to resolve whether males

106 ingest hemolymph from other males, and, if so, whether it passes through their digestive system. The last was to test whether individual males benefit from feeding.

MATERIALS AND METHODS

Experiment 1: Functional Digestive Tract

We collected 40 M. digitata male pupae (recognizable by the lack of compound eyes) developing in a single laboratory culture, isolated them individually in small tightly lidded plastic boxes (50x25x18 mm, Carolina Biological Supply Co., Cat. No. ER-14-

4584), recorded eclosion dates, and randomly assigned the adults either to the control or experimental group. The controls were undisturbed. When wasps in the experimental group were 2 days old, we smeared their mouthparts with either “willow green”,

“cornflower blue”, or “rose petal pink” cake icing dye (Wilton Enterprises).

After the passage of several hours to allow opportunity for treated males to groom and remove icing from their body surfaces, all males were transferred into clean boxes.

We checked the boxes daily and recorded whether colored fecal droppings appeared. We also observed the males under a dissecting microscope to check for dye in their digestive system, and found that it was clearly visible through their translucent cuticle. A 2 test in was used to determine whether individuals in the experimental and control groups differed significantly in passing colored fecal spots vs. undyed spots (Statistica 6.0).

107

Experiment 2: Feeding On Another Male

Because we reasoned that hungry males would be more likely to feed, we produced nutritionally stressed males by isolating individual late male pupae (±1 d until eclosion and providing each with 10 newly eclosed virgin females. Males were allowed to mate ad libitum with these females for up to 5 days post-eclosion. After 3-5 days the males’ gasters became thin and they appeared emaciated.

To produce weakened males with identifiable hemolymph as potential losers, we injected them in the abdomen with blue water-based dye (McCormick & Co., Inc.) using a glass pipette (Soda Lime Glass, 9”, J. & H. Berge, Inc.) that had been stretched while heating it in an alcohol flame. Typically, the dyed male rapidly weakened, and was usually dead in 10 to 15 min.

An emaciated undyed male and a weak freshly dyed male were paired in a deep well projection slide arena (Carolina Biological Supply, Inc.). Because there was only a short window of opportunity for combat, we placed them next to each other to facilitate interaction; even then, most fighting was non-lethal. After lethal fights, most males did not attempt to feed on their defeated brother. However, we continued to dye, expose, and observe the males until we recorded 10 instances of undyed victors that killed their dyed brother and appeared to feed upon them. Each of these victors was placed into a separate observation box and observed for subsequent dye passage in its fecal droppings.

Experiment 3: Benefit From Feeding

To determine whether males benefit from feeding, we gathered 55 M. digitata male pupae from five cultures, isolated each pupa in a glass 1 dram vial, and inspected

108 the vials daily, recording the date on which each male eclosed; 52 pupae eclosed as adults. Males that eclosed on the same day were assigned to an experimental (fed group, n = 26)) or a control (unfed, n = 26) group.

The experimental group was fed insect hemolymph from a Trypoxylon

(Trypargilum) politum Say prepupa. (This is probably the most common of M. digitata’s natural hosts, but in nature it would be inaccessible as a potential food source for adult males, because by the time adult males appear, earlier larval feeding has reduced the host to dry powdery remains.) Using an insect pin to puncture the host prepupal cuticle, we bled one drop of hemolymph onto a glass slide then gently transferred a male to the drop with a fine brush. Males immediately imbibed hemolymph from the drop. When a male did not drink voluntarily, we coaxed its head into the drop. The males invariably fed when their mouthparts touched the hemolymph, and we allowed males to feed to satiation. The control group of males was not fed. We did not give them water or insect saline solution; such resources do not occur in their natural habitat, because males seldom, if ever, leave the pupa case of their host.

All individuals in both groups were individually isolated in 1-dram glass vials and placed in an incubator at 30º C. We recorded how many days each male survived. The difference between the treatment and the control groups was analyzed using a Mann-

Whitney U test and a survival analysis (Statistica 6.0).

RESULTS

In the first experiment, all colors of dye were immediately visible passing through the upper digestive system into the crop of all 20 treated males, and color appeared in

109 their droppings when checked 24 hours later. No control males had droppings of a color similar to the fed males. This difference was very highly significant using a 2 test

( 2=40.0, P<0.001) (Statistica 6.0).

In experiment 2, each of the males that we had suspected of feeding on his brother had blue color moving through the body and into the crop. This was confirmed when we checked 24 hours later that dye was passed in droppings of all 10 individuals.

In experiment 3, individual male adult life spans varied, ranging from 12-16 days for unfed males, and from 13-18 days for fed males (Fig. 8.2). However, the lives of fed

males were1.5 days longer, on average, than those of unfed males (unfed

¡ ¡

±SE=13.2±0.14, =13: fed ±SE=14.7±0.21, =15). Statistically, the difference was very highly significant (Mann-Whitney U test =81.5, N1 = 26, N2 = 26, P<0.001)

(Statistica 6.0).

