REPRODUCTIVE SUCCESS OF BAGWORMS,

OIKETICUS -YI (GUILDIBIG) AND (WALKER)

(:PSYCHIDAE)

Marc Rhainds B.A., Université de Montréal 1986 B.Sc., Universite Laval 1991

THESIS SUBMlTTED iN PARTIAL OF THE REQUlREMENTS FOR THE DEGREE OF =OR OF PHILOSOPHY

In the Department of B iological Sciences

8 Marc Rhainds 1999 SIMON FRASER üNIVERSl'N December 1999

Al1 rights ~served.This work may not be reproduced in whole or in part, by photocopy or other means, without permission of the author. National Library Bibiiothdque nationale du Canada Acquisitions and Acquisitions et Bibliognphic Services services bibliographiques 385 Wdlngtm Street 395, rus Wellington Oîîawa ON K1A ûN4 ûttawaON K1AW CMed. Gonade

The author has granted a non- L'auteur a accordé une licence non exclusive Licence ailowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, disûiiute or seil reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantiai extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otheMrise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. Abfitract. Unusual life history of bagworms (Lepidoptera: Psychidae), with sessile females completing their reproductive activity within a self-constnicted bag, provides an oppominity to evaluate factors infiuencing lifetime reproductive success. Sarnpiing and experimental studies conducted in plantations of oii palms investigated density-dependent processes and life history traits that affect reproductive success and population dynamics of 2 allopatric bagworms, Oiketicu~kirbyi in Costa Rica and itlefisuplmu in Southest Asia. Simiiar within- and between- distribution of M. plana in parental and offspring generations indicated that emergent larvae commonly do not disperse, suggesting that local populations of bagworm exhibit high level of genetic relatedness. Density- and defoliation-dependent dispersal by larvd M. plana contributes to their reproductive success, because larvae on crowded palms attain srnail size at pupûtion, and smali bagworms have low survival during pupal stage (male and female M. plana), mating success (female 0. kirbyi and M. pkuna) and fecundity (femde O. kirbyi and M. plana). Greater incidence of bailooning by femde than male lame rnay be due to females seeking suitable hosts for their progeny. For 0.kirbyi, mortality during pupal stage was greater for females than males; the opposite trend was obsewed for M. plana. For both O. kirbyi and M. plana, females emerged before males; protogyny may be a comrnon life history trait in bagworms that contribute to reduce

inbreeding. Femde O. kirbyi produced a blend of 4 chiral esters to attmct males for mating.

As previously reported for other bagworms, relatively large proportions of female O. kirbyi

and M. plana did not mate as adult. Size-dependent mating success of female bagwoms may

be attributed, in part, to males preferentidy mating with large, moa fecund females.

Density-dependent dispersal by Iwae, as well as small size and low mating success of

females on crowded palms, contribute to stabilize local populations of M. plana. Individual

variations of femaie reproductive success iduenced the population dynarnics of M. pfana. Cette thèse est dédiée b mes parents ACKNOWLEDGMENTS

1 first and foremost express my deepest gratitude to Dr. Gerhard Gries, my senior supervisor, for giving me the opportunity to conduct research in various ports of the world, and for generously providing me with materiai, academic and penonal support throughout my thesis. 1 also deeply appreciated the research freedom Gerhard gave me as a graduate student. I extend my appreciation to Regine Gries, for helping me with pheromone studies, and more importantiy for her generosity and honesty. Gerhard and Regine will remain very special fnends for the rest of rny life.

Mmy thanks go to Da. John H. Borden and Keith N. Slessor, my committee supervisors, for their help, assistance and advice.

This thesis would not have been possible without field assistance and logistic support from several collaborators: D.L. Richardson. R. Escobar, C. Chinchilla, A.C. Oehlschlager,

G. Castrillo, Y. Campos, I. Bulgareili, N. Bamntes and M. Corrales from Palma Tica in

Costa Rica; Y. Sasareila. G. Brown, H.L. Foster, A. Saieh, T. Rhamdan Noor, S. Sutanto, and

H.Z. Abidin fiom London Sumatera in Indonesia; P.S. Chew, K.J. Goh, M.M.Min, B.S. Sim,

P. Valu, and B.N. Ang fiom Applied Agriculture in Malaysia; Ho Cheng Tuck, Dr. Mohd

Hashim Ahmad Tajudin, Dr. Lee Cheng Hee, Fahl Ariffin Saini Shahanidin Bakar,

Mohaimi Mohamad, Abdul Mm,Abdul Latif Rarnlam, Hasan Hahi Saleh, Jaialudin Abdul

Azia, Abdul KaHee Ahmad. Mohamad Rais Ramlam, Adina Rusdi fiom Golden Hope OPRS in Malaysia. The help of R.G. Rodriguez was essential for sampling bagworm populations in Costa Rica. Palma Tica, Appiied Agriculniral Research and Golden Hope generously provided research Ming. 1 acknowledge E. Carefoot for preparation of figures, R.F. Balshaw for assistance with statisticai analyses, J. Li and P. Zhang for chemicd syntheses, G. Owen for mass spectral analyses, and M. W. Basri for shipment of .

I received financiai support from a Natural Sciences and Engineering Research Council

(NSERC) Scholarship, a Fonds pour la Formation de Chercheurs et l'Aide à la Recherche

(FCAR) Scholarship, a Marshall Noble Scholarship, three SFU Graduate Fellowships, an

H.R. MacCarthy Graduate Bursary, a President Ph.D. Research Stipend, and a NSERC research grant to G.G. vii

TABLEOFCONTENTS

TITLE PAGE...... 1 APPROVAL PAGE ...... i i ABSTRACT ...... iii DEDICATION ...... iv ACKNO WLEDGEMENTS...... v TABLE OF CONTENTS ...... vii

LIST OF TABLES ...... Y LIST OF FIGURES ...... xi 1*O INTRODUCTTON ...... 1 Intraspecific variations of reproductive success and population dynarnics..... 1 A short review on Lepidoptera with fligthtless fernales ...... 2 Life history of bagwoms ...... 3 1.3.1 Land stage ...... 6 1.3.2 Pupal stage ...... 6 1.3.3 Adult stage ...... 7 1.3.4 Egg stage and emergent larvae ...... 14 Bagwonns on oil palm ...... 14 Population dynamics of bagworms on oil palm ...... 17 Life history traits related to reproductive successof bagworms ...... 11 1.6.1 Larval stage ...... 21 1.6.1 .1 Density-dependent dispersal ...... 34 1.6.1.2 Size attained at pupation ...... 26 1.6. 1.3 Pupation site ...... 27 1.6.2 Adultstage ...... *...... *...... 31 1 .621 Timing of emergence ...... 31 1.6.2.2 Sexuai communication.. ...*,...... *...... ,*...... *...32 .. Specific research objechves...... 32 1.7.1 kirbyi ...... 33 1.7.2 Metisa plana ...... 33 viii

MATERIALS AND METHODS ...... 35 Study sites ...... Dispenal by larval Metisa plana ...... 2.2.1 Sarnpling studies conducted in plantation of oil palm ...... 2.2.2 Expenrnents conducted in cage environment...... Assessrnent of reproductive success for 0. kirbyi and M. plana ...... 2.3.1 Oiketim kiroyi population in Coto 50 ...... 2.3.2 population in Coto 52 ...... 2.3.3 Metisa plana population in Sungei Merah ...... 2.3.4 Metisa plana population in Ladang Coalfield ...... Size-dependent mating success of female Oiketicirs kirbyi ...... 2.4.1 Field study ...... 2.4.2 Cage study ...... Sexual communication between male and female 0iketicu.s kirbyi ...... 2.5.1 Laboratory andysis ...... 2.5.2 Field expenments ...... Height-dependent captures of male O. kirbyi in pheromone-baited traps ...... Individual variations of reproductive mccess and population dynamics...... STATISTICAL ANALYSES ...... RESULTS ...... 4.1 Inter-site and inter-generational variations of population density ...... 4.2 Dispersal by land Metisa plana ...... 4.2.1 Sampling studies conducted in plantations of oii palm ...... 4.2.2 Experiments conducted in controlled cage environment...... 4.22.1 Recapture of larvae and pupae ...... 4.2.2.2 Rate of within-leaf dispersal ...... 4.2.2.3 Time and rate of ballooning...... 4.2.2.3 Sex-specific ballooning and pupation behaviour ...... 4.3 Size attahed by larvae at pupation ...... 4.4 Vertical distribution of male and female pupae on oil palm ...... 4.5 Mortality during the pupal stage ...... 4.7 Sexuai communication between male and female Oiketicus fiby i...... 103 4.8 Mating success of females ...... 126 4.9 Fecundity of females...... 139 4.10 Inter-generationai variations of population density...... 139 4.10.1 Sampling study conducted in plantation of oi1 paim ...... 139 4.10.2 Experiments conducted in controlled cage settings ...... 154 4.1 1 Individual variations of reproductive success and population dynamics...... 154 4.1 1.1 Cage study with Metisa plana ...... 154 4.1 1.2 Field study with Merisa plana ...... 161 5.0 DISCUSSION ...... 162 Components of reproductive success...... 162 5.1.1 S wival during pupal stage...... 162 5.1.2 Mating success...... 164 5.1.3 Fecundity ...... 166 Life history traits related to reproductive success: larval stage...... 167 5.2.1 Dispersal...... 167 5.2.2 Size attained at pupation ...... 171 5.2.3 Pupation site...... 172 Life history related to reproductive success: adult stage ...... 173 .. 5.3.1 Timing of emergence ...... 173

. f 5.3.2 Sexual communicatron ...... 174 Density-dependent processes and population dynamics...... 176 .. 5.4.1 Oiketicus kbyi...... 177 5.4.2 Metha ...... 178 Individual variations of reproductive mccess and population dynamics ...... 179 6.0 LITERATURE CITED ...... ,.. 181 LIST OF TABLES

Table I . Life history parameters of Oiketicus kirbyi and Metisa plana ...... 20

Table 2 . Characteristics used to evaluate reproductive success of bagworms ...... 22

Table 3 . Examples of sex-specific pupation behaviour in lepidopteran with flightless females and flighted males ...... 29

Table 4 . Site-specitic characteristics of oil palm and sampling procedures for different experimental plantations ...... 40

Table 5. Demographic parameters and components of reproductive success recorded for different populations of Oiketicus kirbyi and Metisa plana ...... 42

Table 6. Statisticd analyses for studies carried out in Coto 50. Coto 52. Sungei Menh. and Ladang Coalfield...... 50

Table 7 . Statistical analyses for cage experiments canied out at OPRS ...... 62

Table 8. Descriptive statistics of population density in various locations ...... 68

Table 9 . Numben of ballooning Metisa plana larvae mnrked and recaptured on oil palms in Ladang Coalfield ...... 74

Table 10. Numbers of Mefisuplana larvae released and recapnired on nursery palms in cage expenments carried out at OPRS ...... 75

Table 1 1. Relationship between length of pupal bags and population density for 5 consecutive genemtions of Metisu plana ...... 92 Table 12. Relationship between population density. Iength of pupal bags and pupd mortality for 5 consecutive generations of Metisa plana ...... 102

Table 13. Parameters of adult emergence for Oiketicus kirbyi and Metisa plana ...... 108

Table 14. Relationship between population density. length of pupal bags and mating success of femaies for 5 consecutive genentions of M .plana ...... 143

Table 15. Relationship between numbers of Metisa plana empty fernde pupal cases and early instars per leaf for 39 oil palrns ...... ISO LIST OF FIGURES

Fig .2 . Winged male Oikeficus kirbyi ...... 9

Fig . 3 . Apterous female Oiketicus kirbyi ...... 11

Fig .4 . Female Oiketinrs kirbyi with pheromone-impregnated scales...... 13

Fig . 5 . Oil palms with moderate and severe defoliation ...... 16 Fig . 6 . Phyllotaxy of oil palm ...... 19

Fig . 7. Nurnbers of pupai bags during 5 consecutive generations of Metisa planu ...... 70 Fig .8 . Relationship between population density and incidence of ballooning arnong lava of Metisa plana ...... 72

Fig .9 . Relationship between dependent and independent variables for cage experiments that investigated dispersa1 behaviour of larval Metisa plana ...... 77

Fig . 1 0. Proportions of larvd Metisa plana that ballooned fiom nursery palms during different I =hour4ntervals...... 80

Fig . 1 1. Relationship between rate of ballooning and nurnber of released larvae for male and female larval Metisa pfunu that attained the pupai stage ...... 82 Fig . 12. Relationship between rate of ballooning and natural defoliation of palms for male and female larvai Metisa plana that attained the pupal stage ...... 84

Fig . 13. Numbers of mole and female pupai Mefisaplana recorded on undefoliated and artificially-defoliated palms ...... 86

Fig . 14 . Length of pupal cases of fernale Oikeiim kirbyi on different leaf groups of oil pahs...... 89

Fig . 15 . Lengths of pupal bags during 5 consecutive generations of Metisa plana ...... 91

Fig . 16 . Numbers of pupal begs on different leaf groups of oil palm ...... 94

Fig .1 7 . Mortaiity of Oiketicus kirbyi pupae on different leaf groups of oil palrn ...... 97

Fig . 18 . Mortaiity of pupal Metisa plana pupae during 5 consecutive generations...... 99 xii

Fig. 19 Lengths of pupal bags of Metisa plana containing live or dead pupae...... 10 1

Fig. 20. Emergence patterns for male and female Oiketicus kirbyi...... 105

Fig. 21. Emergence patterns for male and fernale Me tisa plana...... 107

Fig. 22. Flarne ionization detector and electroantennographic detector responses to pheromone extract of femde Oiketicus krbyi...... 1 10

Fig. 23. Number of male Oiketicus kirbyi capnired in tnps baited with different ratios of(@-and@)-MBD ...... 113

Fig. 24. Number of male Oiketicus kirbyi captured in traps baited with (R)-MBD alone and in combination with optical isomers of MBO, MBN,and MBDD.. .. 1 15

Fig. 25. Number of male Oiketicus kirbyi captured in traps baited with (R)-MBD in binary and pentanary combination with (R)-MBO, (R)-MBN. (R)-MPD and (R)-MBDD...... 1 17

Fig. 26. Number of male Oiketicus kirbyi captured in traps baited with (R)-MBD in ternary and pentanary combination with (R)-MBO, (R)-MBN, (R)-MPD and (R)-MBDD ...... 1 19

Fig. 27. Number of male Oiketicus kirbyi captured in traps baited with (R)-MBD in quatemary and pentanary combination with (R)-MBO, (R)-MBN, (R)-MPD and (R)-MBDD...... 12 1

Fig. 28. Nurnber of maie Oiketicus kirbyi captured in traps baited with (R)-MBD, (R)-MBO, and (R)-MBN in different ratios...... 123

Fig. 29. Number of maie Oiketicus kirbyi captured in traps baited with (R)-MBD alone and in combination with (R)-MI30 and (R)-MBN ai increasing doses.. .. 125

Fig. 30. Lengths of pupal cases of mated and unmated femaie Oiketicus kirbyi...... 128

Fig. 3 1. Lenghts of pupal bags of mated and unmated femde Me tisa plana...... 130

Fig. 32. Proportions of mated femaie Oiketim kirbyi for different weeks and leaf groups of oii palms in Coto 50...... 132

Fig. 33. Proportions of mated female Oiketicus kirbyi for different weeks and leaf groups of oil palms in Coto 52...... 134

Fig. 34. Relationship between length of pupal case and mating success of femaie Oiketims kirbyi on different leaf groups of oil paims...... 136 xiii

Fig. 35. Number of male Oiketicus kirbyi captured in pheromone-baited traps suspended at 3 heights on oil palms ...... 138 Fig. 36. Proportions of mated females during 5 consecutive generations of Me tisa plana...... 141 Fig. 37. Relationship between length of pupal case and weight of egg mass for mated female Oiketicus kirbyi...... 144 Fig. 3 8. Relationship between length of pupal bag and nurnber of eggs laid by mated fernale Metisa plana...... 146 Fig. 39. Numben of Metisa plana empty female pupal cases and early instars on different leaf groups of oil palms...... 149 Fig. 40. Relationship between numbers of hfetisaplana empty female pupal cases and early instars per palm...... 153 Fig. 41. Relationship between numbers of Metisa plana empty female pupal cases and early instars per palm for palms 9 m and > 9 m from heavily infested palms...... 156 Fig. 42. Numbers of enrly instar Metisaplana on nursery palms with or without fernale pupal bags...... 158

Fig. 43. Variations of population density between consecutive generations of Me fisa p[ana...... 160 1.0 INTRODUCTION

1.1 Intnspecific variations of reproductive success and population dynamics

Understanding parameters that regulate spatial and temporal variation of population density in herbivorous insects has long been a fundamental issue in ecology.

Nicholson (1933) first emphasized that populations are balanced in nature by the action of density-dependent factors that inthence birth andlor mortality rate. Life-table analyses

(Morris & Miller, 1954; Corne11 and Hawkins, 1995) provide a useful tool to detect density-

dependent processes among immature insects (Varley & Gradwell, 1960) but may not be adequate for identiQing parameters that uitimately regulate populations of insects, because

they focuss on mortality factors rather than on reproductive output of adult females (Price et al.. 1990) and ignore variations in quality among females (Rossiter 1992, 1995).

Experirnents carried out with herbivorous insects have evaluated the extent and

causes of variation for several components of reproductive success, including longevity of

adults (Elgar & Pierce, 1988; Naugler & Leech, 1994; Ohgushi, 1996), mating success

(Thornhill & Alcock, 1983; Anderson, 1994; Choe Be Crespi, 1997). fecundity (Honek, 1993;

Tamrnaru et al., 1996a), and fitness of progeny (Leonard, 1970; Palmer, 1985; Ohgushi,

199 1; Cushman et al., 1994). Recent studies suggest that variability of reproductive success

among females influences their population dynamics (Elliott, 198 1; Wall & Begon, 1987;

Rossiter, 1992; Tammaru et al., 1996a; Ohgushi, 1996), but testing this hypothesis in naturd

conditions has proven difficult. Because winged insects are srnail and highty mobiie, it is

difficult to follow individuals throughout their lifespan to estimate their reproductive output,

and only few studies have investigated intraspecific variations of lifetime reproductive

success in natural populations of herbivorous imects (Cushman et al., 1994). The reproductive biology of insect species with non-dispersing females provides an opportunity to investigate intraspecific variations of reproductive success in relation to population dynamics. Individual variations of realized fecundity have been quantified in

field populations of insects with fiightless femaies (Barrows, 1974; Mason et al., 1977;

Williams et al., 1990; Sopow & Quiring, 1998); these studies, however, do not fùlly disclose the variation of lifetime reproductive success. because they ignore the tact that some females may not mate during their lifetime.

In my thesis, I report sampling and experimental studies carried out with two bagwomis, Oikeficzcskirbyi (Guilding) and Metisa plana (Wdker) (Lepidoptera: Psychidae),

in commercial plantations of oil palm. My major research objective was to investigate density-dependent processes and life history traits that influence reproductive success and

population dynamics of bagworms. 1 maidy used an empirical approach, addressing the

need for empirically-based theory of population dynamics as one of the most important challenges posed to insect ecologists (Cappuccino, 1995). 1 intended to narrow the wide gap

between theoretically- and empirically-derived theories of population dynamics, as described

by Pnce and Hunter (1995): "theory built on top of theory is an unfortunate feature of much

of the literature on population dynamics.. . We might characterize the current situation as a

theoreticai hippo balanced on the head of an empincal pin". Results presented in rny thesis

rnight help to put the theoretical hippo back on its feet.

1.2 A short review on Lepidoptera with lghtless females

Despite the adaptive significance of flight for colonizing new habitats and locating

mates or oviposition sites, secondary loss of flight capacity by females has evolved independently several times in the Lepidoptera (Hackrnan, 1966; Sattler, 1991 ;Hunter,

199Sa). Interspecific cornparison of life history traits for insects with flighted and flightless

females suggests that loss of flight capacity is adaptive in ecological contexts where the advantage of flightlessness (enhanced fecundity: Ro ff, 1986) outwei ghs the disadvantage of

reduced motility, as is fiequently observed in stable habitats with large canying capacity

(Roff. 1990; Barbosa et ul., 1989). Flightiessness is also cornmon in insects 1) that inliabit woodlands, deserts, the ocean surface, seashores, and mountain habitats; 2) that are active as adults in the winter; and 3) that reproduce parthenogenetically (Roff, 1990). Extreme

reduction of motility is found in lepidopteran species with micropterous or apterous females

that mate and oviposit near their emergence site. Sedentary females are most common in

bagworms, but are also found in Lymanûiidae, and Ardidae (Sattler 199 1).

1.3 Life history of bagworms

The Psychidae represent a highly speciaiized lepidopteran with about 2000

species, most of which have micropterous or apterous females (Sattler, 1991). Psychidae are

known as bagworms because larvae complete development within self-constnicted bags (Fig.

1). Bagworms have a worldw*dedistribution (Sattler, 1991) but may be most diverse in

tropical environments (Barbosa, 1993). Primitive bagworms (Micro-Psychina) are generally

scavengers, and include species with winged or wingless females, and with sexual or

parthenogenetic reproduction. The Macro-Psychina, in conhast, are typically plant feeders

with sexuai reproduction and apterous femaies (Jones, 1927; Davis, 1964; Davis, 1975;

Sattler, 1991). In this section, I review the life history of bagworms, with particular emphasis

on the Macro-Psychina. Fig. 1. Tvto eerirly instar Oiketicus kirbyi (mw).The Iarva on the left was removed fiom its protecting bag.

1.3.1 Larval stage

Upon hatching and generally before feeding, neonatal larvae envelop themselves in a self-constnicted bag made of silk and plant material. As larvae grow, they continually eniarge their bag (Davis, 1964). The ability to construct a bag de novo is restricted to early instars. Mature larvae removed from their bag cannot survive (Kaman, 1968; Bourgogne.

1993). Femaie bagwonns have greater feeding recquirements and longer larval development than males (Morden & Waldbauer, 1971 ; Cruttwell. 1974; Syed, 1978; Ponce et al., 1979;

Thangavelu & Ravinhath, 1985; Villanueva & Granda Paz, 1986; Campos Arce et ul..

1987; Basn & Kevan, 1995). Bagworm larvae are generally polyphagous (Davis, 1964).

Possession of a self-enclosing bag may be of adaptive significance for bagworm larvae in facilitating development and reducing mortality. 1) Bags rnay generate microclimatic conditions thnt fbrther larval development (Barbosa es al., 1983; Smith &

Barrows, 199 1). 2) Enhanced survival of early instars ballooning with, rather than without, a bag suggests that bags alleviate desiccation during dispersal bouts (Cox & Potter, 1986). 3)

Larvae close the antenor part of their bag before moulting or when disturbed (Jones & Parks,

1928; Stephens, 1962; Kaufman, 1968; Wood. 1968; Ameen & Sultma 1977; Syed, 1978;

Mishra, 1978), suggesting that bags physically protect bagworms against pmitoids or predators. 4) The cryptic appearance of bags (Johnson & Lyon, 1976; Ameen & Sultana,

1977; Neal & Santamour, 1990) may also serve as camouflage.

1.3.2 Pupal stage

The Mly grown tightly attaches the anterior part of its bag to a substnte

(usually on the plant), adds more silk inside the bag, and reverses its position within the bag to pupate head-downward (Davis, 1964). This reversai of position is essential for successful emergence of males (Kaufman, 1968) and reproduction of females (Jones, 1927).

Female pupae are larger than male pupae and lack typical lepidopteran appendage sheaths

(Davis, 1964). The pupal stage is shorter for females than males (Stephens. 1962; Cruttwell,

1974; Mishra, 1978; Ponce et al., 1979; Thangavelu & Ravindranath, 1985; Basri & Kevan,

1995). Sex-ratios of Iabontory-reared bagworms are male-biased (Heather, 1975; heen&

Sultana, 1977; Howlader, 1990; Basri & Kevan, 1995).

1.3.3 Adult stage

Shortly before ernergence. the male thsts the head and thoracic segments of his pupal case through the posterior end of the bag. Upon dehiscence of the anterior part of the pupal case, the winged male emerges and is sexually active shortly thereafter (Fig. 2; Davis.

1964). The wingless adult female, in contrat. does not leave her pupal case and bag (Fig. 3;

Davis, 1964). The fernale breaks open the anterior part of her pupal case and releases pheromone-impregnated scales (hairs) into the lower end of the bag and so attracts mates

(Fig. 4; Stephens, 1962; Bosman & Brand, 1971; Zhao, 198 1; Leonhardt et al., 1983; Acosta,

1986; Ned, 1986; Loeb et al., 1989).

Upon landing on the bag of a receptive femde, the male inserts and extends his abdomen through the posterior bag opening into the pupal case to reach the fernale's genitalia for copulation (Jones, 1927). Almost immediately afler mating, the female lays eggs (mixed with abdominal scales) in the postenor section of her pupal case, and drops to the gmund thereafter (Davis, 1964). Unmated females vacate their bag before death. but unlike mated females do not lay eggs in their pupal case (Jones, 1927; Stephens, 1962; Entwistle, 1963; Fig. 2. Winged male Oiketicus kirbyi resting on his empty pupal case shortly der emergence.

Fig. 3. Apterous fernaie Oikrtim kirbyi removed from her pupal case.

Fig. 4. Female Oiketicus kirbyi with pheromone-impregnated scales expelled from her pupd case.

Bosman & Brand, 197 1; Syed, 1978; Mishra, 1978; Thangavelu & Gunasekaran, 1982;

Howlader, 1990; but see Malicky, 1968). Male and female bagworms are usually short-lived and die less than one week after eclosion (Stephens, 1962; Entwistle, 1963; Kaufman, 1968;

Thangavelu & Gunasekaran, 1982; Kuppusarni & Kannan, 1993; Basri & Kevan, 1995).

Males of laboratory-reared bagworms emerge either in synchrony with (Villanueva &

Granda Paz, 1986; Basri & Kevan, 19951, before (Syed, 1968; Heather, 1975; Ponce et al.,

1979), or tifter females (Stephens, 1962; Morden & Waldbauer, 1971; Thangavelu &

Ravindranath, 1985; Gara et ai., 1990; Kuppusamy & Kannan. 1993). However. pattern of adult emergence have not yet been recorded in the field.

1.3.4 Egg stage and emergent Iarvae

Fecundity of female bagworms varies between species fiom less than 100 (Wood,

1968; Wheeler & Hoebecke, 1988; Bourgogne, 1990) to more than 1,000 (Balduf. 1937;

Entwistle, 1963; Wood, 1968; Davis, 1975; Bourgogne, 1990) or even 10.000 eggs

(Cruttwell, 1974). Embryonic development is usually completed in less than one month

(Stephens, 1962; Entwistle, 1963; Syed, 1978; Thangavelu & Ravindranath, 1985; Howlader.

1990; Basri & Kevan, 1995; but see Neal et al., 1987). Neonatal larvae emerge synchronously fiom the posterior opening of their matemal bag.

1.4 Bagworms on oil palm

The oil palm, guineensis (Jacquin) (Palmae), is a tropical monoecious monocotyledon native to Anica and introduced in Southeast Asia and Latin Arnerica (Fig. 5).

A mature oil phis characterized by an efongated stem around which Ca. 40 leaves are Fig. 5. Oil palrns with moderate (left) and severe (right) defoliation by Oiketicus kirbyi. distributed in ascending spirais (Fig. 6). Each leaf consists of a 3-4 m long rachis with 160-

180 folioles on each side. Female inflorescences in the leaf axil develop into large infnctescences that are harvested for extraction of edible (Hartley, 1988). Oil palms are cultivated in > 4.5 million hectares, mostiy in Malaysia and Indonesia. Production of pdm oil exceeds 14.5 million metric tons annually (Oil World, 1993; cited in Ortiz &

FeniBndez, 1994).

