ProQuest Dissertations - The University of Arizona Campus Repository

INFORMATION TO USERS
This manuscript has been reproduced from the microfiim master. UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submiHed. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough. substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and continuing
from left to right in equal sections with small overlaps.
Ptwtographs included in the original manuscript have been reproduced
xerographicaliy in this copy. Higher quality 6' x 9" blade and white
photographic prints are available for any ptiotographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
ProQuest Information and Learning
300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA
800-521-0600

HOST UTILIZATION PATTERNS OF THE WALNUT FLY, RHAGOLETIS
JUGLANDIS, AND THEIR IMPLICATIONS FOR FEMALE AND
OFFSPRING FITNESS
by
Cesar Roberto Nufio
Copyright © Cesar Roberto Nufio 2002
A Dissertation Submitted to the Faculty of the
INTERDISCIPLINARY PROGRAM DM INSECT SCIENCES
In Partial Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2002

UMI Number: 3053915
Copyright 2002 by
Nufio, Cesar Roberto
All rights reserved.
UMI
UMI Microform 3053915
Copyright 2002 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346

THE UNIVERSITY OF ARIZONA ®
GRADUATE COLLEGE
As members of the Final Examination Committee, we certify that we have
read the dissertation prepared by CESAR ROBERTO NUFIO
entitled HOST UTILIZATION PITTERNS BY THE WALNUT
RHAGOLETIS JUGLANDIS ANB THIER IMPLICATIONS
FOR OFFSPRING FITNESS
and recommend that it be accepted as fulfilling the dissertation
requirement for the Degree of
DOCTOR OF PHILOSOPHY
Date
Date
Date
Date
Date
Final approval and acceptance of this dissertation is contingent upon
the candidate's submission of the final copy of the dissertation to the
Graduate College.
I hereby certify that I have read this dissertation prepared under my
direction and recommend that it be accepted as fulfilling the dissertation
requirement.
rtation Director \ ] Date
OB. Daniel R. Papa
2

STATEMENT BY AUTHOR
This dissertation has been submitted in partial fiilfillment of requirements for an
advanced degree at the University of Arizona and is deposited in the University Library
to be made available to borrower under rules of the Library.
Brief quotations from this dissertation are allowable without special permission,
provided that accurate acknowledgment of source is made. Requests for permission for
extended quotation from or reproduction of this manuscript in whole or in part may be
granted by the copyright holder.
SIGNED:
3

ACKNOWLEDGMENTS
This dissertation is the result of not just hours of study and focus, but also of
hours of insightful conversations, discovery and being with a great deal of wonderful
people. It is to those wonderful and important people that I owe a great deal to. I would
Hrst like to thank my advisor Dan Papaj. Dan is a phenomenal scientist and mentor
who has always impressed me by his ability to get to the center of any argument and to
design simple yet elegant experiments. I especially appreciate his wisdom, the many
hours he invested, his company in the field, his great dinners and his friendship. His
wisdom and training will give me great confidence in the years to come.
I also would like to thank Henar Alonso-Pimentel. I was very fortunate to have
overlapped with her while she was in the Papaj lab and I benefited greatly from this
opportunity. Henar was a great mentor who spent a great deal of time picking me up
and placing me on the right track. Henar was patient, always insightful and introduced
me to the importance of mechanisms and the physiology of behavior. I hope to also
carry with me a bit of her wisdom and will think of her when I help others.
I would also like to thank my conmiittee, which was composed of Judie
Bronstein Reg Chapman, Molly Hunter and Bob Smith. I learned a great deal from
their insights and novel perspectives. I am also thankful for their advice and the time
they invested. I would also like to thank the Interdisciplinary Program (IDP) in Insect
Sciences and the Department of Entomology for beings so supportive and always
placing the students first. In particular, I would like to thank Diana Wheeler for going
out of her way to make the IDP Insect Sciences a successful program, one that I will
always be proud to have been a apart of. Sharon Richards, our IDP secretary, was a
phenomenal person that made life so easy for all of us graduate students. Food was put
on the table, papers made it through barriers, numbers where changed in remote
computers and lost things were found because of Sharon.
Special thanks must go to the Entomology, IDP Insect Sciences and Ecology
graduate students. I felt very fortunate to have interacted and learned from such a great
cohort of people. In particular, 1 would like to thank Laurie Henneman, Apama Telang,
Jessa Netting, Karen Ober, Albert Owen and Matt Johnston for hours of entertainment,
support and encouragement. And a special thanks to Susto and Weevil, the two cats that
made me remember that I am not the most important person in the world.
The most important person in this process, however, has been my wife Dena
Smith. She is one of the most driven, talented, intelligent and inspiring persons that I
know and I am very fortunate to have found her. I love her dearly and know that much
of what I have accomplished would not have been possible without her. With her by my
side, I look forward to the future and believe that great things await us.
Financial support for this dissertation was provided by a National Science
Foundation predoctoral fellowship, the University of Arizona's Research Training
Grant for the Analysis of Biological Diversification, the Department of Entomology,
Sigma Xi, Center for Insect Sciences and the Graduate College of the University of
Arizona. Dan Papaj also provided funding and Sheridan Stone provided logistical
support at Garden Canyon, in Fort Huachuca, where this research was conducted.
4

DEDICATION
This work is dedicated to my mother, Maria Elena Arriaza-Nufio. She has always
believed in me, supported my pursuits and showed me that trying to be a good person
one of the most important things in life.

TABLE OF CONTENTS
L ABSTRACT 7
U. CHAPTER 1: INTRODUCTION 9
An explanation of the problem............................................................ 9
Explanation of thesis format.......................................................................... 12
m. CHAPTER 2; PRESENT STUDY 14
IV. REFERENCES 16
APPENDIX A. HOST MARKING BEHAVIOR IN PHYTOPHAGOUS
INSECTS AND PARASITOIDS 19
APPENDIX B. HOST UTILIZATION BY THE WALNUT FLY,
RHAGOLETIS JUGLANDIS TEPHRITIDAE) 43
APPENDIX C. REUSE OF LARVAL HOSTS BY THE WALNUT FLY,
RHAGOLETIS JUCLANDIS, AND ITS IMPACTS FOR FEMALE
AND OFFSPRING PERFORMANCE 54
APPENDIX D. AGGREGATIVE BEHAVIOR IS NOT EXPLAINED BY
AN ALLEE EFFECT IN THE WALNUT-INFESTING FLY,
RHAGOLETIS JUGLANDiS. 112
APPENDIX E. HOST MARKING BEHAVIOR AS A QUANTITATIVE
SIGNAL OF INFESTATION LEVELS IN HOST USE BY THE
WALNUT FLY, RHAGOLETIS JUGLANDIS. 158
6

ABSTRACT
Choosing where offspring will develop is especially important for insects whose
larval stages are restricted to a particular host resource. In such insects, maternal egg
laying decisions may not only involve choosing optimal hosts based on their intrinsic
qualities but also avoiding hosts occupied by conspeciHc brood. The ability to
discriminate between previously exploited and unexploited hosts is often mediated by
the use of a marking pheromone.
Despite engaging in what appears to be host-marking behavior, the walnut fly
Rhagoletis juglandis prefers to deposit clutches into previously exploited hosts. In this
dissertation, I quantiHed host reuse in R. juglandis and assessed its impacts on
offspring fitness. I also explored the role that marking pheromone plays in level of
reuse.
Host reuse by the walnut fly was conunon in the field, where trees were
synchronously infested over a 14 -17 day period. It was not unusual for individual fruit
to bear 40 - 80 eggs; given that females laid clutches of ca. 16 eggs, each oviposition
puncture probably contained ca. 1.6 clutches. The overall number of eggs deposited into
fruit was positively correlated with fruit volume. Field and laboratory experiments
showed that increases in larval densities within fruit reduced larval survival and pupal
weight, the latter being strongly correlated with the number of eggs a female produced
over her lifetime. The temporal staggering of clutches strongly and negatively impacted
survival of later clutches. The effect of spatial patterning of clutches on offspring
ntness depended on the number of clutches in the &uit: at higher densities, clutches
7

perfonned better when deposited into the same puncture than when distributed
uniformly over fruit.
The evidence taken together suggests that host reuse by the walnut fly, R.
juglandis, reduces per capita offspring fitness. Consistent with this inference was a
final set of observations on female host-marking behavior. In field-cage experiments,
fruit that were marked by females for longer durations were less acceptable to other
females. Moreover, the duration of time that a female marked a fruit was positively
correlated with the size of her clutch. These results indicate that, while females
commonly reuse fruit, they nevertheless signal the level of larval competition
associated with a fhiit and adjust allocation of eggs to fruit accordingly.
8

An explanation of tiie problem
CHAPTER 1
INTRODUCTION
The decisions that insects make about where their offspring will develop are
especially important for species with larval stages that are restricted to a particular
environment or host (Thompson 1983, Smith and Lessells 1985, Bemays and Chapman
1994, Mayhew 1997). Because insect larvae that develop within discrete hosts (e.g.
flower buds, seeds, small ftoiit, stems, or other insects) may not be able to acquire more
resources if their natal hosts become depleted, larval development can be affected very
strongly by level of competition for food resources. Researchers have often proposed
that in such insects, females should be under strong selection to choose hosts that are
optimal for larval development (Jaenike 1978, Thompson 1988, Mayhew 1997). These
optimal egg laying decisions may not only involve choosing optimal hosts based on
their intrinsic qualities but also by avoiding hosts previously occupied by conspecific
brood.
It is particularly important that females avoid previously exploited hosts because
brood competition can reduce offspring fitness by increasing the time required for
offspring to reach maturity, as well as by reducing brood survival, size and reproductive
potential (rev. Peters and Barbosa 1977, Fox et al. 1996, Blanckenhom 1998, Sweeney
and Quiring 1998, Allen 2001). This ability for insects to discriminate between
previously exploited and unexploited hosts is often mediated by the presence of a
9

Marking Pheromone (MP), which is placed on the host surface following oviposition by
previous females (Roitberg and Prokopy 1987, Nufio and Papaj 2001).
Frugivorous fruit flies in the family Tephritidae deposit clutches into developing
fruit, where larvae are constrained to feed and develop. These fruit flies are also thought
to possess visual and chemical mechanisms for assessing the quality of available hosts
and for discriminating between previously infested and uninfested hosts (Prokopy et al.
1976, Prokopy and Roitberg 1984, Henneman and Papaj 1999). In the genus
Rhagoletis, for example, females often assess and reject infested fruit on the basis of a
MP (Landolt and Averill 1999). MPs in these systems have been shown to minimize
larval competition by causing females to distribute their clutches more uniformly within
host patches then expected by chance alone (Prokopy 1981, Bauer 1986, Averill and
Prokopy 1989).
Walnut flies in the Rhagoletis suavis group reuse their hosts, a behavior that is
uncommon among other flies in the genus. While reuse of hosts that already bear
conspecific brood is conunonly associated with the lack of available hosts (Roitberg and
Mangel 1983, Papaj et al. 1989), walnut flies actually prefer to oviposit into infested
hosts early in the season when uninfested hosts are still available (Lalonde and Mangel
1994, per. obs.). After deposition of a clutch, females drag their ovipositors on the firuit
surface in a manner known to involve deposition of a MP in congeners. Yet despite
displaying this genus-typical marking behavior, female walnut flies re-infest and often
reuse the actual oviposition sites established by conspecifics (Papaj 1993, 1994).
This dissertation examines the dynamics of host fruit utilization by the walnut
fly R. juglandis. In particular, the following studies examine the context in which host
10

euse occurs in the field, the impacts of host reuse on offspring fitness, and the role of
MP in this system. We found that reuse of walnut fruit is conmion in the field and that
increases in larval densities are associated with reduced larval survival and weight at
pupation. In a laboratory experiment, pupal weight was positively correlated with adult
female size and lifetime fecundity. In this system, reuse of larval hosts negatively
impacts offspring performance. We argue that reuse of previously exploited larval
hosts by the walnut fly, R. juglandis, reflects direct benefits to females that are traded
off against costs in terms of offspring fitness. Because female fitness is a product not
only of offspring quality but also of total number of offspring produced, female walnut
flies may be optimizing their fitness by producing many less fecund offspring.
Consistent with evidence that host reuse reduces per capita offspring fitness was a final
set of observations on the putative host-marking behavior in R. juglandis. In field cage
experiments, the duration of time that a female spent host marking was found to be
positively correlated with the size of her clutch. Additionally, increases in the
aggregate duration of host-marking on a fruit increased the degree to which that ^it
was rejected. In short, despite engaging actively in reuse of hosts, females signaled
level of larval competition and adjusted their level of reuse accordingly.
11

Explanation of dissertation format
The research included in this dissertation investigated the dynamics of host
reuse by the walnut fly, R. juglandis, and its impacts on offspring fitness through a
variety of filed and laboratory experiments. All five appendices represent work that I
conducted and papers that I produced. Dan Papaj served an advisory role throughout
and as such is a coauthor on all papers. Henar Alonso-Pimentel provided technical
support and played on advisory role on Appendix B.
Appendix A, "Host marking Behavior in phytophagous insects and parasitoids"
reviews the marking pheromone literature to provide a basic framework about the
signals insects use to discriminate between exploited and unexploited host and in turn to
avoid previously exploited hosts.
Appendix B, "Host utilization by the walnut fly, Rhagoletis juglandis (Diptera:
Tephritidae)" examined the ecological context under which reuse occurs in the field.
We were particularly interested in understanding how fruit characteristics (such as size
and penetrability) and ecological factors (such as fruit availability over time) influence
the degree to which hosts are reused.
Appendix C," Reuse of larval hosts by the walnut fly, Rhagoletis juglandis and
its impacts for female and offspring performance" examined how the degree to which
hosts are exploited in the Held impacts offspring survival and weight at pupation and
how in turn weight at pupation effects the number of eggs produced by a female over a
lifetime. In particular, we examined how female host reuse patterns directly impact
offspring performance.
12

Appendix D, "Aggregative behavior is not explained by an AUee effect in the
walnut-infesting fly, Rhagoletis juglandis" examined whether different components of
fruit reuse strategies might be associated with larval benefits. In this study we examined
how factors such as increases in larval density, temporal staggering of clutches into a
fhiit and the spatial arrangement of clutches along a fruit might differentially affect
offspring fitness characteristics. These factors were confounded in the previous field
study.
Appendix E, "Host marking behavior as a quantitative signal of infestation
levels in host use by the walnut fly, Rhagoletis juglandis" examined the role marking
pheromone might play in the distribution of clutches into fruit by female walnut flies.
Through a series of lab and field cage experiments we examined the amount of time
females spent marking fruit following egg laying and how ^it marked to different
degrees impact the host acceptance patterns by gravid females.
13

CHAPTER 2
PRESENT STUDY
The methods, results and conclusions of this study are presented in the papers
appended in this dissertation. The following is a summary of the most important
findings of these papers.
Appendix A. Oviposition behavior in phytophagous insects and entomophagous
parasitoids is often modified by the presence of conspecific brood (eggs and larvae).
Often, females avoid laying eggs on or in hosts bearing brood, a behavior that acts to
reduce the level of competition suffered by their offspring. Avoidance of occupied hosts
is typically mediated by cues and/or signals associated with brood. In this article, we
review the role of maridng pheromones as signals of brood presence in both
phytophagous and entomophagous insects.
Appendix B. This field study examined the pattern of host utilization by R.
juglandis and bow fruit variables such as volume and penetrability affect the degree that
hosts are reused. Fruit on trees were synchronously infested and within two to two and a
half weeks all fhiit on these trees were infested. Walnut hosts were commonly multiply
infested and reuse of hosts occurred in as few as 1-2 days after first infestation. Fruit
volume was positively correlated with both the number of punctures on hosts and the
infestation levels within hosts. Penetrability itself did not explain either which fruit
were preferentially utilized throughout the season or the infestation levels within fruit.
Appendix C. The oviposition-preference-offspring-performance hypothesis
states that female insects should prefer to deposit clutches on or in hosts that maximize
14

offspring performance. Counter to the preference-performance hypothesis, we found
that as the reuse of walnut fruit increased larval densities within fruit, these increases
were associated with reduced larval survival and weight at pupation. In a laboratory
experiment, pupal weight was positively correlated with adult female size and lifetime
fecundity. In this system, oviposition preference is therefore negatively, not positively,
correlated with offspring performance. We argue that because female Htness is a
product not only a function of offspring quality but also of total number of offspring
produced, female walnut flies may be optimizing their Hmess by producing many less
fecund offspring.
Appendix D. In this study, we examined whether a pattern of active host reuse
by the walnut fly, Rhagoletis juglandis Cresson (Diptera: Tephritidae), involves an
Allee effect. Increases in larval density strongly reduced pupal weight and to a lesser,
but still significant extent, reduced larval survival. Temporal staggering of clutches into
a host strongly reduced percent survival and, probably owing to competitive release,
increased pupal weight of survivors. Offspring survival and pupal weight were affected
relatively little by whether clutches were deposited within the same oviposition
punctures or evenly distributed along a fruit's surface. Results as a whole, failed to
provide evidence of an Allee effect. We propose that females reuse larval hosts in a
way in which they maximize their own reproductive success at the expense of the per
capita fimess of their offspring. In turn, we also propose that the larval aggregations
that form within multiply infested hosts may provide larvae with a mechanism for
reducing the adverse effects of larval competition.
15

REFERENCES
Allen GR, Hunt J, 2001. Larval competition, adult fimess, and reproductive strategies in
the acoustically orienting Ormiine Homotrixa alleni (Diptera; Tachinidae).
Journal of Insect Behavior 14:283-297.
Averill AL, Prokopy RJ, 1989. Distribution patterns of Rhagoletis pomonella (Diptera:
Tephritidae) eggs in Hawthorn. Annals of the Entomological Society of America
82:38-44.
Bauer G, 1986. Life-history strategy of Rhagoletis altemata (Diptera: Trypetidae), a
fruit fly operating in a non-interactive' system. Journal of Animal Ecology
55:785-794.
Bemays EA, Chapman RF, 1994. Host Plant Selection by Phytophagous Insects. New
York: Chapman and Hall.
Blanckenhom WU, 1998. Adaptive phenotypic plasticity in growth, development, and
body size in the yellow dung fly. Evolution 52:1394-1407.
Fox CW, Martin JD, Thakar MS, Mousseau TA, 1996. Clutch size manipulations in two
seed beetles: Consequences for progeny fitness. Oecologia 108:88-94.
Henneman ML, Papaj DR, Figueredo AJ, Vet LEM, 1995. Egg-Laying Experience and
Acceptance of Parasitized Hosts By the Parasitoid, Leptopilina-Heterotoma
(Hymenoptera, Eucoilidae). Journal of Insect Behavior 8:331-342.
Jaenike J, 1978. Optimal oviposition behavior in phytophagous insects. Theoretical
Population Biology 14:350-356.
16

Lalonde RG, Mangel M, 1994. Seasonal effects on superparasitism by Rhagoletis
completa. Journal of Animal Ecology 63:583-588.
Landolt PJ, Averill AL, 1999. Fruit Flies. In: Pheromones of Non-Lepidopteran Insects
Associated with Agricultural Plants (Hardie J, Minks AK, eds). New York:
CABI Publishing; 3-26.
Mayhew PJ, 1997. Adaptive patterns of host-plant selection by phytophagous insects.
Oikos 79:417-428.
Nufio CR, Papaj DR, 2001. Host marking behavior in phytophagous insects and
parasitoids. Entomologia Experimentalis Et Applicata 99:273-293.
Papaj DR, 1993. Use and avoidance of occupied hosts as a dynamic process in tephritid
fruit flies. In: Insect-Plant Interactions (Bemays EA, ed). Boca Raton: CRC
Press; 25-46.
Papaj DR, 1994. Oviposition site guarding by male walnut flies and its possible
consequences for mating success. Behavioral Ecology and Sociobiology 34:187-
195.
Papaj DR, Katsoyannos BI, Hendrichs J, 1989. Use of fruit wounds in oviposition by
mediterranean fruit-flies. Entomologia Experimentalis Et Applicata 53:203-209.
Peters TM, Barbosa P, 1977. Influence of population-density on size, fecundity, and
developmental rate of insects in culture. Annual Review of Entomology 22:431-
450.
Prokopy RJ, 1981. Oviposition-deterring pheromone system of apple maggot flies. In:
Management of Insect Pests with Semiochemicals (Mitchell EK, ed). New York:
Plenum Press; 477-494.
17

Prokopy RJ, Reissig WH, Moericke V, 1976. Marking pheromones deterring repeated
oviposition in Rhagoletis flies. Entomologia Experimentalis Et Applicata
20:170-178.
Prokopy RJ, Roitberg BD, 1984. Foraging behavior of true fruit-flies. American
Scientist 72:41-49.
Roitberg BD, Prokopy RJ, 1983. Host deprivation influence on response of Rhagoletis
pomonella to its oviposition deterring pheromone. Physiological Entomology
8:69-72.
Roitberg BD, Prokopy RJ, 1987. Insects that mark host plants. Bioscience 37:400-406.
Smith RH, Lessells CM, 1985. Oviposition, ovicide, and larval competition in
granivorous insects. In: Behavioral ecology: ecological consequences of
adaptive behaviour (Sibly RM, Smith RH, eds). Oxford: Blackwell Scientific
Publication; 423-448.
Sweeney J, Quiring DT, 1998. Oviposition site selection and intraspecific competition
influence larval survival and pupal weight of Strobilomyia neanthracina
(Diptera: Anthomyiidae) in white spruce. Ecoscience 5:454-462.
Thompson JN, 1983. Selection pressure on phytophagous insects feeding on samll host
plants. Oikos 40:438-444.
Thompson JN, 1988. Evolutionary ecology of the relationship between oviposition
preference and performance of offspring in phytophagous insects. Entomologia
Experimentalis Et Applicata 47:3-14.
18

APPENDIX A
HOST MARKING BEHAVIOR IN PHYTOPHAGOUS
INSECTS AND PARASITOIDS
19

ppif7^3
pp. 295-302
pp; 30^tT
pp. 313-317
pp. 319-330
pp. 331-339
pp. 341-346
pp. 347-354
pp. 355^
Entomologia Experimentalis et Applicata
Articles
and F»rraitpid8
C(^rh.Niji^,Da
Analysis by DC-EPG of the resistance to Bemisia
tabaci on an Mi-tomato line
Y.X. Jiang, G. Nombeia, M. Muniz
Genetic variiatibn viiMyzus pers^e ^>f)ulertipns
KrSaldZitoudi, John T. Msurgmitopouldsr^skt^l^^
Embryonic development of orchard leafrollers and
the forecasting of egg hatch
LHM Blommers. H.H.M. Helsen. F.W.N.M. Vaal
Temperature and humidity responses of the arctic-
20
alpineseed^gN^'usgroeniandKus
Jens Bdcher, Gdsta Na^maan
Host specificity and comparative foraging behaviour
of Aenasius vexans and Acerophagus coccois, two
endo-parasitoids of the cassava mealybug
Brigitte Dom, Letizia Mattiacci, Anthony C. Bellotti. Silvia Dom
l^arhihg^ot h^irifeste^pl^ vdtertil^ invtha^latval
Juhp'Fuldiskimai Y6

kluwer
the language of science
Dr. C Nufio
Uiriveniiy of Colorado
CU Mnieum of Natural IfisUHy
Boulder, CO 80309
USA
2M«a02
RUiw^ Aca^tBie rabbilMri B.y.
Mn SoiniiicHlait.io "
KAteiy
ISOeM DaMisat
Iht MtlMflindi
Re: Entoiinloifa Expirimeiitalb ct AppHcala CZOOl), T.99, p. 273-293
Dear Dr. Nufio,
T«ji(o>7l^7S-aBg '
I wmattfjil'
f i«ri(E«sSwto»jil
i^fWith refereoce to your request (copy herewith) to reprint material on which Kluwer
Academic Publishers eonnol tbe copy^t, oar permiasioa is granted, fiee of charge, for
tbe use indicated in your enquiry.
iOlib petinisston
• allows you noa-exclnsive lepnduction rights ihrougboot tbe Wodd.
• exdudK use in an eloctroaicrofm. Should you have a specific project in mind,
pieaie reapply for permission.
• penniasion is also obtained ficom the other authors (address is given on die
imprint page, ot with the article);
• tequiiss aiiiH credit (Kluwer Academic Publishers book^oumai title, vohune,
year of publkalion, pagp, chaptoyanicle title, oame($) of authoT(s), figure
niimber(«), original copyright notice) to the pubfication in which the matcdal was
originally published, by adding: with kind permission crfSawer Academic
Publisben.
.^Material may not be repablisbed until at least one year after our puUicaiion dale,
.^fffennisuan fiee of charge on this occasion does not pvejudioe any lights we jnig|ht
have to charge for iqxodnction of our copyrighiod material in the fiitnre.
Sincerely yours,
Iris Agent
Rights and PamiSBions
Tel: •*-31(0)78 6576130
Fu:-1-31 (0)78 6376254
E-mail: irisjagcr^wkapjil
Kluwkr Academic PUBUSHEBS
P&rLEASIS W CniAW TO IfKXUDE OCR UnillNCB 1M AU. aNWBraNODfCB

E.\i*vnutvttiiiii\ ft tiui W: 27."^—2^3. 2lK)l.
0 2l*)I Kiiiuer Anulfiiin Fuhinhcr\. t*rmied m the SviherUutds.
Xfini review
Host marking behavior in phytophagous insects and parasitoids
Cesar R. Nufio'^ & Daniel R. Papaj--'
Deparliiienr of EnioitwIoi'yK Depunment of Eculu^y anil Evoltiliimarv Bioltiny. Center fur Inxecl Science', and
the Inlerdisciplinary Degree Proi>nim in Insect Science^. University of Arizona. Tucson. .42 S5721. USA
ACL-cplcd: January 4. 2)K)I
Ke\ wnrtis: marking pheromone. oviposition deterring pheromone. oviposiiion behav ior, animal communication,
competition, superparasitism. para.sitotds
.AbKlracI
Oviposition behavior m phvtophugous insects and entomophagous parasitoids is iiften muditied by the presence
ot' conspecitic brood (eggs and larvae). Often, females avoid laying eggs on or in hosts bearing brood, a behavior
that acts to reduce the level of competition suffered by their offspring. .Avoidance of uccupicd hosts is typically
mediated by cues and/or signals a.ssociated with broixl. In this article, we review the role of Marking Pheromones
(MPs) as signals of brood presence in both phytophagous and entomophagous in.sects. We place information about
the function and evolution of MPs in theconte.it of recent theory in the field of animal communication. We highlight
the dynamics of host-marking systems and dl.scuss how effects of MPs vary according to factors such as female
experience and egg toad. We al.so examine variation in the form and function of MP communication across a
variety of insect taxa. While studies of MP communication in phytophagous insects have focused on the underlying
behavioral mechanisms and chemistry of MP communication, studies in entomophagous insects have focused
on the functional aspects of MPs and their role in "decision-making' in insects. We argue that an approach that
incorporates the important contributions of both of these somewhat independent, but complementary areas of
research will lead to a more complete understanding of MPs in insects. Finally, we suggest that .MP systems are
model svstems for the studv of animal sienalins and its evolution.
Introduction
Cues versus signals in assessment ofhrood presence
The oviposition behavior of phytophagous and parasitic
insects is often mixiihed by the presence of
conspecific brood (eggs and larvae). Typically, females
avoid depositing eggs on previously exploited
host re.sources. a behavior thought to reduce competition
suffered by their offspring iProkopy. 1981a).
Tlie stimuli permitting females to di.stinguish between
occupied and unoccupied hosts can be categorized as
either cues or signals (Seeley. I99S) i see Table I for
definitions of terms u.sed throughout manuscript). The
distinction is made on evolutionary grounds. Whereas
a signal is presumed to have evolved to convey information
from a sender to a receiver, a cue is a product
of selection on a trait other than communication and
conveys information only incidentally. Females of a
variety of species, for example, as.sess the pre.sence of
conspecitic brood on the basis of vi.sual or tactile stimuli
a.s.sociated with eggs (Rausher. 1979: Williams &
Gilbert. 198k Shapiro. 1981;Takasu & Hirose. 1988)
or larvae (Mappes & Makela. 1993). It is not obvious
that the stimuli involved have been shaped by selection
to enhance their detectability: as such, these stimuli
might best be described as cues of brood presence.
Cues of brood presence need not be directly produced
by the juvenile stages themselves. For some
phytophagous insects, for example, larval fra.ss deters
oviposition on infested hosts. Some investigators have
found that the deterrent compounds are unaltered plant
con.stituents. and not metabolic by-producLs or compounds
actively produced by the larvae them.selves
(Mitchell & Heath. 1985). .As such, thev would be
22
5
r/?
-Ju«
fciar
-4^
•n