DISCUSSION

The results from experiment 1, demonstrating that the digestive tract of male M. digitata is complete and functional, led to the second experiment, which established that males that defeat another male are capable of ingesting hemolymph from the defeated individual. The combination of these two experiments supports the assumption that the apparent feeding behavior that we had previously observed was correctly interpreted because males M. digitata have a working digestive tract and are capable and will imbibe hemolymph from another male.

We showed that M. digitata can and will feed, but this may not settle the controversy completely, because our findings may not apply to all species in the genus.

110

For example, M. femorata Dahms does not appear to have the same propensity for lethal male combat as M. digitata (R.W.M., unpublished data). While an M. femorata male conceivably could feed on a killed female, it would be unlikely to feed on a brother.

Records of feeding by adult male parasitoids are rare. Males of few species have access to hemolymph, and M. digitata seems to take an advantage of an unusual situation.

Nectar is a more usual food source for adult Hymenoptera, but nectar-feeding by parasitoids is also rare, and concentrated in a few families. At the Archbold Biological

Station (Highlands Co., FL), where flower visitors have been studied for many years, there are few records of nectar feeding by male parasitoids. Among Ichneumonoidea, nectar feeding occurs in male Agathis longipalpus (Cresson) (Braconidae); among

Chalcidoidea nectar feeding occurs in male Leucospis affinis Say, L. robertsoni Crawford and L. slossonae Weld (Leucospidae). In contrast, male aculeate Hymenoptera are frequent nectar feeders at the Archbold Biological Station, including numerous species representing 15 families (M.A.D., unpublished data).

Adult male siblicidal cannibalism is a relatively rare phenomenon among those insects which have been studied. A situation somewhat similar to that of Melittobia occurs in ants of the genus Cardiocondyla; ergatoid males engage in lethal combat, usually won by an older male that attacks a recently eclosed sibling (Stuart 1987, Heinze et al. 1998). In this genus, however, workers remove the dead male from the nest or feed it to larvae (Stuart 1987). The situation confronting Melittobia males differs from that of ants in that Melittobia males exist in a closed system, without access to external resources.

111

In mites, female cannibalism has been reported (Schausberger & Croft 2000,

Berndt et al. 2003), but its possible siblicidal nature seems to require further study.

Schausberger & Croft (2000) reported that Phtoseiulus persimilis Athias-Henriot preferentially cannibalized non-siblings, but later Schausberger (2003) reported that if raised without contact with siblings, they preferentially cannibalized siblings.

Cannibalism for its own sake would seem to have several potential disadvantages.

The three most applicable to M. digitata males are the risk of being injured or killed in attacking a similarly capable individual, the risk of contracting a disease from the consumed individual, and the evolutionary cost to fitness (Elgar & Crespi 1992).

However, like M. digitata attacking male pupae, some species seem to avoid the problem of attacking a similar organism when early maturing individuals or individuals of a more advanced developmental stage kill a less capable immature individual (Elgar & Crespi

1992). However, this is not always the case; for example, cannibalism on peers has been recorded in intrauterine sharks (Wourms 1977, Hamlett & Hysell 1998). In M. digitata violent combat, presumably evolved in the context of local mate competition, usually quickly incapacitates the defeated male, thereby removing the risk of further injury. This would leave victorious males free to consume the defeated male without further risk.

Similarly, cannibalism among male Melittobia digitata seems unlikely to transmit disease, as the combatants are usually siblings, having fed off the same host, and lived their entire lives inside a sealed cocoon. The third potential disadvantage, loss of fitness in sibling competition, is a complex issue; kin selection models have endeavored to deal with this problem (Griffin & West 2002). However, cannibalism after combat adds yet another advantage in M. digitata male competition.

112

The third experiment showed that males who fed lived significantly longer than unfed controls. Lengthening one’s adult life by the equivalent of 11% is no biologically trivial matter; presumably, those males that live longer secure more mates, dispatch more rivals, and have increased fitness relative to unfed males. Wiltz and Matthews

(unpublished) found that males are more likely to die before exhausting their sperm, which makes the measurement of longevity a better indicator of increased fitness than sperm production. We have observed early males feeding on males that are just eclosing and are vulnerable. Early males that emerge with the first generation of a few short wing females would benefit greatly in fitness by extending their lives to overlap more with the large emergence of long wing females. Wiltz and Matthews (unpublished) study and our observations expose the possible benefits of males who can extend their lifespan by feeding.

We conclude that male cannibalism in M. digitata may not be rare when the advantages outweigh the disadvantages. The natural history of M. digitata appears to satisfy this criterion. The fact that a single male can potentially inseminate over 200 sisters and is likely to die before exhausting their sperm (B. Wiltz & R. Matthews, unpublished), as appears to occur routinely in some Melittobia species, provides a context in which male feeding and increasing life expectancy would be advantageous. Male M. digitata that defeat and then cannibalize brothers may also obtain nutrients needed to maintain sperm production and sex pheromone production (Consôli & Vinson 2002) for an extended life expectancy, as well as acquire the energy needed to successfully combat newly eclosing brothers (Abe et al. 2005) and repeatedly perform the relatively elaborate

113 courtship displays that characterize the genus (Matthews & Matthews 2003, González &

Matthews 2005).