Bagworms, M. plana and 0.kirbyi, are important defoliators of oil palm in Southeast

Asia (Wood, 1968) and Latin Amenca (Genty et al., 1978), respectively (Fig. 5). Defoliation by bagworms adversely affects productivity of oil palm by reducing both nurnber and size of fruit bunchcs (Wood er al., 1973). Control rneasures against bagworm Iarvae include trunk injections of monocrotophos (Wood et al., 1974), and ground or aeriai applications of lead arsenate, trichlorphon or Bacilh thuritzgiensis (Wood, 1968; Almela Pons et al., 1972;

Villanueva & Granda Paz, 1986).

Both M. plana and 0. kirbyi are typicai bagworms that develop, pupate and (for fernales) reproduce within self-constnicted bags (Stephens, 1962; Wood, 1968). In comparison with M. plana, 0. kirbyi is a relatively large bagworm with a wider range of hosts, greater feeding requirement, longer development, and superior fecundity (Table 1).

1.5 Population dynamics of bagworms on oil palm

Both M. plana and 0. kirbyi are generaily present at low population density in plantations of oil palm, but may occasiomally reach epidemic level (Villanueva & Granda

Paz, 1986; Basri et al., 1988). Factors such as such as dry weather, absence of ground cover vegetation, low incidence of naturai enemies, or road dust, have been associated with the Fig. 6. Phylllotaxy of oil palm (adapted from Hdey, 1988). Leaves are produced in ascending spirals with leaf 1 being the youngest and highest.

Table 1. Life history parameters of Oiketicus kirbyi and Metisa plana on oil palm ' .

- .. Metisa plana Oikeficus kir by i

Number of recorded host-plant species 4 54

Development penod (days) from egg to adult 80 - 107 166 - 320

Fecundity (eggs per female) 1 1 - 300 2,800 - 10,000

Length (mm) of male pupal bag Length (mm) of female pupal bag

Leaf area (cm2) damaged during development 16.8 304.5

Potential defoliation (nurnber of leaves damaged) 0.05 caused by progeny of one mated female

References for M. plana: Wood, 1968; Syed, 1978; Desmier de Chenon, 1982; Syed & Saleh, 991; Basri, 1993; Basri & Kevan, 1995. For 0.kirbyi: Stephens, 1962; Davis, 1964; Cruttwell, 1974; Davis, 1975; Genty et al., 1978; Ponce et al., 1979; Newman, 1980; Villanueva & Granda Paz, 1986; Villanueva & Avila, 1987; Gara et ai,,1990.

For Newman (1980), only species with mean defoliation levels 3.0 included as host- .

Lengths of pupal bag given for 0. kirbyi larvae on banana leaves.

Mean value for male and female larvae (10.6 cm2and 23.0 cm2, respectively)

Assuming: 1) fiond area of 102 800 cm2 per leaf(Basri, 1993); fecundity of 300 (M. danal and 10.000 (0.kirbufi: 31 1 :1 sex-ratio: 4) 100 % survival durine larvai stage. development of outbreaks (Wood. 1968; Syed & Shah. 1976; Basn et al., 1988). However, few field studies have assessed the impact of these factors (but see Ho & Teh, 1999). A sampling study carried out in an oil palm plantation for different tirne penods suggested that natural enemies do not regulate populations of M. plana: although numbers of parasitoids and bagworms were positively correlated, proportions of dead pupae decreased with population density (Basri et al.. 1995). Infestations of .If plunu typically spread from localized centers

(Syed & Shah, L 976). suggesting that dispersal by larvae contributes to spread outbreaks.

1.6 Life history traits related to reproductive success of bagworms

To ceproduce, insects must reach the adult stage, mate, and (for females) oviposit on suitable hosts. Unusual life history of bagworms. with females completing ail reproductive activities within their bag, makes them ideal insects for investigating intraspeci fic variations in lifetime reproductive success. The content of individual bags reveals whether or not the specimen under investigation completed larval development, survived during the pupal stage, emerged as adult, and (for females) mated and oviposited (Table 2). Bagworms also provide a mode1 system to study principles of insect population dynamics (Sheppard, 1975): because pupal bags remain attached to palm leaves even after individuals died, it is possible to assess variations of population density between consecutive generations. In this section. I review life history traits that may affect reproductive success and population dynamics of bagworms.

1.6.1 Larval stage

Life history traits related to reproductive success and population dynamics of bagwonns rnay Uiclude densitydependent dispersai by iarvae (section 1.6.1.1), size attained Table 2. Characteristics used to evaluate reproductive success of bagworms.

Component of reproductive success Characteristic of bags

Pupation ' Anterior part of bag tightly attached to a substrate, usually on the host-plant

Sex of pupae with (male) and without (fernole) appendage sheaths; larger bag and pupal case for males than females

Mortality during pupd stage ' Desiccated, diseased, parasitized, or preyed upon pupa; tom, perforated bag indicating predntion / parasitism

Successful emergence of adults Male: empty pupal case protruding fiom anterior section of bag; dense silking in bag '

Female: rupture of anterior segment of pupal case; presence of pheromone laden-hûin in bag

Mating status of femaies ' Receptive (virgin) female: pheromone Iaden- scales in bag plus live female in pupal case

Mated femde: eggs or neonotes in pupd case; filamentous web spread by dispersing neonates inside pupal case

Unmated female: pheromone-iaden scales in bag plus dead or no female in pupal case

Fecundity of females ' Number or weight of eggs inside pupal case of mated fernale; number of larvae emerging from matemal bag Table 2. (continued)

' Section 1.3.1; Jones, 1927: Figs. 26,27; Jones & Park, 1928: Figs. 4, 16, 17; Wood, 1968; Figs. 39,40; Villanueva & Avila, 1987: Fig. 4.

Section 1.3.2; Jones, 1927: Figs. 16, 17; Jones & Park, 1928: Fig. 5; Stephens, 1962: Fig. 1; Misha, 1978: Figs. 6,7;Newman, 1980: Fig. 4; Fulvia Garcia, 1987: Fig. 4; Basri, 1993: Fig. 1.6.

Barrows, 1974; Sheppard. 1975; Hom & Sheppard, 1979; Ba~tiet ni., 1995.

' Section 1.3.3; Jones, 1927: Fig. 16; Jones & Park, 1928: Figs. 6, 19.4, l9.6,20.2,21.3; Davis, 1964; Figs. 183, 184, 185, 189, 190, 195,202; Wood, 1968: Fig. 34; Davis, 1975: Figs. 64,67,68,71,79, 83, 85,89,95, 100, 103, 104, 106.

Section 1.3.3; Leonhardt et al., 1983: Fig. 1; Villanueva & Granda Paz, 1986: Fig. 3.

Section 1.3 3;Sheppard, 1975; Klun et aL, 1986.

' Sections 1.3.3 and 1.1.4; Jones. 1927: Fig. 28; Jones & Park. 1928: Fig. 11; Wood. 1968: Fig. 36; Villanueva & Avila, 1987: Fig. 1. by larvae at pupation (section 1.6.1.2), and pupation site (section 1.6.1.3).

1.6.1.1 Density-de pendent dispersal

Mechmisms used by insects to alleviate the negative impact of crowding include discriminate oviposition by females (Renwick & Chew. 1994). induction of diapause (Brown et di., 1979; Hagstrum & Silhacek, 1980), and dispersai by adults (Dempster, 197 1; Ziegler.

1978; Hemg, 1995; Fadamiro et ul., 1996) or laivae (Zurlini & Robinson, 1979; Poirier &

Borden, 1992; Berger, 1992; Torres-Vila et al., 1997). Dispersal by larvae is of particular importance in species with females that have limited motility and oviposit on one or few host plants, irrespective of previous colonizntion by conspecifics.

In Lepidoptera with flightless fernales, one major mode of dispersal involves ballooning by early instars. Ballooning larvae suspend themselves from from a silken thread for wind-dispersal to surrounding hosts (Wickman & Beckwith, 1978; Mason & McManus.

1981; Cox & Potter, 1986). Larvae most readily balloon from unsuitable (van der Linde,

197 1; Capinera & Barbosa, 1976; Lance & Barbosa 198 1), least favoured (Lance &

Barbosa, 198 1; Ramachandran, 1987), or severely defoliated hosts (Cox & Potter, 1986;

Harrison, 1997). Wind-dispersed larvae have low capacity to locate a suitable host, and may experience high mortaîity during dispersal (Foltz et al., 1972; Cox & Potter, 1986; Terry et aL , 1989; Price et al., 1990). However, lmae emerging on a suitable host cornmonly do not disperse (van der Linde, 1971 ; Campbell et al., 1975a; Ramachandran, 1987; Harrison, L 994,

1997; Weseloh, 1997; see also Gross & Fritz, 1982: 153).

Interspecific differences in defoliating and reproductive potential (Table 1) suggest that distinct selective pressures affect dispersal behaviour of larvd M. plana and 0.kirbyi. Upon emergence, neonate O. kirbyi either balloon collectively from a long composite of silken thread suspended from their matemal bag, or crawl up to the highest point of plants to be wind-dispersed (Stephens, 1962; Cruttwell. 1974; Newman, 1980). Dispersal of larval O. kirbyi rnay be of adaptive significance even at low population density, because the progeny of only few mated fernales potentially infiicts severe defoliation to their host plant (Table 1).

Dispersai behaviour of larval M. plana has not been studied extensively, but spatial distribution of M. piana populations in oil palm plantations (Syed & Shah, 1976) and

Iaboratory experiments (Basri, 1993) suggest that the incidence of larval dispersal increases with increasing population density. Dispersal of neonate M. plana rnay not be adaptive at low population density, because oil palms can support the development of numerous bagworms (Table 1), and larvae with few hosts (Table 1) rnay not readily encounter suitable food plants during dispersai bouts. Inter- dispersai of M. plana lame may therefore represent a density-dependent adaptation that reflects a trade-off between the cost of dispersal (high mortality, unpredictable rate of encomtering a host-plant) and individuai reproductive success on palms with different population densities.

Dispersal behaviour may be more important for fernale than for male late instars of

M. plana. Apterous adult females oviposit inside a protective bag, and their progeny may not disperse £tom the host where they emerged. As a result, dispersal by female late instars rnay

represent an attempt to seek a suitable host for their progeny. Male larvae, in contrast, are

not nibjected to the same selective pressures because adults are winged (see section 1.6.1.3).

Greater incidence of dispersal by female larvae than male larvae has been demonstrated for

severd insect species with flightless femaies and flighted males (Gross & Fritz, 1982; Moran

et al., 1982; Harrison, 1994). For any spatial scde, temporal variation of population density is ultimately determined by only 4 parameters: birth rate, mortality rate, emigration, and immigration. As a result, dispersal has a strong impact on population dynarnics. and in some insects may be the most important factor influencing population size and stability (Denno & Peterson, 1995).

Theoretical (Nicholson, 1954) and laboratory studies (Gause. 1934; Huffaker, 1958) indicate that, in the absence of dispersai, prey-predator or host-parasitoid systems rnay go extinct. In insect populations with both date and apterous adult morphs. density-dependent dispersal has important consequences for local population dynamics: high mortality among dispersers tends to stûbilize (regulate) local populations (Denno & Petenon. 1995). Successful colonization of host plants by a large proportions of dispeners. in contrast, rnay simpiy redistribute local populations, or lend to colonization of as yet uninfested plants. In herbivorous insects, density-dependent dispersal by larvae (see references above) rnay also strongly influence spatial and tempord variation of densities for local populations. Despite its potential importance, the impact of larvd dispersal on population dynarnics has not been extensively studied (Hunter, 1995).

1.6.1.2 Size attained at pupation

Body size attained by herbivorous insects at pupation or emergence enhances several components of their reproductive success as adults: longevity (Dempster, 1971 ; Elgar &

Pierce, 1988; Marshall. 1988; Torres-Vila et al., 1995; Oberhauser, 1997; but see Carroll &

Quiring, 1993; Tammaru et al., 1996b), mating success (Marshall, 1982; Rutowski, 1982a;

Haukioja & Neuvonen, 1985; Phelaa & Baker, 1986; Tamrnaru et al., 1996b; but see Deinert et al., 1994; Toms-Vila et ai., 1995), ejacdate production (Wiklund & Kaitaia, 1995; Torres-Vila et al., 1995), and fecundity (Dernpster, 197 1; SuzuIci, 1978; Haukioja &

Neuvonen, 1985; Marshall, 1988; Honek. 1993; Tammm et al., 1996a; but see Leather,

1988; Carroll & Quiring, 1993; Wikund & Kaitala, 1995). Resource partitioning is another key component of fitness, because population density and defoliation of host plants adversely affect body size of insects (Haukioja et ai., 1985; Weiss et al., 1985; Lasota & Kok, 1986;

Balliqer el uL , 1990; Carter et ui.. 199 1 ;Ruohomaki, 1992; Poirier & Borden. 19%;

Harrison, 1994; Ohgushi, 1996; but see Haukioja et al., 1988; Collegrave, 1983). Small size and low fecundity of fernales on heavily idested hosts have significant impact on population dynamics. Reviews of literature on density-perturbation experiments or life-table analyses in herbivorous insects indicate that resource-based density-dependence (bottom-up effects) is more prevalent than enemy-based dependence (top-down effects) (Dempster, 1983; Stiling,

1988; Hanison & Cappuccino. 1995).

Examples of size- or density-dependent reproductive success in bagwoms include greater fecundity for large female Thyrihpteryx ephemeraeformis (Haworth) (Cronin & Gill.

1989) and 0.kirbyi (Stephens, 1962; Cmttwell, 1974), reduced parasitism of large pupae of

I: ephemeraeformis (Cronin & Gi11. 1989), density-dependent survival of pupal M. plunu

(Basn et al., 1995), as well as low fecundity and small size for T. ephemerae/ormis and 0. kirbyi, respectively, on severely defoliated host plants (Cox & Potter, 1988; Gara et al.,

1990). Syed (1978) stated that fecundity of female Mahasena corberri (Tarns) is size- independent but provided no data.

1.6.1.3 Pupatioa site

Selection of pupation site by lepidopteran larvae is an important, yet rareiy investigated, component of fitness (Reavey & Lawton, 199 1). Phytophageous insects suffer significant mortality during the pupal stage (White, 1986; Comell & Hawkins, 1995), and pupation site may influence (Cook et al., 1994; Wagner, 1993, parasitism (Gross

& Fritz, 1982; Alghali, 1984; Parry, 1995; Ruszczyk, 1996), and overwintenng survival

(Leather, 1984; Miller & Hart, 1987) of pupae.

in contrast with femaie Lepidoptera that actively seek mates (Kaitala t Wiklund,

1994; Wickman et al., 1995) or oviposition sites (Dernpster et al., 1995; Nash et al., 1995;

Peterson, 1997), sessile female bagworrns complete al1 their reproductive activity within a self-constnicted bag. As with other insects that reproduce near their emergence site, pupation site of female bagworm larvae may affect their rnating and oviposition success as adult. Sexual segregation of pupation site has been reported in severd Lepidoptera with flightless females (Table 3). Distinct habitat partitioning of male and female pupae likely reflects sex-specific constraints affecthg selection of pupation site by liuvae (Gross & Fritz,

1982). Sçlective pressures favour pupation by fernale lame in locations where sessile femdes have greatest potential reproductive success. Male larvae, in contrast, are not subjected to the same selective pressures because adults are winged. Examples of sex- specific adaptive significance of pupation site include enhanced survival of eggs on

(Rossiter, 1987) and branch sections (Lagoy & Barrows, 1989) with proportionately more

female than male pupae.

In several bagworm species, more female than male lawae pupate in the upper crown

of host mes (Stephens, 1962; Cruttwell, 1974; Gross & Fritz, 1982; Rhainds et al., 1997). It

has been hypothesized, yet not experimentally demoastrated, that pupation in tree tops may

be of adaptive advantage for female bagwonns, because it could facilitate both pheromone Table 3. Examples of sex-specific pupation behaviour in Lepidoptera with tlightless fernaies and flighted males.

Orgyia pseudofsugata Female pupae proportionately more abundant than male pupae in the lower cmwn of host plant (Luck & Dahlsten, 1980).

Lymantria dispar Male larvae more likely than female larvae to pupate under bark flaps (Campbell et al.. 1975b) '. Variations in ses-ratios between host plant species (Rossiter. 1987) b.

Orgyia vet usta Male larvae tend to pupate on the host plant where they Fed. Female larvae wander a few feet away before pupation (Harrison, 1994).

Psychidae

Oiket icus kirbyi Female pupae proportionately more abundant than male pupae in the upper crown of host plants (Stephens, 1962; Cruttwell, 1974).

Clanfa cramerii Female pupae proportionately more abundant than male pupae in the upper section of host plants and on non host plants (Rhainds et al., 1997).

Thyrdopteryx Female pupae proportionately more abundant than male pupae ephemerueformis in the upper crown of host plant; upward dispersai on host plant before pupation for female but not male Iarvae (Gross & Fritz, 1982). Female pupae more abundant than male pupae on branches rather than on petioles of plants (Lagoy & Barrows, 1989).

Solenobia rnanni Female pupae proportionately more abundant than maie pupae on the ground rather than on plant (Malicky, 1968).

The difference may be attributed to female pupae king tao large to be accomodated beneath bark flaps that were only adequate to accommodate small maie pupae.

The difference only approached significance (P= 0.13 1). dissemination and wind-borne dispersal of neonatal larvae (Gross & Fritz, 1982; Cox &

Potter, 1988). Captures of male in pheromone-baited traps (McLaughlin et ai., 1 976;

McNally & Barnes, 198 1; Bergh et al., l988), mating success of adults (Kolodny-Hirsch &

Webb, 1993; Weissling & Knight, 1995; see also Rowe II & Potter, 1996), and settling velocity of ballooning larvae (Mitchell, 1979; McManus & Muon, 1983; Cox & Potter,

1986) al1 increase with increasing height within the tree canopy. Because quality of iravrs varies between crown sections (Lawton, 1983; Lowman, 1985; Montgomery, 1989; Rowe 11

& Potter, 1996) and neonate bagworms may remain near their emergence site (section

1.6.3. l), vertical distribution of Fernale pupae rnay also reflect selective pressures favouring pupation by larvae in sites of enhanced development andor survival for their progeny.

Behavioural adaptations contributing to individual reproductive success have important implications at the population level (Hassell & May, 1985; Lomnicky, 1988). For example, discriminate oviposition behaviour by females on host-plants most suitable for the development of their progeny (Pnce et al., 1990; Ohgushi, 1995) and density-dependent dispersa1 (Denno & Peterson, 1995) both contribute to stabilize populations. Indiscriminate oviposition behaviour by fernales, in contrast, is commonly associated with eruptive population dynamics (Price et al., 1990; Pnce, 1994). Selection of pupation site by female bagworm larvae may influence several components of their reproductive success, including survival during the pupal stage, mating success of adults, and fitness of progeny. Thus, pupation behaviour may have a cornplex, multi-facetted impact on population dynamics by simultaneously influencing: 1) mortality rate in parental generation [mortality of late instars foraging for a pupation site; mortality of pupae]; 2) birth rate in parental generation [mating success of adults; reduced fecundity of larvae that use metabolic resources while foraging for a pupation site (Weiss et al., 1987)l; 3) mortality rate in offspring generation [distance

"travelled" by baliooning larvae and probability of host-encounter]; and 4) birth rate in offspring generations [feeding performance and fecundity of non-dispeeing progeny].

Despite its potential importance, the significance of pupation behaviour with respect to population dynamics has largely been ignored (but see Geier, 1964; Campbell, 1967).

1.6.2 Adult stage

Life history traits that rnay influence reproductive success and population dynamics of bagworms include timing of adult emergence (section 1.6.2.1) and sexual communication between moles and females (section 1.6.2.2).

1.6.2.1 Timing of energence

Field observations with sevenl bagworm species suggest that high proportions of femdes never mate (Maiicky, 1968; Magistretti et al., 197 1 ; Barrows, 1974; Sheppard, 1975;

Thangavelu & Gunosekaran, 1982; Kauhann, 1985; Klun et al., 1986). Incidence of unmated females may be attributed, in part, to low abundance of males in protogynous populations (Stephens, 1962; Morden & Waidbauer, 197 1; Thangavelu & Ravindranath,

1985; Gan et al., 1990; Kuppusarny & Kannan, 1993). If that is me, rnating success of female bagworms is expected to increase during the emergence season, as males become progressively more abundant. Theoretical (Wikiund & Fagerstrilm, 1977; Fageatrorn &

Wiklund, 1982; Botterweg, 1982; Zonneveld & Metz, 1991) and experimental studies (Bergh et al., 1988; Elgar & Pierce, 1988; Hastings, 1989; Wang et al., 1990; Wedell, 1992; Dickinson, 1992; but see Baughman, 1991) indicate that timing of adult emergence is an important component of mating success in insects.

Timing of adult emergence in relation to host-plant phenology may also influence reproductive success of females in insects that exploit ephemeral food resource. Emergence date of female checkenpot butterfiies (Lepidoptera: Nymphaiidae) influences the titness of their progeny; in extreme cases. complete asynchrony of adult emergence and availability of host plants may cause local populations to become extinct (Cushman et al., 1994).

1.6.2.2 Sexual communication

Attraction of winged male bagworms for mating is mediated by female-produced sex pheromones (section 1.3.3). (1 R)-1-Methylbutyl decanoate has been identified as a sex pheromone component of T. ephemeraemis (Leonhardt et al.. 1983), but pheromones of other bagworms remain unknown.

In the Lepidoptera, sexual interactions between males and females may have important implications at the population level. The time femaies spend foraging for mates reduce their fecundity (Wickman & Jansson, 1997; see also Kaitala & Wiklund, 1 994), possibly resulting in decreased population size in offspring generation. In Lepidoptera with

flighted females, relativeiy high proportions of mating failures among females has been associated with Iow population density (Shapiro, 1970) or wet conditions hampering flight activity of mate-seeking males (Greenbank, 1963). In Lepidoptera with flightless females

(Doane, 1968; Malicky, 1968; Magistretti et ai., 197 1 ;Barrows, 1974; Sheppard, 1975;

Thangavelu & Gunasekaran, 1982; Harshman & Futuyma, 1985; Kauf'mann, 198 5; Klun et

al., 1986; Sharov et al., 1999, some females never mate during their lifetime. Theoretically, extensive mating failures of females may cause decline in population density or even extinction of local populations. It remains unclear, however, whether losses of fecundity due to mating failures strongly impact population dynamics, because incidence of unmated females ha never been rigorously quantified.

1.7 Specifc research objectives

nie major research objective of my research was to evaluate density-dependent processes and Iife history traits that influence reproductive success and population dynamics of bagworms. In this section, Ireview specific objectives of sampling studies and cage experiments carried out with O. kirbyi and M. plana.

1.7.1 Oiketicus kirbyi

Reproductive success of O. kirbyi was investigated between September 1992 and

June 1993 in two experimentd sites, with each site sampled during one generation. Elements of population dynamics could not be thoroughly investigated, because populations declined to very low density after 1993. Specific research objectives were as follows:

1) CO test the hypothesis that size attained by femde larviie at pupation is positively comlated with their mating success and fecundity as adults (section 1.6.1.2);

2) to test the hypotheses that male and female larvae have distinct pupation behaviour,

and that pupation site selected by femaie larvae influences their mating success as

adults (section 1.6.1 -3);

3) to compare patterns of emergence for males and females, and to test the hypothesis

that timing of emergence infiuences mating success of females (section 1.6.2.1); and 4) to investigate pheromone components that mediate sexual communication between sessile females and winged males (section 1.6.2.2).

1.5.2 Methphna

Specific objectives of sampling studies and cap experiments carried out during consecutive genentions of M. plana were as follows:

to test the hypothesis that larvae commonly remain on their natal host (section

1.6.1.1);

to test the hypotheses that conditions associated with "crowding" of oil paim (high

population density, intense defoliation) promote ballooning by larvae, and that the

incidence of ballooning is greater for female larvae than for male larvae (section

1.6.1.1);

to test the hypothesis that larval ballooning mediates dispersal between palms, and

influences dynamics of local populations (section 1.6.1.1);

to simultaneously assess the impact of body size and population density on

parameten of reproductive success (swival during pupal stage, mating success of

females, fecundity) (section 1.6.1.2);

to assess whethw individual variations of reproductive success affects dynamics of

local populations (sections 1.1 and 1-6.1.2);

to test the hypothesis that male and female larvae have distinct pupation behaviour,

and that pupation sites selected by female larvae infiuence the distribution of their

progeny (sections 1.6.1.1 and 1.6.1.3); and

to compare patterns of emergence for males and femaies (section 1.6.2.1). 2.0 MATERLALS AND METHODS

2.1 Study sites

Field siudies were cmied out in plantations of even-aged oil palms planted in 9-m equilateral triangles. Studies with O. kirbyi were conducted in 2 locations ca 5 km apart in

South Costa Rica (Coto 50 and Coto 52), whereas studies with M. plana were carried out in 2 distant plantations, one located in North East Sumatra. Indonesia (Sungei Merah), and the other in Selangor, Peninsular Malaysia (Ladang Coalfield) (Table 4). Bagworm populations in experimental sites had not been controlled with pesticides before studies were conducted, and remaineci untreated throughout experiments. Cage experiments were conducted at the

Golden Hope Oil Palm Research Station (OPRS)in Banting, Selangor, Peninsular Malaysia.

The distribution of populations of O. kirbyi was extremely patchy in Costa Rican plantations, with most oil palms not or lightly infested while others supported large numbers of bagworms. Populations of M. pluna, in contrast, had relatively unifonn distributions in both Indonesian and Malaysian plantations, with rnost oil palms carrying at lest sorne bagworms. In ail experimental plantations, bagworm populations had discrete, non- overlapping generations, with most individuals being in the same developmental stage.

2.2 Dispenal by larval Metka plana

Sampling and experimental studies investignted the incidence of larval dispersal as a

function of population density, lard sex, and attributes of palm (defoliation, fertilization).

23.1 Sampüng studies conducted in plantation of oil palm

Sampling was carried out in January - Febniary 1996 in Ladang Coaifield, Peninsular Malaysia (Table 3). At the onset of experiments, > 80% of bagworms were in their mid to late instars (bag length of 5-10 mm) (Basri & Kevan, 1995). In the first sîudy, 10 paims were systematically selected for sampling to cover a large range of bagworm densities (20-250 insects per leaf). For each palm, 1recorded numbers of ballooning larvae during two 5-min periods, the first between IOA5 and 13:30 hours and the second between 14:00 and 15:45 hours, for each of 3 consecutive days. To prevent multiple çounting, bailooning iarvae were detached fiom their silken thread and removed from the population. Upon completion of ballooning observations, density of bagworm larvae on each palm was estimated by recording both the number of leaves per palm and the number of live larvae on four leaves

(positions 9- 17; Fig. 6) in the upper palm crown.

In a second study, 7 palms with 150 - 250 bagwoms per leaf were selected for sarnpling; the distance between sampled palms varied between 9.1 and 87.8 m. For each of 4 consecutive days, larvae ballooning from each palm were recorded for one 5-min period per hour between 07:30 - 19:30. During the first 3 days, the bag of each ballooning lama was marked with a dot of oil-based paint, using different color-codes for different days and palms. B allooning larvae dislodged tiom their silken thread during marking were discarded.

Ballooning larvae recaptured at subsequent days were recorded and rernoved from the population. mer 4 days of ballooning reconlings, 10 (25 %) of 40 leaves of each palm were exarnined for the presence of marked larvae.

2.2.2 ErpeRments conducted in cage enviromment

Experiments conducted at OPRS (Apcil - August 1999) used 6 to 18 mo-old nursery paims stocked in polyethylene bags (38 cm circumference x 46 cm high). These palms had low levels of defoliation (0.5%; Basri 1993), and were treated every 2 - 4 weeks with 20 g of fertilizer consisting of N (14%). P (1 3%), K (9%) and Mg (5%). Experimental insects were collected in Sungei Buaya, 6 km fmm OPRS, in a plantation of 3 yr-old, 3 - 4 rn taIl palms.