174
belter categorized as cues than as signals. Other investigators
have found that the deterrent compounds
are actively produced by larvae, and in these coses
deterrent compounds found in feces may be thought of
as signals of brood presence (Hilker & Klein. 1989).
Assessment of conspecitic brood can also be mediated
by chemical and/or physical changes in hosts
induced by the presence of eggs, larvae, or adults.
In phytophagous insects, such changes mclude the
release of plant compounds associated with damage
caused by oviposition (Cirio. 1971) or by tissue destruction
by larvae (Renwick & Radke. 1981; Fitt.
198-J; Landolt, 1993) and/or adults (Heard. 1995).
Similarly, oviposition by enlomophagous parasitoids
may induce changes m a host's hemolymph composition
(Vinson & Iwatsch. 1980; Ferkovich et al..
1983) and these changes may be used to discnminate
unparasitized from parasitized hosts (Rsher &
Canesalingam. 1970).
Induced changes can make it diflicult to determine
whether a particular chemical compound or set
of compounds constitutes a cue or a signal. In Piens
brassicae. for example, it wa.s long believed (hat a
MP was employed by females in order to discriminate
among hosts (Rothschild & Schoonhoven. 1977;
Schoonhoven et al.. 1981). Recent work, however, has
shown that the host plant itself responds to the presence
of P. brassicae eggs and that systemic changes
in leaf surface chemistry may be the primary source
of an oviposition-deterhng effect (Blaakmeer et al..
1994). Is release of such compounds evidence of an
active process that evolved to serve the function of
assessment of conspecific brood? Has P. brassicae
evolved a mechanism by which a systemic change in
leaf chemistry is induced for its benefit? The situation
is complicated further by the finding that feeding
by Pieris caterpillars induces systemic production of
plant volatiles that attract natural enemies (Geervliet
et al.. 1994. 1998) and thus may be favored from
the perspective of the host plant (see also Meiners &
Hilker. 2000). Possibly, a signal evolved by the plant
to attract natural enemies has been commandeered by
the butterfly species as a cue of brood presence.
Cues as progenitors of signals of brood presence
Communication systems must rarely evolve de no\xt.
More likely, communication evolves when existing
cues become amplified by a sender and are. in this
way. transfonned into signals (Hasson et al.. 1992:
Bradbury & Vehrencamp. 1998).
23
The red egg syndrome of pierid butterflies is a possible
cxse that illu.strates the evolution of signals of
brood presence. Females of certain crucifer-feeding
pierid species within an inflorescence-feeding guild
lay red or orange eggs. The visual conspicuousness
of red eggs appears to facilitate assessment of their
presence by females subsequently visiting a host inflorescence
(Shapiro. 1981). Females of species within
a leaf-feeding guild, in contrast, lay yellow to white
eggs and do not assess brood presence. Inflorescences
are evidently more limiting as a food resource for
caterpillars than foliage and as such natural selection
appears to have favored the incorporation of red pigment
in the eggs of the inflorescence feeding guild as a
means of amplifying cues provided by the eggs themselves.
In shon. m inflorescence-feeding pierids. a cue
(eggsi hxs been transformed into a signal (red eggs).
While the details by which chemical cues evolve
into signals of brood presence remain to be worked
out. (here is good evidence for the existence of
actively-produced compounds or sets of compounds
that constitute such signals. Evidence for the use
of actively-produced MPs has been found in five
holometabolous insect orders including Coleoptera.
Diptera. Hymenoptera. Lepidoptera. and Neuroptera.
•MPs are distributed among al least 20 families of
phytophagous insects (reviewed by Prokopy. 1981a.b;
Roitberg & Prokopy. 1987; Landolt & .Averill. 1999).
In hymenopteran parasitoids. evidence of an ability
to discriminate between parasitized and unparasitized
hosts has been gathered for 150 to 200 species
and in nearly every family (van Lenteren. 1981). In
many cases. MPs have been implicated in mediating
host discrimination irevs. King & Rafai. 1970; van
Lenteren. 1976; Hofsvang. l990;Godfray. 1993). The
remainder of this review is devoted to the mechanism,
function and evolution of MPs.
Evidence for the existence of a marking
pheromone
Proof of the existence of a MP requites demonstrating
that a chemical compound (or mix of compounds) is
deposited by a female insect on its host in association
with oviposition and that the compound influences
the behavior of females visiting the host subsequently.
.A thorough characterization of a MP would involve
isolating the active compounds, identifying and synthesizing
(hem and showing that the synthesized compounds
have a behavioral effect equivalent to that of

f.ii'l, i K iiiiN iliituivlMuii l(ic icu
K.iintnume
Si*jn;il LXrIcciuui 1 luft»rv
Deicctmn:
Alarm
E.ivcvlri>ppinc
the putative murk. Additional evidence concerning site
i>t" production and mode of detection buttresii by
...tnN)vvUK> lNpK.»lly pcrcvncti hv a>iuact chcmorcvcption. ihe%e pherc>nionc>
.lie •.•cikt.iIK prtKkiccil hv Iciiialc^ .imi pLi^'cd imlu ik wiihin larv.il rcMUirces
liiIUnAiii'^ cyi: lavintf.
A >iitmilus to \%htch a rcvcivcr responds (h:il cnmeyv mlotnuiiuio ihiIv
iiicuiciUatly C'uo are mil >h.tpcJ hy luiliiral sclcL'tuin lucunvcy inlDriiiaiiun ami
are simpiy hy-pr«xlutts nt beha\ lor perlonncil Uir re:isniis iHher itwn
wiMninunicatioii.
A \innulu>. >iich a> a iiiatkiiitf phcnMiionc. \vhikh has hccii ^hapctJ hy natural
^Hcciion aiKi prfxlticetl hy a M^tuier Npcwiricaily lit convey ininriiiaitiHi ii» a
rcvcivcr
A thcniicai Mihsiance. Mich as a phcrtMiuirK*. which is emuicO by i»nc inthvtdiiai
and MKtdenially evokes a behavioral respiHise in a hctcnisfx'viiic leveivcr thai ts
htfiielictal ii» ihc rccciver btii nui lu (he emitter
A NhJv oI iheory ihai MatCN thai Ihc evuluiiun ol a iuiul in mtluciKeU by a
vompmmtse the bewliin anU ass^toaicil vmh K^ih the piintnwiu'ii
.iiid dLicvtion o( ):iven si^rul. Bcnchts may mcluitc a reiliictuin m compctiiion or
access It) itNid and mates. Cmis may iiictiHle ihc |>hystcal tnakhmcry rci|Uircil lo
produce, emit and dctcct ihc signal, eavesdruppmi:'. and si^milntc emirs, nainclv
nusscd dcicvtions of respi»tHhnu •w taiscalarms «see heluvw»
An event m vwhich a reccivcr faiK loiteiect a siynai tluii has been einuicil I hc
trcijucncy ol such events ami its cnasequence^ lor recet^er htticss is ilescnhed by
'«it:nal dctcv(u>o theory
Ail event in which a receiver rcspntds tit a stitiiulits as ihtniyh ii w^rv a
when in tact no sisnal was civcn. The Irtquency ol such events .ukI its
consequences l«ir rccciver liitios is docnbcd hy signal ilcteciiun theory
Hie e^plottaiMm ol smmih icucn or sicruitsi hy an uninicndcd receiver which is
not ncccssanly heneheial to the emitter. A purasitnid. lor example, may
eavcMlrop on iLs hosis hy usin^ the proence ol a marking phcromnne as an
indic3(nr ol prey prcscnce.
(Prokopy et al.. I982a( and detected with chcmorcceptors
borne on the forelarsi (Crniar •&. Prokopy.
1982).
In most ca.ses. investigators rely on less complete
evidence in making a case that an insect employs a .MP
A reasonably compelling case is often made by showing
that (I) egg-infested hosts are less acceptable lo
ovipositing females than uninfested host.s, and i2) the
dilTerence in acceptance is due to a compound(s) deposited
by a female m a.ssocialion wiih oviposition.
Showing that egg-infested hosts are less acceptable
than uninfested ones is straighttorward and mvolves
the use of routine behavioral assays. Assays of uninfested
hosts treated with chemical extracts of marked
hosts are also straightforward, and demonstrate that an

fubfe 2. Kinds ot cvicknce used lo document the exisience of a marking pheramone
1. Behavtoral assays of response lo NtP.
0. Quaniirication of rejection patterns of females olTered 'marked' hosts.
b. Dt'^tinguishing elTects of a mark' from elTecLs of other stimuli associated with an oviposition event.
I. QuamiHcation of rejection patterns of females offered hosts treated with extracts of marked hcsLs. eggs, or fecal matenal.
II. QuantiHcaiion of rejection paiiems of ferrules offered >urrogate hosts on %bhich females have Jeposiied a putative mark,
c Char:u:tenzaiK>n of the chemical nature of the mark
I. Structural identihcation of the ;u;tivecompourHl(s) m conjunction with behavioral avsays that establish activity
II. Synthesis of the active compound(s) m conjurwtion N^iih behavioral as.says that establish the activity of the synthesized compound(s)
2. Devrnption of a putative hosi^marking behavior
« Determmation of mechanisms for MP production and detection.
X Identmcation of tis.sue involved in prodiKing and storing the MP or MP precursor^.
I Quaniihcaiiun of rejection patterns of females olTercd ho^ts or surrogate hosts treated with extracts ot different tivsues
b Identihcation of mode of MP reception.
I Ablation of candidate semory %tructures and quantincation of changes in rejection patterrts
4 .\sscssment of ecological consequences ot MP a%e.
J- Ouantihcatum ot clutch dislnbution patterns in the tield
cffect of infestation on female behavior is chemicallymediated.
However, neither line of evidence provides
ironclad evidence for a MP. For example, a difference
m either case could be due to a chemical change in the
host it.self that accompanies infestation tSee section
on "Cues versus signals of brood presence" above).
In lieu of direct evidence that the insect itself produces
the chemical compounds involved, behavioral
assays of egg-free surrogate hosts bearing 3 putative
mark are sometimes used. Where investigators are
able to control the surrogate host stimuli precisely,
the effect of a marked egg-free surrogate on host
acceptance can be rea.sonably strongly attributed to
the insect, and probably something deposited by it.
Prokopy et al. (1982b) utilized such marked surrogate
hosts to test for the presence of a MP in the fruit
fly Anasirepha fraierculus. These re.searcheis created
marked surrogate hosts by allowing females to deposit
a clutch into an artilicial agar sphere and (hen. following
egg deposition, transferring the females onto
egg-free surrogate models where they were allowed to
deposit the putative MP. Females e.xpcsed to the eggfree
marked and umarked models rejected the marked
models significantly more, supporting the hypothesis
that A. faterrulus produces and deposits a MP which
deters funher host reuse (Prokopy et al.. 1982b).
The case for the presence of a MP is made more
convincing if the insect e.xpresses a 'marking behavior'.
such as the brushing or dragging of ovipositors
on the host, and more convincing still if the putative
25
marking behavior results in the physical deposition
of some substance on or in the host. In Rhagoletis
Ries. for example, females drag their ovipositor on
the fruit surface and this ovipositor-dragging results
in the obvious deposition of a clear substance on the
fruit surface (.\verill & Prokopy. 1988). However, caution
IS advised as apparent marking behaviors may not
necessarily be associated with the deposition of a MP
(Cirio. 1971; Prokopy & Koyama, 1982; Fitt. 1984).
Support for the hypothesis thai females are utilizing
a MP sometimes takes the form of tield censuses
that show that clutches are not distributed randomly
among available hosts, but rather uniformly among
them (Bauer. 1986; Thiery et al.. 1995). However, a
uniform distribution of clutches might also result if
females arc able to detect changes in hosts that occur
when eggs are deposit.
.Mcchanisms of host-marking pheramone
communication
Effects of MPs on female behavior
The presence of a MP on or in a host may affect female
behavior in multiple ways, both deterring and enhancing
cviposition (Corbet. 1973a; Prokopy. 198 la; Paine
et al.. I99T). The most often reported effect of MPs on
behavior, and the one on which this review focuses,
is a reduction in the number of eggs allocated to pre­

viously marked and uiilized hosis. The rediicluin in
allocation can iKcur m a numher ot W3y>i Frei|iienily.
a MP will decrease ihe lendeiicy tor a temalc visiting
a marked host to lay esgs .in that host. Where
ovipositiiin IS not entirely suppressed by MP. females
may still lay smaller clutches (Ikavva & Okabe. I9S5;
Papaj et al.. IWO). The extent to which clutch si/e is
reduced in a previously utilized host has been shown
in gregarious parasitoids to be a function of the number
of eggs that were deposited previously in that host
iBakkeret al.. 1472; van Dijken l murlwinc pticriMiiniic on the oviptwiiion
hohjMor ol ihc V!cUiicrrjntf.in iniii i!v, > iifjiuitu iMn? ie*i
(t'f t.ltfl;iiK)
sition behavior, namely a tendency for a rty in recent
contact with MP to terminate some stationary behavior
such a.s resting, grooming or oviposition. and to
initiate UKomotion. be it walking or Hying. In the
folklore of fruit fly research, contact with MP is said
to 'irritate' females. .-X generalized "interruptor'. or
irritant. etTect may be a neurally economical way in
which a MP can mediate a suite of functional effects
on oviposition.
Wliv pnultices the murk ami where i.s it plated'.'
In entomophagous Hymenoptera. females may use
either, or both, internal or external marks to indicate
which hosts have been previously exploited (Salt.
1937; revs, van Lenteren. 1976. 1981; Hofsvang.
1990). .According to Bosque & Rabinovich (1979).
whether MPs are deposited on the inside or the outside
of the host depends on the life stage attacked.
Egg parasitoid.s tend to mark hosts externally while
parasitoids utilizing other host stages tend to mark
hosts internally. Bosque & Rabinovich (1979) contend
that this pattern is functional; while an external mark
may be more readily detected by females inspecting
a potential host than an internal one. shedding of a
host larva's cuticle during molting would remove the
external mark. Bosque & Rabinovich also argue (hat
placement of the mark relates to sen.sory contexts for
host examination. Larval and pupal stages are generally
le.ss accessible than eggs to antennal examination.
Larvae may be Ies.s accessible both because they can

278
defend themselves and because ihey sometimes are
covered with hairs or spines (Gross. 1W3). Pupae
are also sometimes protected by coverings of various
sorts and additionally may be found in hard-lo-getto
locations. Larval and pupal examination with the
ovipositor, as opposed to antennae, may thus be more
effective and feasible and. at the same time, internal
examination of hosts provides a ready context in which
an internal mark may be deposited.
In phytophagous insects. .MPs arc again most often
produced by ovipositing females. In contrast to entomophugous
parasitoids. MPs are deposited exclusively
on the outside of the host plant. There are two possible
rea.sons for this difference in placement of the
mark. In contrast to parasitoids. ihe plants on which
phytophagous larvae feed are readily available for external
examination by ovipositing adults. .VIoreover. an
internally-deposited MP might not be as easy to detect
in the ti.ssue of a host plant, as in an insect host where
MP may circulate in the hemolymph.
In some instances, the eggs themselves appear to
be sources of MP (Ganesalingam. 1974; Cauthier &
Monge. 1999). MP production by larvae can be particularly
difficult to distinguish from cues released by
larvae exploiting a host. Despite this difficulty. .MPs
have been shown to be actively produced by larvae in
the form of oral (Corbet. 1973b: Hilker & Weitzel.
1991). anal (Hilker & Klein. 1989; Ruzicka. 1996;
Merlin et al.. 1996) or cxocnne glandular secretions
I Hilker & Weitzel. 1991. Schindek & Hilker. 1996).
Finally, putative MPs deposited by adult males are
uncommon but. where they occur, may be equally or
even more effective at deterring re-use of hosts by
females (Oshima et al.. 1973; Szentesi. 1981). It is
often not clear if male 'pheromones' are signals in
the sense detined above or cues incidentally left by
males during general activities on hosts (cues such as
frass or associated compounds). In one case, however,
a mark produced by males at sites of clutch deposition
has been shown to stimulate female oviposition at
those sites (Papaj et al.. 1996). In this particular insect,
a walnut-infesting tephritid fly. females enjoy direct
benefits by re-using oviposition sites (Papaj. 1994;
Papaj & .Alonso-Pimentel, 1997'). Given a pattern of
re-use. there is a benefit of male-marking to males that
guard those sites in terms of attracting females to them
and a benefit in turn to females m terms of finding
oviposition sites.
Sites of production and modes of detection
27
Although detailed work is lacking for many systems,
sites of MP production and/or storage are typically
associated with either the exocnne. digestive or reproductive
system. In parasitic Hymenoptera. the Dufour's
gland iGuillot & Vin.son. 1972; Mudd et al..
1982; Harrison et al., 1985). poi.son gland (Bragg.
1974; Yamaguchi. 1987'), lateral oviducts (Guillot &
Vinson. 1972) and ovaries (Holler et al.. 1993) have
been implicated as sues of MP production. In other
Hymenoptera. MP production may be associated with
the legs, which are used to mark hosts (Foltyn & Gerling.
1985), or in cases where juvenile hormone is
used as a MP. the corpora allata (Holler et al.. 1994a).
In Coleoptera. the hind gut and possibly Malpighian
tubules are important sources of .VIP (White et al..
1980). but MP IS also produced or stored in prothoracic
and abdominal glands in both adults (Roth. 1943;
Loconti & Roth. 1953) and larvHc (Hilker & Weitzel.
1991). In Oiptera. MPs may be produced either in
the head region and deposited by mouthpans (Quiring
et al.. 1998) or in the midgut and released through
orifices u.sed in defecation (Prokopy et al.. 1982a). In
Lepidoptera, MPs may be produced by paired larval
mandibular glands (Corbet. l973b)orby the accessory
glands that produce egg-coating substances (Thiery
et al.. 1995).
Most MPs are non-volatile and are detected by
contact chemoreceptors (but see van Baaren & Nenon,
1996; Kouloussis Jt Katsovannos. 1991). While some
insects use receptors on antennae to detect MPs (Salt.
1937; Hilker & Klein. 1989; Ferguson et al., 1999).
others use receptors associated with mouthparts or
tarsi (Prokopy & Spatcher. 1977; Cmjar & Prokopy,
1982; Messina et al.. 1987). and/or ovipositors (van
Lcnteren, l972;Ganesalingam, 1974). Tips of ovipositors
are used for either external or internal assessment
of brood presence; they may bear either hairs or
plates designed to detect MPs (King & Rafai. 1970;
Ganesalingam. 1972;Greany &Oatman. 1972).
Costs and benefits of MP use
Benefits of marking hosts
.VIPs allow females to gauge the relative level of competition
that their progeny might suffer in hosts that
have been previously utilized and to adjust allocation
of eggs accordingly. In extreme cases where larval
stages are relatively immobile and where offspring

IcL-il (111 iir ivilliiii Uiscrclc rcMHirce units that support
iiiily line larva ui inalurily. a Icmale that accepted
such a hosi wDuli) expcnoiicc no tiincvs gam or even
a net hlness loss by placmt: her oft'spring in an environment
where it cannot develop successfully. In
such instances, the presence of a MP may be sufficient
to elicit host rejection on the part of the assessing
female. For example, in the apple maggot tly. Rluitioteris
pomonella, the optimal deasity for successful
larval development in hawthorn fruit is one (Averill
& Prokopy. I9!i7a). .After laying a single egg. apple
maggot tly females deposit a .MP that strongly
deters other females from reusing the same fruit. In
this species. MP is believed to function almost exclusively
to allow females to discriminate occupied from
unoccupied fruit.
In instances where more than one larva can develop
m a given resource unit or where larvae are mobile
enough to find new hosts. .MPs may allow females
to assess not just presence versus absence of brood
but also the amount of brood. In such ca.ses. females
may make a graded assessment of the level to which
a host has been previously utilized. Females of the
bean weevil Callnsohruchux macuUttux use chemical
and tactile cues a.s.sociaIed with eggs to a.s.sess not only
whether or not a host bean has been utilized but also
the number of eggs a.ssocialed w ith that host. Females
have been found to selectively re-use hosts that bear a
lower than average number of eggs (Messina & Renw'ick.
1985a.b; Wilson. 1988). Such a.sse.ssmenl is
functional. In most beetle populalion.s. more than one
weevil can develop in each seed; nevertheless, each
additional larva faces both a greater risk of not obtaining
sufticient re.sources for successful development or.
if development is successful, a reduction in fecundity
(Mitchell. I975:Credland et al.. 1986).
A graded assessment of level of infestation implies
that MP contain.s information about the overall
numbers of eggs laid in a host. To the extent that
variation in level of infestation reflects variation in
number of ovipositions. .such information may derive
simply from the accumulation of a con.stant amount of
MP deposited at each oviposition (Papaj et al.. 1992:
Huth & Pellmvr. 1999). Where females lay clutches
of variable size, information about the size of a clutch
may similarly be conveyed in terms of the amount
of MP deposited. For example, the walnut-infe.sting
fly. Rhagolettx juglandis. marks fruit after oviposition
for a duration proportional to the size of its clutch
(D. Papaj. C. Nufio & H. Alonso-Pimentel. unpubl.).
.A similar correlation between marking time and clutch
size deposited within a host has also been found for the
gregarious wasp. Telcnmiuis faruu (Bosque & Rabinovich.
1979). In other cases in w hich females make a
graded assessment of level of infestation, the underlying
mechanism has not been conclusively established
iBakker ct al.. 1972. 1990; van Lenteren & Debach.
1981; van Dijken & Waage. 19X7).
Ciisis of mcirkint;
.-\ny system of communication incurs costs for both
the sender and the receiver of signals (Bradbury &
Vehrencatnp, 1998). These costs include the biological
"machinery" involved in production and release
as well as reception of the signal. From the perspective
of the sender, additional costs may be borne in
the form of eavesdropping. Examples of eavesdropping
are well known in .MP communication sysiems.
In particular. .MPs are sometimes used as kairomones
for parasites of a hosi-marking female's progeny (Corbet.
1973a; Prokopy & Webster. 1978; Roiibcrg iS;
Lalonde. 1991). fhilficcpieni msae. a wasp parasitoid
of R/wi;i>leii\ basioUi. for example, increases (he
amount of time spent searching for hosts and improves
Its efficiency at finding host larvae in the presence
of the host tly's NtP (Roitberg & Lalonde. 1991).
H. mseii may even u.se the fly's .MP trail to localise
the fly's oviposition site (Hoffmeister et al.. 2000).
The risk of larval parasitism associated with eavesdropping
on MP 'dialogue' may favor higher marking
pheromone decay rates and lead to a potential reduction
in signal efficiency (Hoffmeister & Roitberu.
1998).
Another context for eavesdropping is cleptoparasitism.
a form of parasitism in which a parasite species
depends on the host utilization efforts of another parasite
species in order to be able to exploit or gain
access to their hosts. The cleptoparasitic ichneumonid
Temetticha inierniptor. for example, is attracted to
trail odors laid down and u.sed by the braconid
Orgilus nbscurautrio di.scriminate between previously
searched and un.searched areas. This behavior leads
to T. m/errupr«rpreferentially utilizing previously exploited
hosts (Arthur et al.. 1964). Interestingly, the
aphidiid parasitoid Aphiilius iizhekislanictis avoids iLs
own offsprings' parasitism by dispersing away from
host patches previously asses.sed and marked by their
hyperparasitoid .-MIUXYSUI vtcrrix (Micha et al.. 1993;
Holler el al.. 1994b).
Signal detection theory suggests that another cost
of signaling involves errors a.s.sociated with ihe detec-