ACKNOWLEDGEMENTS

We thank and appreciate Jan Matthews and Jessica Beck for reviewing the manuscript,

Jorge M. González for the photograph and stimulating discussion, and Stuart West for his knowledge and insightful ideas. This project was supported by an NSF grant (R. W.

Matthews, P.I.).

REFERENCES

Abe, J., Kamimura, Y., Ito, H., Matsuda H., and Shimada, M. 2003. Local mate

competition with lethal male combat: effects of competitive asymmetry and

information availability on a sex ratio game. J. Evol. Biol. 16: 607-613.

Abe, J., Kamimura, Y. ,and Shimada, M. 2005. Individual sex ratios and offspring

emergence patterns in a parasitoid wasp, Melittobia australica (Eulophidae), with

super-parasitism and lethal combat among sons. Behav. Ecol. Sociobiol. 57: 366-

373.

Balfour-Browne, F. 1922. On the life history of Melittobia acasta, Walker; a chalcid

parasite of bees and wasps. Parasitology 14: 349-370.

114

Berndt, O., Meyhofer, R. and Poehling H.M. 2003. Propensity towards cannibalism

among Hypoaspis aculeifer and H. miles, two soil-dwelling predatory mite

species. Exper. Appl. Acarol. 31: 1-14.

Buckell, E. R. 1928. Notes on the life history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomol. 5: 14-22.

Cônsoli, F. L. and Vinson, S. B. 2002. Larval development and feeding behavior of the

wing dimorphics of Melittobia digitata Dahms (Hymenoptera: Eulophidae). J.

Hymenop. Res. 11: 188-196.

Dahms, E. C. 1984. A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Mem. Queensland Mus. 21: 337-360.

Deyrup, L. D., Matthews, R. W. and González, J. M. 2005. Cooperative chewing in a

gregariously developing parasitoid wasp, Melittobia digitata Dahms, is stimulated

by structural cues and a pheromone in crude venom extract. J. Insect Behav. 18:

293-304.

Edwards, C. J. and Pengelly, D. H. 1966. Melittobia chalybii Ashmead (Hymenoptera:

Eulophidae) parasitizing Bombus fervidus Fabricius (Hymenoptera: Apidae).

Proc. Entomol. Soc. Wash. 96: 98-99.

Elgar, M. A. and Crespi, B. J. 1992. Ecology and evolution of cannibalism, pp. 1-12 In

M. A. Elgar, and B. J. Crespi [eds.] Cannibalism: Ecology and evolution among

diverse taxa. New York: Oxford University Press. 361 pp.

González, J. M. and Matthews, R. W. 2005. Courtship of the two female morphs of

Melittobia digitata (Hymenoptera: Eulophidae). Florida Entomol. (in press).

115

Graham-Smith, G. S. 1919. Further observations on the habits and parasites of common

flies. Parasitology 11: 347-384.

Griffin, A. S. and West, S. A. 2002. Kin selection: Fact and fiction. Trends Ecol. Evol.

17: 15-21.

Hamilton, W. D. 1967. Extraordinary sex ratios. Science 156: 477-88.

Hamilton, W. D. 1979. Wingless and fighting males in fig wasps and other insects, pp.

167-220. In M. S. Blum, and N. A. Blum [eds.] Sexual Selection and

Reproductive Competition in Insects. New York: Academic Press. 463 pp.

Hamlett, W. C. and Hysell, M. 1998. Uterine specializations in elasmobranches. J. Exper.

Zool. 282: 438-459

Hartley, C. and Matthews, R. W. 2003. The effect of body size on male-male combat in

the parasitoid wasp Melittobia digitata Dahms (Hymenoptera: Eulophidae). J.

Hymenop. Res. 12: 272-277.

Heinze, J., Hölldobler, B. and Yamauchi, K. 1998. Male competition in Cardiocondyla

ants. Behav. Ecol. Sociobiol. 42: 239-246.

Hermann, L. D. 1971. The mating behavior of Melittobia chalybii (Hymenoptera:

Eulophidae). M. Sc. Thesis. Athens: University of Georgia. 52 pp.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests and

Associates. Washington DC: Smithsonian Press. 570 pp.

Maeta, Y. and Yamane, S. 1974. Host records and bionomics of Melittobia japonica Masi

(Hymenoptera: Eulophidae). Bull. Tohoku Nat. Agric. Exper. Sta. 47: 115-131.

Malyshev, S. I. 1968. Genesis of the Hymenoptera and the Phases in Their Evolution

(English Translation). London: Methuen and Co. Ltd. 319 pp

116

Matthews, R. W. 1975 Courtship in parasitic wasps, pp. 66-86. In P. W. Price [ed.]

Evolutionary Strategies of Parasitic Insects and Mites. New York: Plenum. 224

pp.