Groups of 2,3 or 4 evenly-shed and -shaped palms were enclosed ca 1 rn apart inside screen cages (3 x 3 x 3 m or 4 x 4 x 2.5 m hi&) and infested with larval M. plana (exps. 1-5) or female pupal bags (axp. 6). One day (exps. 1 - 4) or 13, 18 and 24 days (exp. 6) following infestation, iarvae ballooning from pdms were collected [every 15 min between 8: 1 5 - 17:OO

(exps. 1-2); every 30 min between 08:30 - 17:00 (exps. 3-4); every 30 min between 08:30 -

13:30 (exp. 6)]. Afier termination of bailooning observations, resident larvae on palms were collected. In experiments 3,4 and 6, resident larvae were classified as feeding on folioles or foraging on rachis. The rate of ballooning [number of ballooning larvae / number of ballooning and resident Iûrvae] and of within-leaf dispersal [number of resident larvae on rachis / number of resident larvae on rachis and folioles] was calculated for each palm.

Exp. 1 (2 1 - 27 April; N = 4) investigated the rate of bailooning by male and femaie larvae in relation to infestation level. For each replicate, palms were infested with either 50,

250 or 1000 late instars (bag length: 8 - 13 mm). Ballooning and resident larvae from each palm were reared on enclosed nursery paims or water-immersed palm ledets (Basri, 1993) to determine sex-ratio at pupation. Every 5 - 7 days, pupal bags were collected and sexed. and palm Leaflets replaced.

Exp. 2 (28 - 30 April; N=4) investigated the rate of ballooning by male and female larvae in relation to natural defoliation of palms. Naturai defoliation of 2 1 - 35 % (Basri,

1993) was obtained by enclosing larvae-infested palms for one month; prior to experiment,

al1 larvae were removed. For each replicate, two palms (one defoliated and one undefoliated) were infested with 250 late instars (bag length: 8 - 13 mm). For each palm, ballooning and resident larvae were reared to the pupal stage as described in experiment 1.

Exp. 3 (2 1 June - 15 July; N = 7) investigated the rate of ballooning and within-leaf dispersal by larvae in relation to artificial defoliation of palms. Nursery palms were

'defoliated' by removing either 0.25 or 50 % of folioles from each leaf. For each replicate, each palm was then infested with 400 mid to late imtm (hg length: 4 - 13 mm).

Exp. 4 (27 June - 14 July; N = 7) investigated the rate of ballooning and within-leaf dispersal by larvae in relation to fertilizer treatment. Nursery palrns were either treated with fertilizer every 2 wk or lefi untreated. AAer 3 mo for each replicate, 2 palrns (one fertilized and one unfertilized) were infested with 240 mid to iate instars (bag length: 4 - 13 mm).

Exp. 5 (1 5 July; N = 4) compared abundance of male and fernale pupae in relation to artificial defoliation of palms. For each replicate, 2 palms (with O or 50% of folioles removed) were infested with 240 late instars (bag length: 8 - 13 mm). Larvae were left undisturbed throughout the experiment and allowed to develop to the pupal stage. Every 7th day between 21 Suly - 04 August, pupal bags were collected from each pdm and sexed.

Exp. 6 investigated the rate of ballooning, within-plant fonging, and inter-plant dispersal by larvae in relation to population density in the parental genemtion. Shortly before the onset of emergence of neonatal larvae (21 May), pupal bags of fernales were suspended fiom leaves of 7 out of 14 nursery palms enciosed in screen cages (1 palm with 10 pupd bags plus 1 uninfiested palm in a 3 x 3 x 3 m cage; 1 pah with 100 pupal bags plus 1 uninfested palm in each of hHo 3 x 3 x 3 m cages; 2 pahs with 10 pupal bags plus 2 uninfested palms in a 4 x 4 r 2.5 m high cage; 2 palms with 100 pupal bags plus 2 uninfested palms in a 4 x 4 x

2.5 m high cage). Upon emergence of larvae fiom their maternai bag, numbers of ballooning and resident early instars (bag length: I - 4 mm) were recorded as previously described. A thick coating of grease applied to the polyethylene bag (which enclosed the palm's roots) prevented larvae from walking off plants; the presence of resident larvae on uninfested palms could therefore be attributed to wind-dispersal by ballooning early instars.

2.3 Assessrnent of reproductive success for 0iketicu.s kirbyi and Metka plana

Site-specific characteristics of palms and sampling procedures in field studies assessing parameters of reproductive success are listed in Tables 4 - 5. Distinct procedures were used to sample populations of O. kirbyi and M. plana. In Costa Rican plantations with patchy distribution of 0.kirbyi, paims infested with bagworrns were systematically selected for sampling over ca 30 ha areas. In Southeast Asian plantations with relatively uniforni populations of M. plana, > 40 % of palms were sampled in small plots (< 1.5 ha).

Cornponents of reproductive success for 0. kirbyi and M. plana were assessed using panmeters listed in Table 2. The presence of pupal bags was recorded on various leaves of oil palm (Table 4). Mer incision of bags, male and female specimens were classified as either: 1) live prepupa or pupa; 2) dead prepupa or pupa (dessicated, diseased, parasitized or preyed upon bagworm; tom, perforated bag indicating predation / parasitism); or 3) emerged adult. The presence of (pheromone-Iaden) scales in the bag indicated an emerged female, whereas an empty pupal case protniding nom the bag andlor dense silking inside the bag indicated an emerged male. Adult fernales were classified as either: 1) receptive (scales in bag plus live female in pupal case); 2) mated (eggs or neonates in pupal case); or 3) unmated

(scaies in bag plus dead or no fernaie in pupal case). The presence of a filamentous web spread by dispershg neonates inside the pupal case, or of dead neonates that failed to leave Table 4. Site-specific characteristics of oil palms and sarnpling procedures for different experimental plantations '.

Oiketicus kirby Me tisa plana

Coto 50 Coto 52 Sungei Ladang Merah Coalfield

Location Costa Rica Costa Rica Indonesia Malaysia

Sampling area (ha) 30 0.26

Age of palms (year) 6 4

Height of palrns (m) ' 4-5 3 -4

Sampling dates 30/03193 to 05/05/95 to 08/06/93 1 7/05/95

Number of gene- 1 I rations sampled

Number of palms sampled

Number of palms smpled / time unit

Number of leaves sampled I palm

Phyllotaxie position of leaves sampled Table 4. (continued)

See Table 5 for description of demographic parameters and components of reproductive success recorded in different locations.

Number of years after nursery palms were transplanted in the field.

Height of pdms fiom the ground to the highest point reached by the foliage.

For 20 palms sampled during week IO (O 1 June), the presence of male but not of female pupai bags was recorded.

The sme 95 palms were sampled during 5 consecutive generations of M. pluna 1996 (sampling time for generation 1: 21 February - 07 March; for genention 2: 08 - 23 May; for generation 3: 19 July - 03 August; for genention 4: 28 September - 13 October; for genemtion 5: 14 -29 December).

See Fig. 6. Table 5. Demographic parameters and components of reproductive success recorded for different populations of Oiketicus kirbyi and Merisa plana '.

Oiketicus kirbyi Merisu plana

Coto 50 Coto 52 Sungei Ladang Merah Coal field

Size of bagwoms Yes Yes No Yes

Sex of pupal bags Yes Yes Yes Yes

Mortality of pupae Yes No No Yes

Emergence status Yes Yes Yes Yes

Mating statu Yes Yes No Yes

Fecundity Yes No No Yes

Number of No No Yes No emergent larvae

' See Table 2 for characters used to mess components of reproductive success of bagworms. See Table 4 and section 2.3 for site specific characteristics of palms and sampling procedures in different locations. their materna1 bag (Basri, 1993) allowed to differentiate between mated and unmated femaies. As with several other bagworms (section 1.3.3), virgin female O. kirbyi and M plana enclosed without males did not oviposit in their pupal case.

23.1 Oiketicus kirbyi population in Coto 50

Sampling was initiated shortly before the onset of adult rmergence. The presencc of male and female pupal bags, emergence status of bagworms, and mating success of femaies were recorded on al1 leiives of 253 palms sampled between 24 September - 02 December

1992. The presence of dead pupae was recorded on 1 13 palrns sampled after 26 October.

Lengths of pupal cases and weigths of egg masses were recorded for 207 mated femûles sampled on 19 paims from 28 to 30 October. For andysis, phcrowns were divided in 7 leaf groups, with leaves 2-7 and 3 8-43 being the highest and Iowest, respectively (Fig. 6).

2.3.2 Oiketicus kirbyi population in Coto 52

Sampling was initiated ai the onset of adult emergence. The presence of male and female pupal bags, emergence status, and mating success of females were recorded on 9 leaves in the upper, intemediate, and low crown of 230 palms (leaves 9- 1 1, 17-1 9, and 25-

27; Fig. 6) sampled between 30 March - 8 June 1993. Pupal cases of females sarnpled on 80 palms between 26 Apd - 17 May were memured to the nearest 0.1 mm with a caliper.

2*31 Metha plana po pulatioa in Sungei Mera h

Sampling was carried out in a 0.26 ha plot (72 x 36 m) between 05 - 17 May 1995.

The live population of bagworms in the sampling site consisted only of early instars (bag length < 3 mm; Basri, 1993), indicative of non-overlapping generations. For each of 39 palms within the experimental plot, numbers of early instar M. plana were recorded on 18 leaves in the upper, middle and low crown (leaves 5-10, 18-23, and 3 1-36, respectively; Fig.

6). On the same leaves, pupd bags fiom the previous generation (bag length > 8 mm; Basri,

1993) were opened, and the presence of emerged males and females recorded.

2.3.4 Metka plana population in Ladaag Coalfield

For each of 5 consecutive generations of bagworms in 1996 (sampling time for different generations in Table 4), the same 95 palms were sampled after > 98% of larvae had formed a pupal bog. For each generation, 9 to 25 pdms were sampled every 3 days by recording the number of male and femde pupal bags on one leaf per palm in the upper crown

(leiif5 to 7; Fig. 6). Pupal bags were measured to the nearest 0.01 mm using a digital caliper.

[Lengths of bags and pupal cases were positively correlated for 5 1 male and 97 femaie M. plana sampled during generation 1 (regression analyses, P<0.000 1 )]. Mortality of pupae, emergence statu, and rnating success of females were recorded for each palm and generation. Number of eggs in pupal cases of 84 mated females smpled dunng generation 3

(29 July - 02 August) were counted using a binocular microscope.

2.4 Size-dependent mating success of female Oikeiicus kirbyi

2.4.1 Field shidy

The experiment was conducted in Coto 50 (Table 4) in October 1992. Prior to emergence of adults, 15 palms were cleared of al1 bagworms. Nhe female pupal bags were randomly selected and transferred back ont0 each tree. Each of 3 leaves in the upper, middle and lower crown (leaves 5-7,20-22, and 35-37; Fig. 6) received one bag, suspended 20-30 cm apart from the leaf apex. One month Inter, bags were opened, mating success of females determined (Table 2), and length of pupal cases caliper-rneasured to the nearest O. 1 mm.

2.4.2 Cage study

Field colloçted bags çontaining 5 10 male and 2 10 female pupae were equidistantly suspended (about 5 cm apart) from wires 1 m (males) and 2 m (females) above ground in a large outdoor cage (3 x 5 x 3 m high) (April 1993). For technical rasons, I had to abandon the cage study afier eclosion of 4% (20) of the males and 40% (83) of the females. Bags of emerged females were opened to record their mating status (Table 2) and measure the length of their pupal case to the nearest 0.1 mm.

2.5 Sexual communication between male and fernale Oikrticus kirbyi

2.5.1 Laboratory analysis

In May 1992, pupal bags were collected in Coto, Costa Rica, and sent to Simon

Fraser University. Pupae were removed fiom their bags and kept separately in Petri dishes (9 cm diam.) at 25' C under a photopenod of 12L : 12D. Pheromone present in scdes expelled by a femaie out of the pupal case was extracted in 150 pl of hexane for 10 min. Extracts of

100 females were combined and subjected to gas chromatographie-electmntemographic detection (GC-EAD) analyses (Am et al., 1975), on three hsed silica columns (30 x 0.25 or

0.32 mm ID) coated with DB-23, DB-210 (J&W Scientific, Folsom, California 95630), or

SP-1000 (Supelco, Bellafonte, Pennsylvania 16923). Coupled OC-mass spectmmetry ((MS)

(Hewlett-Packard 5985B) in selected ion monito~g(SM) mode, using a DB-2 10 column with isobutane for chernical ionization (CI), was conducted to coiifirm the identification of

EAD-active components in scale extracts. Full-scan CI mass spectra of synthetic candidate compounds were obtained to select diagnostic ions. In sequence, 200 pg of synthetic compounds, a hexane blank, and a concentmted pheromone extract were then andyzed in

SIM mode, each time scanning for the diagnostic ions. Synthetic candidate pheromone components were fUrtl1er subjrcird to GC-EAD analyses to compare their EAD-activity with those of fernale-produced compounds. Syntheses of dl candidate pheromone components

[(R)-and (S)-1-methylbutyl octanoate (MBO);(R)- and (S)-1 -rnethylbutyI nonanoate (MBB);

(R)-and (S)-l -methylbutyl decanoate (MBD); (R)-and (S)-1-methylpentyl decanoate (MPD);

(R)-and (S)-l -methylbutyl dodecanoate (MBDD)] (Rhainds et al., 1994) were conducted by

Dr. Jianxiong Li, who at the time was a graduate student of Dr. K.N. Slessor from the

Deputment of Chemistry, Simon Fraser University.

2.5.2 Field experiments

Experiments were conducted in plantations of oil palm in Coto 50 and Coto 52 (Table

4). Green Unitraps (Phero Tech Inc., Delta, British Columbia) were suspended frorn oil palrns 3 m above ground in randomized complete blocks, with traps and blocks at 18 to 27-m intervals. Traps were baited with Cotton balls impregnated with hexane solutions of candidate pheromone components. A smail Diclorvos cube (Green Cross, Division of Ciba-

Geigy Canada Ltd., Mississauga, Ontario) placed on the bottom of traps killed captured moths. After I to 2 days, bahwere changed and captured male 0. kirbyi counted. Blocks were rerandomized when a predeterrnined number of males had been captured.

Exp. L compared attractiveness of (R)-and (5")-MBD (1000 pg), alone and in combination. Exp. 2 tested (R)-MBD (10,000 pg) alone and combined with (S)-MBD at respective ratios of 1 :1, 1:O. 1, 1 :0.0 1, 1:0.00 1, and 1 :0.000 1. Exp. 3 tested (R)-MBD (1000 pg) alone and combined with either one or both enantiomers of MBO at 1:O. 1 and 1 :0.0 1 ratios. Exps. 4-6 tested (R)-MBD (1000 pg) alone and in 1 :O. 1 ratio with either one or both enantiomers of MBN (exp. 4), MBDD (exp. 5). and MPD (exp. 6). Exps. 7-9 each tested

(R)-MBD in pentanary combination with (R)-MBO, (R)-MBN, (R)-MPD, and (R)-MBDD, and in ail binary (exp. 7), ternary (exp. 8) and quatemary combinations (exp. 9) with these chiral esters. Exp. 10 tested (R)-MBD (1000 pg) alone and in temary combinations with (R)-

MBN and (R)-MBO at respective ratios of 1 :1, 1:O. 1. 1:0.0 1, and 1:0.00 1. Exp. 1 I tested

(R)-MBD alone and in temary combinations with both (R)-MBO and (R)-MBN at a 1 :1 :1 ratio, employing doses of 10, 100, 1000, and 10,000 pg.

2.6 Height-dependent captures of male 0. kirbyi in pheromone-baited traps

An experiment wuconducted in Coto 52 (Table 4) between 14 Apd and 8 June

1993. Green Unitraps were baited with Cotton balls irnpregnated with 2 sex pheromone components, (R)-MBD (1 000 pg) and (R)-ME30 (100 pg) (section 2.5.2). Three tnps per tree (one at respectively 1.5,2.5, and 3.5 m above ground) were suspended fiom 5 palms >

80 m apart. Captures of males were recorded and baits replaced 3 times per week.

2.7 Individual variations of reproductive success and population dynamics

An experiment was carried out in OPRS, Banting, Malaysia. Shortiy before the onset of larval emergence (25 May 1999), pupal bags of female M. planci were collected fiom 3 yr old, 3-4 m ta11 palms in Sungei Buaya, and suspended from leaves of 9 mo old nursery palms (8 palms with 4 pupal bags, 8 palms with 8 pupal bags, 6 palms with 12 pupal bags).

Nursery paims were then placed individually inside 1 x 1 x 1.5 rn high screen cages which were kept on a 15 x 1 x 1 m high bench; a coating of grease applied to the legs of of each bench prevented ants fiom preying on bagwoms. AAer 3 weeks, early instars on each nursery pdm were collected. and pupal bags opened to determine mortality during the pupal stage, emergence status, and mating success of fernales (Tabla 2). 3.0 STATlSTICAL ANALYSES

Statistical analyses were conducted using SAS statistical package (SAS institute,

1988). Whenever necessary, data were subjected to logarithrnic [x' = In x or x' = ln (x +

OS)], square-root [x' = d x], arcsine [x' = sin-' 4 x], or rank transformations to ensure normal distribution of data and homogeneity of variance. In dl analyses, a = 0.05; for P levels between 0.05 and 0.15, results were considered marginally significant. Independent variables that could be measured quantitatively (number of insects, length of pupal bag, sampling penod. leaf group, level of artificial defoliation) were treated as continuous variables, whereas nominal variables (sex, life stage, survival status, mating status, pheromone treatment, defoliation status, fertilizer treatment, the interval) were treated as class variables. Individual palms were treated as either repiicates or random factors. The following analyses were used: linear regression @roc reg), logistic regression @roc logistic), genenlized linear modelling (proc glm), and analysis of variance followed by Student

Newman Keuls test @roc ANOVA). Statisical analyses for different sampling and experimental studies are listed in Tables 6 and 7. Table 6. Statisiical analyses used for studies carried out with 0.kirbyi (Coto 50 and Coto 52, Costa Rica) and M.plana (Sungei Merah, Sumatera, Indonesia; Ladang Coalfield, Selangor, Malaysia).

-. . Analysis Study Variables SAS procedure Output

Dependent Independent Random Source P

I hl. plana y = number of pupal x l = sex x3 nested within data met I ; xl 0.0093 Ladang Coalfield bags x 1 and x2 input x I x2 x3 y; x2 0.0001 Methds: 2.3.4 x2 = generation y' ;= log (y + 0.5); xl+x2 0.457 Results: 4. I prac glm ; x 3 0.000 1 x3 = palm class xl x2 x3; model y' = x 11x3 x3; random x3;

2 M.planu y = balloonin larvae x = number of larvae data met2; Ladmg Coalfield per palm 9 per palrn input K y; Methods: 2.2 proc reg; . Results: 4.2 model y = x;

3 AI. pluna y = balIoanin larvae x = number of larvae data met3; Ladang Coalfield per palm 9 per palm ' input x y; hlethds: 3.2 y'=ylx; Results: 4.2 proc reg; model y' = x;

4 0,kirhyi y = length of pupal case AI= number of pupal dain oik I ; x 1 0,426 Coto 52 of fernale (cm) bags per palm inpiii xl x2 y; xZ 0.004 1 Methods: 2.3.2 XI*= hl; Results: 4.3 x2 = leaC group ' proc glm; modely =XI*x2; Table 6. (continwd)

Analysis Siudy Variables SAS procedure OUI put ------Dependcnt Independent Random Source P

5 JI$p/u~ y = length of pupal bag x 1 = sex x3 nested withiii data iiiei4; x 1 0.000 1 Ladang Coal field (mm) xl and x2 input x l x2 x3 y; x3 0.000 1 Methods: 2.3.4 x2 = generation proc glni; xl+s2 0,423 Results: 4.3 class KI ~2 x3; K 3 0.000 1 x3 = palm mode1 y' = x 11x2 x3; raiidoni 33;

6 At. plurra y = length of pupal bag x 1 = sex data rne15; Sec=Table 12 Ladang Cwlfield (mm) input XI x2 x3 y; Meihods: 2.3.4 x2 = generation x3' = log x3; Resulrs: 4.3 proc reg; x3 = nuniber of pupal model y = x3'; bags per palin byxl A?;

7 0.kir& y = number of pupal x l = location ' x4 nested within data 0ik2; Co10 50 and Coto 52 bags x2 and x3 input xl x2 x3 x4 y; Methods: 2.3.1 and 2.3.2 x2 = sex proc rmk; Results: 4.4 raiiks y'; x3 = leaf group var y; by xl; x4 = palni proc glrii; class x l x2 x3 x4; rirodel y' = ~31x3x4; rairdom x4; by xl;

Table 6. (coniinued)

Analysis Siudy Variables SAS procedure Output

- Dependent Independent Random

O. kirbyi y - Ex4 / (x4 + x5)) x l = locaiion ' data oikS; s3 0,000 1 Coto 50 and Coio 52 input x l $2 x3 x4 x5; any combinat ion Meihods: 2.3.1 and 2.3.1 x2 = sex proc logisitic; of x 1 aiid X? Resulis: 4.6 mode! x4 /(x4t.x5) = x3; x3 = week by sl x2;

x4 = number of emerged adults

x5 = number of live pupae

AC plana y = lx4 / (x4 + x5)] xi =sex data met IO; x3 0.000 1 Ladang Conlfield input XIx2 x3 x4 xS; any conibination Methods: 3.3.4 x2 = gcneration proc logisitic; ofxl and x3 Resuits: 4.6 model xS /(x4+x5) = x3; x3 = julian date by x l x2;

x4 = nunlber of emerged adults

x5 = numher of live pupae 0- m Ov, Of '?z 9- ?' 00 2 Table 6. (continued)

Analysis Study Variables SAS procedure Output

Dependent l ndependent Random Saurce P

18 O. kirbyi y = lengih of pupal case x 1 = location ' data oik8; SI:] Coto 52 and Coto 50 (cm) input xl x2 x3 x4 y; x2 Meihods: 2.3.2 and 2.4.1 x2 = mating status proc gh; x3 Rcsults: 4.8 class x 1 s2 x3 x-î; ~2~x3 x3 = leaf group model y = ~21x3x4; x4 by x 1; si:? x3 s 3 x2*x3 x4

i9 Aipkrnn y = length of pupal bag x 1 = generatian x3 nested within data met 1 I ; x l 0.0001 Ladang Caai field (mm) x l and x2 input x l x2 x3 y; x 2 0.000 1 Methods: 2.3.4 x2 = mating status ' proc glm; XI*x2 0.971 Rtsulls; 4.8 class xl x2 x3; n3 0.000 1 x3 = palni modrl y = ~11x2x3; random x3;

20 O. kirkyi y = [x4 / (x4+x5)] xl = location ' data oik9; xl:l Coto 50 and Coto 52 input x 1 x2 x3 x4 x5; x2 0.000 1 Methods: 2.3.1 and 2.3.2 x3 = wrek 'O proc logistic; x3 0.000 1 Resulis: 4.8 model x4/(x4 +xS)=x7 x3; x3 - leaf group " by xl; XI:? s2 0.0003 x4 = mated females x3 0.782

1 xS = unniaied females cn .".- -- rn Table 6. (coniinued)

Analysis Siudy Variables SAS procedure Output

De pendent Independent Random Source P

2 1 0.kirbyi y = x3 I (x3+x4) x I = lengtli of feiiiale data oik t O; h; l 0.000 I Coto 52 pupal case (cm) input x l x2 x3 x4; s 2 0.0659 Melhods: 2.3.2 proc logistic; Resubs: 4.8 x2 = leaf group ' modei x3/(x3+x4)=x 1 x2;

x3 = number of mawd feinales

22 0.kirbyi y = x3 / (x3+x4) x l = length of feniale data oik 1 I ; See Fig. 34 Caio 53 pupal case (cm) input x l x2 x3 x4; Meihods: 3.32 proc logistic; Resulis: 4.8 x2 = leaf group model xY(x3+x4)=x 1 ; by s2; x3 = number of mated feinales x4 - nuinber of unmated females

33 0.kidyi y = males captiired in x I = height of iraps dutit oik 1 2; çoto 52 pheromane-baited inpiii x l x2 y; Meihods: 2.6 traps h;2 =- palm Resulis: 48 class xl x2;

iiieans x I / snk; Ln Table 6. (continued)

Analysis Sludy Variables SAS procedure Output

Dependen t Independent Random Source P

24 0.kirbyi y = proportion of mated x I = location ' data oikl3; Co10 50 and Coto 52 females per palm " input u 1 x2 x3 y; Meihods: 2.3.1 and 2.3.2 x2 - week 'O x3' = log s3 Results; 4.8 y' = arsin 4 y; x3 = riumber of pupal proc glm; bags per palm model y' = x2 x3'; by xl;

15 AI. platria y = proportion of mated XI= gcneration data niet 12; Ladang Coalfield females '' input x 1 x2 y; Methods: 2.3.4 x2 = palm y' = aniu 4 y; Results: 4.8 proc glm; inodel y' = x l x2; raiidom x2;

26 A#, pIum y = proportion of mated data niet 13; Ladang Coalfield females per palm '* input x 1 x2 x3 y; Mzthods: 2.3.4 x2 - mean length of pupal x3' = log x3; Results: 4.8 bags per palm (mm) y' - arsin 4 y proc glin; x3 = number of pupal model y' = x2 x3'; bags per palni by xl;

27 0. kkbyi y = weight of egg niass x = lerigth of fernale data oi k 14; Coto 50 (SI pupal case (cm) input x y; Methods: 2,3,1 proc reg; Resulis: 4.9 rnodel y = x; Table 6. (continued)

Analysis Study Variables SAS procedure Owtpu~

Dependent Independent Random

28 hl. plum y = number of egss in x = hgth of female daia me1 14; Ladang Coalfield pupal case pupal bag (mm) input x y; Meihods: 2.3.4 proc reg; Resulis: 4.9 niodel y = x;

29 Al.,plunia y = number of xl = life siage l3 x3 nested wiihin data met 15; x l 0.000 l Sungei Merah bagwoms xl and x2 input XI x2 x3 y; x 2 O .O063 Methods: 2.3.3 x2 = leaf group prac rank; x1W 0.865 Resulis: 4.1 O ranks y'; 83 0.000 1 x3 = palm var y; proc glm; class x 1 x2 x3; model y' = x 11x2 x3; ra~tdoiiis3;

30 kl,.pluw y = number of early x l = number of empty dala niet 16; See Table 16 Sungei Merah instars per leaf female pupal cases input x l x2 y; blethods: 2 -3.3 per leaf proc reg; Rrtsulis: 4.10 model y = xl; by x2;

3 1 AL .phw y = nuniber of early x = number of empty daiû met 17; Sungei Merah instars per palin feniaie pupal cases input x 1 y; Meihods: 2.3.3 per palm proc reg; Results: 4. t O inodel y = x; Table 6. (cont iniied)

Analysis Study Variables SAS procedure Ouiput

De pendent Iiidcpendent Random Sourcr: i'

32 AQlunu y = number of early x J = num ber of ernpty daia met 18; x 1 0.000 I Sungei Merah instars per palm female pupal cases input x 1 x2 y; x3 0.O09 Methoâs: 2.3.3 per palm proc reg; Results: 4. IO mode1 y = x 1 x2;

33 At.plana y = number of early x 1 = number of empty dara met 19; Sungei Merah instars per palrn female pupal cases input x l x2 y; Mohods; Z3.3 per palni proc reg; Resulrs; 4,IO model y = x 1 x2; x2 = distance to heavily infested palrns '' Table 6. (coniinued)

' y = toial nurnber ofballooning kirvae sampkd during iwo 5-min periods for each of three consecuiive days

' x = niean numôer of larvae per leaf * number of lcaves per palm

' x2: I Cor leaves 9- 11 ; x2:l for Ieaves 17- 19; x2:3 for leavcs 25-27

' x 1: 1 Tor Coio 50; x l:2 for Coio 52

y = (dead pupae / (live pupae + dead pupae + emerged adults)) b x3: I if dead pupa; x3:2 if live pupa or emerged adult

7 82:1 for males; x2:2 for females a In Coto 50, x4: 1 for leaves 2-7; x4:3 for leaves 8-13; x4:3 for leaves 14-19; x4:4 for leaves 20-25; x4:S for lcaves 26-3 1; x4:6 for leaves 32-37; x4:7 for leaves 38-43, In Coto 52, x4: 1 for leaves 9-1 1 ;x4:S for leaves 17- 19; x4:3 for leaves 25-27.

x2: I for maied feiiiale; x2:2 for unmated female

In Coto 50, x2: I to x25. In Coto 52, x2: 1 to x2:8. x2 values were soned in chronological order, commencing ai the onset of adult emergence. In both locations, trees sampled afler > 95.8 % offemales Iiad emerged were pooleû and assigned the same x2 valiies (x2:5 in Cato 50 and x2:8 in CotoS2)

" In Coio 5OPw4: I for leaves 2- 13; x4:2 for leaves 14- 19; x4:3 for leaves 20-25; x4:4 for leaves 26-3 1; x4:S for leaves 32-43. In Coio 52, x4: 1 for leaves 9- 1 1 ; x4:2 for leaves 17- 19; x4:3 for leaves 25-27.