280
lion of a signal. Costs associated with error in signal
detection are believed to profoundly influence the
evolution ol communication systems (Reeve, 1989;
Wiley. 1994; Bradbury & Vehrencamp. 1998). Two
kinds of errors in signal detection that are commonly
considered to shape communicalion systems are errors
a.ssociaied with missed detections and errors associated
with false alarms. In the conie.xi of host-marking,
a missed detection would occur when an insect failed
to delect MP when in fact it had been deposited on or
in a host. A false alarm would occur when an insect
responded a.s though a MP had been deposited on or in
a host when in fact it had not.
Aspects of host-marking behavior in some insects
seem designed to reduce signal detection errors due to
mis.sed detection. For example, lephnlid lly females
mark host fruit by dragging their ovipositor on the
fruit surface. Females generally do not restrict hostmarking
to the vicinity of the oviposition site, but
rather generate a trail of MP over signiticant areas of
the fruit surface. Such behavior, though costly in time
and perhaps energy, seems designed to dis.seminate the
•VIP over the surface and thereby decrease the chance
that a female subsequently visiting the fruit might
fail to detect the mark. In Anastrepha fraterculus and
R. poinonella. females mark longer on larger fruit, a
strategy that possibly reflects a tradeoff between the
cost of MP and the cost of missed detection (Prokopy
et al.. 1982b; Averill & Prokopy. 1987b).
In still other in.sccts. females mark substrates
around exploited host patches (Price. 1970; Waage.
1979; Sugimoto et al.. 1986). Such behavior can improve
detection of a MP and reduce the time needed
for host assessment (Hoffmeister Jt Roitberg. 1997).
Rnally. a common strategy in communication systems
for minimizing missed detections is redundancy.
The propensity for certain entomophagous parasitotds
to mark both the inside and the outside of the host
(reviewed by van Lenteren. 1976; Hofsvang. 1990)
may reflect a strategy of redundancy. Other parasitoids
utilize a two-component chemical marking system;
the deployment of more than one component in hostmarking
has many possible explanations, of which
redundancy is one (Holler et al.. 1991).
Examples of errors associated with false alarms
are harder to come by in host-marking systems. On
its face, contact chemoreccption potentially provides
such high resolution in terms of discrimination among
chemical compounds that it seems unlikely that a
chemical compound or set of compounds that were
not a MP would be mistaken for one. However, there
29
are at least hints of the occurrcnce of false alarms in
host-marking communication. Intephritid flies, for example.
extracts of feces deter female oviposition into
fruit or surrogate fruit. Even male feces have such activity
(Prokopy et al., 1982a), causing one to wonder if
fecal deposition on fruit may sometimes be misinterpreted
by inspecting females. Similarly, in the bruchid
beetle, C. maculatus. egg-free beans that have been
'conditioned' by males as they walked or defecated
on seeds receive fewer eggs than unconditioned beans
(Sakai et al., 1986) and, while there are alternative
interpretations, this might again constitute an example
of a false alarm in MP communication.
Documented examples of false alarms in the animal
communication literature are typically intenpecilic
in nature as. for example, in ca.ses of aggressive
mimicry in which one species mimics another species'
signal and preys upon receivers that unwittingly onent
to the mimic signal (Haynes& Yeargan. 1999). Examples
of aggressive mimicry are well known in relation
to sex pheromones and yet wholly unknown in relation
to MPs, perhaps because the effect of a MP (unlike that
of a sex pheromone I is usually deterrent and thus not
of use to a predator. If there are any cases of aggressive
mimicry involving MPs. they might occur in those situations
in which a MP ha.s an aggregative, rather than
a deterrent, effect.-
•Marking pheromones and the value of inrormation
about brood presence
Evidence from strain differences
Use of a signal should vary in relation to changes in the
costs and benefits of information provided by that signal
(Bradbury & Vehrencamp. 1998). If host-marking
is costly in relation to the benefit of information
about brood presence, then we would expect to observe
host-marking only in species or strains for which
such marking has significant value in terms of the fitness
of a female's progeny. In C. maculatus beetles,
avoidance of occupied hosts can lead to considerable
increases in female and offspring fitness (Credland
et al.. 1986) and. in these beetles, degree of avoidance
varies between strains. A series of crosses and
backcrosses between two strains showed that strain
differences were genetically based (Messina. 1989).
Differences in avoidance of occupied hosts may arise
in two ways; first, signalers may change the 'strength'
of marks placed on a host or; second, receivers may

change the degree to which ihey respond to a mark of
a given strength. Experiments conducted by Messina
et al. (1991) demon.strated that both mechanisms contributed
to strain differences in avoidance. Females of
a strain found on relatively smaller hosts where interference
competition was relatively greater responded
more keenly to a mark produced by either strain.
Females of the strain which experienced greater interference
competition also produced a mark that acted
as a greater deterrent to further egg-laying by females
of either strain.
.A genetic change in re.spon.se to a mark may not
nece.s-sarily involve tixed changes in MP reception,
processing or strength. Selection may mediate strain
differences in response via effects on a female's mean
egg load which, in turn, affects behavior iWaaae.
1986; Wajnberg et al.. 1989). Egg load, detined as the
number of mature egg.s a female has available to lay.
can affect respon.ses to host.s m many ways iMinkenbcrg
et al.. 1992). of which respon.se to .MP is but
one. Typically, higher egg loads increase a female's
propensity to lay clutches into previously exploited
hosts (van Randen & Roitberg. 1996). Selection for
higher egg loads may es.sentially lead to females reducing
their propensity to rejected previously marked
host.
Facultative partem^ in MP Jeplovment
Host-marking in phytophagous in.sects is usually
obligatory, with females nearly always depositing MP
when eggs are laid. However, deployment of MP
is occasionally facultative: presumably, facultative
marking reduces the overall costs of marking by deploying
a signal only when it is beneficial lo do so.
For example, females of a Delia (= Hylemya) species
(Family .Anthomy iidae) mark one of their host species
but not the other. On developing flower buds of Polemnnium
folinxsimum. female.s deposit a single egg
and an a.sscK:iated MP. The MP generates an overlydispersed
distribution of eggs within resource patches
(Zimmerman. 1979). The same Delia sp. al.so utilizes
Ipttmopsis aKgregata a.\ a larval resource but. in contrast
to its first host, females often lay more than one
egg on this host. Interestingly. female> do not deposit
a MP when they oviposit on thi.s second host. .According
to Zimmerman (1980. 1982). females fail to
mark I. plants not because these hosts support
relatively more offspring to maturity, but becau.se
egg mortality on /. aggregata is so severe that females
experience little titne.vs los.s&s if consecutive females
250 -I
— 200 '
S I
— ISO ]
I
100 •
50
0 12 3
No. Previous Marks on Fruit
Fiiiuie J. Time spent n);irkinc by lenule AnuMtrfho as
.1 luncium oi ihc niinibcr ul ICIIUICN ihul prcvuHivK h«.»Nl murknl a
Iruil (uciapicU from Pupa| & Alup.
re-use an infested host. In P. folinssimiun. by contrast.
egg monality is low and female re-use leads to
relatively intense larval competition between clutches.
The time spent host marking has also been found
to be correlated with the sequence in which a female
exploits a multiply-infested host. In the fruit fly .4. Intienx.
time spent host marking increa.sed exponentially
as females deposited clutches into hosts previously
marked zero. one. two or three times (Papaj & .Aluja.
1993) (Figure 2). The increa.se in marking time was
proposed to he functional for two rea.sons. First, by increasing
the amount of mark placed on a host, females
might compeasate for partial degradation of previous
marks. Second, increases in marking times might reflect
declines in fitness of progeny in progressively
later-laid clutches.
D\namtcs in iiiteniul \rate and the value nf
mfonnutnm
Female respon.ses to signals from occupied hosts depend
on variation in internal state related to egg load,
age or experience in a manner consistent with dynamical
foraging theory (Mangel & Clark. 1988).
•Avoidance of marked hawthorn fruit, for example, decreases
for a R. pimumella female as her time since
last oviposition increases (Roitberg & Prokopy. 1983i.
Such a pattern is functional. When host fruit are rare
and time between encounters long, females should not
be ;LS choosy about u.se of infested hosts as when host
fruit are common, so long as an egg laid in an infested
fruit has a meaningful chance of surviving to
repaxluce.
Responses to .MP likewise vary with egg load.
Female snowberry flies. R. zephvna. of similar age.
experience and mating status but with higher egg loads
were sianificantiv more likelv to re-use marked hosts

282
than were conspccifics with lower egg loads (van Randen
& Roilberg. 1996). Once encountered, a given
host is of higher value to a female with high egg load
(because the female is more time-limited) and consequently
(he female is less likely lo be deterred by
MP.
There are other contexts in which avoidance of utilized
hosts is dynamic in nature. conte.xts well-studied
in parasitoids. Under conditions where the survival of
ihe second progeny or clutch is greater than zero, parasitism
of an already-parasitized host (a phenomenon
termed 'superparasitism') may be functional if unparositized
hosts are scarce, if search or handling time is
high, if females are time limited or if multiple females
are exploiting a patch simultaneously ivan Lenteren.
1981; van Alphen et al.. 1992: van Alphen & Visser.
1990; Speirs et al.. 1991). .All of these factors influence
the value of the host resource to a female and
thus influence response to MP.
Dynamics in the value of the resource
The extent to which MP deters host use ought to depend
on the relative value of the host resource from
(he perspective of an ovipositing female. In medfly.
for example, (he extent to which females avoid marked
fruit relates both to the size and the ripeness of those
fruit. On ripe fruit, females generally prefer (o lay eggs
in unmarked, uninfested fruit (Prokopy et al.. 1978).
However, degree of avoidance also depends on fruit
size: large egg-infested hosts are avoided less than
small infested hosts (Papaj & Messing. 1996). The
difference in level of avoidance is probably functional,
since the cost of larval competition is greater when that
fruit is small (Averill & Prokopy. 1987a).
•A more dramatic shift in medfly behavior accompanies
changes in fruit ripeness. Whereas females
prefer unmarked (o marked hosts when fruit are ripe,
(he preference is actually reversed when fruit are unripe.
with females preferring to lay eggs in marked
hosts and. in fact, depositing eggs in existing oviposition
punctures (Papaj el al.. 1992: Papaj & Messing.
1996). The preference for marked, infested fruit when
fruit are unripe is also believed to be functional. Females
have a difficult time penetrating unripe fruit
with their ovipositors: re-use of an existing oviposition
puncture saves time and increases the chances of successfully
depositing a clutch. Re-use may also reduce
ovipositor wear. Evidently, when fruit are unripe, such
direct fetnale benettts ore relatively great and offset
the cost of additional competition experienced by a
31
female's clutch; when fruit are ripe, in contrast, the
cost of competition offsets the relatively small benefits
and marked fruit are avoided.
In ca.ses where responses to brood presence change
with changes in the quality of (he host resource, it is
not a given (hat (here is a change in (he response (o
MP per se. In medfly. the response to MP itself does
not appear to change at all with changes in fruit quality.
A given quantity of MP is consistently deterrent,
independent of other fruit quality traits (Papaj et al..
1992). Moreover, the quantity of MP deposited does
not seem to depend on fruit size or ripeness. Instead,
degree of avoidance of marked fruit is governed by a
balance between the deterrent properties of MP. on one
hand, and stimulatory properties of the frui( surface,
on (he other (Papaj et al.. 1992).
Temporal patlerns in responses lo marked hosts
Types of patterns and stimuli involved
Sometimes the pattern of response to a marked, occupied
host remains stable over time. More commonly,
responses change markedly over the time since the
host was previously utilized (revs, van Lenteren. 1976;
Strand. 1986; Hofsvang. 1990). These changes in response
run the gamut of possible forms. Sometimes
females show increased levels of host rejection with
time: some(imes females initially show high levels of
rejection but accept hosts more readily with time. Responses
can also be complex; for example, levels of
rejection may be high at first, then decline to some
asymptote and remain stable (Holler et al.. 1991).
.Mtematively. levels of rejection may be high, then
decline and then increase again (Chow & Mackauer.
1986). In any of these cases, the temporal dynamics
may reflect changes in MP communication, changes in
cues emitted by the host or brood, or cues generated by
an interaction between host and brood (see Hubbard
et al.. 1987: Ueno. 1994; and Gauthier et al.. 1996 for
discussion of complex patterns).
Female behavioral responses to a larval host that
has been occupied for a given length of time may also
not be a hxed species-speciflc pattern, but may be a
condition-dependent one that varies with a female's
experience level (Bosque & Rabinovich. 1979: Hubbard
et al., 1999: Chow & Mackauer. 1986). rearing
density (\1sser, 1996). age and egg load (Vblkl &
Mackauer. 1990). and whether the mark encountered
was produced by the female herself or by conspecifics

(Hubbard et ul.. 1987; Holler el al.. 1991; Gauthier
et al.. 1996). Where L-OSLS and bcnetil.s of host re-
UNC vary under the above conditions, female rejection
patterns are expected to vary accordingly.
Incri'tiM's in reiecluin levels over rune
A pattern in which t'einales show increased level.": of
host rejection with time may not retiect a pattern in
MP communication at all. The patterns may simply
retiect use of cues a.ssociated with the development of
brood and consumption of the host (Strand, 1986). It is
rea.sonable to suppose that such cues require extended
periods of time before they are produced in enough
quantity to be detected. However, increasing levels of
reiection with time. ;i.s observed in certain para.sitoids.
IS at least sometimes due to changes in .VIP activity.
In some cases, changcs m activity probably reflect a
constraint on the amount of time required for a MP
to be produced, activated, or circulated within a host
(van Lenteren. 1976; Cloutier et al.. 1984; Gauthier &
Monge. 1999). In other ca.ses. such a pattern may have
titnes.s value, as when second clutches are associated
with a high rttness gam. but only for some hnite period
of time after the tirst clutch is laid (Gauthier el al..
1996). Where females commit ovicide, for example,
there is likely to be a high payoff for second clutches
until such time as the tirst eggs hatch, after which the
payoff may decline precipitously i Strand & Godfray.
1989; Mayhew. 1997V
Decreases in rejectum levels nver tune
\ pattern in which females initially show high levels
of rejection which then decline with time is often indicative
of a breakdown of MP over time. While it
is likely that degradation of MP sometimes reflects
a constraint on an in.sect's ability to produce a more
persistent signal, it is at least conceivable that the
"half-life" of MP is functional. Roitberg &. Mangel
(1988) addres.sed this possibility in theoretical model.s.
In these models, they assumed that longer-lasting
MPs are costlier to produce. The MP persisLs initially
bccause both the marker and the recipient of the
•MP benefit if the recipient avoids laying eggs in an
already-utilized host. However, as time goes on and
the marker's offspring grow, the probability that her
offspring will be out-compcted by the offspring of a
second female declines. .At some point in time, the
benetii of additional persistence of the MP in terms of
reducing the marker's competitive losses will equal the
cost of the additional persistence. Tliis is the point in
time at which the MP should, by design, break down.
The half-life of MP may retiect not ju.st temporal
changes in the relative payoffs for rirst and second
clutches (Quiring

284
thai (he marker was a mutant for strong marking arising
in a population of weak markers, with both strong
markers and weak markers possessing the ability to
delect marked hosts. A paradigm in which strong
marking evolves from weak marking is consistent with
notions thai host-marking systems evolve via amplilication
of cues of brood presence left incidentally after
ovipositioniFitt. 1984).
Roitberg and Mangel's various simulations yielded
the following observations. First, markers held a
considerable advantage relative to non-markers when
hosts were clumped in distribution and foragers
showed a tendency to concentrate search in areas of
high host density. The greater the tendency of the parasite
to search in area,s of high host density, the greater
the relative fitness of host-marking. This is because
markers tend to end up searching in patches containing
large proponions of unutilized hosts. Non-markers,
by contrast, end up searching in high-density patches
which often contain large numbers of previously utilized
hosts.
Second, the rate of spread of a mutant allele for
host-marking within a population depends on the extent
to which non-mutants rccognize the mark. Conditions
for the evolution of host-marking are more
restnctive under a scenario in which weak markers
recognize the mark of a strong marker. This is because
weak markers gain all of the benetits of recognizing
brood presence while not paying the costs
of host-marking (in the model, a cost expressed in
terms of time required to mark). Host-marking can
still evolve in this context, but only if markers reencounter
marked hosts signiticanily more often than
weak markers. In that case, markers benefit by avoiding
oviposition in hosts that they themselves utilized
and marked (i.e.. by avoiding what is referred to as
"self-superparasitism').
.Although host-marking traits initially spread more
readily when the sender directly benetits from avoiding
self-superparasitism. the models of Roitberg and
.Mangel suggest that, once host-marking is established.
It may be maintained in pan by a benetit gained by
responding to marks deposited by other conspecifics.
Such avoidance is advantageous both to the female
that initially exploited the host and to the female rejecting
that host if it reduces the level of competition
suffered by each set of progeny (Roitberg & Mangel.
1988).
Finally. Roitberg & Mangel's (1988) models assumed
that host-marking evolves under individuallevel
selection. However, the relative degree to which
33
a female's mark evolves to inform herself versus another
female raises a level of selection issue. Host
marking could conceivably evolve more readily to
communicate information to other conspecifics when
the marker is genetically related to those conspecifics.
Host-marking could even evolve as an altruistic trait,
improving the foraging decisions of related individuals
at the expense of the donor's own foraging efficiency.
The evolution of host-marking under kin selection
may require restrictive conditions such as limited dispersal
among related conspecifics (Godfray, 1993).
Such restrictions notwithstanding, kin selection may
account at least m pan for the evolution of trail marking
among social caterpillars (Costa & Pierce. 1997)
or the repellent scent marks placed on recently exploited
flowers by honey bees and bumblebees (Goulsonetal..
1998).
A notion of self in host-marking behavior
Females in some hymenopteran parasitoids are less
likely to superparasitize hosts they themselves parasitized
than hosts parasitized by conspecifics (Volkl &
.Vtackauer. 1990: van Dijken et al.. 1992: van Baaren
et al.. 1994; Danyk & Mackauer. 1993: but also see
Bai & Mackauer. 1990; van Dijken & Waage. 1987:
and van Alphen & Nell. 1982). Discnminaiion of self
and non-self within these systems, often thought to
be mediated by MP. is potentially adaptive because
eggs deposited in a host parasitized by another female
are potential competitors of the superparasitizing
female's offspring, whereas eggs deposited in a host
parasitized by the same female will increase competition
among genetic relatives (van Dijken et al.. 1992).
Whereas self-superparasitism is generally a waste of
time and eggs, conspecitic superparasitism can be
beneficial when there is some probability of eliminating
non-sibling competitors directly, via female
ovicide (Strand & Godfray. 1989; Mayhew. 1997) or
hyperparositism of conspecitic progeny (van Baaren
et al.. 1995). or indirectly via larval competition in the
form of physical combat or physiological suppression
(Podoler& Mendel. 1977; Vinson & Hegazi. 1998).
Discrimination between self and conspecific parasitism
may be facultative. Under conditions where
superparasitism is common, for example, females
may self-superparasitize to insure that their offspring
outcompete potential competitors (Danyk & Mackauer.
1993). Parasitoids may also benefit by reluming
to hosts they previously oviposited into and laying
a second clutch if increasing the density of juve-

nlle stages saturules a host's defenses (van Alphen
& Visser. 19901. Such condition-dependence in selfsuperparasitism
presumably reflects a tlexibility in
response to the MP itself, although data on this point
are lacking.
More generally, the mechanisms by which females
discriminate self from non-self are not well understood.
One means for such discrimination is through
use of a two-component marking system. In such
a system, one of the marking components is shortlived
and allows females to recognize hosts that they
themselves have recently utilized, while the other
component is long lived and provides general information
regarding the host's status i Holler. 1991: Field
& Keller. 1999). Alternatively, recognition of self
versus conspecitic parasitism may be mediated by
variation in .VtP constituents. Where such variation
IS genetically based, the latter mechanism can generate
responses to a parasitized host that are graded
according to decree of relatedness between successive
females (Marns el al.. 1996). Finally, while not fully
explored by researchers, the mechanism tor discriminating
between self and non-self may also involve a
learning component (see Ueno. 1994; Ueno & Tanaka.
1996).
MPs and interspecific ili.tcriminaliim
The di.scussion above suggests that models of the evolution
of host-marking behavior must consider exactly
who is being informed by MPs. In some systems, females
are believed to deposit MPs primarily to inform
themselves as to which hosts have been previously utilized
(Roitberg & Prokopy. 1987). In other systems.
MPs function mainly to convey information from one
conspecitic to another (Prokopy. 1972). Both of these
patterns are intraspecitic in nature. However. MPs
may influence use of occupied hosts by interspecific
competitors as well (Giga & Smith. 1985: .McClure
etal.. 1998). Do MPs evolve to mediate the assessment
heterospecific brood'.'
In parusitoid.s. interspecific discrimination, while
uncommon (Turlings et al.. 1985; Hagvar. 1989), is
most often observed when two species are closely related
(Vet et al.. 1984; McBrien & Mackauer. 1990.
1991; van Baaren et al.. 1994) and to a lesser extent,
when two relatively unrelated species overlap
in their ranges and utilize the same hosts (Bolter &
Laing. 1983; Hagvar. 1988; see also Thiery & Gabel.
1993). The former context reflects effects of phyUv
genetic relatedness. whereas the latter context con­
ceivably reflects an adaptive response to interspecific
competition.
The role of ancestry in cross-recognition of MPs
among species was addressed in work by Prokopy and
colleagues (Prokopy et al.. 1976; .-Vverill & Prokopy,
1981. 1982; reviewed by Prokopy & Papaj, 1999) on
members of three species groups within the tephritid
fly genus RhuKolens. Here the phylogenetic relationships
for the North .American and European species
are well known (reviewed by Smith & Bu.sh. 1999)
and host-marking within the genus well described (reviewed
by .Averill & Prokopy, 1989; Prokopy & Papaj.
1999). Data indicate that females of species from different
species groups are generally not deferred by
each other's MP, Within a species group, however,
females of one species are frequently deterred by the
other's .MP. Since species within a species group do
not specialize on the .same host species, these patterns
of cross-deterrency probably reflect effects of shared
ancc-!r;;.
In general, the conditions under which MPs might
evolve under selection to mediate interspeciHc discnmination
are unclear. The issue ha.s only occasionally
been considered. On the basis of simulation
models that assumed costs of di.scrimination in terms
of mi.s,sed opportunities to lay eggs. Turlings et al.
(1985) concluded that interspecific host discrimination
was unlikely to ari.se de novo, becau.se such discnmination
is disadvantageous to the first species to evolve
to avoid multiparasitism (see also Bakker el al., 1985).
Turlings el al.'s simulations assumed that the
species involved were of approximately equivalent
competitive abilitie.s. Whether or not interspecific di.scnminalion
might arise and be maintained m a situation
in which species differ in competitive abilities
(perhaps a more common situation) has not. to our
knowledge, been considered theoretically. One might
anticipate that an inferior competitor will discnminate
more against u.se of a host occupied by a superior
competitor's offspring than the reverse and. in fact,
this generally appears to be the ca.se (Bolter & Laing.
1983; Ciga & Smith. 1985; McBrien & .Mackauer
1991; Leveque el al.. 1993). In bruchid beetles, an
asymmetry in interspecific discrimination is mediated
by an asymmetry in one species' respon.se to the other
species' .MP. One bruchid species. Callosnhnichus
rlunfesutnux, is deterred by the mark of a superior
competitor. C. maculaiu.i. but the reverse is noi true
(Giga & Smith. 1985). Conspecitic C. macuiatus females
are deterred by their own species' mark, raising
the possibility that C. rhodesiantts is effectively eaves­

286
dropping on ihe MP communication system of its
competitor, a system that evolved in ihe coniext of
informing self or conspecirics. If this is the cisc. it
remains unclear if such eavesdropping constitutes an
evolved trait. While it is possible that C. rhodesUmus
evolved a sensitivity lo its congener's MP. it is al.so
possible that C. macultiius lost a pre-existing sensitivity
to its congener's MP (a sensitivity derived perhaps
from shared ancestry).
In short, it appears thai a respon.se by one species,
or loss of response, to another species' MP can readily
evolve under selection, particularly when ipecies
differ in competitive ability. However, whether a MP
itself can evolve strictly to signal the presence of a
heterospecitic competitor is uncertain. Given the ubiquitousnessof
MPs conveying information about self or
conspecitics. MPs mediating di.scnmmation between
species probably rarely evolve independently of MPs
mediating discrimination within a species. Rather, the
same MP may evolve simultaneously to function at
both intraspecific and interspecific levels.
Phylogenclic perspectives
Eci>li>);ical correlates of host-markin g
The Rhagnietis data on cross-recogniiion reviewed
above suggest that there is something to be gained
from a phylogenetic approach to host-marking evolution.
The Rliaftoleiis data are intriguing; yet it is
noteworthy that data were not collected with phylogenetic
analysis in mind. .Not all of the comparisons
or even the most informative comparisons have been
made, especially in light of revisions of the phylogeny
of Ihe genus. In this group and others, there
is a need for an organized phylogenetic approach to
the evolution of host-marking behavior. In particular,
phylogenetic analyses offer a means of evaluating the
roles of various ecological factors in the origin and
maintenance of host-marking behavior. For example.
It is a truism that MPs are found mainly in insects that
develop as juveniles on resources that are limited in
quantity, such as seeds, fruit, plants or other insects
(Roitberg & Prokopy. 1987). Phylogenetic analysis
would be useful in quantifying the evolutionary gains
and losses in host-marking behavior in relation to
resource limitation, as well as in ruling out other,
correlated ecological factors.
In this regard, a phylogenetic survey of patterns
in host-marking in the tephritid fly genus Rhagoletis
35
represents a beginning. Within the North American
clade of Rhaifoletts. the pattern of host-marking in
one species group, ihe suavis group, differs strikingly
from the pattern in other groups (C. Nutio. D. Papaj.
and H. Alonso-Pimentel. unpubl. data). Whereas
host-marking behavior is present among all members
of at least three other species groups as well as one
unplaced species, host-marking behavior within the
Kiiavis group is spottily distributed. One species in (hat
group. Rhagolelis juglanclis. marks vigorously, but at
least (hree species wi(hin (hat group mark inconsistently
or not at all. Variation among species groups in
host-marking may reflect variation in larval ecology.
Whereas native fruit for most species m other groups
are so small that just one or a few larvae can survive to
pupation in a single fruit, the walnut host fruit used by
all members of the suavis group frequently yield many
pupae. It IS conceivable that some other ecological
factor relating (o life on walnu(s. other (han reduced
resource limitation, accounts for the variable pattern
m host-marking within (he suavis group. Nevertheless,
the pa((ern is in(riguing and deserving of further study.
Another ecological correlate lo consider that may
influence the evolution of marking behavior relates
to the detectability of cues to brood presence. One
might expect the occurrence of host-marking behavior
lo be inversely correlated wi(h Ihe conspicuousness of
non-MP cues (o brood presence such as oviposition
wounds. In Anasrreplw flies, for example, it has been
proposed that the host-marking behavior is correlated
with Ihe degree of latex relea.sed by host fruit during
oviposition which is in turn correlated with deposition
of oviposition within the pulp vs. seed of a fruit (Aluja
et al.. 1999). In A. saggira. whose larvae feed on seeds
of Fnuieria fruit, females oviposit deeply into the fruit,
causing a great deal of late.x to be released, which
could serve as a cue of brood presence. In a closelyrelated
species. .4. serpentina, which feeds on the pulp
of the same fruit species, females oviposit less deeply
inio more-mature fruit, resulting in the release of relatively
little latex. The difference in degree of laie.x
release is associated with a difference in host-marking
behavior whereas .4. serpentina marks the Pouteria
host. .4. saggila does not.
Other ecological correlates worth considering in
a comparative coniext include the ephemerality of
(he host, an insect's host breadth, the paichiness of
the host in space, mobility of the juvenile stages,
and cannibalistic tendencies m the young (Roitberg &
Prokopy. 1987; Diaz-Reischer et al.. 1999).