Matthews, R. W. and Matthews, J. R. 2003. Courtship and mate attraction in parasitic

wasps. pp 59–72. In B. J. Ploger, and K. Yasukawa [eds.] Exploring Animal

Behavior in Laboratory and Field: A Hypothesis-Testing Approach to the

Development, Causation, Function, and Evolution of Animal Behavior. New

York: Academic Press. 472 pp.

Matthews, R. W., Koballa, T. R., Flage, L. R. and Pyle, E. J. 1996. WOWBugs: New Life

for Life Science. Athens: Riverview Press. 318 pp.

Matthews, R. W., Flage, L. R., and Matthews, J. R. 1997. Insects as teaching tools in

primary and secondary education. Ann. Rev. of Entomol. 42: 269-289.

Schausberger, P. 2003. Ontogenetic isolation favours sibling cannibalism in mites. Anim.

Behav. 67: 1031-1035.

Schausberger, P. and Croft B. A. 2000. Kin recognition and larval cannibalism by adult

females in specialist predaceous mites. Anim. Behav. 61: 459-464.

Schmieder, R.G. 1938. The sex ratio in Melittobia chalybii Ashmead, gametogenesis and

cleavage in females and in haploid males (Hymenoptera: Chalcidoidea). Biol.

Bull. Marine Biol. Lab. Woods Hole 74: 256-266.

Stuart, R. J. 1987. Lethal fighting among dimorphic males of the ant, Cardiocondyla

wroughtonii. Naturwissenschaften 74: 548-549.

Wourms, J. 1977. Reproduction and development in chondrichthyan fishes. Amer. Zool.

17. 379-410.

117

SYMBOL LEGEND

= Sample mean ¡

= Sample median

118

FIGURE LEGENDS

Fig. 8.1. A male of Melittobia digitata Dahms that appears to be feeding on a sibling male (Photo courtesy of Jorge M. González).

Fig. 8.2. Longevity of fed and unfed males of M. digitata at 30º C (dotted line and = males that were fed host hemolymph; solid line and = males that were unfed).

119

Fig. 8.1

120

20 s e l a M

e 15 v i l A

f o

# 10

0 12 14 16 18 Days

Fig. 8.2

121

CHAPTER 9

COMPETITION AND AGGRESSION AMONG FEMALE MELITTOBIA FEMORATA,

(HYMENOPTERA: EULOPHIDAE) WITH COMPARISONS TO M. DIGITATA, AND

M. AUSTRALICA

Deyrup, L. D., and R. W. Matthews. To be Submitted to Annals of the Entomological Society of America

122

ABSTRACT

Melittobia is a genus of parasitic wasps well known for high levels of inbreeding and violent male combat. Casual observations of groups of sisters of M. femorata placed with hosts revealed a surprising incidence of body mutilations (broken or missing tarsi, antennae, and wings). Replicated conspecific groups of 1, 2, or 3 foundress females of M. femorata, M. digitata, and M. australica and interspecific groups of M. femorata and M. australica (2:1) were observed over their first ten days in newly established cultures, and the incidence of mutilation recorded. In some groups females were dye-fed, allowing us to subsequently chart their individual activity patterns on or near the host based on patterns of their colored droppings. For M. australica and M. digitata, no conspecific females in any group size ever showed mutilation. However, in M. femorata nearly three- fourths of the females in conspecific groups of 2 or 3 acquired body damage beginning about the time oviposition on the host was first noted. In four of five replicates of the interspecific groups, M. femorata females killed the female of M. australica. Patterns of dyed droppings that developed over several days showed that individual females in groups of both M. femorata and M. australica, increasingly restricted their activities to a small portion of the host. These “micro” territories were non-overlapping and appeared to be actively defended. In contrast, M. digitata females in groups never displayed any obvious territoriality. Possible reasons for these differences in female behavior are discussed.

123

KEY WORDS: Aggression, Melittobia, mutilation, territoriality, intraspecific competition

124

INTRODUCTION

Melittobia is a genus of small, idiobiont, eulophid wasps that are gregarious ectoparasitoids of solitary bees, wasps, and their associates (Balfour-Browne 1922,

Buckell 1928, Dahms 1984, Krombein 1967). Melittobia femorata Dahms is the most commonly collected species of Melittobia in the southeastern United States (Matthews and González unpublished). Yet despite their abundance they are relatively poorly known in comparison to the well-studied sympatrically occurring M. digitata Dahms.

Upon finding a host, a Melittobia female feeds on its hemolymph. This stimulates egg maturation and within two-four days she begins to oviposit clusters of eggs on the host, with hundreds of eggs ultimately laid on larger hosts over the ensuing 10 days.

While males of Melittobia are well known for their lethal combat (Hamilton 1979), females are generally regarded as docile and tolerant of conspecifics. However, females of M. femorata are unusual among their congeners in that when colonies are founded with multiple females in the laboratory, aggression and body damage are commonly observed (Matthews unpublished). Since aggression among female Melittobia does not appear to have been reported, we decided to investigate this phenomenon in greater detail.