" y = (maied fmaks I (mated females + unmated females)]

13 s 1 :I for sarly insiars; x 1:2 for enipty female pupal cases

" xZ: l for trees 9 m îrom heavily infested palms; x2:2 for trees > 9 m apan fiom heavily infesced palms Table 7. Siatistical analyses used for experimenis carried out with Metisupla~~aat OPRS, Banting; expcrimental procedures described in section 2.2.2.

Analysis Experinieni Variables SAS procedure outpiit

Dependent lndependent Source P ------

y = recapiure raie of larvae ' x l = replicate proc gliii; x2 = number of released larvae ' class x 1; madel y = x I x2;

y = raeapture rate of larvae ' x 1 = replicate proc glm; x2 = natural defoliation ofpalm class x 1 x2; model y = x 1 x2;

y = recapnire rate of larvae ' x l = rcplicate proc glni; x2 = level of mificial defoliation class x 1; model y = x 1 x2;

y = recapture rate of larvae ' x 1 = replicate proc gli11; x2 = feriilizer ireaiment ' class x I x2; model y = x 1 x2;

y = recapture rate of puape ' x l = replicate proc gliii; x2 = num ber of released larvae ' class xl; model y = x 1 x2;

y = recapture rate of pupae6 x 1 = replicate proc glrn; x2 = natural defoliation of palm class x l x2; model y = x 1 x2;

Table 7. (continued)

Analysis Expriment Variables SAS procedure Ou[put

Dependeni Inde pendent Source P

y = time of ballooning 'O xl = replicate proc glm; x2 = fenilizr treatment ' class x 1 x3; model y = x l s2;

y = rate of ballooning " xl = replicate proc glin; x2 = number of released larvae class x 1; mode1 y = x 1 x2;

y = rate of ballooning " x l = replicale proc glin; x2 = natunl defoliation of palm ' class x l x2; rnadel y = x 1 x2;

y = rare of ballooning " x 1 = replicate proc glm; x2 = level of anificial defoliation class x l x2; model y = x 1 x2;

y = rate of ballooning " x l = replicate proc glm; XI= fertilizer trtatment ' class x l x2; mode1 y = x 1 x2:

y = raie of ballooning " x I = number of female bags ' proc glin; iiiodel y - x 1 ; CI-

m m *K X - CI CC) r1 -wmri KXXX KXXX

II II II -mm XXK Table 7. (continued)

y = number of recaptured larvae I number of released larvae

xz = O niid 1 For undefoliated and defoliated palms

4 x2 = 0, 0.25 and 0.50

%.ui= O aid 1 for uiifertilized and fertilized palms

y = number of farvae attaining the pupal stage / number of released larvae

' X, = O or 1 for undefoliated and defoliated palnis

y = number of resident larvae on rachis / number of resident larvae on rachis and folioles q=O, 100r100 '" y = niinutes after the onset of ballooning observations

" y = iiumber of ballooning larvae I number of ballooning and resident larvae

" y = number of ballooning lawae collected during lime interval i for day j I number of ballooning larvae collected during day j

If niiie one-liour intervals betwen O8:ûû-09:OO aiid l6:OO- 1 7:00

'' xi = I and xz = 10 for palnis canying 10 fernale bags; xi = 1 and x2 = 100 for palrns carrying 100 feriiale bags; si = O and xz = 10 for uiiinfested palms enclosed in cages containing palnis with 10 female bags; xi = O and x2 = 100 for uninfrsted palnis enclosed in cages containing palms with 100 female bags 4.0 iWSULTS

4.1 Inter-site and inter-generational variations of population density

Nurnbers of pupal bags per leaf per phwere on average greater for M. plana (4.3-

37.3) than O. kirbyi (0.7 - 1.4) (Table 8). Descriptive statistics of population density (mean, variance, coefficient of variation) in different locations can not be compared statistically, because diEerent sampling procedures were used in various plantations (Tables 4,s).

Comparing numbers of pupai bags during 5 consecutive generations of M. piuna revealed: 1) greater abundance of females (N=10,600) than males (N= 7,I 38) (F= 7.00, df =

1,94, P < 0.0095). and 2) significant variations of population density between +generations (F

= 10.90, df =4,94, P < 0.0001) (Fig. 7). Bagworms were most abundant during generations 1 -

2 and least abundant during generations 3-4 (Fig. 7). Some palms carried proportionately more pupal bags than others across generations, as indicated by the significant effect of tree on population density (F= 7.08, df = 94,846; P < 0.000 1) (Fig. 7). Coefficients of variation of mean density of pupal bags per palm for different generations decreased with increasing mean density [correlation analysis: N= 5; r = - 0.994; P = 0.00061, indicating that bagworm infestations were most clumped during generations with low population density.

4.2 Dispersai by larval Mefisaplana

4.2.1 Sampling studies conducted in plantation of oil palms

Between 0-1 10 ballooning larvae were recorded during six 5-min observation periods for each palm. incidence of larval ballooning increased linearly with density of larvae per pah(Fig. 8). The significant negative intercept (H, : = 0; t = - 2.44, P = 0.040) indicated that ballooning is rare at low population density (Fig. 8). Proportions of the estimated Table 8. Descriptive statistics for densities of pupal bags per leaf per palm in various locations '.

Oiketim kir by Me tisa plana

Coto 50 Coto 52 Sungei Menh

Standard 0.44 0.95 deviation

Coefficient 0.65 of variation

See Table 4 and section 2.3 for site specific characteristics of palms and sampling procedures in different locations.

Nimber of pupal bags per palm per generation Fig. 7. Numbea of pupal bags of Metisa plana (males and fernales) sompled during 5 consecutive generations on 95 palms in Ladang Coalfield. Sarnpling procedures sumrnarized in Tables 4,s; statistical analysis 1 in Table 6. NUMBER OF PUPAL BAGS PER PALM (ii + SE) A N O P O O O O Fig. 8. Relationship between the incidence of lmal bailooning (total number of bdlooning larvae sampled during two 5-min penods for each of 3 consecutive days) and population density of Metisu plana (mem number of larvae per leaf * number of leaves per palm) on IO palms sampled in Ladang Coatfield. Sampling procedures sumrnarized in section 2.2.1 ; statistical analysis 2 in Table 6. 3 2 I2O- y = -21 .O4 + 0.024~ r2 = 0.883, p~0.0001 90- t a7 O O4 60- ma L 30- O K O- a .*a I I 1 1 1 1 3: O 1000 2000 3000 4000 5000 Z ESTIMATED LARVAL DENSITY PER TREE larval population that engaged in ballooning (y) increased with density of bagworms per palm (x) (anaiysis 3 in Table 6: y = 0.001 7 + 0.0000035 x; ? = 0.733; P = 0.00 16).

Of 607 marked larvae, 41 (6.8 %) were recaptured while ballooning on subsequent days, and 16 (2.6 %) on the foliage of palms (Table 9). Three of 57 recaptured larvae (5.3

%) were encountered on a pairn different From the one where they had been originally marked; distances between original and new hosts were 9.1,ZJ. 1, and 24.1 m (Table 9).

4.2.2 Experiments conducted in controlled cage environment

4.2.2. Recapture of larvae and pupae

In cage experiments 1 - 4. between 67.1 - 84.6 % of released larvae were recaptured the foilowing day (Table 10). The rate of recapture decreased with increasing number of released larvae, and was greater on undefoliated than natunll y-de foliated palms (Fig. 9).

Artificial defoliation and fertilizer treatment of palms did not afTect recapture of larvae (Fig.

9). The rate of recapture for pupae in experiments 1,2 and 5 varied between 21.8 - 43.7 %

(Table 10). Recapture of pupae decreased with increasing number of released larvae and degree of namal or artificial defoliation (Fig. 9).

4.2.2.2 Rate of within-leaf dispersal

Nurnbers of resident lame recorded on folioles of leaves greatly exceeded those on rachis (Table 10). The rate of within-leaf dispersai in experiments 3-5 varied between 5.9 -

10.0 % (Table 10). The rate of within-leaf dispersal increased with increasing degree of artificid defoliation and number of female bags, but was not affected by fertilizer treatment

(Fig. 9). Table 9. Numbers of ballooning larval Metisaplana rnarked and recaptured on 7 palms in Ladang Coalfirld. The ctxperirneii~al procedure is described in section 2.2.1. Recaptured larvae were classified as ballooning or resident (sampled on a silken thread or on ihe foliage of palms), and as non-dispersing or dispersing (sampled on the host where they had been origi~inllybeen marked or on a new host). Rate of multiple ballooning = number of recaptured ballooning larvae / number of marked lurvae; rate of recapture = number of larvae recaptured on their original host / number of marked larvae; rate of between-plant dispersal = number of Jarvae recapiured on new host / number of larvae recapiured.

Pal m Cumulative

Nuniber of marked larvae 60 11 1 110 68 67 126 65 607

Nun~berof larvae recaptured on original host 2 17 7 4 4 14 fi 54 - Ballooning 1 15 3 3 3 10 5 40 - Resident 1 2 4 1 1 4 1 14

Nuinber of larvae recaptured on new host l a O O O 1 a 1 a O 3 - Ballmning 1 O O O O O O 1 - Resident O O O O 1 1 O 2

Rate of multiple ballooning 0.017 0.135 0.027 0.044 0.045 0.079 0.077 0.068

Reic of recapture 0.033 0,153 0,064 0.059 0.060 0.1 1 1 0,092 0.089

Raie of beiween-plant dispersal 0.333 O O O 0,200 0.067 Q 0.053

a Disioiice to palm oforigin: Tor palm 1,24.1 in; hr palm 5,24.1 rn; for palm 6,g.lm. Table 10. Numbers of larval Methplana released and recaptured on nursery palms in cage experiments conducied at OPRS, lianting; experirnental procedures described in section 2.2.2. The abreviation nr indicates that parameters were not recorded. Recaptuird larvae wrr classified as ballooning or resident. In experiments 3-5, resident larvae were classified as feeàing or foraging (sainpled on folioles or rachis of leaves, respectively). Recapture rate of larvae = number of recaptured larvae / number of released lanue; recapture rate of pupae = number of larvae attaining pupal stage / number of released larvae; sex-ratio at pupation = number of femule pupal bags / number of male plus female pupal bags; rate of ballooning = number of ballooning larvae / number of ballooning plus resident lorvae; rate of within-leaf dispersal = number of resident larvae on rachis / number of resident Iüwie on rachis and folioles.

Experi ment

Length of larval hag (mm) Nuinber of larvae relcased Number of larvae recaptured Nuinber of pupae recaptured Number of ballooning larvae Nuinber of residerit Iürvue - on folioles - on rachis

Rate of lama1 recaptiire 0.846 0.755 0.672 0.67 1 11 r nr Raie of pupal recapture 0.429 0.437 nr lu nr 0.2 18 Ses-ratio at pupat ion 0,808 0.91 1 nr nr nr 0.835 Rate of ballooning 0.2 14 0.274 O. 1 79 O. 129 0.020 nr Raie of wiihin-plant foraging nr nr 0.059 0,098 O. 1 O0 nr Fig. 9. Summary of statistical analyses describing relationships between dependent (yi) and independent (xi) variables for cage experiments (Ei ) that investigated dispenal behaviour of larval

Metisa plana (analyses 1 - 1 9 of Table 7). Experimental procedures are described in section 2.2.2.

Arrows indicate positive or negative relationships between Xi and Yi ; the shade of arrows indicates the significance level. Horizontal lines indicate non-significant relationship between xi and yi .

Dependent variables are as follows: for cells A (Ei~),y! = recapture rate of larvae [number of recaptured larvae / number of released larvae]; for cells B (Ela ), y2 = recapture rate of pupae

[number of larvae attaining the pupal stage / number of released larvae]; for cells C (E3.5), y, = rate of within-leaf dispersal [number of larvae on rachis / number of larvae on rachis and folioles]; for cells

D (E14),y4= tirne of ballooning [min after onset of ballooning observations]; for cells E (Ei4J, y5 = rate of ballooning [number of ballooning larvae / nurnber of ballooning and resident larvae]. lndependent variables as follows: for cells 1 (El.5), ri-5 = replicate; for El , xi = number of reieased larvae [xi :50,250 or 10001; for cells 111 (4),q= status of natural defoliation [x2 :O or 1 for undefoliated or defoliated palrn]; for cells IV (Ei),x~ = degree of artificial defoliation [x3 :O, 0.25 or

0.501; for cells V (&), = Fertilizer treatment [a: O or 1 for unfertilized or fertiliized palm]; for cells

IV (El),xr = status of artificial defoliation [XS: O or 1 for undefoliated or defoliated palm]; for cells II

(E6), Q, = nurn ber of fernale pupal bags per paim 1x6 :0, 1 0 or 1001. S ignificance Ieveis are as fol lows for different cells: A 1: P (ri) = 0.737, P (XI)= 0.0090; A 111: P (r2) = 0.026, P (x2) = 0.0005; A IV:

(r3) = 0.0070, P (x3) = 0.42 1; A V: P (rd) = 0.00 12, P (a)= 0.429; B 1: P (ri) = 0.603, P (xi) = 0.038;

B III: P (n)= 0.587, P (x3 = 0.102; B IV: P (r5)= 0.254, P (x5) = 0.105; C IV: P (r3)= 0.004?, P (q)

= 0.0008; C V: (rd)= 0.455, P (Q) = 0.447, C [I: P (a)= 0.00 19; D 1: P (ri) < 0.0001, P (xi) = 0.210;

D III: P (r2)= 0.039, P (x2) < 0.0001; D IV: P (r3)< 0.0001, P (x3) = 0.033; D V: P (rd) < 0.0001, P

(x() = 0.660; E 1: P (ri) = 0.2 18, P (xl ) = 0.0006, E HI: P (r2)= 0.074, P (4)= 0.0025; E IV: P (r3) =

0.0027, P (x,) = 0.0029; E V: P (rd)= 0.302, P (a)= 0-0 10; E II: P (~g)= 0.010. 1 II III IV v Population Defoliation Fertilization density level level Larvae Adults Natural Artificial

Within-leaf C dispersal

Time of D bailooning f -

Rate of ballooning

Significance levels: - = P>0.15 4.2.2.3 Time and rate of ballooning.

A total of 2659 ballooning larvae was collected in experiments 1- 4. The incidence of ballooning was greatest between 09:00 - 12:OO (Fig. 10). Larvae ballooned earlier duting the day fiom naturally or artificially defoliated than from undefoliated pdms (Fig. 9). Tirne of ballooning was not affected by the number ofreleased lame or fertilizer treatment (Fig. 9).

The rate of ballooning varied behvcen ?.O It (early insrars) and 12.9 - 27.4 % (mid md late instars) (Table 10). The incidence of ballooning increased with increasing number of released larvae and degree of defoliation (Fig. 9). Proportionately more larvae ballooned from dertilized than fertilized palms (Fig. 9). Population density in the parental genention aKected the rate of ballooning (Fig. 9): 49 ballooning larvae were sampled fiom palms infested with LOO female pupal bags, whereas no ballooning larvae were sampled on pdms with O or 10 female pupal bags.

4.2.2.4 Sex-specific ballooning and pupaiion behaviour.

The sex-ratio of larvae that attained the pupal stage in experiments 1.2 and 5 was consistently female-biased, with proportions of femdes exceeding 80.8 % (Table 10). In

experiments 2 and 3, the rate of bdlooning was greater for female than male larvae, and

increased for both sexes with increasing number of released lame (Fig. 11) and degree of

natural defoliation (Fig. 12). In experiment 5, pupae were more abundant on undefoliated

than artificially-defoliated palms; the marginally significant interaction between defoliation

intensity and sex (PO. 123) suggested that female but not male pupae were more abundant

on undefoliated palms (Fig. 13). Fig. 10. Mean proportion (per day) of lamal Mefisaplana that ballooned from nursery palms during 1-hotu-intervais during 12 days (exps. 1 - 4, OPRS, Banting; section 2.2.2). Bars with different letter superscripts are significantly different (statistical analysis 20 of Table 7).

For each date (daylmonth), the index Ei (r,b) summarizes the experiment number (Ei ), number of replicates (r) and nurnber of b~llooningIûrvae (b): 2 1/04?Ez ( 1,265): 22/04. Ei

(1,200); 23/04, E2(1,28 1); 27/04, Et (1,20 1); 28/04, E3(2,159); 29/04, E3(1,137); 30/04, El

(1,118); 22/06, & (3,38 1); 25/06, E4 (2,248); 29/06, Es(4,122); 14/07, Es(3,168); 1 5/07, Ea

(2,386).

Fig. 1 1. Relationship between rate of ballooning [y = number of ballooning lame / number of bailooning plus resident larvae] and population density [x = number of released larvae] for male and female larval Metisa plana that attained the pupal stage (exp. 1 in OPRS, Banting; section 2.2.2). Significance levels as follows: P (replicate) = 0.0 15; P (sex) = 0.0079; P

(density) < 0.0001; P (sex4density) = 0.062 (statisticd analysis 2 1 in Table 7). male

50 250 1O00 NUMBER OF RELEASED LARVAE Fig. 12. Relationship between rate of baliooning [y = number of bdlooning larvae / number of ballooning plus resideni larvae] and natunl defoiiation of palms for male and female lard Metisa plana that attained the pupal stage (exp. 2 in OPRS, Banting; section 2.2.2).

Significance levels as follows: P (replicate) = 0.0065; P (sex) = 0.0004; P (defoliation) <

0.0001; P (sex*defoliation) = 0.072 (statistical analysis 22 of Table 7). fernale male

UNDEFOLIATED DEFOLlATED PALM PALM Fig. 13. Nurnbers of male and female pupal Metisa plana recorded on undefoliated and artificially-defoliatedpalms (exp. 5 in OPRS, Banting; section 2.2.2). Significance levels as

Follows: P (replicate) = 0.158; P (sex) = 0.0003; P (defoliation) = 0.052; P (sex*defoliation)

= 0.123 (statisticûl analysis 23 in Table 7). fernale male

UNDEFOLIATED DEFOLIATED PALM PALM 4.3 Sizc attained by larvae at pupation

Length of pupal cases of femaie O. kirbyi increased fkom upper to lower leaves (t =

2.88, P = 0.004 1) (Fig. 14) but was not affected by population density (t = - 0.80, P = 0.426)

(statistical analysis 4 in Table 6). Comparing tength of pupal bags for different generations of M. plana indicated: 1) significantiy larger bags for females than males (F=3492.02. df=

1,94, P < 0.000 1), and 2) çignificant variations of bag lrngths between generations (F= 16.68, df= 4,94, P < 0.000 1) (Fig. 15). Regression analyses conducted for each of 5 generations

indicated greater intercept values for femdes than for males (Table 1 l), which reflects larger

size for bags of females than males (Fig. 15). Regression analyses also indicated that length

of pupal bags decreased with increasing population density dunng 4 generations for femnles

(generations 1,3,4,5) and during 2 generations for males (generations 1,s); length of male

bags increased with population density during one generation (Table 1 1).

4.4 Vertical distributions of male and female pupae on oii palm

Comparing the abundance of male and femde pupal bags on different Ieaf groups

revealed: 1) signidcmt variations between leaf groups (O. kirbyi in Coto 50: F = 42.59, df =

6,252, P < 0.000 1 ;O. kir& in Coto 52: F = 1 1.5 1, df= 2209, P < 0.0001 ;M. plana in Sungei

Merh: F=5.08, df-2.38, P=0.011),with pupal bags being most abundant in the middle

crown of pdms; and 2) significant interactions between leaf group and sex in al1 locations

(O. kirbyi in Coto 50: F= 10.77, df=6,252, P <0.000 1; 0.kirbyi in Coto 52: F = 22-29, df =

2,209, P < 0.000 1 ;M. plana in Sungei Merah: F= 7.77, df = 2,3 8, P <0.000 1 ), with femaies

being proportionately more abundant than males on upper leaves (Fig. 16). Fig. 14. Relationship between length of pupal cases of 748 fernale Oikeficus kirbyi and th& position within the crown of 80 palrns sampled in Coto 52. For regession analysis, leûf groups were assigned increasing numerical values form upper to lower crown sections on palms (leaves 9-1 1 : 1 ; 17-1 9: 2; 25-27: 3). Regression analysis indicated significant relationship between length of pupal bags and leaf group (P = 0.004), despite low r' value

(0.0 1 I), because the sample size (748 females) was large enough to detect a weak relationship. Sampling procedures summarized in section 2.3.2. LEAF GROUP Fig. 15. Length of mde and female Merisa phpupal bags sarnpled during 5 consecutive genentions on 95 palms in Ladang Coalfield. Sampling procedures summarized in Tables 4,

5; statistical analysis 5 in Table 6. LENGTH OF PUPAL BAGS (mm) (iI SE) Table 1 1. Regression models analyzing the relationship between length of male and female pupal bags (y) and density of pupal bags per palm (x' = In x) for each of 5 consecutive generations of M. plana in 1996 '. N = number of male and female pupal bags sampled during different generations. Parameter estimates with star superscripts are statistically significant ( ' = P<0.05; = PcO.0 1; "* = ~<0.0001) 2.

-- - -p. - Sex of pupae Generation Intercept (Po) Slope (Pi) N

Male

Female

' See Table 4 and section 2.3.4 for site specific characteristics of palms and sampling procedures for different generations.

Statistical analysis 6 in Table 6.

Significance levels for Ho: Po = O and Ho: Pi = O. Fig. 16. Number of male and female pupal bags of Oikericus kirbyi and Metisa pfanu sampled on different leaf groups of oil palm. Leaf groups 2-7 and 38-43 are the highest

(youngest) and lowest (oldest), respectively (Fig. 6). Sampling procedures surnmarized in

Tables 4,s;statistical analyses 7, 8 in Table 6. plana in Sungei Merah T 0 females O males

Coto 52

0.kirbyi in Coto 50

2-7 8-13 14-1 9 20-25 26-31 32-37 38-43 LEAF GROUP high low 1.5 Mortality during the pupal stage

Comparing proportions of dead pupae [dead pupae / (live pupae + dead pupae + emerged adults)] for male and female O. kirbyi sampled on different leaf groups indicated: 1) higher incidence of mortality for female than male pupae (F=30.47, df = 1,112, P < 0.000 1); and 2) marginally significant variations between leaf groups (F= 1.89, df= 6,112, P =

0.0879). with relatively Iow incidence of mortality on uppar laaves (Fig. 17). Comparing proportions of dead pupae for male and female M. plunu sampled during different generations revealed: 1) higher incidence of moriality for male than femaie pupae (F=49.79. df = 1,94, P < 0.000 1), and 2) significant variations between generations (F= 12.52, df = 4.94.

P < 0.000 1) (Fig. 18).

For both male and female M. plana. bags containing dead pupae were significantly srnaller than those containing live pupae or emeqed adults during each of 5 consecutive generations (F= 145.55, df = 1,94. P < 0.000 1) (Fig. 19). The significant interaction between

sex and survival status (F= 5.66, df = 1,94, P = 0.0 194) indicated that differences between

lengths of bags containing dead or live pupae were greater for males than females (Fig. 19).

Incidence of pupal mortality for male and female O. kirbyi was not affected by

density of pupal bags per pdm (statistical analysis 12 in Table 6). Multiple regression

models conducted for each of 5 consecutive generations of M. plana in Ladang Coalfield

indicated: 1) inversely density-dependent levels of mortality for female pupae during 2

generations; 2) density-independent levels of mortality for male pupae during al1 5

generations; and 3) decreasing levels of pupal mortahty with increasing mean length of pupal

bags per tree during 5 and 3 generations for males and females, respectively (Table 12). Fig. 17. Proportions of dead pupae [dead pupae I (live pupae + dead pupae + rmerged adults)] for male and female Oiketicus kirbyi on different leaf groups of 1 13 palms sampled in Coto 50. Leaf groups 2-7 and 38-43 are the highest and lowest, respectively (Fig. 6).

Sarnpling procedures summarized in Tables 4.5; statistical analysis 9 in Table 6. PROPORTION OF DEAD PUPAE PER PALM (X + SE) Fig. 18. Proportion of dead pupae [dead pupae l (live pupae + dead pupae + emerged adults)] for male and female Mefisuplana sarnpled during 5 consecutive generations on 95 palms in

Ladang Coalfield. Sampling procedures summarized in Tables 4,s; statistical analysis 10 in

Table 6. PROPORTION OF DEAD PUPAE PER PALM (k & SE) Fig. 19. Length of pupal bags of male and fernale Metisu plana pupal bags containing dead pupae versus those containing live pupae or emerged adults. Pupal bags were sarnpled during 5 consecutive generations on 95 palms in Ladang Coalfield. Sampling procedures surnmarized in Tables 4, 5; statistical ûnalysis 11 in Table 6. 1 Iive pupae and emerged adults 1 1 dead pupae

females

males Table 12. Multiple regression models anaiyzing the eEects of two variables [xi= density of pupal bags per palrn; x2 = mean length of pupal bags per paim] on mortality of male and femaie Metisa plana pupae [y = dead pupac / (live pupae + dead pupae + emerged adults)] sampled on 95 pahs during 5 consecutive generations of bagworms in 1996 '. To ensure homogeneity of variance, data were subjected to logarithmic [xi'= In x] and arcsin transformations [y'= sin-' 4 y]. Parameter estimates with star supencripts are statisticdly 00 008 - significant ( ' = P<0.05; = P<0.01; - Pc0.0001)2.

Sex of pupae Generation Densiv (Pd Length of bags (Pz)

Male

' See Table 4 and Section 2.3.4 for site specific characteristics of palrns and sampling procedures for different generations.

Statisistical analysis 13 in Table 6.

Significance levels for Ho : Pi = O and Ho: pz = O. Estimates of intercept values (Po) are not epresented because they have no biological meaning (values of x2 never appmach O). 4.6 Emergence of adults

Cumulative proportions of emerged adults increased with time for male and female

O. kirbyi in Coto 50 and Coto 52 (Fig. 20), and for male and female M. plana in Ladang

Codfield during each of 5 consecutive generations (Fig. 2 1). Emergence cycles of male and female bagwoms had similar durations (Table 13). Female O. kirbyi emerged on average

18.66 t 0.23 days before males, whereas emergence of female M. plana preceded that of males by 6.73 f 0.95 days (Figs. 20.2 1; Table 13). Inter-generntionai cornparison ofjulian dates coaesponding to 50% cumulative emergence of fernale M. plana indicoted an average duration of 73.77 * 1.97 days between consecutive generations (Table 13). The location of

O. kirbyi pupae within the crown of oil palms did not affect timing of adult ernergence

(analysis 16 in Table 6). Differences in parameten of emergence patterns for 0.kirbyi and

M. plana (Table 13) should be interpreted carefully because the 2 bagworms were sampled in different locations and time periods.