Role !t Vehrencanip. IWS.
Szamado. I9'W|. In terms ol sheer volume ol el fort
focused on t.he mechanism, function and taxonomic
distribution, work on MP coinmunicalion would seem
to have much to contribute to the held of animal communication.
Many of the key systems studied in .\IP
communication (fruit tlies. bean weevils, panisitoidsi
would seem to be ideal for tests ol theory addressing
a range ol issues in signal evolution, including signal
error, sensory bias, and deception.
Finally, our understanding of both the mechanism
.ind evolution ol hi>st-inarking communication would
benefit greatly trom knowledge of MP chemistry in
both parasitoids and phytophagous in.sects. To date,
few MPs have been identihed. Lack of knowledge of
MP chemistry makes it difficult to construct and evaluate
hyp«)theses of sensory bias or to evaluate the basis
of phyUigenetic patterns m cross-recognition and even
to distinguish between instances in which insects utilize
cues associated w ith brood presence and instances
in which insects utilize an actively-produced signal.
•Acknow ledgcmenls
We acknowledge financial suppt^rt from the Interdisciplinary
Degree Program in In.sect Science, a .National
Science Foundation .Minority Cr;iduate Research Fellowship.
and NRICCP Award No. .'.'>.^I);-IW77. We
thank Judie Bronstein. .Michael Greenfield. Frank
Me.ssina. and Dena Smith for comments on previous
drafts.
Rererenccs
Alphen. J. J- M v;«i H. W. NcU. 19X2. SiipcrparaMii^m and
hir>t discnminaliifn hy A sttixirtt tahuhi i .Mvvi-
36

28S
inac>. 3 larval partsiloiil of Drosophilidae. Netherlands Journal
of Zoology 52; 232-:bO.
Alphen. J. J. M. van & I. Thunntssen. 1983. HOM selection
and sex allocation bv PuLhwirpuiJeus vindrmiaf Rondani
(Piernmalidoe) as a facultative hyperparaMtoid of Asotstra
uibida Nees l Broconidac: Aly^iinae f and L^piopiimti heierutotrut
iCvnipoidea; Eucoilidae». Netherlands Journal of Zoology 33:
497-514.
Alphen, J. J. M. van SL M. E. Visxer. 1990. Supcrpofositism as
an adaptive strategy for insect porasitoids. Annual Review of
Entomolofy 35: 59-79
Alphcn. J J. M. van. M. E. VKscr Jk H. W Nell. 1992- Adaptive
superparasitism and patch time allocation in solitary para-
Hitoids; searching in groups vs sequential patch visits. Functional
Ecology 6: 528-535-
Aluja. M. E F. Bolter. 1992. Host marking phefomone of
Rha^oittts crnut: foraging behavior m rt>pomc (o ^ymheiic
phcromonal isomers. Journal ot Chemical Ecology >8; t299>
1311
AIujx M.. J. Pinero. I. J^ome. F Dta£-Ret.scher & J Siviaski.
1999. Behavior of fruit rties tn the genus Anu-Urepha iTrypetinoc
Tovofrypanimi. In; M. Alu|a & A. L Noabom leds). Fruit Ries
iTephntidaer Phylogeny and Evolution of Behavior CRC Pro.s.
Boca Raton, pp 375—^06.
Arthur. A. P. J. E. R- Stoiner & A. L. Tumbull. I9M The
interaction bcr\fcccn Onfiius ttbscumrur (Ness) (Hymenoptera.
Broconidae) and Temelucha interrupier iGrav.) iHymenoptera;
IchneumonidaeL parasites of the pine shoot moth. Rhvacianiu
butfiuma (SchtfT.i

CUCN. In: J. C. Choc & B. J. Cropi icdsi. The Evolution ot'StKiul
Behavior in IaM»:Ls arui AruchniUs. University Prir^s. CjiiihriU^e.
Uniied Kingdom, pp. 407-442.
Crcdlunil. K K. K. M. DicV. & A. W Wriyhl. l»)86. RclutionNhip between
larval JeaMty. ;)duli Mze and csg production in the cowpea
-iectl beetle. »»mm EcoUv^ical Emomology
II: 41-50
Cmjar. R. M. & R. J. Prukopy. 19S2. Morphological and ciectro>
physiological mapping ot tarsal cheniorecepittfv oi ovipixtiHiin
deterring phen>mone in Hhoyt'tens inmunnrUa tfics. Joumul ol
Inscci Physiology 2S. 395—UX).
Danyk. T. P. & M. Mackaocr. 1993 Diwnmination between selland
cofispecilic-p:iraMti7ed host^ m ihe ;iphid panisitoid Fnum
ffriiiifH/tirum Vicrcck i Hymetioptcni: Aphidiidae). Canadian Enlomologisi
125: 957-964.
Diaz-Fleischer. F. D R. Pupuj. R. J Prokopy SL A. L. Norrb«»m.
19^)9. Evolution ol fruit tly oviposition behavior. In; M. Aluja spccilic marked parasitized hosts by
the solitary parasitoid EfudtitiHani.\ foprtt. Behavtoral Ecology
and .Socii)bioli>gy 30: 77-82.
Ferguson. .\. W.. J. Ztcsinann. M. M. Blight. I. H. Williams. L.
J. Wadhams. S. J. Clark. C. M. Woodcock & A. Mudd, I99«>
Perception of oviposition-deterring pheronuine by cabbage seetl
weevil {Ceuutrhytuhiti ax\iiiuli.%l Journal ol Chemical Ecology
25: 1655-1671).
Ferkovich. S M.. P D. Greany & C Dillard. 1983 ChanccN
m hcmolymph prtneins of the fall urmyworm. SfMnJupteni
fru^tprnfu iJ.E. Smith), avsoctated with parasitism by the braconid
parasitoid Ctue%ui itttir\:tnttftuns »Crcvson». inumal of
Insect Physiology 29: 933-*M2.
Field. S. A. &. M. A. Keller. 1999. Short-term host discnmination in
the parasitoid wasp TnxxoU ux htno/i.x Wollaston »Hymenoptera;
Sceliomdael. Australian Journal ol Zoology 47: 1^28.
Fisher. R. C. A V. K- GancNalingam. 197U. Changes m composuton
ol host haemolymph after attack by an insect parasitoid. Nature
227: 191-193.
Fitt. G. P. 1984. OvipuMtion behavior of two tephntid Iruit ilies.
Oiu us trrniu and Duetts jan-iM. as influenced by the presence ol
larvae in the host fruit. OecnIogia 62: 37-46.
FoUyn. S. St D. Gerling. 198S. The parasitoids of the aleyrodid
tieint.xta iubact in Israel: development, host preference
and discrimination of the Aphelinid wasp ErcnniH rni\ nttwi/u.\
Entoniologia Expenmentalts ct Applicata 38: 255-260.
Ganesalingattt. V. K.. 1972. .Anatomy and histology of sense organs
of ovipositor of Ichneumonid wa.sp. Ori'rrrvr//-283.
Haynes. K. F. & K. V. Yeargan. 1999 ETploitation of intraspecilic
communication systems: illicit signalers and receivers. Annals
of the Entomological Society of America 92: *>60-970.
Heard. T. A.. 1995 Oviposition and feeding prelerrnces ot a rtower
feeding weevil. Ci*el4H eplHtlapuMi la ultutuiu. m relation loconspecilic
damage to its host plant. Entomologia E^perimcntalis ct
Applicata 76. 203-20*)
Hilker. M. & B. Klein. |989 Investigation of ovipositum deterrent
tn larval trass of SiHHtofUent tutnntits (Boisd). Journal iif
Chemical Ecology 15: 929-938.
Hilker. Vt & C Weit/el. I99| Oviposition deterrence by
chemical signals of coaspecihc larvae in Dtprt$in pnu
(Hymenoptera. Dtpnonidaci and FhviltHievtu \ut\:unyMinu
iColeoptera. Chrysomclid;w> Entomologia Generalis 15 29.'^
'01
Holfmeister. T. S. & B. ID Roitberg. 1997 To mark the host or
the patch: OecLsions of a parasitoid searching tor concealed host
larvae. Evolutionary Ecology 11 145-168
Hotfmeistcr. T. S «& B. D Roitberg. 1998. Evolution ol signal
persistence under predator esploitation. Ecoscience 5: 312- »20.
Hoffmeistcr. T S.. B. D Roitberg Jt R.G. Lalonde. 200t) Catching
Ariadne by her thnnid: how a parasitoid exploits ihe herbivore's
marking trails to locate its host. Entomologia EspenmentaliN ct
Applicata 95: 77-85
38

290
Hol'svang. T. 1988. Mechanisms of h(>st discnmioaiion and
iniraspeciHc compeiiiion in the sphid paraMioiil Epheitrmi.tmsiLoia.
Entomologia Expenmeniahs ei Appiicuia 48:
r!9
Hofsvang. T. 1990. Discnmination between unpara2»iiized and parasitized
ho!>tlitary
parasitic wasps. Journal of Animal Ecology 56: 387-401.
Hurler. J.. E. F Boiler. E. Stadler. B. Blattmann« H. R. Buscr.
N- U. Bosshard. L. Oamm. M. W Kozlowski. R. Schoni. F.
Rxschdorf. R. Dahinden. E. Schlumpf. H. Fntz. W. J. Richter &.
J. Schreiher. 1987 Oviposition-detemng pheromone in ff/ruvo
/rn.f certLu L. purification and determination of tte cl^mical
constitution. Expenemia 43: 157-164.
Huth. C. J. &. O Peilmyr. 1999 Yucca moth oviposition and pollination
behavior is affected by past tlower visitors: evidence for a
host-marking pheromone. Oecologia 119: 593-599.
Ikavk-a. T. & H. Oka be. 1985. Regulation of egg number per host to
maximize the reproductive success in the gregarious panuitoid,
Aptmrfles f/nrneranu L (Hymenoptera: Bruonidae). Applied
Entomology and Zoology 20: 331-339.
King. P. E. J Rafai. 1970. Host dLscnminatioo in a gre>
ganous parasitoid yasonia iirnpennts (Walker) (Hymenopcera:
Pleromalidae). Journal of Expenmental Biology 53: 245-254.
KOUIOUSSLS. N. A. & B I. Kat.soyannos. 1991. Host discnmtnarion
and evidence for a host marking pheromone in (he almond
seed wasp. Eurxumut amy tiJuli. Entomologia Expenmentalis et
.Applicata 58: 165-174.
Landnlu P. J.. 1993. Effects of host plant leaf damage on cabbage
looper moth attraction and oviposition. Entomologia Expenmentalis
et Applicata 67: 79-85.
Londolt. p. J. SL A. L Avenll. 1999. Fruit flies. In: J. Bardie & A.
K. Minks tedsi. Pheromones of Non-Lepidopteran lasects As-

Mtcha. S. G- J. Sommcl & C Holler. IW.V 6-Mcihyl-5-hcpicnc-
.1 puiaiivc NC\ jmf Npui.-tii^ phemmune ol ihe aphid
hypcrpuruMluid. Ailoxwui x ntn\ (Hymcnopteru; AliuxyNiidaei
European Journal ol EninnuiiD^y 439-442.
Minkcnbcr?. O P J M.. M. Tauir i J. A. Rosenheim. IW.
Eep loud ON a mufor siiurvc dI vurubilily in in.sect loraginc and
ovipoMiion bchaviur Oikos h5; IM—142.
Mitchell. E- R. & R. R. Heaih, I4H5 (ntUience ol' Ainarunthtn h\hruius
L. atielocheiniCN on ovipoMtiun behavior ot Spinluptem
rtfyiui and Sf*vdttptrFu endtmta (Lcpidoplera: Noctuidoc). Journal
ol Chemical Ecology 11. 609-618.
Mitchell. R . 197.5 Evolution tif oviposiiion lucticv in bcun weevil.
Citlloutbruthti\ mat < F ) Ecology 56; 696--702.
Mudd. A-. R. C. Fi.sher & M. C Smith. I*'82. Volatile hydrocarbons
m the Dufour's eland of the parasite yfinrntis i aneyi rii.> (Gruv i
iMymenopteru: Ichncumonidac). Journal of Chemical Ecology S:
IU35-I(M2.
Nelson. J- M Jk B. D Rottherg. 199^. Factors governing hiv^i
JiNCnmination by Opmi t^Vshmead) i Hymenoptera;
Bnicunidue) Juumut ol ln>cct Behavior 6: 13-24
Okuda. M S i K. V Yearpan. |988 Habitat partitioning by />•
le$jtiiuu.s luniiM and frnuiU u.\ tuMhi.sti (Hymenoptera: Scelionid;>e>
bciN^een herbaceous and wo«xiy host planus. Environmental
Entomology 17: 795-79S.
Oshima. K-. H. Honda &. I. Yamamoto. 1973 iMilation ol an ovipo-
Mtion marker from A/ukt bean weevil. CuUusobruchus chtnrnstx
• L.). Agncullural and Biolocical Chemistry 37: 2679-2680.
Paine. T. D. K. F Raffa & T C Harrtnglon, 1997 Inieractiuns
among scolytid bark beetle>. their asNocintoJ fungi, and live host
contlers Annual Review of Entomology 42: 179-206.
Papaj. D R.. 1994. Ovipnsition site guarding by male walnut llieN
and Its possible consequences lor maimg success. Behavioral
Ecology and Sociobiolugy 54: 187-195
Papaj. D- R. Jt H. AlonstvPimentel. 1997. Why walnut lltes supcrparasui/c:
time savmgs as a pi^ssible explanation. Oecotogia
109: 166-174.
Papaj. D. R. & M. Aluja. 1993 Temporal dynamics of host marking
rn the tropical tephntid tly. AnaUrfpha luttrm Physiological
Entomology 18' 279-2S4.
Pa;xi). D. R. Sc. R. H. Messing. 1996 Functional shifts in the ase
of parasitized hosts by a tephntid lly The role of host quality
Behavioral Ecology 7- 235-242.
Papaj. D R.. B. D. Roitberc & S. B. Opp. 1989. Senal elTecK of
host infestation on egg allocation by the Ntediierranean fruit lly
a rule of thumb and it.s functional MgniHcance. Journal of Animal
Ecolotfv 58: 955-970
Papaj. D. R.. B D Roitberg. S. B. Opp. M. Aluja. R. J Prokopy
A T T. Y. Wong. |9Q0 Effect of marking pheromone on clutch
MZEM the Mediterranean Iruit Hy Physiological Entomology 15
463-i6H.
Papaj. D R.. .A. L. Avenll. R. J- Prokopy & T T. Y. Wong.
IW2- Host-marking pheromone and use ol pre\iou.sly e5iab>
hithed oviposition sites by the Mediterranean fruit fly iDiptera:
Tephntidae). Journal of (asect Behavior 5: 583-598
Papaj. D R.. J. M. Garcia & H. .Moaso-Pimcntel. 1996. Marking
of host truit by male Rhaiittieus Ijincet Cresson llies i Diptera.
Tephntidae» and its effect on egg-lay me. Journal of insect
Behavior 9: 585-598.
Podoler, H. Jfc Z. Mendcf. 1977 Analysis of solitanncs.s tn
a parasite-host -ivstem i\fuu iJtfiiru.\ mprnr. Hymenoptenr
PTeromalidae- Crmiins mpitant, Diptera: Tephntidoci. Ecological
Entomology 2; 153-160
Pncc. P W . 1970. Trail odors - recosnuion hy insects paraMtic on
cocoon.s. Science 170: 546-547
291
Prokopy. R. J.. 1972. Evidence tor a pheromone deterring repeated
uviposition in apple maggot rties. Environmental Entomology I:
326-3.32.
Prokopy. R. J.. 1976. Feeding, malmg. and oviposition activities of
Hhayt'it'm fttuMa (Diptera: Tephntidoe) rtie^ in nature. .AnnaLs of
the Entomological Society of .Amenca 69: 899-904.
Pmkopy. R. J . 1981a. Eptdeictic pherotnono that inlluence spacing
patterns of phytophagous lasectN. In: D.A. Nordlund. R.L.
Jone5. W J. Lewis (eds). Semiochemicals: Their Role m Pe5t
Control. Wiley Pres.s. New York. pp. 181-213.
Prokopy. R. J.. 1981b. Oviposition-deterring pheromone system of
apple maggol Hies. In: E. K. Mitchell (eds). Management of
Insect Peste with Semiochemicals. Plenum Press. New York,
pp. 477-194.
Prokopy. R. J. «& J. Koyama. 1982. Oviposition site partitioning in
Ducu.\ cucurhirae Entomologia Expenmentalis et Applicata 31:
428-J32.
Prokopy. R. J. & D- R. Papa). t999. Behavior ot tlies of the
genera Rhavnietis, Ztmoxttnaut, and Carptmna (Trypetin«;;
Carpomyinai. In: M. Aluja & A. L. Norrbom

29Z
Roitbcfg. B- D it R- G- Laionde. 1991. Hosi marking enhances
poTLMUsm nsW for a truu mfestrng t)v Kha\!t>{ein bmtola. Oikos
til ^89-391
Roiiberg. B. D.

Baker(HynKnnptcni: Aphidiiduel. Cunodinn Eniomnlocist \Z2:
U9-76I.
Wxige. J. K.. h)79. Foraging lur palchily dusinbuted hcnis by ihc
purosiioid. .Vf/fHTim Journal ol" AnitnnI Ecology 48:
*53-^71
Waajic. J. K.. IWS6. Family planmng in parusiioids; adaptive paiicrrtN
ol progeny and sex allocation. In: J. K- Waagc & D. J
Creathead ledM. Uueci ParoMioidN •Vcadeinic Ptcns. London,
pp
Waagc. J. K. & J. A. Lane. 1984. The rcpmductivc Niraiegv of a
parasicic wasp, tl Sex allocation and local male coinpetilton in
Tnthtf\trtinitnii rxone.u rtts. Journal of Animal Ecology 53' 417-
426.
Wainbcre. E.. J Piz/ol A M Bahault. Genetic vanation
in progeny alittcnlion m tru /n>vrtiimmi nuntiix. Entomuiogia
Experimenlalis ct Applicaia 53" 177-187
Wcller, S. J.. N L. Jacoh^on Jfe W E- Conner. I*W>. Theevolulion
ol chemical JelenccN and mating \y>tem\ m tiger muiKs tLcpidoptera
Arcindae). Biological ioumai ot the Lmncan Socieiv
68:557-57S
While. R. A,. M A'^osin. R T. Fmnklm A J W Webb. IMHO
Bark heeiie phcroinone> cvidcncc lor phvMological synihcM>
mechani>m% and (heir ecological implicaiionv. Zeiischril'i fur
Angc>*andce EnioinoU>**»e255-274
Wiley. R. H.. t994. Error, exaggeralion. and deception in antmal
communication. In: L. Real (cd). Behavioral Mcchani.sm.« in
Evolutionary Ecology. University of Chicago Prevs, Chicago.
pp. 157-189.
William.v. K. S L. E Gilbert. 1981. lascctv a.s selective agents on
plant vegciative morphology: egg mimicry reduces egg laying by
bullert1ie%. Science 212; 467-469.
WilNon. K-. 1988. Egg laying decisions by the bean weevil CuUoxo'
hruchtix nuicuUitux- Ecological Efllomology 13: 107-118.
Vainaguchi. H.. 1987. The role of venom m host discrimination of
.\.xcoj(aMer reiit uluitu Wuianabe. Japanese Journal of Applied
Entomology and Zoology } 1: 80-82.
Zimmerman. M.. 1979. Oviposition behavior and ihe existence of
.in oviposiiion-deiemng pherumone m HvlfiftYtt. Environmental
Entomology 8: 2T7~Z79.
Zimmerman. M.. 1980. Selective depoMiion of an ovipositiondeiemne
phcromone by Hxlrtnui. Envirunmentai Entomology
Ul-?24.
Zimmerman. M.. i982. Facuhaiive deposihon t>f an oviposition*
deiemng phen>mone by H\lrm\u. Envtronmenial Entomology
II ^19-522.

APPENDIX B
HOST UTILIZATION BY THE WALNUT FLY,
RHAGOLETIS JUGLANDIS (DIPTERA: TEPHRITIDAE)
43

ENVIRONMENTAL
ENTOMOLOGY
VOLUME 29 OCTOBER 2000 NUMBER 5
PHYSIOLOGICAL AND CHEMICAL ECOLOGY
Influence of Wesiem Com Roocworm (Coleopieri: Chrysomelidae) Larval Injury on Ptioiosynihctic
Rale and Vegetative Grovkih of Different Types of Maize.
M«iuo A. URL«s-L6rrz, LANCE J. MIUNKE, LEON G. HICI.EY, AND FIKRI J. HAII F. 861
Effect of Thermoperiod and Phcxoperiod on Cold Tolerance of Spodoptera exigua
(Lepidoptera: Noctuidae). *
YONCCTRXT«I KIM AND WONME SONC 868
Ntanagement of Osmia tignaria (Hymenoptera; Megachilidae) Populations for Almond Pollination:
Methods to Advance Bee Emergence.
JoMii BOSCH, WILLIAM P. KE.MP, AND STEPHE.N S. PETERSON 874
COMMUNITY AND ECOSYSTEM ECOLOGY
Lepidopteran Communities in Two Forest Ecosystems During the First Gypsy Moth Outbreaks in
Nonhem Michigan.
XtMorm- T. Wonit AND DERORAH G. MCCI LLOVCH 884
Distance-Limited Recoionization of Burned Ceirado by Leaf-Miners and Callers in Central Brazil.
ONILOO J. MARLSI-FILHO 901
Simulation of the Effect of Pollinator Movement on Alfalfa Seed Set.
KAREN STRICKLE* A.M) JOHN WIUMOT VI.NSON 907
POPULATION ECOLOGY
Evaluation of Components of Vegetational Texture for Predicting Azalea Lac^Bug. Stephanitis
pyrioides (Heteroptera: Tingidae). Abundance in Managed Landscapes. *
PAULA M. SHR£%vsRcitv AND MICHAEL J. 919
Effects of Mating on Female LocomoM^ctivity in the Parasitoid Wasp Nasonia vitripennis
(Hymenoptera; Pleromalidae). QQ *
B. H. KM, K. M. GRIMM. A.ND H. E. RENO 927
Developinent and Mona^ of Gceedy Scale (Homopiera: Diaspididae) at Constant
Temperatures. QjQ *
R. H. BLANK, G^.C. GILL, A>O J. M. KELLY 934
A Queen's Worker Attractiveness Influences Her Movement in POlygynous Colonies of the Red
Imported Fire Ant (Hymenoptera; Formicidae) in Response to Adverse Temperatures.
LNDIRA KU*UCHA.N AND S. BIUDLEICH VINSON 943
Ingestion and Defecation of Recombinant and Wild-Type Nucleopolyhedrovinises by Scavenging and
Predatory Aithropods. QQ *
YIHVA LEE AND JAMES R. FCXA 950

Dispersal by Larvae of the Stem Borers Scirpophaga incertulas (Lepidoptera^^lidae) and Chilo
suppreisalix (Lepidoptera: Crambidoe) in Plots of Transplanted Rice. *
MICHAKL B. COHFJS, A.NCELITA M. ROMENA, AND FREO COCLD ..958
Larval Dispersal and Survival of Scirpophaga incertulas (Lepidoptera: Pycalidae) and Chilo suppressatis
(Lepidoptera: Crambidae) on cr>-/.46-transfonned and Non-transgenic Rice.
AHMED M. DIRIE, MICH\EL B. COHEN, AND FRED 972
Distribution and Abundance of Leaf Galling and Foliar Sexual Moiphs of Grape Phylloxera (Hemipteni:
Phylloxeridael and Viiis Species in the Central and Eastern United States. *
DCRCLAS A. DOWME. JEFFREY GR-ANETT, AND JAMES R. FISHER 979
Pea Aphid (Homopiera: Aphididael Fecundity. Rate of Increase, and Within-Plant Distribution
Unaffected by Plant Morphology. *
A.NA LECR.\.NO A.>D PEDRO BARROSA........-.....~~.~..............>....~.~...~.~...~....~~.~~~.~.~~.~.987
Host Utilization by the Walnut Ry. Rhaguletis juglandis (Diptera: Tephritidac). *
CESAR R. Nino, DA.NIEL R. PAPAJ, A.NO HE.NAR ALONSO-PIMENTEL 994
Responses of Ovipositing Moths (Lepidoptera: GeometridaeMoHost Plant Deprivation: Life-Histor\
Aspects and Implications for Population Dynamics. QQ *
TOOMAS TA.MMARL A.NO JL.HA.> JAVOIS.» 1002
Effect of IntraspecificCpmpetition on fVogeny Production of Tonticus piniperda (Coleoptera:
Scolytidae). ^33 *
LAONE AMEZACA AND CARLOS GARBISC 1011
Implications of Larval Motulity at Low Temperatures and High Soil Moistures for Establishment of
Pink Bollworm (Lepidoptera: Gelechiidae) in Southeastern United States Cotton. *
ROBERT C. VE.\ETRE, STEVE.N E. NAR.*.NJO, A.ND W. D. HLTCHISON 1018
PEST MANAGEMENT AND SAMPLING
Seasonal Phenology. Parasitism, and Evaluation of Mowing as a Control Measure for Nezara viridula
(Hemiptera: Pentatomidae) in Australian Pecans.
Overwintering and Comparative Sampling of Seoseiulus fallacis (Acari: Phytoseiidae) on Ornamental
Nursery Plants.
P. D. PRATT A.M> B. A. CROFT — 1034
Time-Concentration Mortality of Pieris brassicae (Lepidoptera: Piendw) and Agroiis segetum
(Lepidoptera: Noctuidae) Larvae from Different D«tiuxins. *
L. THOMSEN AND J. EILENRERC 1041
BIOLOGICAL CONTROL
Effect of Host Plant on Beauveria bassiana- and Paecilomyces fiunosoroseus-lniaced Mortality- of
Trialeurodes vaporariorum (Homopiera: Aleyrodidae).
T. J. PorcAwsKi, S. M. GREENRERC, AND M. A. CIOMFERUK — 1048
Relative Attractiveness of Potra^ Beneficial Insectaiy Plants to Aphidophagous Hoverflies
(Diptera: Syrphidae). *
R. CoiXEY A.ND J. M- LUNA 1054
Thermal Behavior of Two Acridid Species: Effects of Habitatand Season on Body Temperature and
the Potential Impact on Bioconiroi with Pathogens. *
S. BLA.>FORD A.ND M. B. THOMAS 10

Cesar R.. Nufio.
Department of Entomology
Forbes Building 310
University of Arizona
Tucson, AZ 85721
April 23, 2002
Dear Dr. Nufio,
ENTOMOLOGICAL SOCIETY OF AMERICA
The Entomological Society of America (ESA) grants you one-time permission to "include" your
article "Host Utilization by the Walnut Fly, Rhagoletis juglandis (Diptera: Tephritidae)"
published in the Environmental Entomology (29: 994-1001) and co-authored by D. R. Papaj and
H. Alonso-Pimentel as a chapter in your dissertation.
Please include the proper credit for the original work published in the Environmental
Entomology (as a reference in the "Literature Cited" section) or as a footnote.
Best of luck with your dissertation, and we hope you will consider publishing with ESA again.
Sincerely,
fV.
Maggie Meitzler
Managing Editor/Manuscript Editor
46

POPULATION ECOUXTY
Host Utilization by the Walnut Fly, Rhagoletis jugJandis
(Diptera: Tephritidae)
C^SAR R. NUFIO/ DANIEL R. PAPAJ,*- ^ AND HENAR ALONSO-PIMENTEL"- ^
Envifon. Ehtomal 93(S); 9M-1001 (2000)
ABSTRACT BhagaletU jugfandit Cresion is a specialist that deposits its eggs into the husks of
developing walnut fruit. Like other walnut infesting Bies in the R juocit group, A jugtandi* actively
superpaiasitizes its larval hosts. However, little is known regarding the degm to which hosts are
reused and the ecological contest under which host reuse occurs. This field study examined the
pattern of host utilization by KJuglandit and how frait variables such as volume and penetrability
affect the degree that hosts are reused. Fruit on four of five study trees were synchronously infested
and within 2-2.5 wk ail fruit on these trees were infested. Fruit on a fifth tree were significantly less
penetrable than those found among the other trees in the study and this may esplain why fruit on
this tree were larely used thraughout the season. Walnut hosts were commonly multiply infested
and reuse of hosts uccuiied in as few as 1-2 d after first infestion. Infestation levels within fruit
appeared to itahili7C 4-5 d after fruit were first used. Fruit volume was positively correlated with
both the number of punctures on hosts and the infestation leveb within hosts that had been infested
for either 1-2 or 4-9 d. Large fruit were infested more quickly than small fruit, although this trend
was stronger on some trees than others. Finally, despite a size-penetiability correlation among two
of the five trees, penetrability itself did not explain either which fruit were preferentially used
throughout the season or the infestation levels within fruit.
KEY WORDS RhagpUti$juglandis, walnut flies, superparasitisin. marking pheromone, Tephritidae
CHOOSINC WHEBE OFFsnuNc will develop is a simple
form of maternal investment among insects. Oviposita'onai
choices are especially important for insects
whose larval stages are restricted to a particular environment
or host. In such insects, offspring are themselves
limited in their ability to acquire new resources
iftheir natal resources become depleted (Messinaand
Renwick 1985, Smith and LesseUs 1965). Maternal
investment may involve avoiding laying eggi at sites
previously used by conspecifics, a tendency often mediated
by use of a marking pheromone (ftokopy
1981a, Roitberg and Prokopy 1987).
Frugivorous fhiit flies in the fiunily Tephritidae deposit
egg clutches within the husks of devek>ping fhnt
where their larvae are constrained to feed and devek>p.
Females may possess visual and chemical mechanisms
for assessing the quality of available hosts and
for discriminating between previously infested and
uninfested hosts (Prokopy et aL 1976, Prokopy and
Roitberg 1984, Henneman and Papal 19n). Females in
the genus Rhagpletis assess and reject infested fiuit on
the basis of a marking pheromone that is deposited on
the &uit surface after oviposition by previous females.
Thus, marking pheromone in this s]^em is believed to
minimize larval competition by causing females to
' falenllsupUiwy Piugiiiu in bnect Scieiices, Univenity of Arizona.
Tucson. AZ SS721.
'Department of Ecology and Evohilioaaiy Biolaor, Univcisily of
Afizooa. Tucson, AZ SS72L
'Center (or buect Science. Univeisty of Arizona, Tucson. AZ
aS721.
«M6-2ZSX/00/09»4-l(nitQ2J»/0 O 2000 BUomologicJ Society of America
distribute their clutches more uniformly within host
patches than is expected by chance alone (Prokopy
1981a, 1981b; Bauer 1986; Averill and Prokopy 1989).
Rhagoletis jugfandis (Cresson) is a member of the
walnut-infesting fUutgofetis stidois group (Bush 1966).
In southern Arizona this species is found on the Arizona
walnut, Juglans nutfor (Torr.). which can be
fotmd in montane canyons (1,200-2.700 m). These
flies are univohine and females deposit clutches of up
to 30 eggi after puncturing the firuit surface with their
ovipositor. The larval stages feed on the husk of developing
fruit, ptipate in the soil beneath the natal tree,
diapause as pupae through the winter and spring, and
emerge as adults during mid- to late summer. After
deposition of a clutch, female R. juglandis drag their
ovipositors on the firuit siu&ce in a manner suggesting
deposition of a marking pheromone. Despite displaying
the genus-typical marking behavior, female w^ut
flies reinfest and often retise the actual oviposition
sites established by conspecifics (Papaj 1994). Althotigh
superparasitism, the use of hosts that already
bear conspecific brood, is commonly associated with
the lack of available hosts (Roidierg and Mangel 1988,
Papq et aL 1989), walnut flies prefer infested hosts
early in the season when uninfested hosts are still
available (Ladonde and Mangel 1994).
To explain superparasitism by walnut flies, researchers
have proposed that the reuse of the oviposition
sites provides females with direct benefits such
as reduced ovipositor wear (Papaj 1993), reduced
time to deposit clutches (Pap^ and Alonso-Pimentel
47