METHODS

Twenty-eight prepupae of Trypoxylon politum (Say) (Hymenoptera: Sphecidae), a common host for M. femorata, were placed individually in small plastic boxes (50 mm x

125

25 mm x 18 mm, Carolina Biological Supply Co., Cat. No. ER-14-4584). Mated 2-day- old M. femorata females were placed with these hosts in the following design: a single female in six boxes, two females in 13 boxes, and three females in nine boxes. Boxes were maintained at ambient room temperatures and checked daily over the following 10 days noting the females’ behavior and recording the incidence of any body damage.

In order to track individual females and their movements, we established 15 additional cultures with three marked females in each. To mark them we fed females 20% fructose and water dyed with [McCormick®] food coloring. After females imbibe this fluid it is easily visible in their crops through their semi-translucent cuticle; use of different colors served to identify individual females. In addition, because the color is retained in the female’s fecal matter, this technique allowed us to track each female’s activity through the pattern of her droppings on the floor of the plastic box.

For comparisons between species, in a third experiment 20 boxes of three M. digitata, 20 boxes of three M. australica, and five boxes containing one M. australica and two M. femorata were similarly established with all females first fed dyed fructose and water solution as above. These boxes were observed daily for ten days and the incidence of body damage among the females and patterns of their droppings were recorded.

RESULTS AND DISCUSSION

In the 15 colonies containing three dyed M. femorata females, one to four days after being placed on a host the females’ activities began to become increasingly localized, each focused about a particular portion of the host’s body. From the

126 distribution patterns of dyed fecal spots it was apparent that each individual female M. femorata developed a more or less exclusive “micro” territory (Fig. 9.2), and that the boundaries between them were relatively distinct. Undyed females in the groups of two or three in the other set of cultures appeared to behave similarly. Females of M. australica also displayed similar territoriality in all 20 cultures; however, droppings of M. digitata females displayed no grouping pattern in any of the replicates. Interestingly, in the multi-species cultures containing both M. femorata and M. australica females, the onset of territoriality in M. femorata seemed to be delayed (3-5 days after being placed on host), but small sample sizes did not warrant statistical testing.

During the course of oviposition (roughly days 2 -10) the frequency of aggression and incidence of body mutilation (manifested as missing tarsomeres and antennal flagellomeres and wing damage) increased. The number of females displaying body damage also increased as the number of foundresses increased, but no females who were alone on their hosts ever acquired any damage (Table 9.1). We regularly observed females biting at other females and even rolling around in locked combat (Fig. 9.1). In addition, many females were observed to walk about with their wings raised when on the host (normally wings are held flat over their abdomens). Although the wing-raising behavior suggests a form of posturing, this seems counterintuitive since within their host cocoons it is completely dark and tightly confined.

In comparison, none of the females in any of the 20 replicated groups of three M. digitata or M. australica acquired body damage over the ten day period. Periodic observation revealed no indication of agonistic interactions among females of M. digitata. Individual M. australica were sometimes seen to follow or approach other

127 females on the host and appeared to “hassle” the other female with proximity or nudging.

In the mixed species cultures M. australica seemed to sometimes pressure a M. femorata to abandon her territory, and in some instances caused her to move completely off of the host early in their association. However, after the M. femorata became physogastric, the tables turned, and in four of the five replicates the M. australica female exhibited damage and was eventually was found decapitated. In only one case did M. femorata and

M. australica appear to share the same area on the host, with no evidence of any body damage noted.

Territoriality has been widely documented in insects; however, much of the literature focuses on males in various forms of intrasexual selection (Baker 1983).

Territorial interactions among female parasitoids often are mediated via chemical marking of hosts to deter conspecific females from superparasitism (Hoffmeister and

Roitberg 1997, Petersen and Hardy 1996). Territoriality or intense intraspecific competition involving partitioning and defense of resources among conspecific female insects is relatively uncommon in insects, but has been recorded for some tephritid flies

(Prichard 1969, Shelly 1999), water striders (Nummelin 1988), and aphids (Inbar 1998).

In contrast, there are many instances of female territoriality hymenopterans, ranging from ants who actively defend foraging territories (Hölldobler and Wilson 1990) to parasitic wasps (Baker 1988, Griffiths and Godfray 1988). Similar to M. femorata, several other parasitic wasps have been reported defending a host resource or their offspring egg masses from conspecifics (e.g., Field and Calbert 1999, Hardy and Blackburn 1991,

Wilson 1961).

128

Reasons for the extreme degree of female pugnacity in M. femorata and its absence in M. digitata and M. australica are not clear. Perhaps aggressive tendencies evolved to prevent over utilization of the host during mass emergences of nearby broods of M. femorata. Females that successfully defended “micro” territories would drive other females away, and thereby prevent competitors from feeding and egg laying that would limit the resource potential for their offspring. [When 100-200 females of M. femorata were placed on mud dauber hosts in the laboratory, very few offspring were produced and many individual females displayed body damage (Deyrup unpublished)]. Perhaps the context in which we observe the interaction is not the same as the one in which the pugnacity evolved. Given the proximity of mud dauber nests under bridges there is a good likelihood that two or more females emerging from the same clutch might jointly colonize a nearby host. However, high mud dauber nest densities are a relatively recent phenomenon, due to human bridge building. Perhaps other species of solitary bees and wasps were the principal original hosts for M. femorata, and the behavioral ecology of host searching may have been very different from that which occurs today.