4.7 Sexual communication between male and female 0. kirbyi

GC-EAD analyses of female O. kirbyi pherornone extract on a DB-210 column

(section 2.5.1) consistently reveaied five EAD-active compounds (Fig. 22). Coupled GC-MS

of the most abundant and major EAD-active component gave the sarne mass spectrum as

MBD, a previously identified pheromone component of T. ephemeraeforrnis (Leonhardt et

al., 1983). Identical retention and mas spectrornetric characteristics of, and comparable

antemal responses tu, synthetic and female-produced MBD confirmed the presence of this

compound in phetomone extract of female 0. kirbyi.

Retention indices of the four-minor EAD-active compounds (Fig. 22) suggested they Fig. 20. Emergence pattern for male and fernale Oikeficuskirbyi sampled on 253 palms in

Coto 50 (24 September - 02 December 1992) and on 230 palrns in Coto 52 (30 March - 08

June 1993). For each location and sex, cumulative proportions of emerged adults [emerged adults / (live pupae + emerged adults)] per week were modeled using logistic regressions

(andysis 14 in Table 6). Sampling procedures sumrnarized in Tables 4.5. CUMULATIVE PROPORTION OF EMERGED ADULTS Fig. 3 1. Emergence pattern for mde md fernale Me'crisa plana sampled during 5 consecutive generations on 95 paims in Ladang Coalfield in 1996. For each generation and sex, cumulative proportions of emerged adults [emerged adults / (live pupae + emerged adults)]

for different julian dates were modeled using logistic regressions (analysis 15 in Table 6).

Sampling procedures sumrnarîzed in Tables 4,5. CUMULATIVE PROPORTION OF EMERGED ADULTS

Fig. 22. Flame ionization detector (FID) and electroantennographic detector (EAD: male

Oiketim kirbyi antema) responses to pherornone extract of femde 0.kirbyi, chromatographed on a DB-210 column (1 min at 1 OO°C, 15 OC/min to 1 80°C, 2'C/min to

220°C). MBO = 1-methylbutyl octanoate; MBN = 1-rnethylbutyl nonanoate; MBD = 1- rnethylbutyt decanoate; MPD = I -methylpentyl decanoate; 1 -MBDD= 1 -methylbutyl dodecanoate. Retention indices of MBO, MBN,MBD, MPD, and MBDD were respectively:

1643,1745, 1846,1942,2050 (DB-210); 165 1,1756, 1858,1949,2064 (DB-23);1588,

1688, 1788, 1876, 1990 (SP- 1000). Experimental procedures described in sections 2.5.1 and

2.5.2. ' Indicates chiral center of molecule. MBD FI D 4

MBOO MPD i MBO MBN

MBOD /vwdOk

5 6 7 8 9 10 STRUCTURES TlME [min] were homologous MBO, MBN, MPD, and MBDD. Equivalent amounts of synthetic md fernale-produced esters elicited similar antennal responses. GC-MS-CI-SIM analyses of pheromone extract and synthetic MBO, MBN. MPD, and MBDD resulted in retention time and ion ratio matches of synthetic and fernale-produced compounds, except for MBN; synthetic MBO [mlz (%)]: 145 (1 OO), 2 15 (M+1,63), extract: 145 (1 OO), 2 15 (64); synthetic

MPD: 1 73 (100), 257 (M+1,45), extract: 143 (1 00). 257 (40); synthetic MBDD: 20 1 (1OO),

271 (M+1.53), extract: 201 (100). 271 (49). Diagnostic ions m/z 159 and 229 for MBN were detected at the correct retention time. but the ion ratio could not be accurately detemined due to a coeluting compound also containing m/r 229 (but not m/r 159).

In a 12-replicate experiment, (R)-MBD attracted on average 11.7 male bagworms per trap, while the (S)-enantiorner was unattractive and in combination with (R)-MBD completely inhibited response. Unbaited traps did not attract any males. Attraction of males to (R)-MBD pmportionally decreased as the arnount of (S)-MBD in the lure increased (Fig.

23). The (R)-but not (S)-enantiomen of either MBO, MBN. or MBDD strongly enhanced attractiveness of (R)-MBD (Fig. 24). (S)-MBO and (9-MBDD were inactive, whereas (8-

MBN added to its antipode reduced attractiveness (Fig. 24). (R)-,(5')-, and racemic MPD were behaviorally benign. A five-ester blend was more attractive than any of four binary blends containing (R)-MBD (Fig. 25). Blends of (R)-MBD in temary combination with either (R)-MBO and (R)-MBN or (R)-MBN and (R)-MBDD were as attractive as the five- ester blend (Fig. 26). Five- and four-ester blends were equally attractive (Fig. 27).

Increasing the amount of synergistic (R)-MBO and (R)-MBN relative to (R)-MBD increased attractiveness (Fig. 28). Lure attractiveness increased as the arnount of phemmone increased fiom 10 ta 10,000 pg (Fig. 29). Fig. 23. Number of male Oikericlrs kirbyi captured in Unitnps baired with different ratios of

(R)-and (8)-MBD. 5 - 9 April 1993, Coto, Costa Rica, N=8. Bars with the same letter superscript are not significantly different (analysis 17 in Table 6). Expetimental procedures described in section 2.5.3. - RATIOS OF (FI)- and (9-1-METHYLBUTYL DECANOATE (MBD) - Fig. 24. Number of male Oikcricus kirbj? captured in Unitnps baited with (R)-EUIBD alone and in combination with optical isomers of MBO (1 - 15 November 1992, N = 1 j), MBN (15

- 23 October 1993, N = 1O), and MBDD (1 4 - 28 October 1993, N = 1 O), Coto, Costa Rica.

Bars with the same letter superscript are not significantly different (anaiysis 17 in Table 6).

Expenmental procedures described in section 2.5.3; compound abbreviations as in Fig. 22. 1 (R)- or (S)-1 -methylbutyl octanoate (MBO)

-- tom 1000 1000 1000 1000 lm W-MBD Iwl 0 9 9 10 100 10 100 (R)-MBO [ri91 - 10 100 œ - 10 1W (S)-MBO[re]

4 (R)- or (S )-1-methylbutyl nonanoate (MBN) a

- - 1O0 100 (R)-MBOD L~J - 100 - too (SI-~eoo BLENDS OF CANDIDATE PHEROMONE COMPONENTS - Fig. 25. Number of male Oiketicus kirbyi capttsred in Unitraps baitrd with (R)-iLLBD in binary and pentanary combinations with (R)-MBO, (R)-MBN, (R)-MPD, and (R)-MBDD. 1

- 8 November 1993, Coto, Costa Rica. N = 10. Bars with the same letter supencript are not significantly different (analysis 17 in Table 6). Experimental procedures described in section

2.5.3; compound abbreviations as in Fig. 22. MBD MBO MBN MPD MBDD BLENDS OF CANDIDATE PHEROMONE COMPONENTS - Fig. 26. Number of male Oiketicus kirbyi captwed in Unitraps baited with (R)-MBD in ternary and pentanary combinations with (R)-MBO, (R)-MBN, (R)-MPD. and (R)-MBDD. 2

- 22 November 1993, Coto, Costa Rica, N = 20. Bars with the sarne letter superscript are not significantly different (analysis 17 in Table 6). Experimental procedures desctibed in section

2.5.3; compound abbreviations as in Fig. 22. a

M8D MBD M80 MBD MBD MBD MBO MBO - MBO - MBO MBN MBN MBN - MBN - MPD - - -' MPD MPD MBDD - MBDD MBDD - - BLENDS OF CANDIDATE PHEROMONE COMPONENTS Fig. 27. Number of male Oikeficuskirbyi captured in Unitmps baited with (R)-MBD in quaternary and pentanary combinations with (R)-MBO, (R)-MBN, (R)-MPD, and (R)-

MBDD. 4 - 15 November 1993, Coto. Costa Rica, N = 10. Bars with the same letter superscript are not significantiy different (analysis 17 in Table 6). Experimental procedures descnbed in section 2.5.3; compound abbreviations as in Fig. 22. a

MBD MBD MBD MBD MBO O MBO MBO MBN MBN 0 MBN MPD MPD MPD 0 L RENDS OF CANDIDATE PHEROMONE COMPONENTS - Fig. 28. Number of male Oikrticus kirbyi captured in Unitraps baited with (R)-MBD. (R)-

MBO,and (R)-MBN in different ratios. 12 - 2 1 November 1993, Coto, Costa Rica, N = 10.

Bars with the same letter superscript are not significantly different (analysis 17 in Table 6).

Experimental procedures described in section 2.5.3; compound abbreviations as in Fig. 22. RATIOS OF CANDIDATE PHEROMONE COMPONENTS Fig. 29. Number of male Oiketicus kirhyi captured in Unitraps baited with (R)-MBD alone and in combination with (R)-MBO and (R)-MBN (1: 1 : 1) at increasing doses. 19 - 27

November 1993, Coto, Costa Rica, N= 10. Bars with the same letter supencript are not significantly different (anaiysis 17 in Table 6). Experimental procedures described in section

2.5.3; compound abbreviations as in Fig. 22. MBD+MBO+MBN MBD 4.8 Mating success of females

Size of females was a significant component of mating success for 0.kirbyi [Coto 50:

F=28.03, dp1,69, P<0.0001; Coto 52: F=66.41, df= 1,397, P < 0.000 1; cage experiment: t =

3.63, df = 8 1, P = 0.0005 (section 2.4.2)] and M. plana (F= 77.00, df = 1,94, P < 0.000 1). with mated females being larger than unmated females (Figs. 30,3 1). Non-significant interactions behveen mating statu and crown section [O. kirbyi in Coto 50: F= 1.4 1, df = 2,69, P = 0.252;

O. kirbyi in Coto 52: F = 1.10, df= 2,397, P = 0.3351, and between mating status and generation [M. plana in Ladang Coalfield: F = 0.13, cl!= 4.94, P = 0.97 11 indicated that size of females influence their mating success independently of either their position within the canopy of palms (Fig. 30) or the generation during which sampling was cmied out (Fig. 3 1).

Proportions of mated female O. kirbyi [rnated fernales / (rnated females + unmated females)] increased with time (Coto 50: 2 = 53.05. P < 0.0001 ; Coto 52: 2= 13.98, P c

0.0001) and decreased from upper to lower palm leaves in Coto 50 w' = 20.79. P < O .O00 1) but not Coto 52 &=0.08, P =O.782) (Figs. 32,33). Non-significant variations of mating success between leafgroups in Coto 52 may have been caused, in part, to small females on the upper crown of palms having relatively low mating mccess (Figs. 14,30). Assessing proportions of mated femdes in relationship with both crown section and length of pupd case indicated rnarginally significmt differences between leaf groups & = 3.3 8, P = 0.0659), with females on upper leaves having slightly greater mating success than femdes on lower leaves (Fig. 34). Captures of male 0.kirbyi in pheromone-baited traps suspended on oil palms 1.5,2.5 and 3 -5 m above grouad increased with increasing trap height (Fig. 35).

Density of pupal bags per palm did not significantly affect mating success of female O. kirbyi

(Coto 50: F=0.84,df= 1,165, P=0.361; Coto 52: F=0.30, df= 1,145, P=0.582). Fig. 30. Length of pupal cases for mated and unmated fernale Oiketim kirbyi sampied on different leafgroups of 80 palms in Coto 52 (section 2.3.2) and 15 palms in Coto 50 (section

2.4. l), and in a cage expenment (section 2.4.2). Leaf groups 5-7 and 35-37 are the highest and lowest, respectively (Fig. 6); statistical analysis 18 in Table 6. LENGTH OF PUPAL CASES (cm) (X + SE) Fig. 3 1. Length of pupal bags for mated and unrnated fernale Metisa plana sampled during 5 consecutive generations on 95 palms in Ladang Coalfield. Sampling procedures summarized in Tables 4,s; statistical analysis 19 in Table 6.

Fig. 32. Percentage of mated female Oiketicus kirbyi [mated femaies / (mated kmales + unmated females) * 1001 for different weeks (w) and le& groups (g) of oil pdms sampled in

Coto 50. w:l for 17 palms sampled between 05-10 October; w:2 for 21 palms sampled between 13-1 7 October; w:3 for 29 palms sarnpled between 19-24 October; w:4 for 25 pdms sampled between 26-3 1 October; w:5 for 16 pdms sarnpled between 02-07 November; w:6 for 19 polms sampled between 09-14 November; w:7 for 20 palms sampled between 16-2 1

November; w:8 for 13 palms sampled between 23-28 November; w:9 for 22 palms sampled between 30 November - 05 December. g: 1 for leaves 2- 13; g:2 for leaves 14- 19; g:3 for leaves 20-25; g:4 for leaves 26-3 1; g:5 for Ieaves 32-43. Oil palms sampled after more than

97.3% of females had emerged (w5 to w9) were pooled. Leaf groups 2-13 and 3243 are the highest and lowest, respectively (Fig. 6). Smpling procedures sumrnarized in Tables 4, 5; statisticai analysis 20 in Table 6. week 3 week 4 weeks 5-9

LEAF GROUP Fig. 33. Percentage of mated female Oiketicus kirbyi [meted females I (mated females + unmated females) * 1001 for different weeks (w) and leafgroups (g) of oil palms sarnpled in

Coto 52. w: 1 for 20 palms sampled the 05 April; w:2 for 20 palms sampled the 12 April; w:3 for 25 palms sampled the 19 April; w:4 for 20 palms sampled the 26 Apd; w:5 for 20 palrns sampled the 03 May; w:6 for 20 palms sampled the 10 May; w:7 for 20 palms sampled the 17

May; w:8 for 20 palms sampled the 24 May; w:9 for 20 palms sampled the 08 June. g: 1 for leaves 9-1 1; g:2 for leaves 17-19; g:3 for leaves 25-27. Oil palms sampled ;ifter more than

95.8% of females had emerged (w7 to w9) were pooled. Leaf groups 9-1 1 and 25-27 are the highest and lowest, respectively (Fig. 6). Sarnpling procedures summarized in Tables 4,s; statistical analysis 20 in Table 6. week 3 week 4 week 5 week 6

LEAF GROUP Fig. 34. Logistic regressions (analysis 22 in Table 6) modeling the relationship between proportion of mated females [mated femdes / (mated females + unmated females)] and

length of pupal case for fernale Oiketicus kirbyi sampled on different leafgroups of 80 pdms

in Coto 52. Leaf groups 9-1 1 and 25-27 are the highest and lowest, respectively (Fig. 6).

Sampling procedures described in section 2.3.2. leaf group

LENGTH OF PUPAL CASE (cm) Fig. 35. Nurnber of male Oiketicus kirbyi captured in pheromone-baited traps suspended at 3 heights on 5 palms in Coto 52. Bars with the same letter superscript are not significantly different (analysis 23 in Table 6). Experimental procedures described in section 2.6. 1.5 2.5 3.5 TRAP HEIGHT (m) Proportions of mated female M. plana per palm varied between generations (F=

23.87, df= 1,93, P < 0.000 1), with females having greatest mating success during generations

3 to 5 (Fig. 36). Multiple regression models chedout for each of 5 generations of M. plana indicated: 1) inversely density-dependent mating success during 4 generations; and 2) increasing mating success with increasing mean Iength of pupal bags per palm during 3 generations (Table 14). The effect of pupation site or emergence time on mating success of female M. planu could not be investigated because only one leafper palm was sarnpled after the majority of females had emerged (Table 4; Fig. 21).

4.9 Fecundity of females

Weight of egg masses of 207 female 0. kirbyi was positively correlatrd with length of pupal cases (Fig. 37). Fecundity of Rmale M. plana was positively correlated with length of pupal bag (Fig. 38); mated females laid on average 155.7 t 5.5 eggs (N = 84) (Fig. 38).

4.10 Inter-generational variation of population density

4.10.1 Sampling study coaducted in plantation of oil palm

The Iive population of M. plana in Sungei Merah (section 2.3.3) consisted of only early instars, indicative of non-overlapping generations. Previous surveys revealed low

nurnbea of bagwoms before Febniary 1995, indicating that possible accumulation with time

of empty female bags on paim leaves was insignificant. This justified the use of regressions

between numbers of female bags (generation 11 and early instars (generation i+ 1 ) as a means

to investigate intra- and inter-tree distributions for successive generations of bogworms.

The abundance of early instars aad empty pupal bags varied between leaf groups of Fig. 36. Proportion of mated fernale Metisaplonn [mated females / (mated femaies + unmated females)] sampled during 5 consecutive genentions on 95 pdms in Ladang

Coalfield. Sampling procedures summarized in Tables 4,s; statistical analysis 25 in Table 6. PROPORTION OF MATED FEMALES PER PALM (XISE) Table 14. Multiple regression models analyzing the efTects of two variables [xi = density of pupai bags per palm; xz = mean length of pupal bags per palm] on mating success of fernale Metisa plana [y = mated females 1 (mated females + unmated females)] for each of 5 consecutive generations '. To ensure homogeneity of variance, data were subjected to logarithmic [xi'= ln x] and arcsin transformations [y'= sin " d y]. Pnnimeter estimates with O* II' star superscripts are statistically significant ( = Pc0.05; = Pc0.0 1; = P<0.000 1) '.

Generation Density (P ,) ' Length of bags (Pz) '

' See Table 4 and Section 2.3.4 for site specific characteristics of paims and sampling procedures for different generations.

Statisistical analysis 26 in Table 6.

Significance levels for Ho :Pi = O and Ho: pz = O. Estimates of intercept values (P.) are not represented because they have no biological meaning (values of x2 never approach O). Fig. 37. Relationship between length of pupal case and weight of egg rnass for 207 mated fernale Oiketicus kirbyi sampled on 19 paims in Coto 50. Sampling procedures described in section 2.3.1; statistical andysis 27 in Table 6. LENGTH OF PUPAL CASE (cm) Fig. 38. Relationship between length of pupal bag and number of eggs in pupd case for 84 mated female Metisa pkuna smpled on oil palms in Ladang Coalfield. Sarnpling procedures are describeci in section 2.3.4; statistical analysis 28 in Table 6. f I L 1 I 0 L 1 10 Il 12 13 14 15 16 17 LENGTH OF PUPAL BAG (mm) palm (F=5.8 1, df= 2,38, P =0.0063), with largest numbes of bagworms being present on the middle strahun of palm crowns (Fig. 39). The non-significant interaction between life stage and leaf group (F= 0.1 5, df = 2,3 8, P = 0.865) indicated similar within-tree distributions for female bags and early instars (Fig. 39).

If emergent larvae commonly do not disperse, leaves with high numbers of female pupal cases would be expected to cary high numbers of larvae. Significant positive regressions between numbers of female bags and early instars per leaf were observed for most (36 of 39) palms (Table 15). supporting the hypothesis that larvae commonly remain on the same leaf where they emerged. Significani positive intercepts for most (34 of 39) palms

(Table 15) indicates that leaves not infested with female pupd cases cmied significant numben of larvae; this may be attributed, in part, to larvae having dispersed From leaves with high density of fernales.

The relation between numben of enrly instars (y) and female bags (x) per palm \vas

3 more adequately described by non-linear (y = Po + Pl x + pz x-, with Po= 4.14. P = 0.984; P ,=

5 1.7. P<0.000 1; P2= - 0.175, P = 0.0009; 2 = 0.867, P < 0.000 1) than linear regression model

(y = Po + Pi x, with Po= 604.6, P c 0.0001 ; P ,= 27.46. P < 0.000 1; 2 = 0.8 18, P c 0.000 1) (Fig.

40). The significant negative P2 value resulted in a decreasing dope with increasing density of female bags, suggesting (among othet possibiliiies) density-de pendent dis persal by larvae.

If tme, palms surroundhg heavily infested palms would be expected to carry proportionately more larvae than other palms. This hypothesis wûs tested by regressing numbers of female bags against numbers of early instars on palms 9 m and > 9 m apart fiom the 4 most heavily infested palms (> 1 10 female bags per ph). A reduced regression model, excluding the non-significant interaction between numbers of female bags and distance to heavily infested Fig. 39. Numbers of empty female pupal cases and earlirly instvs of Mefisaplunu on different leaf groups of 39 palms sampled in Sungei Merah. Leaf groups 5-1 O and 3 1-36 are the highest and lowest, respectively (Fig. 6). Sampling procedures swnmarized in Tables 4,s; statistical analysis 29 in Table 6. NUMBER OF EMPTY FEMALE PUPAL CASES PER PALM (a)(cm) (Tc i SE)

NUMBER OF EARLY INSTARS PER PALM (O) (Xf SE) Table 15. Coefficients of detennination ($), intercepts (a) and slopes @) of regressions between numbea of empty female pupal cases (x) and early instars (y) of Metisa plana per leaf on 18 leaves of 39 oil palms sampled in Sungei Merah '. Parameter estimates with star mperscnpts 80' "1' are statistically significant ( : PQ).05; ' : Px0.0 1; : P<0.00 1; : P<0.000 1) '.

Palm r2 Intercept (Pa) s lape (P 1 Y Table 1 5. (continued)

See section 2.3.3 for site specific chancteristics of palms and sampling procedures.

' Statisistical analysis 30 in Table 6.

Significance levels for Ho : Po = O and Ho : Pl = 0. Fig. 40. Relationship between numbers of M. plana empty fernale pupal cases and early instars per palm for 39 palms sarnpled on in Sungei Merah. Sampling procedures summarized in Tables 4, 5; statistical analysis 32 in Table 6 NUMBER OF EMPTY FEMALE PUPAL CASES palms (F= 0.05, df= 1JO, P = 0.8 19), revealed different intercepts (F=12.00, df = 1,3 1, P =

0.001 6) for trees 9 m and > 9 m apart fiom heavily infested palms. The difference in intercept values indicated that trees adjacent (9 m) to heavily infested palms carried proportionately more early instars per female bag than trees merapart (> 9 m) (Fig. 4 1).

1.10.2 Expcrimcnts conducted in controlled cage settings

Larvae were more abundant on palrns infiested with LOO nther than 10 female bags.

In each cage, there were more larvae on palms Uifested with female bags than on uninfested palms. Uninfested pdrns carried more larvae when enclosed with palms infested with 100 rather than 10 female bags (F ig. 42).

4.11 Individual variations of reproductive success and population dynamics

4.1 1.1 Cage study with Metka plana

Comparing numbers of early instars on nursery palrns infested with either 4.8, or 12

female pupal bags (section 2.7) indicated that densities of larvae and female bags were

positively correlated, although data were scattered and the coefficient of determination was

relatively low (r2 = 0.345; P = 0.0040) (Fig. 430). Dissecting pupal bags to determine

reproductive success of females greatly improved the ability to predict infestation level in the

offspring generation, as indicated by a highly significant regession between densities of

larvae and mated females (? = 0.760; P c 0.000 1) (Fig. 43b). Reproductive output per

femde in the offspiing generation (nurnber of larvae per female bag per palm) increased with

reproductive success in the parental generation (number of mated femdes per number of

female bag per palm) (Fig. 43c). Fig. 41. Regressions between numben of empty female pupal cases and eariy instars of

Metisa plana per palm for oil pdms 9 rn (N = 18) and > 9 m apart (N= 17) from the 4 most heavily infested palms (> 1 10 female pupal cases) in Sungei Merah. Sampling procedures summarized in Tables 4, 5; statisticd anaiysis 33 in Table 6. NUMBER OF EMPTY FEMALE PUPAL CASES Fig. 42. Nurnbea of early instars of Metisa plana (y) sampled on palms non-infested or infested with femaie pupal bags [xi = O or 1, respectively]; palms were infested with either

10 or 100 female bags [x2 = 10 or 1001 [xi = 1 and x2 = 10 for pdms cmying 10 femde bags; xi = 1 and xz = 100 for palms carrying 100 female bags; xi = O and XI = 10 for uninfested palms enclosed in cages containing palms with 10 female bags; xi = O and xl=

100 for uninfested palms enclosed in cages containing palrns with 100 female bags].

Significance levels as follows: P (xi) = 0.750; P (xi) = 0.0036; P (xl*x2) = 0.0064.

Experirnental procedures described in section 2.2.2; statistical analysis 24 in Table 7. NUMBER OF LARVAE / PALM$c+SE) Fig. 43. A - C. Regression models comparing variations of population density between consecutive generations of Metisa plana on 22 palms enclosed individually in screen cages in

OPRS, near Banting; experimental procedures are listed in section 2.7. A) Relationship between numbers of fernde bags ruid emly instars per pdm. B) Rclationship between nurnber of mated femaies and early instars per palm. C) Relationship between reproductive success of fernales in parental generation (number of mated females per palm I total number of females per palm) and reproductive output per female in offspnng generation (number of

lame per palm I total nurnber of females per palm). For regression models B and C. female bags containing live larvae that had not yet dispersed were excluded from anaiysis. D - F.

Regression models comparing variations of population density between consecutive

generations of Metfsa plana sampled on 95 palms in Ladang Coalfield; experimental

procedures are listed in section 2.3.4. D) Relationship between nurnber of females in

parental generations and number of males plus females in offspnng generations. E)

Relationship between number of mated females in parental generations and number of males

plus females in offspring generations. F) Relationship between reproductive success of

females in parental generations (number of mated femaies I total number of females) and

reproductive output per female (number of males and females in ofispring generation / total

number of femaies in parental generation). 12 NUMBER OF FEMALES

NUMBER OF REPROOUCTIVE FEMALES

0.25 0.50 0.75 1.00 PROPORTION OF REPROOUCTlVE FEMALES 4.1 1.2 Field study with Metka plana

The impact of female reproductive success on population dynamics of bagworms was assessed by quantiQing variations of population density between consecutive generations of

M. plana in Ladang Coalfield (section 2.3.4). Results of regression models presented below should be interpreted carefùlly, because: 1) densities of bagwoms in consecutive generations are not independant data; 2) mortality during the lamal stage?as well as immigration and emigration of larvae fiom the experimental plot, were not recorded; and 3) each model has low statistical power because it relied on only 4 data points. The models are based on the assurnption that numbers of pupal bags recorded (destroyed) for different generations provide an unbiased estimûte of population density in the study site.

In a first regression model. numben of females in parental generation (i)were compnred with numbers of males plus fernales in offspring generation (i+l),using 4 sets of inter-generational cornparisons (generation 1 versus generation 2,2 vernis 3.3 versur 4, and

4 versus 5). Densities of pupal bags in ofTspnng generations were weakly conelated with densities of female bags in parental generations (# = 0.338; P = 0.419) (Fig. 43d). A second regression model, comparing numbea of mated femaies in parental generations with nurnbea of males plus fernales in offspring generations, disclosed a relatively high but insignificant coefficient of determination (r2= 0.560; P = 252) (Fig. 43e). A third regression model, comparing reproductive output per female in offspring generations (number of males plus females in offspring generation I nurnber of fernales in parental generations) versus reproductive success of females in parental generations (number of mated fernales / nurnber of females), indicated a significant positive correlation (2= 0.974; P = 0.01 3) (Fig. 430. 5.0 DISCUSSION

Sampling and experimentai studies conducted in plantations of oil paim revealed individual variations of lifetime reproductive success for female O. kirbyi and M. plana.

Relatively large proportions of females did not reproduce, either because they did not survive as pupae (section 5.1.1) or mate as adults (section 5.1.2). Assessments of fecundity mer indicated variations of reproductive output among mated femdes (section 5.1.3). Life history traits that affect reproductive success of females include: dispersal by larvae (section

5.2.1), size attained at pupation (section 5.2.2), selection of pupation site (section 5.2.3), timing of adult emergence (section 5.3. l), and sexual communication between sessile

females and winged males (section 5.3 2). Density-dependent dispersa1 and inversely density-dependent reproductive success of females contxibute to stabilize populations of M. plana on individual palms (section 5.4). Variations of reproductive success among female

M. plana affect their population dynamics (section 5.5).

5.1 Components of reproductive success

In order to reproduce, female bagworms must first survive as pupae (section 5.1.1)

and then mate as adults (section 5.1.2). Females may also enhance their reproductive success

by laying large number of eggs (section 5.1.3).