996 ENVIRONMENTAL
I
1
s
w
0.S •
"8 0.4
§
\ «•*
I
IS 17 It 21 23 28 27 2t 31
Jaly
Fig. I. Proportion of study fruit punctured over die season
(n = 232). By 31 July, all &uit on each of the study trees
werr punctured.
festation levels found in fhiit that had been infested
for either 1-2, 4-5 or 8-9 d.
Results
Although five trees were in the original design, one
tree (labeled A4 and located in lower Garden Canyon)
was eventually eliminated because we observed
few R Juglatdis individuals in mid-July and almost no
flies in this tree during any other census date. By the
end of the season, only 12 fhut. which included tagged
and untagged fruit, were found to be punctured. Even
at the end of the season the fniit on this tree were
smaller (F= 46.6; dr= 4.180; P< 0.0001; Tukey HSD,
P < 0.05 for each of four comparisons) and less penetrable
(F = 38.2; df = 4,70; P < O.OOOl; Tukey HSD.
P 0.05 for each of four
within tree comparisons).
A weak but significant negative correlation between
fhiit volume and fhiit penetrability was found for both
the unpunctured haphazard samples and fhiit punctured
on the fint two census dates within trees A1 and
A2 (regression coeCBcient = -0.0173, t = —481, df =
52, r' = 0.31, P = 0.0004; regression coefficient =
-0.01, t = -2.5, df = 55, r* = 0.10, P = 0.015; trees A1
and A2, respectively). There were no significant differences
between the slopes of the fhiit volumes versus
penetrability data within these trees for the unpunctured
sample versus the recently punctured fhiit
(F = 2.00; df = 1,50; P = 0.79; F= 2.85; df= 1.53;P =
0.10, trees A1 and A2, respectively). There was also no
significant differences in the slopes of the (hiit volumes
venus penetrabih'ty data between trees A1 and
A2 for the unpunctured sample and the recendy punctured
(hiit (F = 0.96;df = 1,92; P = 0.33).To simplify
the graphic representation between penetrability and
fhiit volume within trees A1 and A2, we pooled the
unpunctured haphazard sample and recently punctur^
fruit data from both trees (regression coefficient
= -0.02; »= -S82,df=94.r* = 0.26,P 0.05).
Infestition Level and Fruit Volume. On-estimate of
the average (±SE) dutch size deposited by females
during a singje oviposition event was 15.7 ± 1.5. The
average infestation level of hosts increased with age of
fruit The number of eggs deposited in a single ovqiosttion
event was s^iificantiy less than the infestation levels
found in fruit collected 1-2,4-5. or 8-9 d afW they were
fint infested (F = 33; df= 3,24; P

October 2000 NUFIO ET AL.; HOST UTIUZAHON BY R. Jtigfandis
Al.
A2.
8
I
iB
6
a
:>
2
Cb
22
20
18
16
1> 34 1<
27 22 II
21 23 25 27 30
A3.
A5.
24
ZD
16 1
12
29 14
997
-i#- Punctured Sample
'•&' Unpunctured Sample
^
19 21 23 25 27
Fi^ t. Volume of both Gruit fixim each of the study trees that were punctured and those of a haphazard sample that
remained unpunctured throughout the season. Sample sizes are given for each point (*, significant difference between the
puncturcd and unpunctured fruit volume during a particular census date; Tukey honestly significant difference (HSD) for
within tree comparisons, P < 0.05).
analyses aiui increase the sample size per treatment, was no difference between the infestation levels and
-Rie appropriateness of n posteriori pooling strut- volumes of the day 4-5 and 8'»9 d cohorts suid these
cgf we used was supported by our findings that there latter cohorts had significantly different infestation
levels than their respective 1-2 d cohort (Fig. 4).
(n • Ml (n «t7)
All ar Fnil CMam (4qi)
FnritVoiMM(citf)
F^4. Median number ofeggs deposited during a single
3. Rdationsh^ between fruit volume and penetra- oviposition and the median infestation levels (±SE) as a
bility using pooled data from trees Al and A2. Regression function of fruit cohort age. Bats sharing the same letter are
lines drawn for diagrammatic purposes. not significantly different (Tukey HSD. P < 0.06).
49

FniitV«kraic(ca')
EtrvnoNKCMTAL ENTOMotxxnr
Fi^ 5. Relatioiuhip between fruit volume and infestation
levels. (A) PooM 1-2 d cohorts. (B) Pooled 4-9 d
cohorts. Note the difference in scale between the infestation
levels of the two cohorts. Regression lines drawn for diagrammatic
purposes.
Hereafter, the pooled 1-2 d cohorts and the pooled 4-5
and 8-9 d cohorts wiU be referred to as the 1-2 d cohoits
and 4-9 d cohoits respectively.
We found a positive correlab'on between fhiit vollime
and infestation levels for the 1-2 d and for the 4 -9
d cohorts (r, = 0.384. n = 91, P = 0.0001; r, = 0.60, n =
141, P = 0.0001; 1-2 and 4-9 d cohorts, respectively)
(Fig. 5A and B).
Hie niunber of oviposition punctures was positively
correlated with infestation leveb within the 1-2 d
cohorts (r. = OSt. n = 91, P < 0.0001) and 4-9 d
cohorts (r, = 0.70,n = 141,P16 eggs within 1-2 d
of the initial oviposition event Therefore, many fruit
are reused within 1-2 d of first being attacked.
The exact degree to which the 1-2 d fruit cohoits are
being reused is difficult to calculate because clutch
size varies among females and we have no way of
distinguishing clutches that are laid at the same site.
Our study abo did not address whether females deposit
larger clutches into larger fruit, a process that
could contribute to an early positive relationship between
fruit volume and infestation level Even if females
adjust clutch size to fruit size, our data demonstrates
that reuse of hosts by multiple females must still
be an important fjKtor le^ng to increases in infestatian
lewis. Because the mean infestation leveb of
fruit that remained on the tree 4-9 d were significandy
greater than that estimated for asingle clutch and that
found in fruit which were infested for only 1-2 d.
females had to be reusing many fhiit to some extent.
Our tree censuses showed that all fruit on four of
five trees were infested within 2-2.5 wk (Fig. 1). Fruit
on a fifth tree were virtually untouched. This result
suggests that fruit within a given tree are either nearly
50

October 2000 Nufio et au: Hoct UnuzATiON BY R Jtiglandis QOO
all acceptable or all unacceptable during the flight
season and further that although our sample size is
admittedly limited to Just five trees most trees fall into
the 'all fruit acceptable' category. Finally, our study
finds that 'all fruit acceptable' trees are synchronously
infested and that all fhiit on each of the trees are
infested within 2-2.5 wk.
Fruit Characteristics and Host Utilization. Fruit
volume appears to influence not only fhiit that are
used throughout the season, but also the degree to
which fruit are superparasitized. Our field data show
that in two of four trees, for all but the last census
dates, the mean volume of fruit that were used exceeded
the mean volume of fhiit that remained unpunctured
(Fig. 2). In the remaining two trees, we
found that the mean fhiit volume of recently punctured
fhiit was greater than that of unpunctured fruit
only during the first census. Although not consistent
among trees, it still appears that fhiit volume may
sometimes influence fhiit use.
Fruit volume appears to not only influence which
fhiit are used by walnut flies but also the degree to
which hosts are superparasitized. In our study we
found a positive correlation between fhiit volume and
their respective infestation levels (Fig. 5). This positive
correlation was not only found for fruit that remained
on the tree 4-9 d alter they were first infested
but abo fhiit that had only been infested for 1-2 d.
Although females appear to superparasitize larger
fhiit to a greater extent than smaller fhiit, densitydependent
factors leading to higher offspring mortality
in smaUer fhiit could also explain the relationship
between fruit size and infestation levels. Our measure
of infestation level was the number of eggs and larvae
present within a host. By not being able to count eggs
as soon as they were laid, we may have inadvertently
neglected to count individuals that had hatched but
died as early instars. To address this latter issue, we
conducted a field experiment the following year in
which we specifically defined infestation levels as the
number ofeggs and egg husks (the latter produced by
individuak that had hatched and moved away from
their egg chorion) present at oviposition sites. In this
second study we confirmed our previous finding that
the number of eggs placed within a host is significantly
correlated with its volume (C.RN. and D.R.P., unpublished
data).
During the first two censuses in each of our study
trees we did not find that recently punctured fhiit
were significantly more penetrable than sampled fhiit
that remained unpunctured on the same trees. This
finding suggests that fhnt penetrability was not a factor
that determined which fhiit were preferentially
used by the walnut flies early in the season. However,
akhou^ 6uit ripeness was not found to influence
which fhiit within a tree were selectively punctured,
it does seem to influence which trees will be used by
the walnut flies. In our study, for example, few flies
were surveyed and few walnuts were used in one of
the study trees, a tree that consistendy contained &uit
that uniformly were less penetrable than those found
on other trees. These findings for the walnut fly appear
to be consistent with those found for the apple maggot
fly. This is because although fhiit ripeness influences
which trees are preferentially'used by the apple maggot
flies (Averill and Prolcopy 1989, Murphy et aL
1991), once in a tree these flies may not discriminate
among hosts on the basts of ripeness (Prokopy and
Papaj 1989).
Why Do Flies Superparasitize Their Larval Hosts?
Reuse of walnut hosts by R Jti^andis may be influenced
by three factors. First, reuse of walnut hosts may
be related to their size. Most Rhagoletis species use
relatively small hosts (e.g., hawthorn berries, cherries,
blueberries, and dogwood berries (Bush 1966|) that
appear to offer fewer resources for developing offspring
than do walnut fhiit Studies of R pomonetta
(Walsh), for example, have shown that rarely do more
then three or four pupae emerge from a hawthorn
berry, even when more are deposited within them
(Averill and Prokopy 1987; Feder et aL 1995). In contrast.
it is not unusual for a walnut host to yield several
dozen R jugUmdis or R boycei (Cresson) pupae
(C.R.N. and D.R.P., unpublished data). The ability of
walnut hosts to support greater infestation levels may
also explain why members of the R suaois clade deposit
clutches rather then single eggs at oviposition
sites, the latter being the rule in most other species
within the genus.
Within walnut hosts, variation in size may determine
the degree to which these hosts can be reused
with minimal or acceptable costs associated with larval
competition. Measured as either infestation levels
within a fhiit or the number of punctures on a fruit (a
conservative estimate of host reuse), we found that
larger fhiit were superparasitized more often then
smaller fhiit (Fig. 5). We also found that 4-5 d after
fhiit were initially infested, fhiit were of^en no longer
reused. Thus, we hypothesize that by 4-5 d, infestation
levels reach a point at which the costs of larval
competition, due to a reduction in available larval
resources and age asymmr*"rnger by detecting changes in marking
pheromone concentration, changes in host quah'ty
associated with the presence of conspecific broods
(Fitt 1984), or a combination of the two. Although it
is unlikely that a host's response to being infested
requires 4-5 d to ac-cumulate this process might also
help to explain why females do not reuse hosts previously
infested 4-5 d or more.
If the availability of larval resources within hosts
is important to larval survival or fitness, we might
expect that females would preferentially use larger
fruit. In two of four study trees we did find that
larger fhiit were consistently more heavily attacked
over most census dates (Fig. 2). The preference for
large fhiit does not seem to be explained by a tendency
for large fhiit to be more penetrable: in two
of the four trees in which larger fruit were preferentially
used, we did not find a correlation between
fruit size and penetrability. Furthermore, penetra­
51

1000 ENVIRONMENTAL ENTOMOLOCT VoL 93. no. 5
bility did not explain much of the variation in the
degree to which fruit were reused.
The second factor that may influence the reuse of
hosts by walnut flies is the ephemeral nature of these
larval resources. We propose that because nearly all
walnut hosts within an area will be synchronously
used within 2-2.5 wk, there will be both a spatial and
temporal limit on the total amount of larval resources
available to a population of walnut flies. On
an individual level, this may mean that females are
time limited and must maximize the number of
clutches deposited within the limited window of
larval resource availability. One way to maximize
the number of clutches deposited within the allotted
time may be to superparasitize hosts as they
ripen and become accessible to females. Superparasitizing
hosts to maximize the number of clutches
deposited may again be a viable strategy for walnut
flies because walnut husks can support the development
of more than a few clutches (C.R.N. and
D.R.P., unpublished data).
The third factor that may influence superparasitism
by walnut flies concerns the benefits that females
may gain by not simply reusing a host fruit but
by reusing the actual oviposition punctures created
by previous females. By reusing oviposition punctures,
females may save time (Papaj and Alonso-
Pimentel 1997), decrease the wear to their ovipositors
(Papaj 1993). or gain access to fruit that are
relatively impenetrable (Lalonde and Mangel
1994). lliese benefits have been proposed to increase
the number of clutches a female can lay over
a lifetime. Although not designed to test the benefits
associated with reusing oviposition punctures, our
study suggests that reuse of oviposition sites is common,
with each puncture containing 1.5-1.7
clutches on average. Benefits associated with reuse
of oviposition sites, therefore, may be commonly
experienced by females throughout a season. Still,
although reuse of oviposition sites seemed to be
occurring, it explained only a portion of reuse of a
fruit and the establishment of new oviposition sites
appears to contribute more to total infestation levels
within a host.
Our fieM study was designed to examine the patterns
of host utilization by R ju^andis. The results of
our study suggest that fniit characteristics, namely
fhnt volume and penetrability are important fiKton
that influence host utilization by walnut flies. To understand
how host use patterns emerge it is important
to directly examine female oviposition behavior (van
Lenteren 1981) and how f^ors such as ihiit characteristics
influence the choices females make. Another
important (actor to examine is the use of a potential
marking pheromone in this system. Preliminary field
cage assays suggest that R-jugjiimdis utilizes a marking
pheromone which decreases reuse (CJLN. and
D.ILP.. unpublished data). Future studies will directly
examine female oviposition behavior and how marking
pheromone may influence patterns of host utilization
in the field.
Acknowledgments
We thank Zac Forsman. Laurie-Henoeman. Jessa Netting,
and Dena Smith for comments and discussioii. Sheridan
Stone of the Foil Huachuca Wildlife Management office of
the U.S. Anny provided pennisaon and logiitical support for
field work in Garden Canyon. We alio thank T. L. Lysyk and
an anonymous reviewer for their critical reviews and suggestions.
Research was supported by NRICX^F grant no. 93-
37302-9126 to DJLP.
References Cited
Avcrill, A.U,andR.J. Prokopy. 1987. Intraspecific competition
in the tephritidfhiit fly, RAiigobtispoiiunieUa. Ecology
68; 878-886.
AveriIl,A.L.,andR.J.Prokopy. 1989. Distribution patterns
of RhagpUttf pomoneUa (Diptera: Tephritidae) eggs in
hawthorn. Ann. EntomoL Soc. Am. 82: 38-44.
Bauer, C. 1986. life-history strategy of AAogDlctft ditCTTMUa
(Diptera: Trypetidae), a fhiit fly operating in a 'noninteractive'
system. |. Anim. EcoL 55: 785-794.
Bush, G. L. 1966. The taxonomy, cytology and evolution of
the genus RhagoUtis in North America (Diptera. Tephritidae).
Butt. Mus. Comp. ZooL Harv. Univ. 134- 431-
562.
Feder, J. L, K. Beyndds, W. Go, and E. C. Wang. 1995.
Intra- and interspecific competition and host race formation
in the apple maggot fly. RhagoUtis pomontOa
(Diptenu Tephritidae). Oecologia 101: 416-4^.
Fitt, G. P. 1984. Oviposition behavior of two tephritid
fruit Bies. Daaa tryoni and Daaa Jaroisi, as influenced
by the presence of larvae in the host fruit. Oecologia 62:
37-46.
Henneman, M. L., and D. B. Papa|. 1999. Role of host fhiit
color in the behavior of the w^ut fly Rhagaletit fugUtndi$.
EntomoL Exp. AppL 93: 249-258.
Lalonde, B. G., and KL Mangel. 1994. Seasonal effects on
supeipafasitism by RhagpleUs completa. J. Anim. EcoL 63:
583-588.
Lentcren, C., van. 1961. Host discrimination by parasitoids.
pp. 153-179. fn D. A. Nordlund, B. L. Jones, and
W. J. Lewis [eds.|. Semiochemicals: their role in pest
controL Wiley. New York.
Messina, F. J., and J.A.A. Benwick. 1985. Ability of ovipositing
seed beetles to discriminate between seeds with
differing egg loads. EcoL EntomoL 10; 225-230.
Murphy,B.C.,L.T.Wilson,andB.V.D0WCII. 1991. Quantifying
apple maggot (Diptera: Tephritidae) preference
for apples to optimize the distribution of traps among
trees. Environ. EntomoL 20:981-987.
Pkp^D.B. 1983. Use and avoidance ofocctipied hosts as a
dynamic process in tephritid fruit flies, pp. 25-46. M E. A.
Betnays [ed.|. Insect-plant interactions. voL 5. CRC,
Boca Baton. ^
Pait^D.B. 1994. Oviposition-site guarding by male walnut
flies and its possible consequences for mating success.
Behav. EcoL SociobioL 34:187-195.
Papq, D. B., and B. Alonso-PimentcL 1997. Why walnut
flies superparasitize; time savings as a possible explanation.
Oecologia 109: 166-174.
Papaj, D. B., B. D. Boitbcrit and S. B. Opp. 1989. Serial
effects of host infestation on egg allocation by the Mediterranean
fruit fly; a rule of thumb and its functional
significance. |. Anim. EcoL 58: 955-970.
Prokopy, B. J. 1981a. Epideitic pheromones that influence
spacing patterns of phytophagous insects, pp. 181-21X fn
D. A. Nordlund. R. ll Jones, and W. J. Lewis. (eds.|.
52

October 2000 NUFIO ET AL4 HOST UIHIZAIION BV R ju^andis 1001
Semiochemicab. their role in pest control Wiley, New
York.
Proka|ty,ILJ. 1981b. Oviposition-detenring pheromone system
of the apple maggot flies, pp. 477-497. fo E. R. Mit^eU
(ed.|. Management of insect pests with semiochemicals.
PImum. New York.
Prakopy.lLJ.,aii4lD.R.Pivaj. 1988. Can ovipositing iUiag»letti
pamoneUa females (Diptera: Tephritidae) learn to
discriminate among different ripeness stages of the same
host biotype? Fla. Entomol 72: 489-494.
Prokopy R. J., and B. D. Roitberg. 1984. Foraging behavior
of true ftiiit flies Am. Sci. 72: 41-49.
Prokopy, R. J., J. R.Zieg|er, and T.T.Y. Wong. 1978. Deterrence
of repeated oviposition by ihiit marking pheromone
in CemlUit coptfatt (Diptera: Tephritidae).
J. Chem. EcoL 4:55-63.
Praioipy R. J., W.H. Retain, and V.Moericke. 197S. Marking
pheromones deterring repeated oviposition in RfcogoleHt
flies. Entomol Eip. AppL 20:170-178.
Smith R. B.,andC.M. 1985. Oviposition. ovicide,
and larval competition in gnnivorous insects, pp. 423-
448. Ill R. M. Sibly and R. H. Smith {eds.|. Behavioral
ecology: ecoloKical consequences of adaptive behaviour.
Blackwell. Oxford.
Roitheif B. D., and M. ManceL 1988. On the evohitionacy
ecology of marking pheromones. EvoL Ecol 2: 289-315.
Roitberg B. D., and R. J. Prokopy. 1987. Insects that mark
host plants. Bioscience 37: 4QO-406.
Reoeified ybr publication 26 May 1989; aatpted 20 June
2000.
53

APPENDIX C
REUSE OF LARVAL HOSTS BY THE WALNUT FLY,
RHAGOLETIS JUGLANDIS, AND ITS IMPACTS FOR
FEMALE AND OFFSPRING PERFORMANCE

Reuse of larval hosts by the walnut fly, Rhagoletis jugUmdis^ and its
Implications for female and offspring performance
Cesar R. Nufio'-*
and
Daniel R. Papaj""^
Department of Entomology', Department of Ecology and Evolutionary Biology", Center
for Insect Science^, and the Interdisciplinary Degree Program in Insect Science"^,
University of Arizona, Tucson, AZ 85721, USA
55

ABSTRACT
The oviposition-preference-offspring-perfonnance hypothesis states that female insects
should prefer to deposit clutches on or in hosts that maximize offspring performance.
An important assumption of this hypothesis is that female and offspring fitness are
tightly correlated. In this study, we evaluate offspring performance in the walnut fly,
Rhagoletis juglandis Cresson (Diptera: Tephritidae), in relation to a previously
described oviposition preference for previously exploited host fruit. In particular, we
examined how reuse of walnut hosts influences offspring survival and weight at
pupation under Held conditions. We found that reuse of walnut fruit increased larval
densities within fruit, and that these increases were associated with reduced larval
survival and weight at pupation. In a laboratory experiment, pupal weight was
positively correlated with adult female size and lifetime fecundity. In this system,
oviposition preference is therefore negatively, not positively, correlated with offspring
performance. We argue that pattems of female preference in this system reflect direct
benefits to females that are traded ol^ against costs in terms of offspring fitness.
Because female fitness is a product not only a function of offspring quality but also of
total number of offspring produced, female walnut flies may be optimizing their fimess
by producing many less fecund offspring. We hope that this study encourages future
studies examining the preference-performance to consider reproductive conflicts
between parents and offspring as potential mechanisms that influence the congruence
between parental preference and offspring performance.
56

KEYWORDS
Life history" marking pheromone' offspring fimess oviposition-preference-offspring-
performance" parent-offspring conflict" reproductive trade-offs Rhagoletis juglandis'
Tephritidae walnut flies
57

INTRODUCTION
The oviposition-preference-offspring-performance hypothesis was proposed to explain
patterns of host specificity in herbivorous insects (Jaenike 1978, Thompson 1988a,
Mayhew 1997). The hypothesis states that, in insects that utilize discrete host resources
or environments, and in which progeny are limited in their ability to disperse to other
hosts, females should be under strong selection to choose hosts that are optimal for
larval development. As a result of such selection, a female insect's oviposition
preference is expected to correspond to patterns of host suitability that optimize larval
performance.
In support of the preference-performance hypothesis, a good correspondence
between female preference and offspring performance has been found in some systems
(Copp and Davenport 1978, Wiklund 1981, Williams 1983, Leather 1985, Thompson
1988b, Craig et al. 1989, Rossi and Strong 1991, Price and Ohgushi 199S, Nylin and
Janz 1996). However, in many other systems, the correspondence has been found to be
poor or nonexistent (Messina 1982, Karban and Courmey 1987, Valladares and Lawton
1991, Fox 1993, Larsson et al. 1995, rev. Mayhew 1997).
A failure to find a strong correspondence between female preference patterns
and offspring performance has sometimes been attributed to physiological constraints
that prevent females from making optimal choices. For instance, ovipositing females
may simply be limited, at a sensory level, in their capacity to discriminate between
juvenile-suitable and unsuitable hosts (rev. Courmey and Kibota 1990, Bemays 2(X)1).
58

Alternatively, a weak correlation between preference and performance may reflect an
incomplete measure of performance. For example, hosts associated with high juvenile
growth and survival under controlled lab conditions may be associated with high levels
of juvenile predation or parasitism in the field (Gratton and Welter 1999). Analogously,
the correspondence between female oviposition preference and juvenile performance
may be influenced by the occurrence of competition among juveniles for host resources
(Hanks et al. 1993, Ekbom 1998). In many insect-host systems, the deposition of eggs
on a host reduces the quality of that host in terms of future oviposition owing to
potential competition among conspeciflcs. Many females recognize the presence of
juveniles on hosts and avoid laying eggs on such hosts (Nuflo and Papaj 2001). Hence,
as females lay more and more eggs on an otherwise-optimal host, there will be
increasing incentive, from the standpoint of larval performance, to allocate eggs to
alternative hosts (McMillin and Wagner 1998, Cronin and Abrahamson 1999). A
correlation between preference and performance might be weak or even non-existent.
Less often acknowledged in the insect-host literature, particularly in the context
of juvenile competition, is the possibility that oviposition patterns that maximize
parental fltness are not the same as those that maximize the fltness of any one offspring
(rev. Mayhew 1997, Scheirs and DeBruyn 2002). A female may potentially increase
her fitness by increasing the number of offspring she produces over a lifetime (either by
increasing her reproductive life span or rate of oviposition), even at the expense of some
decrease in offspring fitness. For example, juvenile-optimal hosts may be less common
than other hosts (Williams 1983, Etges and Heed 1987, Mayhew 1997) or are associated
with higher female mortality (Weiser et al. 1994); in such cases, a female may opt to
59

search selectively for a juvenile-suboptimal host because it permits her to increase the
number of eggs laid over her lifetime (Rausher 1980). Similarly, juvenile-optimal hosts
may be associated with higher offspring predation (Price et al. 1980, Denno et al. 1990,
Yamaga and Ohgushi 1999, Ballabeni et al. 2001). So long as the gains that a female
accrues in terms of numbers of offspring produced more than offset her losses in terms
of per capita offspring performance, her decisions may be optimal for herself, yet in
conflict with the interests of any one offspring. In certain situations, the conflict
between parent and offspring interests can even generate a negative correlation between
female preference and offspring fitness (Schiers et al. 2000).
In this paper, we evaluate the correspondence between preference and
performance in a frugivorous fly in which a peculiar pattern of female preference has
been described previously, but in which offspring performance has not been previously
evaluated. The walnut fly, Rhagoletis juglandis, the focus of this study, is a specialist
tephritid species that utilizes the husks of developing walnut fruit as a larval resource.
Like many other tephritid flies, R. juglandis engages in host-marking behavior
following the deposition of a clutch of eggs into a fruit. Paradoxically, while marking
pheromones in Rhagoletis species that use other hosts cause arriving females to reject
occupied hosts in order to minimize larval competition (reviewed in Prokopy 1981,
Landolt and Averill 1999), R. juglandis and other walnut-infesting Rhagoletis species
conmionly reuse hosts in the field, often depositing eggs into existing oviposition
cavities (Papaj 1994, Nufio et al. 2000). Despite possible costs associated with reusing
larval hosts (in terms of increased larval competition), walnut flies prefer to lay eggs in
60

previously attacked hosts early in the season when unattacked hosts are still available
(Lalonde and Mangel 1994, Nufio et al. 2000).
If larvae are competing within host fruit, as they do in congeneric species, R.
juglandis' early-season preference for infested fruit would likely be negatively
correlated with offspring performance. Alternatively, it is possible that larvae do not
compete within fhiit but gain some advantage from being deposited together (for
example, an advantage due to a sharing of microbial symbionts thought to generate the
rot on which larvae feed; Howard et al. 1985, Howard and Bush 1989). In that case, the
correlation between preference and performance might be positive, at least over a range
of larval densities. In this study, we quantify Held patterns of host utilization by the
walnut fly, R. juglandis (Tephritidae), and examine offspring performance in relation to
the females' unusual preference for previously exploited fruit.
61