Genetic analysis of foundress relatedness, experimental manipulation of host searching cues, and further life history research may ultimately help us to better understand what selects for one species of Melittobia to behave aggressively but not another. Certainly, M. femorata has the potential to yield much new information because they are so abundant and easily reared and manipulated. And because different members of the genus Melittobia appear to display the full continuum of aggressive interactions in both sexes, further study will likely be rewarding.

129

ACKNOWLEDGEMENTS

Funding that supported part of this research was provided by a National Science

Foundation grant to R. W. Matthews. We thank Kathryn Hauth who first noticed the mutilation occurring among groups of M. femorata females during an undergraduate independent study.

REFERENCES

Baker, R. R. 1983. Insect territoriality. Annual Review of Entomology. 28: 65-89.

Balfour-Browne, M. A. 1922. On the life-history of Melittobia acasta Walker, a chalcid

parasite of bees and wasps. Parasitology 14: 349-369.

Buckell, E. R. 1928. Notes on the life-history and habits of Melittobia chalybii Ashmead

(Chalcidoidea: Elachertidae). Pan-Pacific Entomologist 5: 14-22.

Dahms, E. C. 1984. A review of the biology of species in the genus Melittobia

(Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum 21: 337-360.

Field, S. A., and G. Calbert. 1999. Don’t count your eggs before they’re parasitized:

contest resolution and the trade-offs during patch defense in a parasitoid wasp.

Behavioral Ecology 10: 122-127.

Griffiths, N. T., and H. C. J. Godfray. 1988. Local mate competition, sex ratio and clutch

size in bethylid wasps. Behavioral Ecology and Sociobiology 22: 211-217.

130

Hamilton, W. D. 1979. Wingless and fighting males in fig wasps and other insects. In:

Sexual Selection and Reproductive Competition in Insects (ed Blum, M. S. and N.

A. Blum) Academic Press. New York.

Hardy, I. C. W., and T. M. Blackburn. 1991. Brood guarding in a bethylid wasp.

Ecological Entomology 16: 55-62.

Hoffmeister, T. S., and B. D. Roitberg. 1997. Counterespionage in an insect herbivore-

parasitoid system. Naturwissenschaften 84: 117-119.

Hölldobler, B., and E. O. Wilson. 1990. The Ants. The Belknap Press of Harvard

University Press. Cambridge, MA.

Inbar, M. 1998. Competition, territoriality and maternal defense in a gall-forming aphid.

Ethology Ecology and Evolution 10: 159-170.

Krombein, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories, Nests, and

Associates. Smithsonian Press, Washington, DC.

Nummelin, M. 1988. The territorial behavior of four Ugandan waterstrider species

(Heteroptera: Gerridae): a comparative study. Annales Entomologici Fennici 54:

121-134.

Petersen, G., and I. C. W. Hardy. 1996. The importance of being larger: parasitoid

intruder-owner contests and their implications for clutch size. Animal Behavior

51: 1363-1373.

Pritchard, G. 1969. The ecology of a natural population of Queensland fruit fly, Dacus

tryoni. II. The distribution of eggs and its relation to behaviour. Australian Journal

of Zoology 17: 293-311.

131

Shelly, T. E. 1999. Defense of oviposition sites by female oriental fruit flies (Diptera:

Tephritidae). Florida Entomologist 82: 339-346.

Wilson, F. 1961. Adult reproductive behavior in Asolcus basalis (Hymenoptera:

Scelionidae). Australian Journal of Zoology 9: 739-751.

132

Table 9.1. Incidence of damage among foundresses of M. femorata in different sized foundress groups during the first 10 days of their being placed with a prepupal

Trypoxylon politum host.

Initial # of Females With Damage Without Damage Replicates

1 0 6 6

2* 16 9 13

3* 20 5 9

* Three females escaped or were accidentally killed.

133

FIGURE LEGENDS

Fig. 9.1. Two egg-laden female M. femorata locked in combat. Although these scuffles do not tend to be lethal, females often mutilate one another.

Fig. 9.2. The distribution patterns of accumulated dyed fecal spots of three M. femorata females on a Trypoxylon prepupa four days after being placed together. The essentially non-overlapping “micro”territories or spheres of concentrated activity for each of these females is apparent.

134

Fig. 9.1

135

Fig. 9.2

136

CHAPTER 10

CONCLUSIONS

People often misunderstand how much is known about the behavior and life history of insects. It appears to some that much of the work has been done. Perhaps this is the result of a focus on a few model species such as the fruit fly (Drosophila melanogaster). However, most insects are not well studied. Even in model or economically important species there sometimes is a lack of basic behavioral information. For example, how does a pest insect find its host or nesting site? Some aspects of these questions may be known, such as heat/sun cover or age/tree type, but many details are often poorly known. Knowledge of such details could be useful in devising ways to control a pest.