5.1.1 Survival durhg pupal stage

For both O. kirbyi and M. plana, large proportions of individuals did not survive as

pupa (Figs. 17,18). As reported for T. ephemeraefomis (Gross & Fritz, 1982; Cox & Potter,

1988; Cronin & Gill, 1989), moaality of male pupae of M. plana exceeded that of female pupae (Fig. 18). Mortality of female pupae of 0. kirbyi, in contnst, was greater thm that of male pupae (Fig. 17). Sex-specific differences in mortality may be attributed to either larger size of female pupae (Rotheray & Barbosa, 1984; Cronin & Gill. 1989), longer pupal stage of males (Stephens, 1962; Ccuttwell, 1974; Mishm, 1978; Ponce et al., 1979; Thangavelu &

Ravinciranath, 1985; Basn & Kevan, 1995), sex-specific susceptibility of pupae to natural enemies, or any combination thereof.

Size attained by iarval M. piana at pupation enhanceci survival during pupal stage

(Fig. 19). Size-dependent survivd of pupal T. ephemeraeformis cornes about because parasitoids are unable to penetrate large bags with their ovipositor (Cronin & Gill, 1989).

Pmsitism is an important mortality factor among pupal M. plana (Basn et al., 1995).

The location of pupd O. kirbyi within palm crowns had a marginally significant

effect on pupal rnortality, with proportions of deûd pupae increasing from upper to lower

leaves (Fig. 17). Low incidence of pûnsitisrn among pupal T. ephemerueformis in the upper crown of redcedars has been attributed to parasitoids restricting host searching behaviour to

the bottom strata of trees (Gross & Fritz, 1982; but see Cronin & Gill, 1989).

Mortdity of neither male nor female pupae of O. kirbyi was afYected by population

density (analysis 12 in Table 6). Similarly, mortality of male pupae of M. plana was density-

independent during each of 5 generations (Table 12). Mortality of female pupae of M. plana,

in contrast, was inversely density-dependent during 2 generations and density-independent

during 3 generations (Table 12). Mechauisms underlying inter-generational and sex-specific

variations of densitydependent processes remain unclear. 5.1.2 Mating success

Because females generdly invest more time and energy in their offspring than do males, most insect mating systems are characterized by males competing for access to females and by females discriminating between potential mates (Thornhill & Alcock, 1983;

Anderson, 1994; Choe & Crespi. 1997). Sex-role reversal has been reported in insects with either parental investment of males exceeding that of femdes (Gtvynne, 198 1; Smnsson B

Petersson, 1987; Simmons & Bailey, 1990), strongly female-biased operational sex-ratios

(Shelly & Bailey, 1992), or both. However, few field studies have estimated proportions of adult females that do not mate (but see Greenbank. 1963; Shapiro, 1970).

As demonstrated or suggested for the b~gworms0. kirbyi (Magistretti et al., 197 1), T. ephemeraefurmis (Bmows, 1974; Sheppard, 1975; Kaufmann, 1985; Klun et al., L 986),

Clania crumerii (Wemood) (Thangavelu & Gunasekaran, 1982), Solenobia manni (2.)

(Malicky, 1968), and other insect species with flightless females (Doane, 1968; Hmhman &

Futuyma, 1985; Wing, 1991; Sharov et al., 1995). large proportions of female O. kirbyi and

M. plana did not mate as adults (Figs. 30-34.36). Occurrence of unmated female bagwoms

may be attributed to: 1) complex mating procedure [up to 20% of males that insen their

extensible abdomen into bags of receptive females fail to inseminate them (Kaufman, 1968;

Thangavelu & Gunasekaran, 1982)l; 2) short longevity and sexual attractiveness of females

(Stephens, 1962; Entwistle, 1963; Kaufman, 1968; Thangavelu & Gunasekaran, 1982;

Kuppusami & Kaanan, 1993; Basn & Kevon, 1995); 3) low mating capacity of short-lived

males that may be active only a few hours each day and invest large amount of time per

copulation (Thangavelu & Gunasekaran, 1982; Campos Arce et al., 1987); 4) female-biased

operatioaal sex-ratios in protogynous populations with non-overlapping generations (Figs. 20,2 1 ; see also Stephens, 1962; Morden & Waldbauer, 197 1 ;Thangavelu & Ravinciranath,

1985; Gara et al.. 1990; Kuppusamy & Kannan, 1993); 5) flightlessness per se constraining mating success of females (Bell. 1982; Roff, L990); or any combination of the foregoing.

Size-dependent mating success of female O. kirbyi and M. plana (Figs. 30-3 1) may be due, in part, to males preferentially mating with large females. Mate choice by males has been reported in many insects (Thonihill & Alcock, 1983; Andeason, 1994) and is üdaptiw when 4 conditions are met (Johnson & Hubbell. 1984). 1) Males have limited mating capacity and cannot mate al1 receptive females. Mating bears hi& costs for males. either in terrns of ejaculate production (Boggs & Gilbert, 1979; Sims, 1979; Gwynne, 198 1 ;

Dewsbury, 1982; Markow & Ankney, l984), time investment associated with copulation

(Hieber & Cohen, 1983), or mate guarding (Johnson & Hubbell, 1984). 2) Femdes differ in quality as sexual resource. Positive correlations between size and fecundity of fernales have been demonstrated in several insects (Honek, 1993). 3) Males are able to discriminate between females of different quality. Visual or chernical cues may mediate mate discrimination by males in Lepidopten (Rutowski, 1982b). 4) Possibilities for mate choice exist. Mate discrimination by males is adaptive when the operationai sex-ratio is skewed

toward females (Lawrence, 1986; Gwyme & Simmons, 1990; Shelly & Bailey, l992), or

when males encounter large aggregations of receptive females (Johnson & Hubbell, 1984).

Al1 4 conditions for mate choice by males seem to be met for 0. kirbyi and M. plana.

1) Signifïcant proportions of unmated female bagworms (Figs. 30-34,36) indicate that males

are a limiting =source. 2) Females daer in quality as sexual cesource, with large females

being most fecund (Figs. 37,38). 3) Superior mating success of large female O. kirbyi in a

confïned, cage environment (Fig. 30) indicates that males discriminate between and are more strongly attracted to large females. 4) In protogynous populations with non-overlapping generations (Figs. 20,21), males encounter large aggregations of non-dispersing females that emerge in relative synchrony.

Decreasing mating success of female M. plana with increasing density of pupal bags per palm (Table 14) rnay be attributed to either mate choice by males or "dismption" of mating on crowded palms. 1) Offspring of female M. plana on densely populated palnu possibly have relatively low fitness, because neonatal larvae rnay suEer significant mortality during dispersal bouts (Cox & Potter, 1986) or attain small size at pupation (Table 1 1).

Males rnay therefore increase their fitness by discriminating between host plants and by mating preferentially with females on scarcely infested palms. 2) Female bagworms attnct mates by expelling pheromone-impregnated scales from their pupal case into the lower part of the bag (section 1.3.3). Because pheromone dissemination fiom these scdes is independent of the fernale's mating status, females rnay remain attractive for some time after mating, unless repellent pheromones mediate cessation of sexual attractiveness (section

5.3.2). Low mating success of females on crowded palms rnay therefore stem fiom numerous bags of mated females either disorienting or repelling mate-seeking males.

5.1.3 Fecundity

As with several other Lepidoptera (Honek, 1993), fecundity of female O. kirbyi (Fig .

37) and M. plana (Fig. 38) is size-dependent. Additional factors ihat rnay affect fecundity of fernale bagworms include age of females at mating (Rogers & Marti, 1994, 1997; Vicken,

1997) and quaiity of males (Royer & McNeü, 1993; Delisle & Bouchard, 1995). 5.2 Life history traits related to reproductive success: larval stage

Bagworm larvae may enhance their reproductive success as adults by dispersing fiom crowded palms (section 5.2. l), attaining large size before pupation (section 5.2.2), and selecting a suitable site for pupation (section 5.2.3).

5.2.1 Dispersal

Upon leaving their matemal bag, neonate bagworms may either remain on the host where they emerged or disperse to sunounding hosts. Similar distribution of bagwoms in parental and offspring generations. both within and between palms (Figs. 39,40,42; Table

13, indicate that M. plana larvae commonly remain near their site of emergence. These results suggest that populations of bagworms on individual palms comprise coexisting families of sibling larvae, with each family defined as the non-dispersing progeny of one mated fernale. Localized populations of bagworms that subsisted on the same host for more than two generations may further be viewed as complex demographic structures that include families of sibling mated females and subfamilies of larval offspring. High levels of genetic relatedness within and between families would suggest that cooperative interactions among aggregated larvae are adaptive. Numerous examples of social behaviour have been reported in lepidopteran species with ciustered eggs and non-dispersing Iarvae (Wellington, 1957,

1965; Fitzgeraid. 1993, 1995; Costa and Pierce, 1997), but interactions between foraging bagwonn laivae have not yet been investigated.

Population density is positively correlated with incidence of larval dispersal in several lepidopteran species (Poirier & Borden, 1992; Berger, 1992; Torres-Villa et al., 1997).

Resource partitioning is a key component of fitness, because high density of insects and defoliation of host plant adversely affect body size and reproductive success of adults

(Haukioja & Neuvonen, 1985; Cox Br Potter, 1988; Carter et al., 199 1; Ruohomaki, 1992).

Crowded conditions on palms negatively affect the reproductive success of M. plana, as indicated by small size of pupal bags (Table 11) and low mating success of fernales (Table

14). Sampling studies conducted with M. plana provided evidence that larvae most readily disperse fiom crowded pnlms: 1) proporîions of early instars per fernale bag decreased with increasing population density (Fig. 40); 2) palms surrounding heavily infested palms carried proportionately more larvae per female bag (Fig. 41); and 3) incidence of ballooning by mid to late instars increased with increasing population density (Fig. 8). These studies. however. did not discnminate between effects of population density per se and intense defoliation of heavily infested palms. Experiments carried out in controlled cage environment (section

2.2.2) revealed that both increasing population density (number of larvae; number of female

bags in parental generation) and defoliation level (naturd or artificial) prornote dispersal by

larvae, as indicated by low recapture rates for Iarvae and pupae, high rates of within-leaf

dispersa1 and ballooning, as well as early timing of ballooning (Fig. 9).

Density-dependent dispersal (Figs. 9, 11.42) indicates that the presence of

conspecifics influences larval behaviour. possibly resulting from more frequent contacts

between individuals (Mariath, 1984; Berger, 1992). Defoliation-dependent dispersal (Figs. 9,

12, 14) may result fiom either shortage or detenorated palatability of leaf folioles. Feeding

activity of M. plana larvae decreases with increasing level of leaf tannins (Basri, 1993).

Positive relationships between defoliation level and tannin content has been dernonstrated for

some plant species (Karban & Myers, 1989), but this relationship has not yet been

investigated for oil palms. The higher rate of larval bailooning fiom unfertilized than fertilized palm (Fig. 9) rnay reflect low nutrient content in the former. Pale green to yellow color of dertilized leaves in experiment 4 in OPRS (section 2.2.2) was symptomatic for nitrogen deficiency (Turner, 198 1), which rnay have adversely affected feeding by M plaru larvae (Basri, 1993).

Rate of within-leaf dispersal was similar for early and mid to late instars; rate of ballooning, in contrast, was much lower for early than for mid to late instars (Table 10).

Dispersal by ballooning rnay be most advantageous late during larval development for several reasons: 1) benefiis of larval aggregation (social facilitation of feeding, low levels of predation or parasitism) rnay be most signifiant for early instars (Fitzgerald, 1993; Costa &

Pierce, 1997); 2) cornpetition for food is most intense among late instars (Basri, 1993); 3) large larvae rnay tolerate long penods of starvation during dispersal bouts (Stockoff, 199 1); and 4) in cornparison with early instars, large ballooning larme that land on the ground rnay walk faster and Mer(Reavey, 1993), therefore having greater probability of locating a suitable host. Future studies need to assess the distance over which early and late instars cm walk without food.

The mark-recapture study indicates that a large proportion of lame undertake multiple episodes of ballooning (Table 9). Occurrence of larvae on palms different fiom the one where they had been marked (Table 9) demonstrates that ballooning mediates between- host dispersal by larvae. Large numbers of marked larvae recaphired on their original host

(Table 9), however, also indicate that bdlooning larvae commonly do not disperse between hosts. Larvae transferred on nursery palms (section 2.2) fiequentiy ascended and descended on their siken thread, sometimes reRuning to the leaf fkom which they originally ballooned

(personal observation). Ballooning behaviour of M. plana larvae rnay therefore represent a stmtegy for foraging both within hosts ("tethered flight") and between hosts (''free flight").

As reported for other Lepidoptera with flightless females (Table 3), dispeaal and pupation behaviour of male and femaie late instar M. plana differ (Figs. 11,12,13. 16). The incidence of ballooning was greater for femaie than male larvae, and increased for both sexes with increasing population density and defoliation level (Figs. 1 1, 12). Greater abundance of

Fernale than mde pupae on undefoliated palrns (Fig. 13) further suggens scx-specific impact of host defoliation on pupation behaviour of larvae. Differences in dispersal and pupation behaviour of male and femde Iarvae rnay partly be explained by findings that crowdedness mostly affects reproductive success of fernales. Fernales on crowded pnlms attain small size at pupation (Table Il) and have low mating success (Table 14) and fecundity (Fig. 38) as adults. For males, in contrast, crowdedness does not consistently affect size attained at pupation (Table 1 l), and is not expected to influence mating success becûuse winged adults can disperse. Unexpectedly, similar rates of ballooning were observed for male and female larvae at high population density (Fig. 11). This result suggests that sex-specific impact of population density is not the only factor affecthg dispersa1 behaviour of females and males.

Altematively, this result may be an experimentai artifact: densities of up to 1000 late instars per nursery palm rnay have exceeded naniral Mestation levels (personal observation).

Life history of bagworms mggests that dispersai is particularly adaptive for female larvae. Because neonatal M. pluna larvae commonly initiate development on their natal host, dispersal and selection of pupation sites by female late instars influences the distribution and

(possibly) fitness of their progeny (Figs. 39,40,42; Table 15). Greater incidence of ballooning by female than male larvae (Figs. 1 1, 12), therefore, may be amibuted to femdes seeking suitable hosts for their friture progeny. Assuming that dispershg late instars most effectively locate suitable hosts (see above), natural seiection rnay have favoured dispersal by late instar females rather than their emergent progeny. Future studies need to quanti@ trade- offs betweeen the costs of dispersa1 [e.g. mortality] and benefits of colonizing uncrowded hosts [e.g. enhanced reproductive success of adults] for early and late instar M. plana.

53.2 Size rttained nt pupation

Size attained by larvae at pupation is an important component of reproductive success, with large bagworms having greatest swival during pupal stage (male and female

M. plana: Fig. 19), mating success (female O. kirbyi and M. plana: Figs. 30,3 1) and fecundity (female O. kirbyi and M. plana: Figs. 37,38). Lengths of pupal case of female O. kirbyi were density-independent (analysis 4 in Table 6), possibly due to a srnall range of population densities (0.4 - 5.6 larvae per leaf per ph)within the experimentd plot. In contrast, lengths of pupal bag of male and female M. planu were negatively correlated with population density during 2 and 4 generations, respectively (Table 1 1). Unexpectedly, lengths of male pupal bags increased with population density during one generation (Table

1 1). Sex-specific impact of population density on length of pupai bags may have resulted from male M. plana having lower feeding requirements (Basri & Kevan, 1995) and therefore being Iess affected by intraspecific cornpetition in cornparison with females.

It may seem surprishg that lengths of M. plana pupal bags decreased even for population densities (1 to 176 pupal bags per leaf) well below the theoreticai carrying capacity of oil palm (Basri, 1993). This phenomenon rnay be attributed to nipid deterioration of leaf quality in response to herbivory, or to shortage of apical leaf folioles, which are most commonly fed upon by defoliators (Basri, 1993; Rhainds et ai., 1993, 1996). 5.2.3 Pupatioa site

Greater proportions of mated female 0. kirbyi in the upper than lower leaf groups of palms (Figs. 32,34; but see Fig. 33) support the hypothesis that pupation site of females influences the mating success of emergent adults. Increasing captures of male bagworms in pheromone-baited tmps with increasing trap height (Fig. 3 5) indicntes that females on upper lewes most cffectively attnct males for rnating. Enhanced mating success of remde 0. kirbyi in tree tops rnay be attributed to effective dissemination ofpheromone €rom high sites. or to males foraging predominantly in the upper crown of oil palms.

Similar within-tree distributions of female pupol cases and enrly instars (Fig. 39) and positive correlations between numbers of female pupal cases and early instars per leaf(Tab1e

15) support the hypothesis that pupation site of female M. plana influences the distribution of emergent larvae. The quality of leaves as a food resource for bogworms may Vary within the palm crown. Feeding activity of larvai M. plana increases with low leaf tannin and high nitrogen contents, both of which are characteristics of young leaves in the upper palm crown

(Basri, 1993). Because early instars commonly remain on the same leaf where they emerged

(Table 1 S), pupation of females on most palatable leaves in the upper palrn crown mûy contribute to the fitness of their progeny.

Large nurnbers of pupd bags present on heavily defoliated palm leaves (unpublished observations) suggest that bagworm larvae feed and pupate on the same le&. Late instars, however, were aiso observed to walk on palm trunks and to pupate on leaves weakly defoliated (unpublished observations), suggesting that some larvae forage between leaves when seeking suitable pupation sites. Selection of pupation site by late instar bagworms rnq represent a trade-off between the cost [energetic expenditure, enhanced larval mortality (Weiss et al., l987)] and benefits of inter-Ieaf movements before pupation. Sexual segregation of pupation site for 0.kirbyi and M. plana (Figs. 13, 16) likely resuits fiom sex- specific constraints affecting pupation behaviour of larvae.

5.3 Life history traits reiated to reproductive succeas: adult stage

Timing of emergence (section 5.3.1) and sexud communication behveen mdes and females (section 5.3.2) influence the reproductive success of adult bagworms.

5.3.1 Timing of emergence

Generation cycles of 70-80 days for M. plana (Table 13) approximate developmental tirnes recorded under labontory conditions (82-90 days; Barri, 1993). Developmental times for M. plana reared in captivity suggest that males and females emerge in synchrony (Basri

& Kevan, 1995). In Ladang Coaifield, however, females emerged before males during each of 5 generations (Fig. 2 1). Similarly, Fernale O. kirbyi ernerged before males in both Coto 50 and Coto 52 (Fig. 20). Increasing proportions of mated fernale O. kirbyi during the emergence period (Figs. 32,33) likely resulted fiom an increased availability of males (Fig.

20), as has been reported for the noctuid Spodopiera lirura (F.) (Ôtake & Oyama, 1973).

Protancky is cornrnon in the Lepidoptera (Thomhill & Alcock, 1983). Protogyny, in contrast, has rarely been reported (lankovic, 1960; Solomon & Neel, 1972; Mikkola, 1987;

Taylor & Shields, 1990; Shaw, 1993) but may be common in bagworms (Figs. 20,2 1; see also Stephens, 1962; Morden & Waldbauer, 197 1; Thangavelu & Ravinchnath, 1985; Gara et al., 1990; Kuppusamy & Kannan, 1993). In species with singly mating females and multiple-mating mdes, natural selection should favour protandry because early emergent males have greatest mating success (Wiklund & Fagerstr6m, 1977; but see Baughman,

199 1). Protogyny in bagwoms is therefore unexpected and may have evolved as a strategy to reduce inbreeding (Thomhill& Alcock, 1983). Because neonatal larvae cornmonly remain on their natal host (M. plana; Figs. 39,40,42; Table 15) or are wind-dispersed in groups of siblings to surrounding hosts (0.kirbyi; Newman, 1980), localized populations of bagworms possibly exhibit hi& levels of genetic relatedness (section 5.21).

In uniform plantations of oil palm with discrete. non-overlapping generations of bagwoms, large proportions of females never mate as adults (Figs. 32,33,36). Greater reproductive success of late emerging female O. kirbyi (Figs. 20,32,33) may generate strong selective pressures to synchronize emergence of males and females. If, however, bagworm developmental times and stages vary in diverse microhabitats and climates of tropical rainforests, protogynous female bagworms may attract males from adjacent populations before males of their own local population emerge. In such habitats, protogyny would represent an unusud but stable evolutionary strategy that furthers outbreeding.

5.3.2 Sexual communication

Following eclosion, female bagworms attract mates by expelling pheromone- impregnated scales out of the pupal case into the lower pari of their bag (Stephens, 1962;

Bosmand & Brand, 197 1; Zhao, 198 1 ; Leonhardt et al., 1983; Acosta, 1986; Neal, 1986;

Loeb et al., 1989). GC-EAD analysis (Fig. 22) and field bioassays (Fig. 23) indicated that

(R)-MBD is the major pheromone component of O. kiroyi. This cornpound was previously reported for T. ephemeraeformis (Leonhardt et al., 1983). but sex pheromones of 0.kirbyi and T. ephemeraeformis are clearly distinct Although (S)-MBD is behaviourally benign in T. ephemeraeformis (Leonhardt et al., 1983). it strongly inhibits behavioural response of male O. kirbyi (Fig. 23). Contrasting with the single ester sex pheromone of T. ephemeraeformis, female O. kirbyi produce a pheromone blend that includes MBD plus 3 synergistic chiral esters: MBO. MBN and MBDD (Figs. 22,24). The more complex

pheromone of O. kirbyi may be attributed to a more diverse bagworm fauna in the tropical

Amencas (Barbosa, 1993), providing seleetive forces for 0.kirbyi to evolve a multiple component blend of chiml esters for species-specific communication (Cardé & Baker, 1984).

Enhanced atûactiveness of (R)-MBD when combined in either two-, three-. four-, or

five-cornponent blends with (8)-MBO. (R)-MBN, (R)-MPD and (R)-MBDD (Figs. 24-27)

indicated that al1 compounds except MPD are sex pheromone components. Equal

attractiveness of 2 temary and al1 four- and five-component blends (Figs. 26,27) suggest

redundancy of pheromone components for attraction of males. Redundancy of pheromone

components was first documented in the cabbage looper, Trichoplusia ni (Hübner), with

various pheromone blends compensating for the Iack of one or more components (Lim et cd.,

1984). Although individual components may be redundant for the attraction of male O.

kirbyi, they may serve to inhibit cross-attraction of sympatric bagworms.

During calling, most female moths actively release sex pheromone fiom abdominal

glands for attraction of males (Percy-Cunningham & MacDonald, 1987). Cessation of

calling behaviour and decline of pheromone production &er mating (Richerson & Cameron,

1974; Raina, 1984; Webster & Cardé, 1984; Giebultowicz et al., 199 1) has been associated

with ûansfer of sperm [Lymaniria dispar (L.) (Giebultowicz et al., 199 l)] or male accessory

gland secretion [Heliothis zeu (Boddie) (Raina, 1984)l into the spermetheca Receptive

female 0.kirbyi, in contrast, amct males by expelling most of their pheromone-impregnated scales at once fiom their pupal case into the lower part of the bag (unpublished observations).

As pheromone dissipation From thes2 scales is independent of the femaie's mating status, it is unclear whether or how mate-seeking males discriminate between bags containing virgin or rnated females. Mechanisms involved in cessation of sexual attractiveness may include production of anti-pheromone by mated femaies, rnarking of mated fernales by males, andor npid disseminntion ofpherornone components From scales. Some female O. kirbyi exprlled hairs from their pupd case iifter mating (unpublished observations), but the chernical composition of these hais was not determined.

Intraspecific variations of emission rate or composition of pheromones are comrnon in female moths (Shorey & Gaston, 1965; Miller & Roelofs, 1980; Mistro Pope rr ai., 1982;

Haynes et al., 1984; Collins & Cardé, 1985; Barrer er al., 1987; Du et al.. 1 987; Foster et cil.,

1989; Witzgail& Frérot, 1989), but no studies have yet tested whether individual variations in phemmone blends influence cornpetitive interactions between females for access to males

(Phelan, 1992). Greaier arnounts of pheromone-impregnated scales released by larger fernale

0.kirbyi (unpublished observations) and pheromone-dose-dependent attraction of males

(Fig. 29) mggest that quantitative variations in pheromone production by female 0. kirbyi mediates cornpetition for access to males. If that is hue, greater mating success of large females (Fig. 30) may be attributed to either large females most effectively attracting males for mating, or males being preferentially attracted to pheromone plumes of large fernales.

5.4 Density-dependent processes and population dynamics

Although O. kirbyi and M. plana inhabit similar habitats (uniform plantations stocked with evenly-aged, eveniy-spaced oil palms), they exhibit distinct population dynamics. Fecundity of female 0. kirbyi greatly exceeds that of M. plana (Table l), but densities of pupal bags per leaf were greater for M. plmthan O. kirbyi (Table 8). Spatial distribution in plantations of oil palm was extremely patchy for 0. kirbyi and relatively uniform for M. plana (section 2.1). In this section, 1 review density-àependent processess that may influence dynamics of local populations.

5.4.1 Oiketicus kirbyi

Neither size attained by lame at pupation, survival during pupal stage, nor mating success of females were affected by population density (analyses 4, 12 in Table 6; section

4.8), suggesting that density-dependent processes do not contribute to stabilize populations of

0.kirbyi on individual palms. This conclusion, however, should be interpreted carefùlly because: 1) the range of population density within experimental plots was relatively small

(0.4 to 5.6 pupal bags per leaf per palm); and 2) palms infested with O. kirbyi were systematicaily selected for sampling over large (CU 30 ha) areas (section 2.3), and large-scale spatial patterns rnay have "masked" density-dependent processes.

Density-independent dispeaal by land O. kirbyi may be the most important parameter affecting dynamics of local populations. Progeny of one female O. kirbyi has the potential to almost completely defoliate one oil palm (Table 1), and larval dispersal may be adaptive even at low population density. Clumped distribution of O. kirbyi in plantations of oil paim (section 2.1) may reflect either dispersa1 by larvae in groups of siblings, or spatial variation in incidence of larval mortality. Future stuâies, assessing demographic trends for consecutive generations of O. kirbyi, are needed to detennine the impact of dispersal on population dynamics. 5.4.2 Metisa plana

The progeny of one fernale M. plana causes minor defoliation of oil palm (Table l), and dispersai by larvae may not be adaptive at low population denstiy. Sampling studies and cage experiments indicated that neonatal larvae commonly remain on their natal host (Figs.

39,40,42; Table 19, and that "crowding" of oil palm (hi& population density, intense defoliation) promotes biillooning by larvae (figs. 8,9, 1 1. 12). Density- and defoliation- dependent ballooning by larval M. plana strongly impacts the dynamics of local populations, because: 1) ballooning mediates dispersal between palms (Fig. 42; Table 9); 2) lame commonly undertake multiple episodes of ballooning (Table 9); and 3) incidence of ballooning is greater for female than male Inrvae (Figs. 11, 12). Density-dependent dispersal may generate relatively unifocm distributions of M. plana in plantations of oil palm, by simultaneously stabilizing populations on heavily infested palms and redistributing larvae on lightly infested palms.

Mortality of female pupae likely did not have a regdatory impact on populations of

M. plana, because: 1) incidence of mortality slightly decreased with increasing population density (Table 12); and 2) natural enemies removed mainly small pupae (Fig. 19) that have

low expected reproductive success as adults (Figs. 3 1,38). However, the regulatory impact of naniral enemies on populations of M. plana remains unclear, because mortality during the

larval stage was not assessed in my studies.

Size attained by female lame at pupation decreased with increasing population

density (Table 1I), suggestiag that resource availability influences population dynarnics.

Small size of females on heavily infested palms likely has a stabilizing efTect on local

populations, because srnail females have low reproductive output (Figs. 3 1,38). Low incidence of mated females on heavily infested palms (Table 14) may ais0 contribute to stabilize populations of M. plana.