MATERIALS AND METHODS
Natural History
R. juglandis is a member of the walnut-infesting R. suavis group (Bush 1966). In
southern Arizona, this species is found on the Arizona walnut, Juglans major Torr,
which is conunon in montane canyons between 1200 and 2700 meters. These flies are
univoltine and females deposit clutches of ca. 16 eggs (+S.E.I.5) (Nufio et al. 2000)
after puncturing the fruit surface with their ovipositor. The larval stages feed on the
husk of developing fruit, pupate in the soil beneath the natal tree, diapause as pupae
through the winter and spring and emerge as adults in mid to late summer.
Field patterns of host reuse and offspring fitness
In mid-June, 1996, five Juglans major trees in lower Garden Canyon (ISOOm altitude)
in Cochise County, Arizona were selected for study. Trees were selected for their
relatively large fruit yield, and the accessibility of the fruit to manipulation from ground
level. On each tree, twenty five to sixty fruit that were accessible from ground level
were chosen and tagged for the study. The fruit from a given tree used in this study
constituted roughly 2 to 5% of the total fruit yield of that tree.
Walnut flies were first observed on a study tree on 2 July. Every few days
thereafter, fruit were censused for the presence of oviposition punctures, which are
created by females when depositing clutches within their hosts. After 8 July, when the
first punctures were observed, study trees were censused every two days as follows.
From 0900 to 1100 hours, tagged firuit within each tree were examined for signs of
62

walnut fly oviposition punctures. The minimum and maximum axes of each of the
recently punctured fruit were recorded with digital calipers. These measurements were
later used to estimate the volume of a given walnut, by assuming a walnut was
spherical, taking the average of the axes measurements as an estimate of sphere
diameter, and then computing fruit volume as 4/3 k r^.
To measure the impact of host reuse on offspring fimess, we manipulated the
period of time over which cohorts of infested fruit were exposed to additional
oviposition. At the end of a census day, newly-infested firuit were grouped according to
size and location and then haphazardly assigned to one of three treatments. Fruit in the
first treatment were bagged inmiediately with thin bridal veil, and thus represented fruit
that had been infested that day and or the previous day and thus was considered exposed
to female reuse for 1-2 days. Fruit placed into the second and third treatments were
bagged 2 and 4 days later and thus represented fruit that had been infested and exposed
to female reuse for 3-4 and 5-6 days, respectively.
Assessing the precise number of ovipositions in a host was not possible in our
study because females add clutches to existing oviposition cavities and it was not
possible to distinguish individual clutches deposited within a single cavity. We instead
used total number of eggs deposited within a host as an indication of the degree to
which hosts were reused. We estimated the number of eggs deposited in a fruit as
follows. On each census date, all study fruit not previously bagged were examined, and
new punctures on previously infested fruit were circled and dated. After a given
oviposition site was at least 6-7 days old, it was removed by excavating a cylinder
around the puncture, roughly 6mm long and 8mm wide. Because larvae larvae move
63

towards the seed and away from the area surrounding their oviposition sites shortly after
hatching, only very rarely were larvae found or noticeably damaged while making this
excavation. To keep the fruit from drying out and larvae from prematurely leaving the
host fruit, the space previously occupied by the oviposition cylinder was covered with a
piece of parafilm over which was placed a 15 by 20 mm strip of black electrical tape.
Excavated oviposition cavities were placed in vials with alcohol and brought to the lab
where they were dissected and the hatched and unhatched eggs counted. The number of
hatched eggs was used as an estimate of the number of larvae initially present within a
fruit.
Ten days after a given study fruit was initially punctured on a tree, it was
removed and brought to the lab and placed individually into an 'incubator' and stored in
a growth chamber set at a constant temperature of 30°C and 50% humidity. The
incubators were 48 ml plastic cups with plastic petri dishes inserted onto their wide
tops. The original bottoms of the cups were then removed and cups were then inverted.
Infested fruit were placed within the cups, resting on a 4 cm long by 3.5 cm wide PVC
tubing inserted into a 3 cm deep bed of mixed vermiculite and sand. The vermiculite/
sand layer was periodically kept moist by adding water until Day 15, when larvae began
to emerge from the fruit. Fruit were placed on PVC tubing to keep them from becoming
moldy and from absorbing water. The live and dead pupae associated with a fruit were
counted, and live pupae weighed.
64

Effects of offspring weight on adult reproductive potential
Pupal weight and adult size
Pupae used in this experiment were subsets of those that successfully pupated during
the 1996 field experiment. Pupae were grouped into one of 6 weight classes ranging
from 0.20-0.31 to 1.01-1.10 mg. All members of a given weight class were placed into
one of six rearing cups. After adults emerged, they were sexed. Using a dissecting
microscope and an ocular micrometer, adult size was estimated by measuring the length
of the medial vein bordering the anterior portion of the discal medial cell. This wing
measure could be made rapidly and previous laboratory experiments indicated that this
measure was strongly correlated with other indicators of female size such as thorax
length, head width and femur length. Adult discal medial cell length was regressed
against pupal weight for each sex separately.
Female size and reproductive success
We conducted the following laboratory experiment to examine the relationship between
female size and lifetime fecundity. Adult flies used in each of two replicates of the
experiment, which were conducted in separate years, were collected as larvae from fruit
collected in Garden Canyon in the Huachuca Mountains in southern Arizona. These
pupae from which the adults in the study emerged were collected from the field one
year prior to their emergence in the lab and stored in daricness at 4C until needed. Ca. 4-
5 weeks prior to the experiment, pupae were removed from cold storage and warmed to
28 C under a 14; 10 h lightrdark cycle. Under such conditions, adults typically emerge
4-6.5 weeks later. Approximately 24 hours after emergence, a female and male were
65

placed into a clear 16-fl oz (473-ml) plastic Solo cup, fitted with a petri dish lid, in
which they were provided with ad libitum water, sugar and a yeast hydolysate and sugar
mixture. If the male died before the female, he was replaced with a reproductively
mature male that was 10 - 20 days post-emergence. Fly pairs placed into the rearing
cups were stored in a room with a 14: lOh light:dark cycle and a day temperature of 28
C.
Starting on Day 1, when females were first placed into the rearing cups, a ripe
walnut fruit hung from the top of the cup. Walnut fhiit were collected from a variety of
localities in southern Arizona and refrigerated for up to two weeks prior to use in the
experiment. All fruit provided to females bore 4 punctures made with a no. 00 insect
pin, placed equal distances apart on the fruit surface. Female walnut flies are known to
reuse oviposition punctures in the field, to an extent probably determined by fruit
hardness and fly size; our artificial punctures thus allowed gravid females access to host
fruit regardless of fly size and fruit hardness. While females in this experiment actively
oviposited into the artificial punctures provided, females in this experiment also
conunonly deposited clutches into female made oviposition cavities.
Every two days, a new fhiit replaced the older fruit. The old fhiit was dissected
and the eggs deposited within the host over the previous two days were counted. After
an experimental female died, her size was estimated under a dissecting microscope by
measuring the length of the discal medial cell as above.
Female size was regressed against the number of days that she lived, the number
of days until she laid her first clutch, the size of her first clutch, the total number of eggs
66

she deposited over a lifetime and the average number of eggs deposited per day that she
was alive. We also explored patterns of adult survival over time.
Statistical Analysis
In order to identify factors that influence % total survival (from egg deposited to
successful pupation), we conducted a multiple regression analysis wi± % total survival
as the dependent variable and fruit cohort, fruit volume and the number of eggs
deposited within a ^it as potential independent variables. We conducted analogous
analyses for the survival of specific life stages but we also included the number of eggs
that hatched and the number of larvae that emerged from a fhiit as dependent variable
under the appropriate life stage. Multiple regression models used were forward iteration
models with p-value thresholds for entering a variable set at 0.25 and thresholds for
retaining a variable set at 0.10.
Percent data were arcsine-transformed to meet normality assumptions. In no
case did inferences resulting from such analyses differ from those resulting from
regression analyses of the continuous untransformed variable. For consistency, we
report only the regression analyses for all continuous variables. The stage-specific
sxirvival data for % egg hatch were, however, bimodal, with one mode at 100%. For that
variable, we generated a new nominal variable that took a value of 1 when the variable
was 100% and 0 when it was not. We then conducted logistic regression analyses on
the nominal variable.
67

In order to identify factors that influenced pupal weight, we conducted a
multiple regression analysis with pupal weight as the dependent variable and number of
larvae emerging from a fruit, fruit volume, fruit cohort (number of days fruit was
available for reuse), the number of larvae that hatched within a fruit, and the number of
eggs deposited within a fruit as independent variables. I^ipal weights were compared
among the 1-2, 3-4 and 5-6 day cohorts and were regressed against both their respective
fruit volume per hatched egg and fruit volume per larva that emerged from the fruit.
We examined the relationship between pupal weight and adult size by use of a
linear regression model. Separate analyses of variances were used to determine whether
there were differences between female size, lifespan, and number of eggs produced by
the two cohorts used in the lifetime fecundity study. We also conducted multiple
independent regression analyses to examine the relationship between the independent
variable female size and the dependent variables days till first clutch was deposited, size
of first clutch, total eggs deposited, lifespan, and number of eggs produced per day.
All statistical analyses were conducted with JMP-IN statistical software for
Macintosh and IBM platforms (JMP-IN version 4, 1989-2(XX)).
68

RESULTS
Levels of host reuse in the field
That egg density is a reasonable measure of host reuse is suggested by the pattern of egg
density for fruit exposed for varying lengths of time since first infestations (Fig. I). The
number of eggs in a fruit clearly increased as the duration over which it was exposed to
females increased; on average, roughly 20 new eggs were placed into fruit every two
days.
The number of eggs deposited within a fruit was positively correlated with fruit
volume for each cohort (r, = 0.45, n = 56, P = 0.0004; r, = 0.61, n = 61, P < 0.0001; r, =
0.57, n = 46, P < 0.0001; 1-2, 3-4 and 5-6 d cohorts, respectively; Fig. 2). Because more
eggs were placed into larger fruit, available fruit volume per egg was not significantly
different for offspring placed into small versus large fruit (r^ = -0.005, n = 156, P =
0.95). The number of oviposition punctures, a conservative estimate of reuse levels, was
also positively correlated with fruit volume within each of the cohorts (r, = 0.33, n = 57,
P = 0.012; r, = 0.65, n = 55, P < O.OOOl; r, = 0.89, n = 42, P < 0.0001; for 1-2,4-5 and
6-7 d cohorts, respectively). The median number of punctures found on a fruit was 1 ~
0.11,2 " 0.15 and 3 ~ 0.21 (X + 1 S.E.) for the 1-2, 3 '4 and 5-6 d cohorts respectively.
The number of eggs deposited within a fruit was in turn correlated with the number of
punctures on a fruit within each of the cohorts (r, = 0.78, n = 57, P < O.OOOl; = 0.81, n
= 54, P < 0.0001; r, = 0.88, n = 46, P < 0.0001; 1-2, 3-4 and 4-6 d cohorts, respectively).
On average, each oviposition puncture was associated with a net increase of 26 ' 7.6
69

eggs. As noted above, how many clutches are represented by this increase in egg
numbers could not be determined in this study, however, a previous study did find that
on average this walnut fly deposits clutches of ca. 16 eggs in the field (Nufio et al.
2000). If average clutch sizes are consistent between years, each oviposition puncture in
our study contained ca. 1.6 clutches.
Host reuse and offspring survival
Total survival
Offspring survival declined with increases in the number of eggs deposited within a
fruit and increased as fruit volume per egg increased (Figure 3 A & B). On average, 52 ±
0.02% of eggs deposited into a fruit hatched, emerged from the fruit and successfully
pupated (number of fruit = 156). A multiple regression analysis indicated that the
number of eggs deposited within a fruit and finit volume were significant predictors of
the percentage of eggs surviving to pupation (% total survival), while fruit cohort was
not (Table 1). Percent total survival declined with egg density (r* = 0.30, F = 65.35, df =
1, 156, P < 0.0001, slope different from zero) (Figure 3A) and increased with fruit
volume (Table 1).
Percent total survival was positively related to the amount of available fruit
volume per deposited egg (r* = 0.38, F = 94.77, df = 1, 156, P < O.OOOl, slope different
from zero t = 9.74) (Figure 3B). The number of punctures in a fruit (a conservative
estimate of the number of clutches placed within a fruit) was also negatively related to
total offspring survival (r' = 0.22, F = 44.03, df = I, 156, P < 0.0001, slope different
from zero, t = -6.64).
70

Survival from egg to egg hatch
Of the eggs deposited into the 156 fruit, on average 88% + 0.16 of the eggs hatched. A
multi-factorial logistic model using egg hatch as a nominal response (either 100% egg
hatched or not) found that egg hatch within a fruit was negatively related to the number
of eggs deposited within a fruit, but that fruit cohort and fruit volume did not account
significantly for variation in % hatch (Table 2). We sought to determine if the
relationship between % hatch and egg density, where analyzed on a per-fruit basis, was
also found on a per-oviposition puncture basis. For this analysis we used only fruit with
a single oviposition puncture to attempt to control as much as possible for variables
such as the temporal spacing between clutches that might alter density affects. For fruit
with a single puncture, there was again an inverse relationship between percent egg
hatch and the number of eggs within the puncture (logistic regression, r* = 0.06, X' =
4.99, df = I, 66, P = 0.03).
Survival from egg hatch to larvae emerging from a fruit
Of the eggs that hatched 79% + 0.02 (n = 156 fruit) of these larvae emerged from the
host fruit. Offspring survival from egg hatch to emergence from a friiit was negatively
correlated with egg density, but positively correlated with fruit volume (Table 1).
Neither the number of eggs that hatched within a fruit nor fruit cohort provided further
explanatory power for the number larval survival. Offspring survival from egg hatch to
emergence from a fruit did not peak at intermediate egg densities.
71

Successful pupation
Finally, of the larvae that emerged from a fruit, a mean 77% + 0.02 (n = 184 fruit)
successfully pupated. The number of individuals that emerged from a fruit and fruit
volume best explained the percent of emerging larvae that survived to pupate
successfully. Percent larval survival to pupation was positively related to fruit volume
but negatively related to the number of larvae emerging from a fruit (Table 2). One
possible interpretation of the latter result is that many larvae emerging may have
interacted and impacted their chances to successfully pupate in the relatively small
cups. The number of eggs deposited within a fruit, number of eggs that hatched and
fruit cohort did not further explain larval emergence to pupal survival.
Host reuse and pupal weight
A multiple regression analysis showed that the median weight of viable pupae that
emerged from a single fruit was best explained by the number of larvae that emerged
from a fruit, fruit volume and fruit cohort (Table 1). Neither the number of eggs
deposited within a fhiit nor the number of eggs that hatched within a fruit improved the
model's explanatory power. In general, larvae emerging from fruit that were only
exposed to female reuse for 1-2 d weighed significantly more than individuals from
fruit that were exposed to further reuse for 3-4 and 5-6 d (F = 11.63, df = 2, 174, P <
0.0001) (Figure 4). The negative impact of larval infestation level on offspring weight,
however, appears to be reduced when brood are placed into larger fruit (Table 3).
Independent of the number of individuals placed into a fruit, fruit volume was
72

significantly correlated with pupal mass (r = 0.06, F = 11.84, df = 1, 175, P = 0.0007,
slope different from zero, t = 3.44).
While fruit volume and the number of larvae within a fruit independently
affected pupal weight, they also have an interactive effect. Median pupal weight
declined as a function of both fruit volume per egg hatched and fruit volume per
emerging larva (Figure 5). Variation in pupal weight is however, better explained by
fruit volume per emerging larva (2*^ power polynomial regression, r' = 0.44, F = 68.32,
df = 2, 174, P < 0.0001) than by fruit volume per hatched egg (2* power polynomial
regression, r" = 0.32, F = 35.15, df = 2, 148, P

eplicates, but not the other, larger females laid their first clutch significantly earlier
than did smaller females (Table 5).
In both the 1997 and 1999 replicates, there was a significant relationship
between our index of body size and the size of the flrst deposited clutch, with relatively
larger females laying relatively larger first clutches (Table 5). Furthermore, in both
replicates larger females also deposited significantly more eggs than did smaller ones (r'
= 0.17, F = 6.63, df = 1, 34, P = 0.02, slope different from zero, t = 2.26; r" = 0.29 F =
19.7, df = 1,50, P = 0.0001, slope different from zero, t = 4.44; the 1997 and 1999
replicates, respectively) (Table 5, Figure 9). Larger females appeared to lay more eggs
during the course of these experiments not because they lived longer than smaller
females, but because they deposited significantly more eggs per day (Table 5).
74

DISCUSSION
While the femaie-preference-offspring-performance hypothesis was initially designed to
evaluate factors that affect host ranges in phytophagous insects (Jaenike 1978,
Thompson 1988a, Mayhew 1997), this hypothesis addresses a fundamental issue
regarding how female egg-laying decisions are expected to evolve with respect to
offspring performance. In the last decade or so, for example, researchers have began to
examine how female preference for given host varieties (Thompson et al. 199S, Ahman
et al. 2000), genotypes (Larsson et al. 1995, Cronin and Abrahamson 2001, Fujiyama
and Katakura 2001, Harris et al. 2001) or hosts of a given species that vary in some
potentially discemable way impact offspring performance (Craig et al. 1989, Bjorkman
et al. 1997, Sweeney and Quiring 1998, Steinbauer 1999, Leyva et al. 2000, Wilson and
Faeth 2001).
Given an ability to discriminate and ample evolutionary time to adjust to the
novel hosts within their range, it is reasonable to expect that mothers would do best if
they preferred only hosts that maximize their per offspring fitness. However, models of
parent-offspring conflict, progeny size-number trade-offs and even optimal foraging
predict that, under a variety of conditions, females should devalue offspring fimess if
such behavior increases their own reproductive success. In insects, such models are
used primarily to understand the egg laying decisions made by parasitoids and seed
predators (Chamov and Skinner 1985, Smith and Lessells 1985, Godfiray 1987, Spiers
et al. 1991, Mayhew 1998) and generally not acknowledged in empirical studies of
75

insect-plant systems in which the preference-performance hypothesis has been most
often addressed (but see Rausher 1980, Larsson and Ekbom 1995, Nylin et al. 1996,
Scheirs and De Bruyn 2002).
Our present results indicate that female walnut flies commonly deposit multiple
clutches into larval hosts and host reuse negatively impacts offspring performance,
measured in terms of offspring survival and weight at pupation. The negative ejects of
reuse on offspring weight at maturation were further shown in laboratory assays to
translate into a reduction in expected female lifetime fecundity. Because even slight
increases in infestation levels appear to negatively affect offspring size and weight at
pupation, the walnut fly female's previously-described preference for previously
exploited hosts appears to run counter to expectations of the preference-performance
hypothesis. Whereas the preference-performance hypothesis would predict that female
walnut flies should avoid reusing larval hosts so as to maximize both larval and female
fitness, females instead prefer to deposit eggs in hosts associated with low offspring
performance. If females were behaving in a manner that improved their own fitness at
the expense of offspring performance, our results would imply that, with respect to re­
use of infested fruit, female fitness and per capita larval fimess are negatively, not
positively, correlated.
Preference for previously exploited hosts
An important point to emphasize with respect to the preference-performance hypothesis
in this system is that females are not simply reusing hosts because they cannot
76

discriminate between hosts that have been previously exploited versus those that have
not been previously exploited. In previous experiments, gravid females were not only
found to selectively visit fruit containing artificial punctures, these females also
appeared to be depositing clutches directly into the punctures (Papaj 1994, Lalonde and
Mangel 1994). In our experience, punctures such as those placed on fruit in the above
experiments are always associated with previous egg-laying events of conspecifics and
never with those created by other insects (such as pentatomid and coreid bugs) or other
natural causes. In general, in this study and a previous field study, we found that reuse
of walnut hosts by R. juglandis appears to be conunon (Nufio et al. 2000). In this
experiment, we found that females placed nearly 20 new eggs (vs. 23 eggs estimated in
a previous study, Nufio et al. 2000) into a given fruit every two days. Another important
point to make is that R. juglandis females, like other fruit flies (Landolt and Averill
1999), also deposit a marking pheromone on the fruit surface following clutch
deposition. This mark has been found to deter reuse, with the degree of deterrency being
directly proportional to the amount of time that previous females marked the fruit
(Nufio and Papaj, APPENDIX E).
Offspring perfomiaiice as a consequence of host reuse
Host reuse and offspring survival
When we examined the components of offspring survival affected by reuse, we found
that offspring survival declined with egg density during nearly each of the life stages
examined. To our surprise, even percent egg hatch in a fruit declined with egg density
77

(Table I). The mechanism underlying this latter result is unclear. Possibly, newly
hatched larvae spoil the cavity environment for eggs yet to hatch. Alternatively, the
pattern in egg density may be confounded with changes that occur in the fruit over the
time it takes eggs to accumulate. Such changes could be induced by egg deposition or
be independent of it. It is also possible that females damage previously laid eggs while
adding eggs to an existing cavity.
Increases in the number of eggs in a fhiit were also associated with reduced
survival after hatch (Table 1). This pattern presumably reflects competition for
resources among offspring. While our field study did not permit us to determine how
clutch position in a fruit (that is, when an egg was laid) affects survival, laboratory
experiments (Nufio and Papaj, appendix 4) suggest that second clutches may hatch but
may be less likely to survive than first clutches, an asymmetry that depends on the time
between clutch deposition.
Host reuse and pupal weight
The time over which a fruit was exposed to reuse also affected offspring weight. As
time passed, females deposited more eggs into fhiit and the increase in offspring
numbers led to a marked reduction in pupal weight (Figures 2 & 4). The number of
larvae emerging from a fruit was, however, a better predictor of pupal weight than was
the number of eggs deposited within a fiiiit or the number of eggs that hatched within a
fiiiit (Table 1). This suggests that because volume per individual is an imponant factor
determining pupal weight (Figure S), larval mortality may be higher in newly emerged
larvae than in established older larvae. The death of earlier instar larvae means that they
78

will consume fewer resources and thus more will be available for the remaining larvae.
The death of older larvae, on the other hand, may have a greater impact on the
availability of resources for other older larvae, because they consume a larger potion of
larval resources. However, it is not clear whether early instar larvae might be more
susceptible to the negative effects of competing with same age conspecifics (either due
to direct competition for resources or perhaps to the build up of waste products [Huang
et al. 1971]) or if perhaps early instars from later laid clutches are less able to compete
for resources with older larvae or are less able to deal with toxins that build up from the
cider more developed clutches.
Pupal weight and its impact on offspring reproductive potential
Both total percent larval survival and median pupal weight increased with fruit volume
per egg or per larva in a negatively accelerating fashion (Figures 3b and 5). The
tendency for pupal weight to asymptote as available resources increase may reflect an
upper limit on body size in these flies. In our laboratory experiments, we found that
larvae that pupated at lower weights became smaller adults (Figure 6). In both replicates
of our laboratory experiments, we found that smaller females laid fewer eggs over time
(Figure 8). Smaller females produced fewer eggs over the course of the experiment not
because their life spans were shorter, but because larger females appeared to produce
more eggs per day (Table 5). Similar reproductive advantages of female size have been
noted in other &uit flies, such as Rhagoletis pomonella Walsh (Averill & Prokopy
1987), as well as in other systems (Credland et al. 1986, Wickman and Karlsson 1989,
Hirschberger 1999, Bonduriansky and Brooks 1999, Mills and Kuhlmann 2000,
79

Amezaga and Garbisu 2000). While not measured in our study, egg size and not just the
number of eggs produced over a lifetime could also be an important character that is
affected by female size (Fox and Savalli 1998, Visser 1994) and males may also also
experience costs associated with being small (Beeleret al. 2002, Bissoondath and
Wiklund 1997, Robertson and Roitberg 1998).
Overall patterns showed that female walnut flies live an average of 26 days and
typically begin to lay their first clutches when they are 7-8 days old (median 9 + 0.4 for
each replicate) but some females do not begin to deposit clutches until they are 17-18
days old. In the 1999 replicate of our study, larger females laid their first clutches
signiHcantly earlier than did smaller females. This relationship, however, was not
significant for the 1997 cohort. Interestingly, in the 1999 replicate, where we did find a
correlation between female body size and days till a first clutch was deposited, females
also laid significantly fewer eggs than did females from the 1997 replicate. It is not
known why the second replicate produced fewer eggs as females originated from the
same locality and as female size, rearing conditions and fruit collecting localities all
appeared to be similar. If there were differences in any variable that induced stress or
quality differences in the food provided to females, our experiment may show that
reproductive advantages associated with an increase in body size may be particularly
important under stressful or marginal conditions.
What benefits do females receive from reusing hosts?
As we have argued previously, reuse of walnut hosts by R.juglandis may be influenced
by three factors (Nufio et al. 2000). First, reuse may be influenced by the benefits that
80

females gain not by simply reusing a host fruit but by reusing the actual oviposition
punctures created by previous females. Females appear to show an active oviposition
preference for fhiit bearing oviposition punctures, frequently depositing eggs directly
into these previously-established punctures (Papaj 1994, Lalonde and Mangel 1994). In
this study, we estimated that an oviposition puncture on a fruit was reused an average of
1.6 times. By reusing oviposition punctures, females may save time (Papaj 1993, 1994;
Papaj and Alonso-Pimentel 1997), decrease the wear to their ovipositors (Papaj 1993),
or gain access to fruit that are relatively impenetrable (Lalonde and Mangel 1994).
These benefits have been proposed to increase the number of clutches that a female can
lay over a lifetime. Time savings might also reduce a female's risk of predation if, as
seems likely, females are especially vulnerable to predators while ovipositing (Papaj
1993).
Another reason that females may reuse fruit, independent of use of punctures,
may be related to the host fruit's size. Most Rhagoletis species utilize relatively small
hosts (e.g. hawthorn berries, cherries, blueberries and dogwood berries [Bush 1966] that
appear to offer fewer resources for developing offspring then do walnut fruit. In studies
of R. pomonella, for example, rarely did more than 3 or 4 pupae emerge from a
hawthorn berry, even when more were deposited into the berry (Averill and Prokopy
1987; Feder et al. 1995). Our present study indicates that individual walnut fruit can
yield dozens of walnut fly pupae. The ability for walnuts to support greater infestation
levels than other Rhagoletis hosts may also explain why members of the walnut-
infesting R. suavis species group deposit eggs in clutches rather than singly. Deposition
of single eggs is the rule in most other species within the genus. With respect to the
81

preference-perfonnance hypothesis, in this system, the cost to larvae forced to compete
with conspecifics, while meaningful, is not as severe as it could be in smaller hosts and
this may in part explain why females may prefer to reuse fruit.
The third factor that may influence the reuse of hosts by walnut flies is the short
temporal and limited numerical availability of larval hosts in the field. Since nearly all
walnut hosts within an area will be synchronously utilized within two to two and a half
weeks, there will be a limit on the total amount of larval resources available to a
population of walnut flies (Nufio et al. 2000). On an individual level, this may mean
that females are time limited and must maximize the number of clutches deposited
within the limited window of larval resource availability. One-way to maximize the
number of clutches deposited within the allotted time may be to reuse hosts as they
ripen and become accessible to females. The eventual limitation of available hosts may
also produce a game in which females that reuse hosts early on experience greater
fitness returns than females that wait until all or most fruit are utilized before they begin
to reuse hosts. Reusing hosts to maximize the number of clutches deposited during the
short time hosts are available may be a viable strategy for walnut flies because, within
limits, walnut husks can support the development of more than a few clutches.
Preference-performance and parent-ofTspring conflicts
As previously stated, central to the preference-performance hypothesis is the
expectation that a female should choose to oviposit on hosts that maximize the per
capita fimess of her offspring because this in turn maximizes her own fitness. However,
as the reproductive success of a female is a product not only of the reproductive quality
82

of her offspring but also of the number of offspring she produces, a female should be
expected to choose to produce a number and quality of progeny that maximizes the
number of progeny that can be produced by her offspring as well as the number of
progeny she produces (Godfray 1987, Lloyd 1987, Forbes 1991, Winemiller and Rose
1993). While reproductive conflicts between parents and offspring have been explored
in a variety of systems (Chamov and Skinner 1985, Spiers et al. 1991, Smith and
Lessells 1985, Sinervo 1999, Einum and Fleming 2000) they have not often been
explored in regards to the preference-performance hypothesis, perhaps because
researchers typically couple the outcomes of female decisions as they affect per capita
offspring performance and not the different components of female performance (rev.
Mayhew 1997, but see Rausher 1980, Larsson and Ekbom 1995, Nylin and Janz 1996,
Scheirs et al. 2000, Scheirs and De Bruyn 2002).
The walnut flies R. juglandis appears to illustrate the reproductive conflicts that
can arise when egg-laying strategies that maximize a female's reproductive success may
not to be the same strategies that maximize an offspring's performance. We thus
believe that a lack of a correlation between female preference and o^spring
performance should not always be thought of as less than optimal decision making on
the part of the female or sub-optimal conditions that limit offspring (and thus female)
fimess, but alternatively, that such decision making can be optimal from the perspective
of its impacts on female Htness. In other words, conditions may exist where even if a
female is given a choice between where offspring perform best or where offspring do
less well, females may, in the long run, actually do best choosing the latter host. We
hope that investigators examine how alternative explanations, such as those provide by
83

parent-offspring models or optimal foraging models (Scheirs and De Bruyn 2002), may
impact their interpretations and expectations regarding the evolution of female
preference for oviposition sites and larval performance.
Summary
Host reuse by walnut flies in the fleld is conunon and this behavior negatively impacts
larval fimess characteristics (namely survival, size at pupation, and potential life-time
fecundity). While reuse negatively impacts offspring fitness, reuse appears to be best
explained by the direct benefits imparted to females that not only reuse actual
oviposition punctures but also that simply reuse fhiit that can typically support several
broods to pupation. Because hosts are limited in quantity and available for a short time
during a season, reuse of fruit may optimize female fimess by increasing the number of
clutches that can be deposited over the season. In other words, reuse behavior in this
system appears to maximize the product of offspring numbers and performance. Our
present results should encourage others to question the assumption that parental and
offspring fitness should be tightly coupled.
84