Melittobia is in the middle of the spectrum of behavioral knowledge. Some aspects of their behavior have been investigated because of their use in educational curricula (Matthews et al. 1996). They also have been cited as examples in theories relating to the evolution of sociality, kin selection, and sex-ratio determination (Deyrup et al. 2005, Hamilton 1967, Matthews et al. 2005). Nonetheless, there is still much to learn about its behavior and life history. Thus, my research has focused on investigating little known or disputed aspects of Melittobia behavior.

137

This dissertation represents a journey into the behavior of Melittobia digitata. It began with an investigation of host utilization differences between mated and virgin females. If simultaneously offered two hosts (flesh fly pupae), mated M. digitata females used both. In contrast, virgin females tended to use only one of the hosts. Curiously, the unused second host developed to adulthood significantly less often than did controls.

These results had certain implications. One was that females can discriminate one flesh fly host from another, and choose to avoid oviposition on both. Virgin females laid only a few eggs that always developed into males. This behavior utilizes very little of the host resource, and appears adaptive in that potential hosts remain available, but developmentally arrested, for later full exploitation by the same female (now mated by her offspring). The developmental arrest suggested that the female was physiologically manipulating the host, and this aspect seemed worthy of further investigation.

It seemed likely that the mother was the cause of the developmental delay, since it is common for a female parasitoid wasp to sting a host. However, I felt that there was no need for Melittobia to have paralyzing venom since the hosts we find them on are immobile. Consequently, mealworms were chosen as a host that the female would sting, but which could be easily tested for movement as well as development. I found that mealworms exposed to females were not likely to move when tested, and that they were also unlikely to complete adult development. These results confirmed my suspicions that the mother is responsible for developmental delay. They also indicated that the female can paralyze the host.

It was likely that venom injected by the sting was responsible for the paralysis and developmental delay. It is the first thing that most of us suspect when thinking about a

138 wasp, yet I did not actually isolate the venom and inject it. In addition, the previous study was conducted using a factitious host. Therefore to conclusively demonstrate that the venom was responsible for the paralysis and developmental delay, I developed a technique to milk the venom of the wasp which could then be injected into the mealworm. In addition to the mealworm, I used the alfalfa leaf cutter bee to test for paralysis since it is a natural host of the genus. The venom caused paralysis in both of the hosts, and developmental delay in the mealworm. This work finally settled the controversy in the literature by demonstrating that developmental delay or paralysis can be caused by the venom alone. Equally important, it was a first step in venom characterization for this genus, demonstrating the feasibility of examining M. digitata venom for specific compounds responsible for causing paralysis or delaying development and determining their modes of action.

Having exhausted some of the behavioral avenues for examining host paralysis and developmental delay, I then decided to explore an unusual behavior witnessed during venom milking. After milking a wasp, other conspecific females sometimes appeared to investigate the milked venom. In addition, I noticed that there were sting marks in the center of unfinished cooperatively chewed holes in the plastic culture containers. I decided to utilize the technique I had developed for milking venom, to investigate possible cues for the cooperative chewing.

Using the rearing chambers that the females normally attempt to chew out of, I experimented with several possible cues that I thought could be stimulating the chewing.

A combination of milked venom and artificial pits on the inner surface of rearing containers elicited a stronger chewing response than either stimulus alone. This suggested

139 that the venom-associated cue may be a pheromone that facilitates mutual focused chewing and that these behaviors may have arisen from behaviors associated with the initial stages of host attack. The next step was to find out if the chewing in response to venom was shared by other species, and if the chemicals were active across species groups. For the first objective, I used a closely related species, M. femorata, and tested the role of its venom as a stimulant. This was a little more complicated since M. femorata did not exude venom from the milking technique. As an alternative I dissected out the venom reservoir and alkaline gland from M. femorata females. Contents of these structures were then tested on cultures of M. femorata in the same manner as the test for

M. digitata milked venom. Compared to controls, there was significant chewing. I also wanted to know if the venom chemicals from another species group would elicit chewing in a female M. digitata. To do this I milked venom from M. australica, and presented it to

M. digitata. Almost surprisingly, this species’ venom also elicited chewing. This result suggested that the chemicals involved in the chewing response may be conserved between the species groups.

I was still intrigued by the apparent similarities between the feeding and the chewing behaviors. The initial sequences of the behaviors seemed similar, but then diverged at a certain point. Evidence to support or refute the case for the two behaviors being similar could be obtained through experimentation. Identifying a cue that could switch both behaviors on or off in both contexts would show that the initial parts of the behavior were closely linked, even though they occur in quite different contexts in the wasp’s life cycle. Dahms (1984), working with M. australica, had contrasted the situation when a female lays eggs through a fly host’s puparia, and when she decides to chew into

140 a cocoon, thereby gaining direct access to the host tissues. Although Dahms (1984) did not mention it, I felt that these situations could be separated by whether hemolymph can be fed on through the host puparia or not. This led to my testing hemolymph as the key cue involved in comparison with a non-nutritious liquid (water). By forming a small well in the culture container wall, filling it with one of the liquids, and applying venom to stimulate chewing at that spot, I showed that the wasps would switch from chewing to feeding only if they encountered hemolymph.