5.5 Individual variations of reproductive success and population dynamics

Inteapecific differences in oviposition behaviour of females (Price et ai., 1990; Price.

1994) or environmentally-based matemal effects (Rossiter, 1995) havc been hypothesized to strongly influence population dynamics of latent and eruptive insect species. Intraspecific variations of female reproductive output, associated with either size-dependent fecundity

(Tammaru et al., 1996a), size-dependent swival (Ohgushi. 1996), population quality induced by maternai effects (Rossiter, 1992), emergence time (Cushman et ai., 1994). or density-dependent decline of reproductive output (Wall & Begon, 1987). have further been hypothesized to influence demographic trends of insect populations. However, no studies carried out under aatural conditions have yet quantified whether and to what extent variability of female reproductive success influences population dynamics.

Monality during pupal stage and mating failure of fernale bagwoms represent a net reproductive loss in offspring generations, and may therefore influence demographic trends.

The relationship between individual variations of reproductive success and population dynamics could not be assessed for 0.kirbyi, because sampling was not replicated accross generations. Sarnpling studies and cage experiments carried out for consecutive generations of M. plana allowed me to quanti@ the relationship between reproductive success of females and population dynamics. Regression models comparing female reproductive success in parental generations and reproductive output in oflspring generations, both accross generations (regression analysis using 4 inter-genenitionai comparisons; section 4.7.2) and accross individual palms (regression analysis using 22 individually cage palms; section

4.7.1), revealed significant positive relationships (Fig. 43), indicating that variations of female reproductive success influence both temporal and spatial fluctuations of population density .

Significant van0ationsof population density, bag length and pupal mortality between trees (Figs. 15, 18, 56) suggest that attributes of oii paim (such as quality of leaf folioles as food resource for larvae, or incidence of natural enemies) ûffect population dynamics of M. plana. Seasonal fluctuations of rainfall or temperature may also have contributed to variations of population density (Fig. 7), length of pupal bags (Fig. 15), rnortality during the pupal stage (Fig. 18) and mating success of femdes (Fig. 36) dwing different genentions.

Inter-generational variations of density-dependent processes (Tables Il, 12, 14) may also be attributed, in part, to seasonal fluctuations of abiotic factors. The impact of abiotic facton, however, remains dificult to assess because my study was not replicated across locations or years. Long-term studies are needed to determine whether and to what extent seasonal fluctuations of abiotic facton, attributes of oil palm, and inter-genentional variations of reproductive success influence population dynamics of M. plana in plantations of oil palm. 6.0 LITERAW CITED

Acosta, A.G. (1986) Estudios altemos al control de Oikericus kirbyi Guild. en palmeras de la Costa S.A. IV Mesa Redonda Latinoamericana sobre Palma Aceitera. Valledupar, Departamento del Ctsar. Colombia. pp. 99- 102.

Alghaii, A.M. (1 984) The selection of pupation sites by the stalk-eyed fiy Diopsis thoracica (Diptera: Diopsidae) and pupal pamitism in sorne rice cultivars. Annals of Applied Biology 105: 189-194.

Almela Pons, G., H. Domenech & N.U. Martini (1972) Accion del Bacillus thurigiensis Berliner sobre Oiketicus kirbyi Guilding en test de Iaboratorio. Revista de la Facultad de Ciencias Agrarias, Univeaidad Nacional de Cuyo. 18: 55-60.

Ameen, M. & P. Sultana (1977) Biology of the bagworm crumerii Westwood (Lepidoptera: Psychidae) from Dacca, Bangladesh. Journal of Natural History 1 1 : 17- 24.

Andenson, M. (1 994) Sexual selection. Princeton University Press, Princeton. 599 pp.

Am, H., E. Stiidler & S. Rauscher (1975) The eiectroante~ographicdetector - a selective and sensitive tool in the gas chromatographie analysis of insect pheromones. Zeitschrifi Rlr Naturfonchung 30c: 722-725.

Baiduf. W.V. (1 937) Bionomic notes on the common bagworm, ephemeraeformis Haw. (Lepid., Psychidae) and its insect enemies (Hymen., Lepidoptera). Proceedings of the Entomological Society of Washington 39: 169- 1 84.

Barbosa, P. (1993) Lepidopteran foraging on plants in agroecosystems: constraints and consequences. In N.E. Stamp & T.M.Casey (eds). : ecological and evolutionary constraints on foraging. Chapman & Hail, New York.

Barbosa, P., V. Knschik & D. Lance (1989) Life-history traits of forest-inhabiting flightless Lepidoptera. American Midland Naturalist 22: 262-274.

Barbosa, P., M.G. Waldvogel & N.L. Breisch (L983) Temperature modification by bags of the bagworm Thyridpteryx ephemeraefirmis (Lepidoptera: Psychidae). Canadian Entomologist 1 15: 855-858.

Barrer, P.M., MJ. Lacey & A. Shani (1987) Variation in relative quantities of airborne sex pheromone components fiom individual female Ephestiu cmtella (Lepidoptera: ). Journal of Chemical Ecology 13: 639-653,

Bamws, E.M. (1974) Some factors affecting population size of the bagwonn, Thyridopteryx ephemeraeformis (Lepidoptera: Psychidae). Environmental Entomology 3 :929-93 2. Basri, M.W. (1993) Life history, ecology and economic impact of the bagworm, Metisa plana Walker (Lepidopten: Psychidae) on the oil paim, Jacquin (Palmae), in Malaysia. Ph. D. thesis, University of Guelph, Canada. 217 pp.

Basri, M.W.,A.H. Hassan & 2. Masijan. 1988. Bagwoms (Lepidoptera: Psychidae) of oi palms in Malaysia. PORIM Occasional Paper, Palm Oil Research Institute of Malaysia.

Basri, M.W. & P.G. Kevan (1995) Life history and feeding behaviour of the oil palm bagworm, Mrrisa plana Walker (Lepidoptera: Psychidae). Elaeis 7: 1 8-34.

Basn, M.W.,K. Norman & A.B. Harndan, A.B. (1995) Natural enemies of the bagworm, Metisa plana Walker (Lepidoptera: Psychidae) and their impact on host population regulation. Crop Protection 14: 637-66.

Baughman, J.F. (1991) Do protandrous males have increased mating success? The case of Euphydryas editha. American Naturalist 13 8 : 536-542.

Bell, G. (1982) The masterpiece of nature: the evolution and genetics of sexuality. University of California Press, Berkeley. 635 pp.

Bellinger, R.G., F.W.Ravlin & M.L. McManus (1990) Predicting egg mass density and fecundity in field populations of the gypsy moth (Lepidoptera: Lymantriidae) using wing length of male moths. Environmental Entomology 19: 1024- 1028.

Berger, A. (1992) Larval movements of Chilo partellus (Lepidoptera: Pyralidae) within and between plants: timing, density responces and survival. Bulletin of Entomologicd Research 82: 441 -448.

Bergh, J.C., E.S. Eveleigh & W.D.Seabrook (1 988) The mating statu of field-collected male spruce budwom, Choristoneurafinziferma (CLem.) (Lepidoptera: Tortncidae), in relation to trap location, smpling method, sampling date, and aduit emergence. Canadian Entomologist 120: 82 1-830.

Boggs, C.L. & L.E. Gilbert (1 979) Male contribution to egg production in buttedies: Evidence for transfer of nutrients at mating. Science 206: 83-84.

Bosman, T. & J.M. Brand (1971) Biological studies of the sex pheromone of junodi Heyl. (Lepidoptera: Psychidae) and its partial purification. Journal of Entomological Society of South Afiica 1: 73-78.

Botterweg, P. (1982) Protandry in the pine looper, Bupalus piniarius (Lep., Geometridae): an explanatory model. Netherlands Journal of Zoology 32: 169-1 93. Bourgogne, J. (1990) Observations sur les premiers états et le comportement des Psychides. Premiére partie: l'oeuf et le fourreau primaire [Lep.] Bulletin de la Société Entomologique de France 95: 145- 159.

Bourgogne, J. (1993) Observations sur les premiers 6tats et le comportement des Psychides. IV Études expérimentales (Lepidoptera: Psychidae). Bulletin de la Société Entomologique de France 98: 343-350.

Brown, G.C., A.A. Berryman & T.P. Bogyo (1979) Density-dependent induction of diapause in the codling moth, Laspeyresiu pomonella (Lepidoptera: Olethreutidae). Canadian Entomologist 1 1 1 : 43 1-43 3.

Campbell, R.W. (1967) The analysis of numericd change in gypsy moth populations. Forest Science Monognph 15: 1-33.

Campbell, R.W., D.L.Hubbard & R.J. Sloan (197Sa) Patterns of gypsy moth occurrence within a sparse and numerically stable population. Environrnental Entomology 4: 535-542.

Campbell, R.W., D.L. Hubbard & R.J. Sloan (1975b) Location of gypsy 1110th pupae and subsequent pupal survival in sparse, stable populations. Environrnental Entomology 4: 597-600.

Campos Arce, J.J., O.F. Peres & E.F. Berti (1987) Biologia do bicho co cesto Oiketicus kirbyi (Lands-Guilding, 1827) (Lepidoptera: Psychidae) em folhas de Ei~calyptusspp. Anais de Escuela Superiore Agricultura Luiz de Queiroz 44: 341-358.

Capinera, J.L. & P. Barbosa (1 976) Dispersal of first-instar gypsy moth larvae in relation to population quality. Oecologia 26: 53-64.

Cappucinno, N. (1 995) Novel approaches to the snidy of population dynamics. Ln N. Cappuccino & P.W.Price (eds). Population dynamics: New Approaches and Synthesis. Academic Press, San Diego. pp. 3-16.

Cardé, R.T. & T.C. Baker (1984) Sexual communication with pheromones. In W.J. Bell & R.T. Cardé (eds), Chemicd ecology of insects. Sinauer Associates, Sunderland. pp. 355-383.

Carroll, A.L. & D.T. Quiring (1993) Interactions between size and temperature influence fecundity and longevity of a tortricid moth, Zeirapheru canadensis. Oecologia 93: 233-241 .

Carter, M.R. F.W. Ravlin & M.L. McManus (1991) Changes in gypsy moth (Lepidoptera: Lymantriidae) fecundity and male wing length remlting fiom defoliation. Environmentai Entomology 20: 1042-1 047. Choe, I.C. & B.J. Crespi (eds) (1977) The evolution of mating systems in insects and arachnids. Cambridge University Press, Cambridge. 387 pp.

Collegrave, N. (1993) Does larvai cornpetition affect fecundity independently of its effect on adult weight? Ecological Entomology i 8: 275-277.

Collins, R.D. & R.T. Carde (1985) Variation in and hentability of aspects of pheromone production in the pink bollworm moth, Pectinophora gossrpiella (Lepidoptera: ). Annals of the Entomological Society of Arnerica 78: 229-234.

Cook, S.P., F.P. Hain & H.R. Smith (1 994) Oviposition and pupai survivai of gypsy moth (Lepidoptera: Lyrnanûiidae) in Virginia and North Carolina pine-hardwood forests. Envuonmental Entomology 23 : 360-366.

Comell, H.V.& B.A. Hawkins (1 995) Survival patterns and mortality sources of herb ivorous insects: some demographic trends. Amencan Naturalist 145: 563-593.

Costa, J.T. & NE. Pierce (1997) Social evolution in the Lepidoptera: ecological context and communication in larval societies. In J.C. Choe & B.J. Crespi (eds). The evolution of social behavior in insects and arachnids. Cambridge University Press. Cambridge. pp. 407-442.

Cox, D.L. & D.A. Potter (1986) Aerial dispersal behaviour of larval bagworms, Thyridopteryx ephemeraeformis (Lepidoptera: Psychidae). Canadian Entomologis t 1 18: 525-536.

Cox, D.L. & D.A. Potter (1988) Within-crown distributions of male and femaie bagworm (Lepidoptera: Psychidae) pupae as affectcd by host defoliation. Canadian Entomologist 120: 559-567.

Cronin, J.T. & D.E. Gill(1989) The influence of host distribution. sex, and size on the level of parasitism by ltoplectis conquisitor (Hymenoptera: Ichneumonidae). Ecological Entomology 14: 163-1 73.

Cruttwell, R.E. (1974) The bagwocms (Lep.: Psychidae) of Trinidad and their natunl enemies. Technical Bulletin of the Commonwealth Institute of Biological Control 17: 127- 159.

Cushman, J.H.,C.L. Boggs, S.B. Weiss, D.D. Murphy, A.W. Muphy, A.W. Harvey & P.R. Ehrlich (1994) Estimating female reproductive success of a ihreatened butteffly: influence of emergence time and host-plant phenology. Oecologia 99: 194-200.

Davis, D.R. (1964) Bagworm moths of the Western Hemisphere. United States National Museum Bulletin 244,233 pp. Davis, D.R. (1 975) A review of the West Indian moths of the family Psychidae with descriptions of new taxa and immature stages. Smithsonian Contributions to Zoology 188,66 pp.

Deinert, E.I., J.T. Longino & L.E. Gilbert (1994) Mate competition in . Nature 370: 23 -24.

Delisle, L & A. Bouchard (1 995) Male larval nutrition in Choristoneura rosaceuna (Lepidoptera: ): an important factor in reproductive success. Oecologia 104: 508-51 7.

Dempster, J.P. (197 1) The population ecology of the cinnabar moth, Tyriajacobaeae L. (Lepidoptera, Arctiidae). Oecologia 7: 26-67.

Dempster, J.P. (1983) The nanusil control of populations of butterflies and rnoths. Biological Review of Cambridge Phylosophical Society. 58: 46 1481.

Dempster, J.P., D.A. Atkuison & M.C. French (1995) The spatial population dynamics of insects exploiting a patchy food resource. II. Movements between patches. Oecologia t 04: 3 54-362,

Denno, R.F. & M.A. Peterson (1995) Density-dependent dispersal and its consequences for population dynamics. In N. Cappuccino & P.W. Pnce (eds). Population dynamics: New Approaches and Synthesis. Academic Press, San Diego. pp. 1 13-130.

Desmier de Chenon, R. (1982) Field guide for oil paim pests and plantation sanitary protection. Medan, Indonesia. 244 pp.

Dewsbury, D.A. (1 982) Ejaculate cost and male choice. American Natunlist 1 19: 60 1 -610.

Dickinson, J.L. (1992) Scramble competition polygyny in the milkweed leaf beetle: combat. mobility, and the importance of being there. Behavioral Ecology 3: 32-41.

Doane, C.C. (1968) Aspects of mating behaviour of the gypsy moth. AM~Sof the Entomological Society of Arnerica 6 1: 768-773.

Du, J.W., C. Lofstedt & I. Lofqvist (1987) Repeatability of pherornone emissions fiom individual female ermine moths Yponomeuta padellus and Yponomeuto rorellus. Journal of Chernical Ecology 13: 143 1-1441.

Elgar, MA.& NE. Pierce (1988). Mating success and fecundity in an ant-tended Lycaenid . In T.H. Clutton-Brock (ed). Reproductive success: studies of individual variations in contrasting breeding systems. University of Chicago Press, Chicago. pp. 59-75. Elliott, W.M.(198 1) The relationship of embryo counts and suction trap catches to population dynamics of Macrosiphum euphorbiae (Homoptera: Aphididae) on tornatoes in Ontario. Canadian Entomologist 1 13 : 11 13- 1 122.

Entwistle, P.F. (1 963) Observations on the biology of four species of Psychidae (Lepidoptera) on L. in Western Nigeria. Proceedings of the Royal Entomological Society of London (A) 38 : 145- 152.

Fadamiro, H.Y., T.D.Wyatt & M.C. (1996) Flight activity of Prostephunus truncutus (Hom) (Coleoptera: Bostrichidae) in relation to population density, resource quality, age, and sex. Journal of Insect Behaviour 9: 339-35 1.

Fagestrom, T. & C. Wiklund (1982) Why do males emerge before fernales? Protandry as a mating stmtegy in male and fernale butterflies. Oecologia 52: 164- 166.

Fitzgerald, T.D. (1993) Sociality in caterpillars. In N.E.Starnp & T.M. Casey (eds), Caterpillars: ecologicai and evolutionary constraints on foraging. Chapman and Hall, New York. pp. 372-403.

Fitzgerald, T.D. (1995) The tent caterpillars. Corne11 University Press, Ithaca.

Foltz I.J., F.B. Knight & D.C. Allen (1972) Numerical analysis of population fluctuations of the jack pine budwom. Annals of the Entomological Society of Arnerica 65: 82-89.

Foster, S.P., J.R. Clearwater & S.J. Muggleston (1989) Intraspecific variation of two components in sex pheromone gland of Planotortrix excessana sibl ing spec ies. Jomal of Chernical Ecology 15: 457-465.

Fulvia Garcia, R. (1987) Aspectos biologicos y manejo del gusano canasta Oiketicus kirbyi. Institut0 Agropdcuario. Boletin Tecnico 149.23 pp.

Gara, R.I., A. Sarango & P.G. Cannon (1990) Defoliation of an ecuadorian mangrove forest by the bagworm Oiketicus kirbyi Guilding (Lepidoptera: Psychidae). Journal of Tropical Forest Science 3 : 18 1 - 186.

Gause, G.F.(1 934) The Struggle for Existence. Williams & Wilkins, Baltimore.

Geier, P.W.(1964) Population dynamics of codling moth Cydiapornonella (L.). Australian humal of Zoology 12: 38 1-416.

Genty, P., R. Desmier de Chenon & J.P. Morin (1978) Las plagas de la palma aceitera en America Latina. Oléagineux 33: 395-396.

Giebultowicz, LM., A.K. Raina, E.C. Uebel & RL. Ridgway (1991) Two-step regulation of sex-pheromone decline in mated gypsy moth femaies. Archives of Insect Biochemistry and Physiology 16: 95-105. Greenbank, D.O. (1963) The anaiysis of moth survivai and dispersal in the unspniyed area. In R.F. Moms (ed). The Dynamics of Epidemic Spruce Budworm Populations. In Memoirs of the Entomologicai Society of Canada, 3 1: 87-99.

Gross, S.W. & R.S. Fritz (1982) DBerential stratification, movement and parasitism of sexes of the bagwonn Thyridopteryx ephemeraeformis on redcedar. Ecological Entomology 7: 149-1 54.

Gwynne, D.T. (1981) Sexual seiection theory: Mormon crickets show role reversal in mate choice. Science 2 13: 779-780.

Gwy~e,D.T. & L.W.Simrnons (1990) Experimental reversal of courtship roles in an insect. Nature 346: 172-174.

Hackman, W. (1 966) On wing reduction and loss of wings in Lepidoptera. Notuale Entomologica 46: 1- 16.

Hagstrum, D.W.& D.L. Sihacek (1980) Diapause induction in Ephestia cautelfa:an interaction between genotype and crowding. Entomologia Experimentalis et Appiicata 28: 29-37.

Hanison, S. (1994) Resources and dispenal as factors limiting a population of the tussock moth (Orgvia vetusta), a flightless defoliator. Oecologia 99: 27-34.

Harrison, S. (1 997) Persistent, Iocdized outbreaks in the western tussock moth Orgyia vetusta: the roles of resource quality, predation and poor dispersal. Ecological Entomology 22: 158-166.

Harrison, S. & N. Cappuccino (1995) Using density-manipulation experiments to midy population regdation. in N. Cappuccino & P.W. Price (eds). Population dynarnics: New Approaches and Synthesis. Academic Press, San Diego. pp. 13 1- 147.

Harshman, L.G. & D.J. Futuyma. (1985) Variation in population sex-ratio and mating success of asexual lineages of Alsophila pometaria (Lepidoptera: Geometridae). hnais of the Entomological Society of America 78: 456-458.

Hartley, C. W. (1988) The Oil Palm (Elaeisguineensis (Jacqu.)). Longman Scientific & Technical, New York. 761 pp.

Hassell, M.P. & KM. May (1985) From individual behaviour to population dynamics. In RM. Sibly & RH. Smith (eds). Behavioural Ecology. Blackwell, Oxford. pp. 3-32.

Hastings, J. (1 989) Protandry in western cicada Wler wasps, (Spheciusgrandis, Hymenoptera: Sphecidae): an empirical study of emergence tirne and mating opportunity. Behaviorai Ecology and Sociobiology 25: 255-260. Haukioja, E. & S. Neuvonen (1 985) The relationship between size and reproductive potential in male and fernale Epirrita autumnata (Lep., Geometridae). Ecological Entomology 10: 267-270,

Haukioja, E., E. Pakarinen, P. Niemela & L. ho-Iivari (1988) Crowding-triggered phenotypic responses aileviate consequences of crowding in Epirrita autumnata (Lep., Geometridae). Oecoiogia 75: 549-558.

Haukioja, E., J. Suomela & S. Neuvonen (1 985) Long-term inducible resistance in birch foliage: triggering cues and efficacy on a defoliator. Oecologia 65: 363-369.

Haynes, K.F., L.K. Gaston, M. Mistrot Pope & T.C. Baker (1984) Potential for evolution of resistance to pheromones: interindividual and interpopulational variations in chernical communication system of pink bollworm moth. Journal of Chernical Ecology 10: 1551-1565.

Heather, N. W. (1975) Life history and biology of the leaf bagworm, Hyalarcta huebneri (Westwood) (Lepidoptera: Psychidae). Journal of Ausenlian Entomologicai Society 14: 353-361.

Herzig, A.L. (1995) Effects of population density on long-distance dispersa1 in the goldenrod beetle Trihabh virgata. Ecoiogy 76: 2044-2054.

Hieber, C.S. & J.A. Cohen (1983) Sexual selection in the lovebug, Plecia neariica: The role of mate choice. Evolution 37: 987-992.

Ho, C.T.& C.L. Teh (1999) The use of Euphorbia heterophylla (L.) for natural reduction of leaf pests damage to oil palm. In Proceedings PORIM International Palm 0i1 Congress: Emerging Technologies and Opportunities in the Next Milleniurn. PORiM, Kuala Lumpur. pp. 139-164.

Honek, A. (1993) Intmpecific variation in body size and fecundity in insects: a general relationship. Oikos 66: 483-492.

Hom, D.J. & R.F. Sheppard (1979) Sex-ratio, pupal parasitism, and predation in two declining populations of the bagworm, Thyridopteryx ephemeraefrmis (Haworth) (Lepidoptera: Psychidae). Ecological Entomology 4: 259-265.

Howlader, M.A. (1990) Biology of the bagwom moth, Pteromaplagiophleps Hamps. (Lepidoptera: Psychidae) fiom Dhaka, Bangladesh. Bangladesh huma1 of Zoology 18: 1-9.

Huffaker, C.B. (1 958) Experimental studies on predation: dispersion factors and predator- prey oscillations. Hilgardia 27: 343-3 83. Hunter, A.F. (1995a) The ecology and evolution of reduced wings in forest rnacrolepidoptera. Evolutionary Ecology 9: 275-287.

Hunter, A.F. (1995b) Ecology, life history, and phylogeny of outbreak and nonoutbreak species. In N. Cappuccino & P.W.Price (eds). Population dynamics: New Approaches and Synthesis. Academic Press, San Diego. pp. 4 1-64.

Jankovic, L. (1960) Vertikalna distribucija gubara (Lymaniria dispar L.) i njegovo ponasanje. Zast Bilja 57/58: 203-209.

Johnson, L.K. & S.P. Hubbell(1984) Male choice: experimental demonstration in a brentid weevil. Behaviourd Ecology and Sociobiology 15: 183- 188.

Johnson, W.T.& H.H. Lyon (1976) Insects that feed on trees and shnibs: rn illusmited guide. Comell University Press, Ithaca. pp. 150- 153.

Jones, F.M. (1 927) The mating of the Psychidae (Lepidoptera). Transactions of the Arnerican Entomological Society 53 :293-3 12.

Jones, F.M.& H.B. Parks (1928) The bagworms of Texas. Bulletin of Texax Agncultural Experiment Station 382: 5-26.

Kaitda, A. & C. Wiklund (1994) Polyandrous female butterflies forage for matings. Behaviod Ecology and Sociobiology 35: 385-388.

Karban, R. & J.H. Myers (1989) Induced plant responses to herbivory. Annual Review of Ecology and Systematics 20: 33 1-348.

Kaufinan, T. (1968) Observations on the biology and behaviour of the evergreen , Thyridopteryx ephemeraeformis (Lepidoptera: Psychidae). Annals of the Entomological Society of Arnerica 61 : 38-44.

Kaumian, T. (1985) Dicymolomiu julianalis (Lepidoptera: Pyralidae) as an endoparasite of the bagworm, Thyridopteryx ephemeraeformis (Lepidoptera: Psychidae): its relation to host, life history and gonad development. Pan-Pacific Entomologist 6 1: 200-209.

Klun, LA., J. W. Neal Jr, B.A. Leonhardt & M. Schwarz (1 986) Suppression of female bagworm, Thyridopteryr ephemeraeformis reproduction potentiai with its sex pheromow, 1-MBD. Entomologia Experimentalis et Applicata 40: 23 1-238.

Kolodny-Hirsch, D.M. & R.E. Webb (1993) Mating disruption of gypsy moth (Lepidoptera: Lymantriidae) following ground application of high rates of racemic disparlure. Journal of Economic Entomology 86: 8 15-820. Kuppusamy, A. & S. Kannan (1 993) Life history of Eumeta cramerii (Westwood) (Lepidoptera: Psychidae) and its natural enemy Sinophorus psycheae Sonan (Hymenoptera: Ichneumonidae). Phytophaga 5: 109- 12 1.

Lagoy, P.K.& E.M. Banows (1989) Larval-sex and host species eEects on location of attachent sites of kt-instar bagworms, Thyridopteeryx ephemeraeformis (Lepidoptera: Psychidae). Proceedings of the Entomologicd Society of Washington 9 1 : 468-472.

Lance. D. & P. Barbosa (198 1) Host tree influences on the dispersal of first instar gypsy moths. L-vmczntrir dispar (L.). Ecological Entornology 6: 4 1 1-4 1 6.

Lasota, LA. & L.T. Kok (1 986) Larval density effects on pupal size and weight of the irnported cabbageworm (Lepidoptera: ), and theu suitability as hosts for Pteromalus puparum (Hymenoptera: Pteromalidae). Environmental Entomology 15: 123401236,

Lawrence, W.S.(1 986) Male choice and cornpetition in Tetraopes tetraophralmus: effects of local sex ratio variation. Behavioural Ecology and Sociobiology 18: 289-296.

Lawton, J.H. (1983) Plant architecture and the diversity of phytophageous insects. Annual Review of Entomology 28: 23-39.

Leather, S.R. (1984) Factors affecthg pupal survival and eclosion in the pine beauty rnoth, Panoiisflamrnea (D&S). Oecologia 63: 75-79.

Leather, S.R. (1 9 88) Size, reproductive potential and fecundity in insects: things aren 't as simple as they seem. Oikos 51: 386-389.

Leonard, D.E. (1970) intrinsic factors causing qualitative changes in populations of Portheria dispar (Lepidoptera: Lymantriidae). Canadian Entomologist 102: 239-249.

Leonhardt, B.A., J.W. Neal Jr, LA. Klun, M. Schwarz & LR. Plimmer (1983) An unusual lepidopteran sex pheromone system in the bagwonn moth. Science 21 9: 3 14-3 16.

Lim, C.E., L.B. Bjostad, J.W. Du & W.L.Roelofs (1984) Redundancy in a chernical signal: behavioral responses of male Trichopiusia ni to a six-component sex pheromone blend. burnal of Chemicai Ecology 10: 1635-1658.

Loeb, J.R., LW.Neal Jr & LA. Klun (1989) Modified thoracic epithelium of the bagworm (Lepidoptera: Psychidae): site of pheromone production in adult fernales. A~alsof the Entomological Society of America 82: 2 L 5-2 19.

Lornnicky, A. (1988) Population Ecology of Individuals. Princeton University Press, Princeton. Lowrnan, M.D.(1985) Temporal and spatial variability in insect grazing of the canopies of five Austraiian minforest tree species. Australian Journal of Ecology 10: 7-24.