ACKNOWLEDGEMENTS
We thank Henar Alonso-Pimentel, Judie Bronstein, Reginald Chapman, Laurie
Henneman, and Dena Smith for comments and discussion. Sheridan Stone of the Fort
Huachuca Wildlife Management offlce of the US army provided permission and
logistical support for fieldwork in Garden Canyon. A Pre-graduate National Science
Fellowship supported this research, NRICGP grant no. 93-37302-9126 to D.R.P., Sigma
Xi.
85

REFERENCES
Ahman I, Tuvesson S, Johansson M, 2000. Does indole alkaloid gramine confer
resistance in barley to aphid Rhopalosiphum padil Journal of Chemical Ecology
26:233-255.
Amezaga I, Garbisu C, 2000. Effect of intraspecific competition on progeny production
of Tomicus piniperda (Coleoptera; Scolytidae). Environmental Entomology
29:1011-1017.
Averill AL, Prokopy RJ, 1987. Intraspecific competition in the tephritid fruit fly
Rhagoletis pomonella. Ecology 68:878-886.
Ballabeni P, Wlodarczyk M, Rahier M, 2001. Does enemy-free space for eggs
contribute to a leaf beetle's oviposition preference for a nutritionally inferior
host plant? Functional Ecology 15:318-324.
Heeler AE, Rauter CM, Moore AJ, 2002. Mate discrimination by females in the burying
beetle Nicrophorus orbicollis: the influence of male size on attractiveness to
females. Ecological Entomology 27:1-6.
Bemays EA, 2001. Neural limitations in phytophagous insects: Implications for diet
breadth and evolution of host affiliation. Annual Review of Entomology 46:703-
727.
Bissoondath CJ, Wiklund C, 1997. Effect of male body size on sperm precedence in the
polyandrous butterfly Pieris napi L. (Lepidoptera: Pieridae). Behavioral
Ecology 8:518-523.
86

Bjorkman C, Larsson S, Bommarco R, 1997. Oviposition preferences in pine sawflies:
A trade-off between larval growth and defence against natural enemies. Oikos
79:45-52.
Bonduriansky R, Brooks RJ, 1999. Reproductive allocation and reproductive ecology of
seven species of Diptera. Ecological Entomology 24b:389-395.
Bush GL, 1966. The taxonomy, cytology and evolution of the genus Rhagoletis in
North America (Diptera: Tephritidae). Bulietine of the Museum of Comparative
Zoology of Harvard University 134:431-562.
Chamov EL, Skinner SW, 1985. Complementary approaches to the understanding of
parasitoid oviposition decisions. Environmental Entomology 14:383-391.
Copp NH, Davenport D, 1978. Agraulis and passiflora. 1. control of specificity.
Biological Bulletin 155:98-112.
Courtney SP, KibotaTT, 1990. Mother doesn't know best: selection of host by
ovipositing insects. In: Insect-plant interactions (Bemays EA, ed). Boca Raton,
FL: CRC; 161-188.
Craig TP, Itami JK, Price PW, 1989. A strong relationship between oviposition
preference and larval performance in a shoot-galling sawfly. Ecology 70:1691-
1699.
Credland PF, Dick KM, Wright AW, 1986. Relationship between larval density, adult
size and egg production in the cowpea seed beetle, Callosobruchus maculatus.
Ecological Entomology 11:41-50.
Cronin JT, Abrahamson WG, 1999. Host-plant genotype and other herbivores influence
goldenrod stem galler preference and performance. Oecologia 121:392-404.
87

Cronin JT, Abrahamson WG, 2001. Goldenrod stem galler preference and performance:
effects of multiple herbivores and plant genotypes. Oecologia 127:87-96.
Denno RF, Larsson S, Olmstead KL, 1990. Role of enemy-free space and plant quality
in host-plant selection by willow beetles. Ecology 71:124-137.
Einum S, Beming lA, 2000. Highly fecund mothers sacriflce offspring survival to
maximize fitness. Nature 405:565-567.
Ekbom B, 1998. Clutch size and larval performance of pollen beetles on different host
plants. Oikos 83:56-64.
Etges WJ, Heed WB, 1987. Sensitivity to larval density in populations of Drosophila
mojavensis: influences of host plant variation on components of fltness.
Oecologia 71:375-381.
Feder JL, Reynolds K, Go W, Wang EC, 1995. Intraspecific and interspecific
competition and host race formation in the apple maggot fly, Rhagoletis
pomonella (Diptera: Tephritidae). Oecologia 101:416-425.
Forbes LS, 1991. Optimal size and number of offspring in a variable environment.
Journal of Theoretical Biology 150:299-304.
Fox CW, 1993. A quantitative genetic analysis of oviposition preference and larval
performance on two hosts in the bruchid beetle, Callosobruchus maculatus.
Evolution 47:166-175.
Fox CW, Savalli UM, 1998. Inheritance of environmental variation in body size:
Superparasitism of seeds affects progeny and grandprogeny body size via a
nongenetic maternal effect. Evolution 52:172-182.
88

Fujiyama N, Katakura H, 2001. Variable correspondence of female host preference and
larval performance in a phytophagous ladybird beetle Epilachna pustulosa
(Coleoptera: Coccinellidae). Ecological Research 16:405-414.
Godfray HCJ, 1987. The evolution of clutch size in parasitic wasps. American
Naturalist 129:221-233.
Gratton C, Welter SC, 1999. Does "enemy-free space" exist? Experimental host shifts
of an herbivorous fly. Ecology 80:773-785.
Hanks LM, Paine TD, Miliar JG, 1993. Host species preference and larval performance
in the wood- boring beetle Phoracantha semipunctata F. Oecologia 95:22-29.
Harris MO, Sandanayaka M, Griffin A, 2001. Oviposition preferences of the Hessian
fly and their consequences for the survival and reproductive potential of
offspring. Ecological Entomology 26:473-486.
Hirschberger P, 1999. Larval population density affects female weight and fecundity in
the dung beetle Aphodius ater. Ecological Entomology 24:316-322.
Howard DJ, Bush GL, 1989. Influence of bacteria on larval survival and development in
Rhagoletis (Diptera: Tephritidae). Annals of the Entomological Society of
America 82:633-640.
Howard DJ, Bush GL, Breznak JA, 1985. The evolutionary significance of bacteria
associated with Rhagoletis. Evolution 39:405-417.
Huang SL, Singh M, Kojima K, 1971. A study of frequency dependent selection
observed in the esterase-6 locus of Drosophila melanogaster using a conditioned
media method. Genetics 68:97-104.
89

Jaenike J, 1978. Optimal oviposition behavior in phytophagous insects. Theoretical
Population Biology 14:350-356.
Karban R, Courtney S, 1987. Intraspecific host plant choice: lack of consequences for
Streptanthus tortuosus (Cruciferae) and Euchloe hyantis (Lepidoptera: Pieridae).
Oikos 48:243-248.
Lalonde RG, Mangel M, 1994. Seasonal effects on superparasitism by Rhagoletis
completa. Journal of Animal Ecology 63:583-588.
Landolt PJ, Averill AL, 1999. Fruit Flies. In: Pheromones of Non-Lepidopteran Insects
Associated with Agricultural Plants (Hardie J, Minks AK, eds). New York:
CAB I Publishing; 3-26.
Larsson S, Ekbom B, 1995. Oviposition mistakes in herbivorous insects : Confusion or
a step towards a new host giant. Oikos 72:155-160.
Larsson S, Glynn C, Hoglund S, 1995. High oviposition rate of Dasineura
marginemtorquens on Salix viminalis genotypes unsuitable for offspring
survival. Entomologia Experimentalis Et Applicata 77:263-270.
Leather SR, 1985. Oviposition preferences in relation to larval growth rates and survival
in the pine beauty moth, Panolis flammea. Ecological Entomology 10:213-217.
Leyva KJ, Clancy KM, Price PW, 2000. Oviposition preference and larval performance
of the western spruce budworm (Lepidoptera: Tortricidae). Environmental
Entomology 29:281-289.
Lloyd DG, 1987. Selection of offspring size at independence and other size-versus-
number strategies. American Naturalist 129:800-817.
90

Mayhew PJ, 1997. Adaptive patterns of host-plant selection by phytophagous insects.
Oikos 79:417-428.
Mayhew PJ, 1998. The life-histories of parasitoid wasps developing in small gregarious
broods. Netherlands Journal of Zoology 48:225-240.
McMillin JD, Wagner MR, 1998. Influence of host plant vs. natural enemies on the
spatial distribution of a pine sawfly, Neodiprion autumnalis. Ecological
Entomology 23:397-408.
Messina FJ, 1982. Food plant choices of two goldenrod beetles: relation to plant quality.
Oecologia 55:342-354.
Mills NJ, Kuhlmann U, 2000. The relationship between egg load and fecundity among
Trichogramma parasitoids. Ecological Entomology 25:315-324.
Nufio CR, Papaj DR, 2001. Host marking behavior in phytophagous insects and
parasitoids. Entomologia Experimentalis Et Applicata 99:273-293.
NuHo CR, Papaj DR, 2002. Host marking behavior as a quantitative signal of infestation
levels in host use by the walnut fly, Rhagoletis juglandis. Appendix E.
NuHo CR, Papaj DR, Alonso-Pimentel H, 2000. Host utilization by the walnut fly,
Rhagoletis juglandis (Diptera: Tephritidae). Environmental Entomology 29:994-
1001.
Nylin S, Janz N, 1996. Host plant preferences in the comma butterfly {Polygonia c-
album): Do parents and offspring agree? Ecoscience 3:285-289.
Nylin S, Janz N, Wedell N, 1996. Oviposition plant preference and offspring
performance in the comma butterfly: correlations and conflicts. Entomologia
Experimentalis Et Applicata 80:141-144.
91

Papaj DR. 1993. Use and avoidance of occupied hosts as a dynamic process in tephritid
fruit flies. In; Insect-Plant Interactions (Bemays EA, ed). Boca Raton; CRC
Press; 25-46.
Papaj DR, 1994. Oviposition site guarding by male walnut flies and its possible
consequences for mating success. Behavioral Ecology and Sociobiology 34; 187-
195.
Papaj DR, Alonso-Pimentel H, 1997. Why walnut flies superparasitize; time savings as
a possible explanation. Oecologia 109; 166-174.
Price PW, Bouton CE, Gross P, McPheron BA, Thompson JN, Weis AE, 1980.
Interactions among three trophic levels; influence of plants on interactions
between insect herbivores and natural enemies. Annual Review of Ecology and
Systematics ll;41-65.
Price PW, Ohgushi T, 1995. Preference and performance linkage in sl phyllocolpa
sawfly on the willow, Salix miyabeana, on hokkaido. Researches On Population
Ecology 31:23-2S.
Prokopy RJ, 1981. Oviposition-deterring pheromone system of apple maggot flies. In;
Management of Insect Pests with Semiochemicals (Mitchell EK, ed). New York;
Plenum Press; 477-494.
Rausher MD, 1980. Host abundance, juvenile survival, and oviposition preference in
Battus philenor. Evolution 34;342-355.
Reid ML, Roitberg BD, 1995. Effects of body-size on investment in individual broods
by male pine engravers (Coleoptera: Scolytidae). Canadian Joumal of Zoology-
Revue Canadienne De Zoologie 73:1396-1401.
92

Robertson IC, Roitberg BD, 1998. Duration of paternal care in pine engraver beetles:
why do larger males care less? Behavioral Ecology and Sociobiology 43:379-
386.
Rossi AM, Strong DR, 1991. Effects of host-plant nitrogen on the preference and
performance of laboratory populations of Cameocephala floridana (Homoptera:
Cicadellidae). Environmental Entomology 20:1349-1355.
Scheirs J, De Bruyn L, 2002. Integrating optimal foraging and optimal oviposition
theory in plant-insect research. Oikos 96:187-191.
Scheirs J, De Bruyn L, Verhagen R, 2000. Optimization of adult performance
determines host choice in a grass miner. Proceedings of the Royal Society of
London Series B-Biological Sciences 267:2065-2069.
Sinervo B, 1999. Mechanistic analysis of natural selection and a refinement of Lack's
and Williams's principles. American Naturalist 154:S26-S42.
Smith RH, Lessells CM, 1985. Oviposition, ovicide, and larval competition in
granivorous insects. In: Behavioral ecology: ecological consequences of
adaptive behaviour (Sibly RM, Smith RH, eds). Oxford: Blackwell Scientific
Publication; 423-448.
Speirs DC, Sherratt TN, Hubbard SF, 1991. Parasitoid diets - does superparasitism pay?
Trends in Ecology & Evolution 6:22-25.
Steinbauer MJ, 1999. The population ecology of Amorbus dallas (Hemiptera: Coreidae)
species in Australia. Entomologia Experimentalis Et Applicata 91:175-182.
93

Sweeney J, Quiring DT, 1998. Oviposition site selection and intraspecific competition
influence larval survival and pupal weight of Strobilomyia neanthracina
(Diptera: Anthomyiidae) in white spruce. Ecoscience 5:454-462.
Thompson DC, Knight JL, Sterling TM, Murray LW, 1995. Preference for specific
varieties of woolly locoweed by a specialist weevil, Cleontdius trivittatus (Say).
Southwestern Entomologist 20:325-333.
Thompson JN, 1988a. Evolutionary ecology of the relationship between oviposition
preference and performance of offspring in phytophagous insects. Entomologia
Experimentalis Et Applicata47:3-14.
Thompson JN, 1988b. Variation in preference and specificity in monophagous and
oligophagous swallowtail butterflies. Evolution 42:118-128.
Valladares G, Lawton JH, 1991. Host-plant selection in the holly leaf-miner: Does
mother know best? Journal of Animal Ecology 60:227-240.
Visser ME, 1994. The Importance of being large: the relationship between size and
fitness in females of the parasitoid Aphaereta minuta (Hymenoptera:
Braconidae). Journal of Animal Ecology 63:963-978.
Weisser WW, Houston AI, Volkl W, 1994. Foraging strategies in solitary parasitoids:
The trade-off between female and offspring mortality risks. Evolutionary
Ecology 8:587-597.
Wickman PO, Karlsson B, 1989. Abdomen size, body size and reproductive effort of
insects. Oikos 56:209-214.
94

Wiklund C, 1981. Generalist vs. specialist oviposition behavior in Papilio machaon
(Lepidoptera) and functional-aspects on the hierarchy of oviposition
preferences. Oikos 36:163-170.
Williams KS, 1983. The coevolution of Euphydryas chalcedona butterflies and their
larval host plants .m. Oviposition behavior and host plant quality. Oecologia
56:336-340.
Wilson D, Faeth SH, 2001. Do fungal endophytes result in selection for leafminer
ovipositional preference? Ecology 82:1097-1 111.
Winemiller KO, Rose KA, 1993. Why do most fish produce so many tiny offspring?
American Naturalist 142:585-603.
Yamaga Y, Ohgushi T, 1999. Preference-performance linkage in a herbivorous lady
beetle: consequences of variability of natural enemies. Oecologia 119:183-190.
95

Table 1. The influence of various factors on total survivorship (from eggs deposited to
successful pupaption) and the influence of various factors on survivorship from egg
hatch to larvae emerging from a fruit different life history stages.
Total percent survival
(deposited egg to successful pupation)
Number of eggs deposited
Fruit volume
Fruit cohort (number of days
fruit available for re-use)
R'= 0.35
% Larval Survival from hatch to
Number of eggs deposited
Fruit volume
Fruit cohort (number of days
fruit available for re-use)
Number of eggs that hatched
0.12
96
MULTIPLE REGRESSION ANALYSES
F DF ESTIMATE P
79.16
11.46
from fruit
I -0.006

Table 2. The influence of various factors on egg hatch (either 100% hatch or not) and
survival from larvae emerging from a fruit and successfully pupating (either 100%
survival or not).
% Egg hatch
97
MULTI-FACTORIAL LOGISTIC REGRESSION
WALD X' DF ESTIMATE P
Number of eggs deposited 27.00 -0.07

Median pupal weight
Number of larvae emerging from a fruit 74.54
Fruit volume 69.12
Fruit Cohort 6.59
Number of eggs deposited
Number of eggs hatched
R'= 0.45
98
MULTIPLE REGRESSION ANALYSES
•F RATIO" DF ESTIMATE
I -0.5668

Table 4. Analysis of variance comparing both the 1997 and 1999 replicates of our
study on female size and lifetime fecundity.
1997 Replicate (iis34) 1999 Replicate (nsSl)
Mean ±SE Mean ±SE F P
Female Size (mm) 1.57 0.02 1.55 0.02 0.16 ns
Female lifespan (days) 26 2.14 25.55 1.87 0.04 ns
Number Eggs Deposited 126 17.52 64 8.10 12.64 0.0006
ns, P > 0.05
99

nCURE CAPTIONS
Figure 1. Mean (± SE) number of eggs deposited as a function of fhiit cohort age.
Bars sharing the same letter are not significantly different (Tukey HSD. P < .05).
Figure 2. Relationships between fhiit volume and the number of eggs deposited
within. (A) Fruit infested for only 1-2 d. (B) Fruit infested for 3-4 d. (C) Fruit
infested for 5-6 d. Note the difference in scale between the number of eggs
deposited within the different cohorts. Regression lines drawn for diagrammatic
purposes.
Figure 3. (A) Percent total survival of offspring as a consequence of the number
of eggsdeposited within a fhiit. (B) Percent total survival as a consequence of
available fhiit volume per egg deposited within a fhiit (%survival = -0.95 +
0.24x Log Fruit volVegg).
Figure 4. Average median (± SE) weight of pupae emerging from fhiit as a fimction
of fhiit cohort age. Bars sharing the same letter are not significantly different
(Tukey HSD, P < .05).
100

nCURE CAPTIONS - Continued
Figure 5. Relationship between median pupal weight of offspring emerging from a
given fruit as a function of the (A) available fruit volume per egg that hatched
(polynomial regression; r2 = 0.32;Y= 42.08 + 0.03X +1.35x10-6X2) and (B)
available fhiit volume per larva that emerged from the fhiit (polynomial
regression; r2 = 0.44;Y= 46.31 + 0.03X +1.47x10-6X2).
Figure 6. The relationship between pupal weight and anterior discal cell length
(an estimate of adultg size) for both females (discal cell length = 1.05 + 7.2
xIO-4 X pupal weight) and males (discal cell length = 1.02 + 6.6x10-4 x pupal
weight).
Figure 7. The proportion of females alive as a function of female age for the (A)
1997 (n=35) cohort and the (B) 1999 cohort (n=51).
Figure 8. The age at which females from the two experimental cohorts deposited their
first clutches.
101

nCURE CAPTIONS - Continued
Figure 9. The relationship between female size(estimated by measuring the length
of portion of the medial vein that makes up the anterior discal cell length)and
lifetime fecundity for the (A) 1997 (total eggs = -354.54 +313.0 x wing length)
and (B) 1999 (Total eggs = -214.67 + 180.05 x wing length) female cohorts.
Note difference in scale in the total number of eggs produced.
102

Figure 1.
Fruit Cohorts
(Number of days a fruit was exposed)

OB > i
Number of Eggs Deposited Number of Eggs Deposited Number of Eggs Deposited
•I
n
K»

Figure 3.
A.
B.
A
>
on
•m
e
« 50
s
w
k
a.
98
>
S
V
100
75 -
25 -
100
75 -
« 50 -
25
4.5
40 80 120
—
5.5
Number of eggs deposited
.-•h
I • •
6.5 7.5
Log Fruit volume (em3) per egg deposited
I
8.5
105

Figure 4.
«8
a
0.8 -
OA
S
>=- 0.6 -
OA
9i
0.4 -
0.2 -
(n = 61) (n=61) (n = 54)
1-2 3-4 5-6
Fruit Cohorts
(Number of days a fruit was exposed to reuse)
106

Figure 5.
B.
es
&
1 1
0.8 •
0.6 -
• MB S 0.4 -
"2
ju
'S
0.2
1 -I
0.8 -
"i 0.6 -
•5 0.4
9i
IS
0.2
# • •
0 .5 1 1.5 2 2.5
Fruit Volume per Hatched Egg (cm^
L • •
— m ••
-1—
.5
1-
1 1.5 2.5
Fruit Volume per Larvae Emerging from Fruit (cm^
107

Figure 6.
Mean pupal weight (mg)
Females
Males
108

Figure 7.
B.
>
09
"S
s
e
s
o
•Ml
k
S.
s
Om
0.6 -
Median 27 + 2.2
(SE)
0 5-6 15-16 25-26 35-36 45-46 55-56 65-66
Days since individuals first emerged
Median 25 + 1.8
(SE)
5-6 15-16 25-26 35-36 45-46 55-56 65-66
Days since Individaals first emerged
109

Proportion
of females
depositing
their first
clutch
1997 Cohort (n = 35)
Median = 9 ±.(SE) .44
1999 Cohort (n = 48)
Median = 9 ±.(SE) .43
7-8 9-10 11-12 13-14 15-16 17-18 19-20
Female age (days)

Figure 9.
400
9t
W
s
1
t 300
«
Of)
Dll
U

APPENDIX D
AGGREGATIVE BEHAVIOR IS NOT EXPLAINED BY AN
ALLEE EFFECT IN THE WALNUT-INFESTING FLY,
RHAGOLETIS JUGLANDIS
112

Aggregative behavior is not explained by an Allee effect in the wahiut-
infesting fly, Rhagoletis juglandis
Cesar R. Nufio'-*
and
Daniel R. Papaj^
Department of Entomology', Department of Ecology and Evolutionary Biology", Center
for Insect Science^, and the Interdisciplinary Degree Program in Insect Science"^,
University of Arizona, Tucson, AZ 85721, USA

ABSTRACT
Component Allee effects are considered to be a driving force in the origin and
maintenance of aggregative behavior. In this study, we examine whether a pattern of
active host reuse by the walnut fly, Rhagoletis juglandis Cresson (Diptera: Tephritidae),
involves an Allee effect. We examined how the density of eggs deposited within a fruit
influences survival to pupation and pupal size, the latter a strong indicator of lifetime
female fecundity. We also examined the role of egg density in relation to the temporal
pattern in which successive clutches are deposited and the spatial distribution of
clutches over a fruit surface. Increases in larval density strongly reduced pupal weight
and to a lesser, but still significant extent, reduced larval survival. Temporal staggering
of clutches into a host strongly reduced percent survival and, probably owing to
competitive release, increased pupal weight of survivors. Offspring survival and pupal
weight were affected relatively little by whether clutches were deposited within the
same oviposition punctures or evenly distributed along a fruit's surface. Nevertheless,
in three-clutch treatments, percent total survival in evenly distributed clutches was
significantly lower and median pupal weight higher than in clutches placed in the same
oviposition cavity. It appeared that the placement of clutches did not significantly
affect egg hatch or pupal survival, rather that there was a trend towards higher larval
survival when clutches were placed together. Results as a whole, however, fail to
provide evidence of an Allee effect. We put forward a scenario by which females
appear to reuse larval hosts in a way in which they maximize their own reproductive
success at the expense of the per capita fitness of their offspring. In tura, we propose
114

that the larval aggregations that form within multiply infested hosts may provide larvae
with a mechanism for reducing the adverse effects of larval competition.
Key words Allee effect' marking pheromone reproductive trade-offs' parent-offspring
conflict "Rhagoletis juglandis Tephritidae plant-insect interactions
115

INTRODUCTION
Host specific insects such as phytophagous insects or entomophagous parasitoids can be
categorized roughly into those that avoid use of hosts previously exploited by
conspeciHcs and those that actively aggregate at hosts (rev. Nufio and Papaj 2001,
Prokopy and Roitberg 2001). The former group is generally characterized by intense
competition among conspecifics for hosts, even at low conspecific densities, and
frequently involves use of a marking pheromone (MP) that signals occupation of the
host resource (rev. Nufio and Papaj 2001). Females of the bean weevil Callosobruchus
maculates, for example, use chemical marks and physical cues associated with the
deposition of clutches on hosts to assess not only whether or not a host has been utilized
but also the number of eggs associated with that host. Females have been found to
selectively re-use hosts that bear a lower than average number of eggs (Messina and
Renwick 1985, Wilson 1988). This rejection of previously exploited hosts is thought to
be functional as each additional larva typically faces both a greater risk of not obtaining
sufficient resources for successfiil development or, if development is successful, a
greater reduction in fecundity (Credland et al. 1986).
Species that actively aggregate at resources are frequently characterized by the
occurrence of an AUee effect in which larval fitness, or some component of fitness,
decreases at intermediate conspecific densities. Species that beneHt from an AUee
effect typically utilize aggregation pheromones or other mechanisms to attract and
regulate the degree to which a host is exploited by conspecifics (Corbet 1973, Prokopy
1981, Hedland et al. 1996, Paine et al. 1997, Prokopy and Roitberg 2001, Wertheim
116

2001). The shifting benefits and costs associated with group size and the mechanisms
that regulate group size have been well explored in bark beetles (Raffa et al. 1993,
Kirkendall et al. 1997, Pureswaran et al. 2000). In the genus Dendroctonus and Ips, for
example, in order to attract mates and overcome host defenses, pioneer beetles arriving
at potential hosts release pheromones that attract conspecifics. Initially individuals
benefit greatly from an increase in the number of colonizing conspecifics, which jointly
help to overcome a host's defenses. However, overcrowding eventually leads to a
decreases in per capita clutch size and increases in o^spring mortality (Raffa 1993,
Reeve et ai. 1998). As overcrowding begins to affect colonizers adversely, they begin to
emit chemical and acoustic signals that deter other conspecifics from colonizing (Byers
1989, Paine et al. 1997, Reeve et al. 1998).
In this paper, we investigate a system involving a specialist tephritid fly,
Rhagoletis juglandis, and its host walnut fruit, which does not fit neatly into either of
these two categories. In this system, females deploy a host-marking pheromone after
oviposition into the walnut husk that has a deterrent effect on oviposition by conspeciHc
females (Nufio and Papaj, APPENDIX E). Yet, despite the deterrent effects of the MP,
females actively deposit clutches where eggs have been laid previously, often placing
those clutches directly within oviposition cavities previously excavated by another
female. Owing to a significant degree of multiple oviposition into walnut fruit and even
into individual oviposition cavities, larvae typically feed in aggregations originating
from multiple clutches (Papaj 1994; Nufio et al. 2000, Nufio and Papaj, APPENDIX C).
While the number of eggs deposited within a single clutch of eggs is ca. IS, within 4-5
days of being first attacked, a fhiit may contain ca. 4S eggs, and it is not unusual to find
117

fruit into which multiple females have deposited 80 or more eggs (Nufio et al. 2000,
Nufio and Papaj, APPENDIX C).
Reuse of oviposition cavities apparently provides females with direct benefits
including reduced time to deposit clutches (Papaj and Alonso-Pimentel 1997), reduced
ovipositor wear (Papaj 1993), and increased access to relatively impenetrable fruit
(Lalonde and Mangel 1994). Females might reuse fruit because these direct benefits
more than offset costs in terms of competition suffered by their larvae. However, it is
also possible that females reuse fruit because their offspring benefit themselves from
being deposited into fruit in the company of other larvae. A larval benefit would
account in particular for reuse of fruit that does not involve reuse of oviposition
cavities.
In tephritid flies, there is reason to expect a beneHt to larvae of being in large
groups. Larvae feed not on fresh walnut husk but on a rot that is in part a consequence
of microbes transferred to the fruit by the female during oviposition (Howard et al.
1985, Howard and Bush 1989). Therefore, by reusing fruit, females might provide their
offspring with a 'prepared medium' (Hausmann and Miller 1989). Alternatively,
intermediate larval densities might furnish optimal conditions for growth of the
microbial flora (Sang 1956, Howard et al. 1985, Courtney et al. 1990).
In the following study, we evaluated components of larval fimess in relation to
the number of clutches deposited into host fruit. In doing so, we were particularly
interested in the possibility that larval fimess is optimal at some intermediate larval
density. We fiirther explored the effect of larval density in relation to the timing and
118

spatial patterning of clutches deposited successively into the same fruit. A failure to
fmd evidence of an Allee effect in this system would imply that aggregative behavior in
this system is not a consequence of direct benefits to larvae.
119

MATERIALS AND METHODS
Natural History
Rhagoletis juglandis is a member of the walnut-infesting Rhagoletis suavis group (Bush
1966). In southern Arizona, this species is found on the Arizona walnut, Juglans major,
which occurs in montane canyons between 1200 and 2700 meters. These flies are
univoltine and females deposit clutches of ca. 16 eggs (Nufio et al. 2(XX)) after piercing
the fruit surface with their ovipositor and hollowing out a small cavity in the walnut
husk. The larval stages feed on the husk of a single fruit, after which they emerge and
pupate in the soil beneath the natal tree. Pupae diapause through the winter and spring
and adults emerge during mid to late summer of a subsequent year.
General protocol
Adult flies used in the laboratory experiments originated as larvae from fruit collected
1-2 years earlier in Garden Canyon in the Huachuca Mountains in southern Arizona.
Until emergence, pupae were stored at 4°C. When drawn from refrigeration, adults
emerged in 4-5 weeks. Upon emergence, adults were maintained in 3.79-liter plastic
containers and provided with ad libitum water, sugar, and slips of a yeast hydrolysate
and sugar mixture. Flies were held in a room with a 14; lOh light:dark cycle and a
day/night temperature of 32''C and 28 °C.
Flies used in the laboratory experiments were 12 to 30 days old, post-eclosion.
For each of the experiments, 10 to IS females and S to 7 males were placed into clear
120

16-fl oz (473-ml) plastic Solo cups, fitted with petri dish lids, in which they were
provided with water, sugar, and a yeast hydrolysate and sugar mixture. Because males
often interfered with female oviposition behavior, males were removed from cages at
the beginning of each day on which experiments were conducted. So as to increase the
likelihood that females deposited fertile eggs, males were reintroduced into cages at the
end of the day (Telang et al. 1996). A walnut fruit was hung from the top of the cage
and females permitted to oviposit into the fruit. After a female initiated oviposition,
both fruit and fly were gently removed from the cup. Any other females present on that
fruit were removed. The fruit was then hung from the top of an empty cup cage and the
female allowed to complete oviposition into the fruit. After oviposition was completed,
the female was removed from the cup and discarded. Depending on experimental
treatment (see below), a fruit was sometimes re-exposed to females that were allowed to
reuse the host or the previously created oviposition puncmre, depending on the
particular experiment. Any females that were discarded were replaced with new
females, keeping the number of females per cup constant. All oviposition sites on fruit
were circled with a felt-tip marking pen. For each fruit in each study, we measured the
minimum and maximum length in order to estimate the volume of a given host. This
was done by assuming that a walnut was spherical in shape, taking the average of the
length measurements as an estimate of sphere diameter, and then computing fruit
121
volume as 4/3 ic r^, where r is the radius of the sphere. The volume of a fhiit was used as
a rough indication of the amount of larval resources provided by that firuit.