Turning on feeding was only half of the story. Since any insect might stop and feed when it encountered food, I felt that in order to have any confidence in the test, it would be necessary to demonstrate that chewing could be turned on when hemolymph was not encountered. Rising to the challenge, I made artificial hosts of Parafilm®, and filled them with either water or hemolymph. In the hemolymph hosts, feeding was present and no chewing occurred, yet when hemolymph was not present, the females chewed large holes. In some cases they chewed right into the water and drowned. The ability of the presence or absence of a single cue to turn both behaviors on or off in both contexts indicates that, at the very least, the first parts of the behaviors are interchangeable, with one likely derived from the other.

There has been a long-standing controversy in the literature as to whether males feed. Incidental observations of wasps in our laboratory cultures provided evidence that males may be feeding on other males in the pupal stage or as adult that they had killed in combat. After long hours of observation I witnessed several incidences of what clearly appeared to be feeding. In one case a victor male had bitten into the head capsule of his defeated brother. His mouthparts moved as a female’s do when they are feeding on

141 liquid, and the male’s abdomen appeared to swell. This convinced me that the males do indeed feed; however, experimental proof was needed to put aside the doubts in the literature.

First, to show that males had a working functional digestive system, I smeared dyed sugar on their mouth parts. Sure enough, the dye was passed in their droppings a day later. Next, to confirm whether hemolymph was being ingested from a defeated male I injected males with dye, and let them fight with a non-injected males. Although there were difficulties, I finally obtained instances of males killing and feeding on the dyed male. The dyed fluid could be seen through the cuticle going into the mouth and ending in the swelling crop. This experiment showed that males will feed on a vanquished rival.

Finally, there was the issue of whether a male gains any benefit from feeding?

Wiltz (unpublished) had done an experiment to determine how many females a male was capable of mating. Surprisingly, the data showed that males generally die before exhausting their sperm. This suggested that longevity might be a better metric of feeding benefits than sperm production. Accordingly, I compared longevity of fed and unfed male, and found not surprisingly, that males who fed, did indeed live longer.

The last project in my dissertation involved less observation on my part. An undergraduate student of Dr. Matthews had noticed that females of M. femorata frequently suffered body damage when placed in groups together on a host. In addition, the females appeared to be confining themselves to a relatively small portion of the host.

Since M. femorata seemed a bit odd in these respects compared to other species in the genus, it seemed worthwhile to attempt to confirm and extend these observations.

142

Accordingly, colonies with differing numbers of M. femorata foundresses were set up to assess the incidence of body damage. Solitary foundress females did not sustain any body damage, yet many of the females in the foundress groups, started missing tarsal or antennal segments as well as acquiring wing damage. I observed several conflicts and watched females bite at others who entered their space. It also appeared that females were confining their activity to a small part of the host and defending it. To get a better idea of how females were spending their time, I fed females of three species M. femorata, M. australica, and M. digitata a 20% fructose solution colored with dye. I then placed three females of different colors on a host. Their droppings left a record of where the female had stood, confirming territoriality in two species, but not in M. digitata. In addition behavioral observation and records of damage showed that although M. australica and M. femorata were both territorial, M. australica did not violently defend their territory.

My research has hopefully increased in some small amount our knowledge of this very interesting wasp. Some of my projects served to simply record unusual or interesting behaviors, while others were designed to resolve controversies. Despite the time consuming nature of behavioral research, my journey has been a thoroughly enjoyable experience. Along the way I got to watch the most exciting reality shows around: Sisters fighting turf wars, brothers killing and eating brothers, chemical subterfuge against other species, and even sex! All this excitement, and no guilt, since I had the satisfaction of knowing that I have added to our knowledge of the how the world works.

143

REFERENCES

Dahms, E. C. 1984b. A review of the biology of species in the genus Melittobia

Hymenoptera: Eulophidae) with interpretations and additions using observations

on Melittobia australica. Memoirs of the Queensland Museum 21: 337-360.

Deyrup, L. D., R. W. Matthews, and J. M. González. 2005. Cooperative chewing in

Melittobia digitata Dahms, a parasitoid wasp, is stimulated by structural cues and

a pheromone in crude venom extract. Journal of Insect Behavior 18: 293-304.

Hamilton, W. D. 1967. Extraordinary sex ratios. Science 156: 477-488.

Matthews, R. W., T. R. Koballa, Jr., L. R. Flage, and E. J. Pyle. 1996. WOWBugs: New

Life for Life Science. Riverview Press, LLC, Athens, GA.

Matthews, R. W., L. D. Deyrup, and J. M. González. 2005. Increased male sex ratio

among brachypterous progeny in Melittobia femorata, a sib-mating parasitoid

wasp (Hymenoptera: Eulophidae). Insect Science (In press).