Luck, R.F. & D.L.Dahlsten (1980) Within and between tree variation of Live and parasitized Douglas-fir tussock moth, Orgyia pseudotsugata (Lepidopten: Lymantriidae), cocoons on white fir in Central Cdifomia and its implications for sampling. Canadian Entomologist 112: 23 1-23 8.

Magistretti, O., A.R. Mallea, G.S. Mhcola, J.G. Garcia Saez, J.H. Suarez & LA. Bahamondes (197 1) Estudios sobre partenogenesis en "bicho del cesto" cornun. Revista de la Facultad de Ciencias Agrnrias Universidad Nacional de Cuyo 17: 1-5.

Malicky, H. (1968) Untersuchungen über die Tendenz zur parthenogenetischen Fortpflannuig bei Solenobia manni 2. (Lepidopteni: Psychidae). Revue Suisse de Zoologie 75: 999- 1004.

Mariath, H.A. (1984) Factors affecting the dispersive behaviour of larvae of an Australian geometrid moth. Entomologia Erperimentaiis et Applicuta 35: 159- 16%

Mnrkow, T.A. & P.F. Ankney (1984) Drosophifa males contribute to oogenesis in a multiple mating species. Science 224: 302-303.

Marshall, L. (1982) Male courtship penistence in Colias philodice and C. ruwhemc (Lepidoptera: Pieridae). Journal of the Kansas Entomologicd Society 5 5 :729-736.

Marshall, L.D. (1988) Intraspecific variation in reproductive effort by fernale Parapediasiu teterrella (Lepidoptera: Pynlidae) and its relation to body size. Canadian Journal of Zoology 68: 44-48.

Mason, R.R., R.C. Beckwith & H.G. Paul (1977) Fecundity reduction during collapse of a Douglas-fir tussock moth outbreak in northeast Oregon. Experimental Entomology 6: 623-626.

Mason, R.R. & M.L.McManus (1981) Larval dispersal of the gypsy rnoth. In C.C. Doane & ML. McManus (eds). The Gypsy Moth: Research toward Integrated Management. Technical Bulletin L 584, United States Department of Agriculture, Washington DC. pp. 16 1-202.

McLaughlin, I.R., R.E. Doolittle, C.R. Gentry, E.R. Mitchell & J.H.Tumiinson (1976) Response to pheromone ûaps and disruption of pheromone communkation in the lesser peachtree borer and the peachtree borer (Lepidoptera: ). Journal of Chernical Ecology 2: 73-8 1.

McManus, M.L. & C.I. Mason (1983) Determination ofthe senling velocity and its significance to larval dispersal of the gypsy moth (Lepidoptera: Lymmtriidae). Environmental Entomology 12: 270-272. McNaily, P.S. & M.M. Bames (1981) Effects of codling moth pheromone tnp placement, orientation and density on trap catches. Environmental Entomology 10: 22-26.

Mikkola, K. (1987) Protogyny during an outbreak in mendica (Lepidoptera: Arctiidae). Annales Entomologici Fennici 53 : 30-32.

Miller, F.D.& E.R. Hart (1987) Overwintenng survivorship of pupae of the mimosa webworm, Homadmla anisocentra (Lepidoptera: ), in an ur ban landscape. Ecological Entomology 12: 4 1-50.

Miller, J.R. & W.L. Roelo fs (1980) Individual variation in sex pheromone component ratios in two populations of the redbanded leafioller moth, Argyrotuenia velurininanri. Environmental Entomology 9: 3 59-363.

Mishra, S.C. (1978) Observations on the biology of bagworrn - CIania cramerii Westwood (Lepidoptera: Psychidae). The Indian Forester 104: 135- 14 1.

Mistrot Pope, M., L.K. Gaston & T.C. Baker (1982) Composition, quantification, and periodicity of sex pheromone gland volatiles from individual Heliothis virescens temales. Journal of Chernical Ecology 8: 1043- 1055.

Mitchell, K.G. (1979) Dispersal of early instars of the Douglas-fir tussock moth. Annals of Entomological Society of Arnerica 72: 29 1-297.

Montgomery, M.E. (1 989) Relationships between foliar chemistry and susceptibility to Lymrintria dispar). In W.E. Wallner & K.A. McManus (eds). Lymantriidae: a cornparison of features of new and old world tussock moths (general technical report NE- 123). US Department of Agriculture. Wassington, pp. 339-3 50.

Moran, V.C., B.H. Gunn & G.H. Walter (1982) Wind dispersal and settling of first-instar crawlers of the cochined insect Dactylopius austrinus (Homoptera: Coccoidea: Dactylopiidae). Ecological Entomology 7: 409-4 19.

Morden, R.D. & G.P. Waldbauer (1971) The developmental rates of Thyridopteryx ephemeraefomis fiom four latitudes and notes on its biology (Lepidoptera: Psychidae). Entomological News 82: 15 1- 156.

Morris, R.F. & C.A. Miller (1954) The development of life tables for the spnice budworm. Canadian Journal of Zoology 32: 283-30 1.

Nash, D.R.,D.J.L. Agassiz, H.C.J. Godfray & J.H. Lawton (1995) The pattern of spread of invading species: two leaf-mining moths colonizing Great Britain. Journal of Ecology 64: 225433. Naugler, C.T. & S.M. Leech (1 994) Fluctuating asymmetry and suMval ability in the forest tent moth Malacosorna disstria: implications for Pest management. Entomologia Experimentalis et Appiicata 70: 295-298.

Neal, LW. (1986) Saiient vestiture and morphologicai characters of the pharate female bagwom, Thyridopteryx ephemeraeformis (Lepidoptera: Psychidae). Annals of the Entomological Society of Arnerica 79: 8 14-820.

Neal Jr, LW.,M.J. Raupp & L.W.Douglass (1987) Temperature-dependent mode1 for predicting larval emergence of the bagworm, Thyridopteryx ephemeruejormis (Haworth) (Lepidoptera: Psychidae). Environmental Entomology 16: 1 111 - 1 1 11.

Neal Ir, J. W. & F.S. Santarnour Ir (1 990). Biotic indicators of host preference by the bagworm (Lepidoptera: Psychidae). Journal of Economic Entomology 83: 2393-2397.

Newman. D. (1980) Susceptibilidad de algunos arboles al ataque defoliador del gusano canasta (Oiketicus kirbyi Guilding) en el valle del Cauca. Investigacion Forestal, Carton de Colombia S.A. Informe de Investigacion 63. 10 pp.

Nicholson, A.J. (1 933) The balance of animal populations. Journal of Animal Ecology 2: 132-1 78.

Nicholson, A.J. (1954) An outline of the dynamics of animal populations. Australian Journal of Zoology 2: 9-65.

Oberhauser, K.S. (1997) Fecundity, lifespan and egg mass in buneflies: effects of male- derived nutrients and female size. Functional Ecology 1 1 : 166-1 75.

Ohgushi, T. (1991) Lifethe fitness and evolution of reproductive pattern in the herbivorous lady beetle. Ecology 72: 2 1 10-2122.

Ohgushi, T. (1995) Adaptive behavior produces stability in herbivorous lady beetle populations. In N. Cappuccino & P.W.Price (eds). Population dynarnics: New Approaches and Synthesis. Academic Press, San Diego. pp. 303-3 19.

Ohgushi, T. (1996) Consequences of adult size for survival and reproductive performance in a herbivorous ladybird beetle. Ecological Entomology 2 1: 47-55.

Ortiz, R.A. & 0. Femhdez (1994) Cultivo de la palma aceitera. Editorial Universidad Estadal a Distancia, San ksé, Costa Rica. 19 1 pp.

Ôtake, A. & M. Oyama (1973) Influence of sex-ratio and density on the mating success of Spodopteru litura F. (Lepidoptera: ). Applied Entomology and Zoology 8: 246-247. Palmer, J.O. (1985) Life-history consequences of body-size variation in the milkweek Ieaf beetle, Labidomera clivicollis (Coleoptera: Cluysomelidae). Annals of the Entomologicd Society of America 78: 603-608.

Parry, D. (1995) Larval and pupal parasitism of the forest tent caterpillar, Malacosorna disstria Hübner (Lepidoptera: ), in Alberta, Canada. Canadian Entomologist 127: 877-893.

Percy-Cunningham, J.E. & LA. MacDonald (1 987) Biology and ultrastructure of sex pheromone-producing glands. In G.D. Prestwich & G.I. B lomquist (eds), Pheromone biochemistry. Academic Press. Orlando, pp. 27-75.

Peterson, M.A. (1997) Host plant phenology and butteffly dispersal: causes and consequences of uphill movement. Ecology 78: 167- 1 80.

Phelan, P.L. (1992) Evolution of sex pheromones and the role of asymmetric tracking. In B.D. Roitberg & M.B.Isman (eds), Insect chernical ecology: an evolutionary approach. Chapman & Hall, New York. pp. 265-3 14.

Phelan. P.L. & T.C.Baker. (1986) Male size related courtship success and intersexual selection in the tobacco moth, Ephestiu elutella. Experientia 42: 129 1- 1293.

Poirier, LM.& J.H.Borden (1992) Some efTects of population density on the life history of the obliquebûnded leafroller Choristuneura rosaceana Harris (Lep., Tortricidae). loumai of Applied Entomology 1 13 : 307-3 14.

Ponce, O., H. Pelhez & J.L. de la Cruz (1979) Estudio biologico del gusano canasta Oiketims kirbyi Lands Guilding (Lepidoptera: Psychidae) en platano y reconocimiento de sus principales parasitoides. Acta Agronomica 29: 4 1-46.

Price, P. W. (1 994) Phylogenetic constraints, adaptive syndromes, and ernergent properties: hmindividuals to population dynamics. Researches in Population Ecology 36: 3- 14.

Price, P.W., N. Cobb, T.P. Craig, G. Wilson Fernandes, J.K. [tami, S. Mopper & R.W. Preszler (1990) Insect herbivore population dynamics on trees and shbs: new approaches relevant to latent and eruptive species and life table development. In E.A. Bemays (ed), Insect-plant interactions, vol. II, CRC Press, Boca Raton. pp. 1-38.

Price, P.W. & M.D. Hunter (1995) Novelty anci synthesis in the development of population dynamics. In N. Cappuccino & P.W. Price (eds). Population dynamics: New Approaches and Synthesis. Academic Press, San Diego. pp. 389-412.

Raina, A.K. (1984) Male-induced temination of sex pheromone production and receptivity in mated females of Heliothis zea. huma1 of Insect Physiology 35: 821-826. Ramachandran, R (1987) Infiuence ofhost-plants on the wind dispersal and the survival of an Australian geometrid caterpillar. Entomologia Experimentalis et Applicata 44: 289-294.

Reavey, D. & J.H.Lawton (199 1) Larval contributions to fitness in leaf-eating insects. In W.J. Bailey & J. Ridsdill-Smith (eds), Reproductive behaviour of insects: individuals and populations. Chapman & Hall, New York. pp. 293-329.

Reavey, D. (1993) Why body size matten to caterpillars. In N.E. Stamp & T.M.Casey (eds) Caterpillars: Ecological and Evolutionary Constrahts on Foraging. Chapmm and Hall, New York. pp. 248-279.

Renwick, J.A.A. & F.S. Chew (1994) Oviposition behaviour in Lepidoptera. hnuai Review of Entomology 39: 377-400.

Rhainds, M., C.M. Chinchilla & G. Castnllo (1 993) Desarollo de un metodo de muestreo para las de Opsiphunes cassim Felder en palma aceitera. Manejo Integrado de Plagas 30: 15-1 8.

Rhninds, M., G. Gries, J. Li, R. Gries, K.N. Slessor, C.M.Chinchilla & A.C. Oehlschlager (1 994). Chiral esters: sex pheromone of the bagwonn, Oiketicus kirbyi (Lepidoptera: Psychidae). Journal of Chernical Ecology 12: 3083-3096.

Rhainds, M., G. Gries, & C. Chinchilla (1 996) Development of a sarnpling rnethod for tint instar Oiketicus kirbyi (Lepidoptera: Psychidae) in oil palm plantations. Journal of Economic Entomology 89: 396-40 1.

Rhainds, M., G. Gnes & A. Saleh (1997) Sex-specific habitat utilization by bagwoms (Lepidoptera: Psychidae). Canadian Entomologist 129: 199-200.

Richerson, J.V. & E.A. Cameron (1974) Differences in pheromone release and sexual behavior between laboratory-reared and wild gypsy moth adults. Environmental Entomology 3 : 475-48 1.

Roff, D.A. (1986) The evoiution of wing dimorphism in insects. Evolution 40: 1009- 1020.

Roff, DA. (1990) The evolution of flightlessness in insects. Ecological Monographs 60: 389- 421 .

Rogers, C.E. & O.G. Marti (1994) Effects of age at first mating on the reproductive potential of the fa11 armyworm (Lepidoptera: Noctuidae). Environmental Entomology 23 :322-325.

Rogers, C.E. & O.G. Marti (1997) Once-mated beet arrnywomi (Lepidoptera: Noctuidae): effects of age at mating on fecundity, fertility, and longevity. Environmental Entomology 26: 585-590. Rossiter, M.C. (1987) Use of a secondary host by non-outbreak populations of the gypsy moth. Ecology 68: 857-868.

Rossiter, M.C. (1992) The impact of resource variation on population quality in herbivorous insects: a critical aspect of population dynamics. In M.D. Hunter, T. Ohgushi & P.W. Price (eds). Effects of Resource Distribution on Animal-Plant Interactions. Academic Press, San Diego. pp. 13-42.

Rossiter, M.C.(1 995) Impact of life-history evolution on population dynamics: predicting the presence of matemal effects. In N. Cappuccino & P.W. Pnce (eds). Population dynamics: New Approaches and Synthesis. Academic Press, San Diego. pp. 25 1-375.

Rotheray, G.E. & P. Borbosa (1 984) Host related factors affecting oviposition behavior in Brachymeria intermedia. Entomologia Experimentdis et Applicata 35: 14 1-1 45.

Rowe II, W.J.& D.A. Potter (1996) Vertical stratification of feeding by Japanese beetles within Iinden tree canopies: selective foraging or height per se?. Oecologia 108: 459- 466.

Royer, L. & LN. McNeil(1993) Male investment in the european corn borer. Ostrinia nubilalis (Lepidoptera: Pyrnlidae): impact on femde longevity and reproductive performance. Functional Ecology 7: 209-2 15.

Ruohomaki, K. (1992) Wing size variation in Epirrita auiumnata (Lep., Geomeuidae) in relation to larval density. Oikos 63: 260-266.

Ruszczyk, A. (1 996) Spatial patterns in pupai mortality in urban pdm caterpillars. Oecologia 107: 356-363.

Rutowski, R.L. (1 982a) Epigamic slection by males as evidenced by courtship partner preferences in the checkered white buaerfly (Pieris prolodice). Animal Behaviour 30: 108-1 12.

Rutowski, R.L. (1 982b) Mate choice and lepidopteran mating behavior. Flonda Entomologist 65: 72-82.

SAS Institute (1988) SATISTAITM. User's guide, release 6.03. SAS Institute, Cary,NC.

Sattler, K. (199 1) A review of wing reduction in Lepidoptera Bulletin of British Museum NadHistory (Entomology) 60: 243-288.

Shapiro, A.M. (1 970) The role of sexual behavior in density-related dispersal of pierid butterflies. American Naniralist 104: 367-372. Sharov, A.A., A.M. Liebhold.& F.W. Ravlin (1995) Prediction of gypsy moth (Lepidoptera: Lymantriidae) mating success fiom pheromone trap counts. Environmental Entomology 24: 1239-1244.

Shaw, M.R. (1 993) Protogyny in Lasiocampu quercus callunae Palmer (Lepidoptera: Lasiocampidae), with notes on some parasitoids. Entornologist's Record 105: 28 1- 282.

Shelly, T.E. &W.J. Bailey (1992) Experimental manipulation of mate choice by male katydids: The effect of female encounter rate. Behavioural Ecology and Sociobiology 30: 277-282.

Sheppard, R.F. (1975) The bagworm, Thyridopteryx ephemeraefomis: a mode1 system for studying the pnnciples of insect population dynamics. Bulletin of the Entomological Society of America 2 1: 153-1 56.

Shorey, H.H.& L.K. Gaston (1965) Sex pheromones of noctuid rnoths. VIL Quantitative aspects of the production and release of pheromones by fernales of Trichoplusiu ni (Lepidoptera: Noctuidae). Annals of the Entomological Society of Amenca 58: 604- 608.

Simmons, L.W.& W.$.Bailey (1990) Resome influenced sex roles of Zaprochiline tettigoniids (Orthoptera: Tettigoniidae). Evolution 44: 1853- 1868.

Sims, S.R. (1979) Aspects of mating frequency and reproductive maturity in Papilio zelicaon, American Midland Naniralist 102: 36-50.

Smith, M.P. & E. Banows (199 1) Effects of larval case size and host plant species on case intemal temperature in the bagworm, Thyridopteryx ephemeraeformis (Haworth) (Lepidoptera: Psychidae). Proceedings of the Entomologicai Society of Washington 93: 834-838.

Solomon, J.D. & W.W. Neel (1972) Emergence behavior and rythms in the carpenterworm moth, Prionoxystus robiniae. Annals of the Entornological Society of America 65: 1296-1299.

Sopow, S.L. & D.T. Quiring (1998) Body size of spnice-galling adelgids is positively related to realized fecundity in nature. Ecological Entomology 23: 476-479.

Stephens, C.S. (1962) Oiketiczu kirbyi (Lepidopteni: Psychidae) a pest of bananas in Costa Rica Ioumai of Economic Entomology 55: 38 1-386.

Stiling, P. (1988) Density-dependent processes and key factors in Uisect populations. Journal of Animal Ecology 57: 581-593. Stockhoff, B.A. (199 1) Starvation resistance of gypsy moth, Lymantria dispar (L.) (Lepidoptera: Lymantriidae): tradeoffs arnong growth, body size and survival. Oecologia 88: 422-429.

Sd,Y. (1978) Adult longevity and reproductive potential of the srnall cabbage white, Pieris rapae cmcivora (Boisduval) (Lepidoptera: Pieridae). Applied Entomology and Zoology 13: 3 12-3 13.

Svenaon, B.G. & E. Petersson (1987) Sex-role reversed courtship behaviour, sexual dimorphism and nuptial gifts in the dance fly, Empis borealis (L.) Annals of Zoology Femici 34: 323-334.

Syed, R.A. (1 978) Bionomics of the three important species of bagworms on oil palm. Mdaysian Agriculture Journal 5 1 :392-398.

Syed, R.A. & A. Sdeh (1 99 1) Management of insect pests of oil palm in PTPP London Sumatra Indonesia plantations in Sumatera, Indonesia. PORIM International Palm Oil Conference - Agriculture. Kuala Lumpur, Malaysia. pp. 45 1-457.

Syed, R.A. & S. Shah (1976) Some important aspects of insect pest management in oil palm estates in Sabah, Malaysia. PORIM International Palrn Oil Conference - Agriculture. Kuala Lumpur, Malaysia. pp. 577-590.

Tammani, T., P. Kaitaniemi & K. Ruohornaki (1996a) Realized fecundity in Epirrira uutumnata (Lepidoptera: Geometridae): relation to body size and consequences to population dynamics. Oikos 77: 407-4 16.

Tamrnani, T., K. Ruohomtki & K. Saikkonen (1996b) Components of male fitness in relation to body size in Epirrifa autumnata (Lepidoptera, Geometridae). Ecological Entomology 21: 185-192.

Taylor, P.S. & E.J. Shields (1990) Development of the amyworm (Lepidoptera: Noctuidae) under fluctuating daily temperature regimes. Environmental Entomology 19: 1427- 1431.

Terry, L., J.R. Bradley & J.W. Van Duyn (1989) Establishment of early instar Heliothis zea on soybeans. Entomologia Experimentalis et Applicata 51 :233-240.

Thangavelu, S. & K. Gunasekaran (1982) Aspects of mating behaviour in bagworm moth cramerii (Westwood) (Psychidae: Lepidoptera). Entomon 7: 457-462.

Thangavelu, S. & M.H. Ravinciranath (1 985) Morphology and life-history of the bag-worm moth CZania cramerii (Westwood). Journal of Natural History 19: 1- 19.

Thomhill, R & I. Alcock (1983) The Evolution of insect Mating Systems. Harvard University Press, Cambridge, Massachusetts. 547 pp. Torres-Vila, L.M., I. Stocke1 & R. Roehnch (1995) Le potentiel reproducteur et ses variables biotiques associées chez le mâle de I'Eudémis de la vigne Lobesia botranu. Entomologia Experimentalis et Applicata 77: 105- 1 19.

Torres-Vila, L.M.,J. Stockel, R. Roehnch & M.C. Rodriguex-Molina (1997) The relation between dispenal and survival of Lobciria bofrana larvae and their density in vine inflorescences. Entornologia Experimentalis et Applicata 84: 10% 1 14.

Turner, P.D.(198 1) Oil Palm Diseases and Disorders. Oxford University Press, Oxford. pp. 234-23 5. van der Linde, V.R.J. (197 1) Der Schwebeflug der jungen Rnupen des Schwammspinners (Lymantria dispar L.) und der EinfluP der Nahningspfianze auf dm Entstehen desselben. Zeitschnfi RLr Angewandte Entomologie 67: 3 16-324.

Vdey, G.C. & G.R. Gradwell(1960) Key factors in population studies. Journal of Animal Ecology 29: 399-401.

Vickea, R.A. (1997) Effect of delayed mating on oviposition pattern, fecundity and fertility in codling rnoth, Cydia pornoneila (L.) (Lepidoptera: Tortricidae). Austnlian Journal of Entomology 36: 179-1 82.

Villanueva, A.G. & M. Avila (1987) El gusano canasta Oiketicus kirbyi, Guild. Fedepalma, Colombia. 28 pp.

Villanueva, A.G. & E. Granda Paz (1 986) Experiencias con el Oiketicus kirbyi, Guild, en palmeras de la Costa S.A. IV Mesa Redonda Latinoamencana sobre pdma aceitera. Valledupar, Departemento del Cdsar. Colombia, pp. 94-98.

Wall, R. & M. Begon (1987) Population density, phenotype and reproductive output in the grasshopper, Chorhippw bnrnneus. Ecological Entomology 12: 33 1-3 39.

Wagner, D. (1995) Pupation site choice of a North Amencan lycaenid butterfly: the benefits of entering ant nests. Ecological Entomology 20: 384-392.

Wang, G., M.D. Greenfield & T.E. Shelly (1990) Inter-male cornpetition for high-quality host-plants: the evolution of protandry in a territorial grasshopper. Behavionl Ecology and Sociobiology 27: 19 1- 198.

Webster, R.P. & R.T. Cardé (1984) The effects of mating, exogenous juvenile hormone and a juvenile hormone analogue on pheromone titre, calling and oviposition in the ornnivorous leafioller moth (Plafynotastultana). Journal of Insect P hy siology 30: 113-1 18.

Wedell, N. (1992) Protandry and mate assessrnent in the wartbiter Dectius verrucivow (Orthoptera: Tettigonüdae). Behavioral Ecology & Sociobiology 3 1: 30 1-308. Weiss, M.J., K.P. Seevers & 2.9.Mayo (1985) Influence of western corn rootworm larval densities and damage on corn rootwonn survival, developmental tirne, size and sex ratio (Coleoptera: Chrysomelidae). Journal of the Kansas Entomological Society 58: 3 97-402.

Weiss, S.B., R.R. White, D.D. Murphy & P.R. Ehrlich (1987) Growth and dispersal of larvae of the checkerspot butteffly Euphydryas editha. Oikos 50: 161-166.

Weisslhg, T.J. & A.L. Knight (1995) Vertical distribution of codling moth aduits in pheromone-treated and unwated plots. Entomologia Experimentaiis et Applicata 77: 271 -275.

Wellington, W.G.(1 957) Individual differences as a factor in population dynamics: the development of a problem. Canadian Journal of Zoology 35: 293-323.

Wellington, W.G. (1965) Some natural influences on progeny quality in the western tent caterpillar MaIacosoma pluviale (Dyar). Canadian Entomologist 97: 1- 14.

Weseloh, R.M.(1997) Evidence for limited dispersai of larval gypsy moth, Lymrintriu dispar L. (Lepidoptera: Lymantriidae). Canadian Entomologist 129: 355-361.

Wheeler Jr, A.G. & E.R. Hoebecke (1 988) Apteronu helix (Lepidoptera: Psychidae), a palearctic bagwom moth in North Amerka: new distribution records. seasonal history, and host plants. Proceedings of the Entomological Society of Washington 90: 20-27.

White, R.R. (1 986) Pupal mortality in the bay checkerspot butteffly (Lepidoptera: ). Journal of Research on the Lepidoptera 25: 52-62.

Wickman, B.E. & R.C. Beckwith (1978) Life history and habits. In M.H. Broukes, R.W. Stark & R.W.Campbell. The Douglas-fir tussock moth: a synthesis. Technical Bulletin 1585, United States Department of Agriculture. Washington, DC: 30-3 7.

Wickman, P., C. Garcia-Barros & C. Rappe-George (1995) The location of landmark leks in the small heath butterfly, Coenonympha pumphilus: evidence against the hot-spot model. Behavioral Ecology 6: 39-45.

Wickman, P. & P. Jansson (1997) An estimate of female mate searching costs in the lekking butteffly Coenonympha pamphiius. Behavioural Ecology and Sociobiology 40: 321- 328.

Wiklund, C. & T. Fagentrom (1977) Why do males ernerge before females? A hypothesis to explain the incidence of protandry in buttedies. Oecologia 3 1: 153-1 58. Wiklund, C. & A. Kaitala (1995) Sexual selection for large male size in a polyandrous butterfly: the effect of body size on male venus femafe reproductive success in Pieris napi. Behavionl Ecology 6: 6-1 3.

Williams, D.W.,RW. Fuester, W.W.Metterhouse, R.J. Balaam, R.H. Bullock, R.J. Chianese & R.C. Reardon (1 990) Density, size, and mortality of egg masses in New Jersey populations of the gypsy moth (Lepidoptera: Lymantriidae). Environmental Entomology 19: 943-948.

Wing, S.R. (199 1) Timing of Photinus colhsfranr reproductive activity: Finding a mate in time (Coleopten: Lynpyridae). Coleopterist Bulletin 15: 57-71.

Witzgall, P. & B. Frérot (1989) Pheromone emission by individual females of carnation torûix, Cacoecimorphapronubana. Journal of Chernical Ecology 15: 707-7 17.

Wood, B.J. (1 968) Pests of oil pdms in Malaysia and their control. Incorporated Society of Planters, Kuala Lumpur, 204 pp.

Wood, BJ., R.H.V.Corley & K.H. Goh (1973) Studies on the effect of pest damage on oil paim yield. In R.L. Wastie Br D.A. Earp (eds) Advances in oil palm cultivation. Incorporated Society of Plantes, Kuala Lumpur. Malaysia. pp. 360-374.

Wood, B.J., S.S. Liau & J.C.X. Knecht (1974) Trunk injection of systemic against the bagworm, Metisa piana (Lepidoptera: Psychidae) on oil paim. Oléagineux 29: 499-504.

Zhao, B. (1 98 1) Studies on the glands and release mechanism of the female sex pheromone in Cfaniavariegata Snell. Journal of Nanjing Technical College Forestry Products 4: 97-1 02.

Ziegler. J.R. (1 978) Dispenal and reproduction in Tribolium: the influence of inital density . Environmental Entomology 7: 149- 156.

Zonneveld, C. & J.A.J. Metz (199 1) Models on butterfly protandry: virgin females are at risk to die. Theoretical Population Biology 40: 308-321.

Zurlini, G. 8r A. Robinson (1979) Laddispersal as related to density in wild and laboratory strains of Delia (=Hyiemiu) antiqua. Netherlmd Journal of Zoology 29: 40241 7.