After fruit received the appropriate number of clutches in the appropriate
locations, they were wrapped in parafilm so as to minimize water loss. Fruit were then
placed individually into 'incubators' and stored in a growth chamber at a constant
temperature of 3(fC and 50% humidity. The incubators consisted of 16-fl oz Solo brand
plastic cups that were inverted with plastic petri dishes inserted as tops. The original
bottoms of the cups were removed and fruit were placed within the cups on top of a 3
cm long PVC tubing inserted into a 3 cm deep bed of mixed vermiculite and sand. The
vermiculite/sand layer was kept moist by adding water periodically until Day IS when
larvae might potentially begin to emerge from the fruit. After 6-7 days, the parafilm
covering the fruit was removed and oviposition cavities (if they were 6-7 days old) were
removed by excavating a cylinder roughly 6mm long and 8 mm wide around the
puncture which contained the oviposition cavity. To keep the fruit from drying out and
larvae from prematurely leaving the host fruit, the space previously occupied by the
oviposition cylinder was then covered with a piece of parafilm over which was placed a
IS by 20 mm strip of elastic bandage tape. Excavated cavities were stored individually
in alcohol in vials and later dissected. The unhatched eggs and egg husks present within
each oviposition cavity were counted. The number of egg husks, which are left by
individuals that had hatched and migrated into the husk, was used as an estimate of the
number of larvae hatching within the fhiit.
122
In order to examine effects of different patterns of reuse on offspring Htness, we
calculated total percent survival, as the percentage of eggs deposited into a given fruit
that successfiilly developed to pupation. We also calculated percent survival of the
different life stages, including the percentage of eggs deposited that hatched, the

percentage of hatchlings that emerged from the fruit as mature larvae, and the
percentage of emerging larvae that pupated successfiilly. To assess the impact of reuse
patterns on offspring size, viable pupae associated with a fruit were weighed. The
usefulness of pupal weight as a predictor of lifetime female fecundity has been
documented in Nufio & Papaj (APPENDIX C).
Experiment 1. The effect of offspring density on offspring fitness
In the first laboratory experiment, we manipulated egg density within fruit to examine
its impact on offspring fitness. Egg density was manipulated by distributing walnuts
into sets of five that were similar in size and ripeness, then exposing these fruit to
females, as described in the general protocol above. Over the course of 1 to 3 hours,
females were allowed to place a total of 1, 2, 3,4 or 5 clutches into each of the fruit in a
set. Females were free either to deposit clutches into new cavities or to reuse previously
made oviposition sites. The number of clutches that a given fruit received was a
function in part of how many females attempted oviposition. In order to control for
among-cage variation in oviposition activity, fhiit were often transferred among cages.
One factor, which we could not completely control, was the effect of the fruit itself on
oviposition activity. While fruit receiving different numbers of clutches within a set did
not appear to differ in known determinants of female preference such as size or
ripeness, it is possible that there were unknown determinants of preference that
influenced clutch deposition. If such a preference reflected a disposition to put clutches
into fruit in which larvae perform better, then high quality fruit would receive many
clutches and low quality fruit would receive few clutches. Such a pattern would tend to
123

educe the effect of egg density on offspring fitness and hence our method should tend
to be conservative with respect to effects of egg density. After fruit were removed from
a cage, they were wrapped in paraAlm and stored in a growth chamber, as above, and
their cavities removed after 6-7 days. Their cavities were dissected and number of
unhatched and hatched eggs, number of larvae emerging from a fruit and weights of
viable pupae were recorded.
Experiment 2. Effects of temporal patterning of clutches on offspring fitness
In order to determine effects of time between clutch deposition on offspring
fimess characteristics, we manipulated the temporal spacing of clutches into fruit by
matching sets of four fruit for size and ripeness and exposing these fruit to females, as
described above. In one of the four fruit, two clutches were deposited on the same day
(i.e., within a 1-3 hour period) into a single oviposition site. In a second fruit, four
clutches were deposited on the same day (i.e., within a 1-3 hour period) in sets of two
clutches at each of two different sites. In a third fruit, two clutches were deposited into a
single puncture on the same day and roughly two days later, two more clutches were
deposited into a second site. In a fourth fruit, two clutches were deposited into a single
site on the first day and, roughly four days later, two more clutches were placed into a
second site.
In the treatments involving two sets of two clutches, the last two clutches were
deposited at a single site roughly 1.5 cm from where the Hrst two clutches were
deposited. For replicates that required multiple clutches to be deposited within a single
site, gravid females were induced to reuse sites by gently brushing them towards a
124

previously created oviposition puncture or towards pricks made in the fruit with a No.
00 pin roughly 1.5 cm from the first puncture.
After both clutches were deposited in a fruit on a given day, the fruit was
wrapped in paratilm and stored in "incubators' in a growth chamber as described above.
Fruit were than unwrapped either 48 or 96 hrs later, as determined by the protocol, at
which time two additional clutches were deposited. Fruit were rewrapped in parafilm
and placed again into the growth chamber. After 6-7 days, the parafilm was removed,
oviposition cavities (once they themselves were 6-7 days old) were excavated as
described above, and unhatched and hatched eggs within the cavities counted. We later
counted the larvae that emerged from a fruit and weighed viable pupae.
Experiment 3. Effects of clutch spacing on larval fitness
In order to determine whether walnut fly brood benefit from being placed within the
same oviposition site, we manipulated the spacing of multiple clutches on fruit. Clutch
spacing was manipulated by matching sets of Ave fruit for size and ripeness and then
exposing these fruit to females which were allowed to either lay a single clutch within a
fruit, two clutches within the same oviposition puncture, two clutches each placed on
opposite sides of the fruit, three clutches within the same puncture, or three clutches
spaced regularly around a fruit's perimeter.
Ovipositions were obtained generally as described above for egg density. In
order to ensure that clutches were distributed around a ftuit according to treatment
designation, we ourselves initiated oviposition punctures by lightly pricking the fruit
125

surface with a no. 00 insect pin. Punctures were made along an equatorial line which
was perpendicular to the stem-calyx axis of the fruit.
In order to ensure that clutches were placed within the same puncture in two-
clutch and three-clutch treatments, we gently brushed gravid females towards the
previously created punctures. After one or several attempts at new sites, females usually
encountered the previously made oviposition site and reused the puncture. For the two-
clutch and three-clutch replicates in which oviposition sites were placed on opposite
sides of the fruit, females were also brushed to pin pricks as soon as they attempted to
oviposit into the fruit. After a clutch was deposited into a site initiated by a pin prick,
the oviposition site was temporarily covered with tape to prevent further reuse and a
new pin prick on the opposite side of the fruit was created before being exposed to the
next set of females. All clutches were deposited within the fruit over a I to 3 hour
period. Fruit were then wrapped in parafilm and stored in a growth chamber, as above,
and their cavities were removed after 6-7 days. Oviposition cavities were dissected after
6-7 days and unhatched and hatched eggs within the cavities counted. Later we counted
the larvae that emerged from a fruit and weighed viable pupae.
126

RESULTS
Experiment 1. Effect of offspring density on offspring Htness
Despite substantial variation in clutch size (clutch size in 1-clutch fruit =16.23 + (SE)
0.86 eggs; range = 7-36 eggs; N = 65), a progressive increase in the number of
clutches within fruit was associated with a consistent increase in egg density in a fruit
(R" = 0.58, F = 86.20, df = 1,64, P < 0.0001; test for slope different from zero, t = 9.28;
P < 0.0001; Figure lA). Mean fruit volume, 18.99 + 9.71 cm^ overall, did not differ
signiHcantly among clutch number treatments (F = 0.29, DF = 4,64, P = 0.88).
Effect of density on offspring survival
On average, 55 + 0.03 % (N = 65 fruit) of eggs deposited within a fruit hatched,
developed as larvae and successfiilly pupated. Multiple regression analysis revealed that
total percent survival from egg deposition to successful pupation declined marginally
significantly with number of clutches placed within a fruit, but did not change
significantly with number of eggs deposited within the fruit, or fruit volume (Figure IB;
Table 1).
On average, egg hatch (88 + 0.03%) was highly variable over all clutch number
127
treatments, a pattern that was possibly due to variation in female mating status (umnated
females will lay eggs but such eggs are inviable). We therefore conducted an analysis of
percent survival from egg hatch to successfiil pupation. Substituting number of eggs
hatched for number of eggs deposited doubled the model's R~, although it remains low.

The decline in percent survival with clutch number became more significant statistically
(Table 1). In this analysis, the factor 'number of eggs hatched' entered the model after
the factor 'clutch number', and is weakly but significantly positively related to percent
survival to pupation. Taking the clutch numbers and egg density results together,
adding clutches to a fruit appears to increase levels of competition, whereas increasing
the number of eggs per clutch may not be a precise estimate of overall larval
competition because egg hatch was highly variable among fruit. The effect of clutch
number on percent survival remains negative on balance.
Effect of density on median pupal weight
In contrast to larval survival, effects of clutch number and egg density on pupal weight
were particularly striking (Figure IC; Table 2). In exploratory data analysis, the pattern
of pupal weight and egg number appeared to be nonlinear (cf. Figure 2); for this reason,
we log-transformed egg number variables and confirmed that these yielded more robust
patterns. Multiple regression analysis explained over 50% of variation in median pupal
weight. Median pupal weight decreased highly significantly with log (no. eggs
deposited). Fruit volume entered the model after log (no. eggs deposited) and was
significantly positively related to median pupal weight. Number of clutches did not
enter the model.
Owing to the variability in egg hatch discussed above, we reanalyzed pupal
weight, using clutch number, fhiit volume and log (number of eggs hatched) as factors
(Table 2). Substituting log (number of eggs hatched) for log (number of eggs
128

deposited) again improved the model's R", albeit only slightly (from 0.52 to 0.56).
Results are essentially identical to the first analysis.
The non-linear relationship between median pupal weight and egg numbers is
illustrated in Figure 2A. Also shown is the change in median pupal weight as a function
of per capita ftuit volume (calculated as fruit volume divided by number of eggs
hatched)(Fig;ure 2B). It appears that as the number of eggs placed into a host increases,
median pupal weight decreases in a negatively accelerating fashion. Median pupal
weight may not decline further than about 0.2 mg because this may reflect the minimal
size at which a larva might be able to successfully emerge from a fruit and pupate. As
may be expected, the negative affects caused by increases in larval infestation levels
are, in turn, tempered by an increase in the amount of available resources per larva.
Experiment 2. Effect of temporal spacing of clutciies on offspring fitness
In this experiment, mean clutch size in a two-clutch fruit was 13.00 ± 0.40 eggs (N s
90). Mean egg density in the three treatments receiving four clutches did not differ
significantly among treatments (F = 2.88, DF = 2, 66, P = 0.06). In contrast, mean egg
density in the two-clutch treatment was significantly lower than mean egg density in
each of the four-clutch treatments (two-clutch vs. four-clutch contrast of least squares
means; F = 57.24, DF = 1, 86, F < 0.0001; Figure 3A). Mean fhiit volume, 24.02 + 0.31
cm^, did not diffisr among treatments (F = 0.67, DF = 3, 89; P = 0.67). Within the four
clutch treatments, there was no difference in egg hatch (F = 0.47, DF = 2,68; P = 0.63)
or pupal survival (F = 1.95, DF = 2,67; P = 0.15), but there was a difference in the
129

number of larvae that hatched and emerged from a fruit (F= 9.0, DF = 2,67; P =
0.0004).
Effect of timing on offspring survival
Total percent survival from egg deposition to successful pupation differed significantly
among ureatments (R" = 0.34, F = 15.27, DF = 3, 83; P < 0.0001; Figure 3B). This level
of significance is due in part to a density effect, percent survival in the lone two-clutch
treatment being significantly higher than in the three four-clutch treatments combined
(two-clutch vs. four-clutch treatment contrast of least squares means; F = 32.66, DF = I,
80, P < 0.0001). However, the statistical significance is due also to a difference in
percent survival between staggered-four-days treatment and either the staggered-two-
days treatment or the no-stagger treatment (Tukey HSD tests, P < 0.05, Figure 3B).
Percent survival in the staggered-two-days treatment and the non-staggered treatment
were not significantly different (Tukey HSD test, P > 0.05, Figure 3B).
While our methods did not permit us to distinguish the fates of individuals in
earlier-laid vs. later-laid clutches, it is reasonable to suppose that members of the earlier
clutches were more likely to survive than members of the later clutches. If so, our
results indicate that later arrival into fruit is associated with reduced survival.
Effect of timing on median pupal weight
Median pupal weight differed significantly among treatments (F = 21.65, DF = 3,85; P
= 0.0001; Figure 3C). This level of significance is due in part to a density effect.
130

median pupal weight in the lone two-clutch treatment being significantly higher than in
the three four-clutch treatments combined (two-clutch vs. four-clutch treatment contrast
of least squares means; F = 24.95, DF = 1, 85, P < 0.0001). However, the statistical
significance is due also to a statistically significant difference in median pupal weight
among the four-clutch treatments. Median pupal weight increased progressively from
non-staggered to staggered-two-days to staggered-four-days treatments (Tukey HSD
tests, F < 0.05, Figure 3B).
The significant difference in median pupal weight between the two-days-
staggered and the four-days-staggered treatments squares well with the significant
difference in percent survival betv.'een the same treatments (Figure 3B & C). The
relatively greater reduction in numbers of competitors in the four-day treatment
probably provided survivors with relatively more food resource and thus resulted in
higher median pupal weights. The significant difference in median pupal weight
between the non-staggered and two-days-staggered treatments is not so readily
explained, because total percent survival did not differ signiticantly between those two
treatments. In general, it appears that the greater the temporal spacing between clutches
the greater the decrease in larval survival of the latter clutches and this may in turn lead
to earlier clutches experiencing a reduction in larval competition and attaining higher
larval weights.
131

Experiment 3. Effect of clutch spacing on offspring fitness
Despite substantial variation in clutch size (clutch size in 1-clutch fruit = 14.75 ± (SE)
0.77 eggs; range = 5-34 eggs; N = 60), number of eggs deposited increased
significantly with the number of clutches in fruit (R~ = 0.51; lest for slope different from
0, t58 = 7.83, P < 0.0001; Figure 4A). In pairwise contrasts, number of eggs deposited
did not differ significantly within either two-clutch treatments or three-clutch
treatments. Mean fruit volume, 25.33 ^1.05cm^, also did not differ among ureatments
(F = 0.94, DF = 4, 60; P = 0.45).
Effect of clutch spacing on offspring survival
Mean total percent survival from egg deposition to successful pupation was 61% ± 0.03
and did not differ overall among treatments (R" = 0.14; F = 2.11, DF = 4, 51; P = 0.10;
Figure 4B). A least squares means conurast of percent survival of the two two-clutch
treatments was not significant (F = 0.057, DF = 1, 51; P = 0.81; Figure 4B). However,
a contrast of percent survival for the two three-clutch treatments indicated that percent
survival for three separated clutches was signiHcantly lower than for three clutches
placed together (F = 7.63, DF = I, 51; P = 0.008; Figure 4B). Total survival was lower
132
for the three separated clutches not because egg hatch or pupal survival were lower (F =
2.68, DF= 1, 24; P = 0.31; F = 0.31, DF = 1, 24; P = 0.58, egg hatch and pupal weight,
respectively), but perhaps primarily because survival from egg hatch to larvae emerging
from a fhiit were lower (F = 3.6, DF = 1,24; P = 0.07). We next conducted a two-way
analysis of variance of effects of clutch number and clutch placement on % total

survival, using two-clutch and three-clutch treatment data only. In that analysis, the
effect of clutch number on percent survival was not signlHcant (F= 0.58, DF = 1,42; P =
0.45). The clutch spacing effect was significant (F= 4.20, DF = 1,42; F = 0.046), but
the interaction between clutch .spacing and clutch number was not (F= 2.87, DF = 1,42;
P = 0.09). Perhaps the range of clutch numbers was not large enough to be able to
measure a clutch number related effect.
Effect of clutch spacing on median pupal weight
An analysis of variance of median pupal weight across clutch spacing treatments was
highly significant (R" = .27, F= 4.95, DF = 1,54; P = 0.002). A substantial portion of
this effect was due to clutch number. When data were pooled over spacing treatment,
median pupal weight declined significantly with number of clutches (1, 2 or 3) (linear
regression, R" = .22, estimate of slope = -12.03, test of slope different from zero, t57 = -
4.17, P = 0.0001; Figure 4C).
133
Pairwise contrasts of least squares means within the original analysis of variance
failed to support the hypothesis that pupal weight depends on clutch spacing. Pupal
weight did not depend on clutch spacing within either the two-clutch treatments (F=
0.07, DF = 1,54; P = 0.80) or the three-clutch treatments (F= 2.58, DF = 1,54; P = 0.11).
Finally, we conducted a two-way analysis of variance of effects of clutch
number and clutch placement on median pupal weight, using two-clutch and three-
clutch treatment data only. In that analysis, number of clutches again had a highly
significant effect on median pupal weight (F= 9.97, DF = 1,45; P = 0.003), with pupal

weight for three-clutch fruit being lower than for two-clutch fruit. However, neither
clutch spacing effects (F= 1.50, DF = 1,45; P = 0.23) nor the interaction between clutch
spacing and clutch number (F= 0.73, DF = 1,45; P = 0.40) were significant.
134

DISCUSSION
Absence of an AUee efTect on offspring fitness
Density effects
In our study, offspring survival and pupal weight, the latter a strong predictor of lifetime
female fecundity (Nufio and Papaj, APPENDIX C), declined monotonically with
increases in conspecific density over a range of conspeciHc densities. Both survival and
pupal weight were highest when fruit were infested by just a single clutch. In short, for
the components of total individual fitness estimated here, there was no obvious Allee
effect. It remains possible that an Allee effect might have been found if fitness were
measured more completely or measured in nature. For example, as has been proposed
for other systems, there may occur an Allee effect that manifests itself in terms of an
inverse density dependence in parasitism rates (Price 1988, Abrahamson and Weis
1997). An increase in aggregation size may lead to individuals being less susceptible to
being found or reached by parasitoids. Larvae of R. juglandis are attacked by a larval-
pupal parasitoid, Biosteres juglandis-, however, we have no reason to think that larvae
would be less at risk of parasitism in larger groups and it is even possible that per capita
risk of parasitism increases with group size (van Alphen and Galis 1983, Hoffineister
and Rohlfs 2001).
135

Temporal spacing of clutches
The temporal spacing of clutches significantly affected total survival and pupal mass of
progeny (Figure 3 B «& C), in a manner wholly consistent with a pattern of competition
between clutches. The effect of temporal spacing of clutches on total percent survival
was due mainly to an effect on larval survival from post-hatch until emergence from the
fruit. Because we did not distinguish pupae from first and second clutches,
interpretation of pupal weight data was necessarily less straightforward. Staggering two
pairs of clutches four days apart in a given fruit, for example, increased median pupal
weight, not decreased it as might be expected intuitively. Our explanation for this effect
is as follows. When four clutches are placed into a fruit simultaneously, larval survival
is relatively high. The high survival means that the amount of resources available per
offspring is small, which leads to low pupal weight. When clutches are spaced four days
apart, early larval survival, in contrast, is probably high for members of early clutches
and, assuming a competitive advantage to larger larvae, low for members of the later
clutches (Figure 4). The earlier-laid, surviving offspring enjoy a relatively greater per
capita amount of fruit volume and consequently achieve a higher median pupal weight
than offspring of clutches deposited simultaneously.
The lower survival of later clutches could be a result either of the first clutches
consuming ail available resources, or of a decline in fhiit quality associated with larval
infestation (Huang et al. 1971), or of a competitive superiority experienced by earlier
laid clutches over later laid clutches (Averill and Prokopy 1987). As the two clutches
that emerged from the four-clutch treatments fruit separated by four days attained
weights no different than those attained by two clutches being deposited into fruit
136

simultaneously. It would seem that the clutches deposited four days after the first two
had little effect on the ability of members of the first two clutches to acquire their share
of resources. While we did not determine directly whether the larvae that emerged from
the treatment in which clutches were separated by four days, were from the first or
second clutches, it is reasonable to suppose that emerging larvae would be from the first
two clutches as these larvae were about the same size as those emerging from fruit that
contained only two clutches. If the first two clutches had initially developed but were
usurped by the latter two clutches, the first clutches would have been able to consume
resources for at least a few days and the later two clutches, though dominant, would
have to emerge at a smaller size than if they were the only consumers. Still, we cannot
exclude the possibility that some individuals from later clutches survived and that they
replaced some early larvae that failed to survive.
The explanation for why the four-clutch treatment separated by two days
resulted in heavier pupal weights than clutches deposited simultaneously is a bit
difficult to explain. This result conceivably reflects an advantage to clutches that are
laid two days or so after the first clutches are laid. However, this result is also
consistent with a scenario of competitive disadvantage for later-laid clutches. The
137
relatively greater pupal weight of 2-days-staggered survivors may mean that, despite the
similarity in overall percent survival, survivors in that treatment had more food
available to them than survivors in the non-staggered treatment. This food differential
might arise if larvae fated to die in the non-staggered treatment consumed relatively
more food before djdng than larvae fated to die in the staggered treatment. Such a
difference is not unreasonable to expect, if the two-day head start for members of

earlier-laid clutches U-anslated to a competitive advantage over members of later-laid
clutches. In the non-staggered treatment, all larvae are growing at similar rates over
most of their development; however, some fail to grow fast enough to reach pupation
before food runs out. bi the 2-days-staggered Ureatment, larvae in the earlier-laid
clutches get a 2-day head start on larvae in the later clutches. Owing to this advantage,
larvae are larger than and growing much more rapidly than larvae in the later-laid
clutches. As a consequence, larvae in the later-laid clutches are more likely to die at a
smaller size than larvae in the non-staggered treatment. Possibly, they also die at an
earlier age than larvae in the 2-days-staggered treatment. Either or both patterns in
mortality will give survivors in the 2-days-staggered treatment potentially more food
per capita than survivors in the non-staggered treatment.
In our opinion, the latter explanation is more parsimonious than an alternative
explanation that supposes that survivors get a bang out of being deposited into fruit 2
days apart but get their ass kicked when deposited into fruit 4 days apart.
All in all, it appears that four clutches deposited on the same day had higher
survival than an equal number of clutches separated by two or four days, though
offspring survival and size were still not greater then when fewer (two) clutches were
138
placed into a fruit. It appears that an increase in density decreases offspring survival and
weight, while temporal staggering of clutches benefits earlier-laid clutches, bi short, it
appears our temporal staggering experiment does not support a mechanism by which an
Allee effect may be achieved by clutches being simultaneously deposited into a fruit.

Spatial distribution of clutches
In regards to the spatial distribution of clutches experiment, we may have found
evidence that suggests that clutches placed in the same oviposition cavity perform better
than clutches spaced evenly apart on a fruit. Whether clutches were placed together or
spaced apart did not seem to differentially affect total survival or median pupal weight
of the two clutch treatments, but the three clutch treatments experienced greater
offspring survival when the clutches we placed together (Figure 4B). Possibly, the
trend becomes yet more pronounced as more clutches are added to existing cavities.
Still, while reuse of oviposition puncmres may temper larval survival, offspring
themselves are not experiencing higher survival or weight gains as the number of
clutches placed within an oviposition cavity increases. Our current density experiment
and a previous field experiment (Nufio and Papaj, APPENDIX C), for example, show
that, in general, increases in larval density decrease both larval survival and pupal
weights (Nufio Papaj, APPENDIX C).
From an offspring's perspective, it thus appears that levels of competition are a
function of the overall number of eggs in a fruit and of exactly the time at which an egg
is laid relative to previously-laid eggs (see below) and, perhaps when densities are high
139
enough it may also be a function of degree of spatial proximity to conspeciHcs. Perhaps
larvae being placed together experience increased developmental rates or perhaps a
localized and moving larval front might degrade the host at a more manageable rate that
when clutches degrade the fruit from multiple points. While the first possibility would

allow larvae to access available host resources more rapidly, the latter may make
available resources available for a longer period of time.
The absence of an Allee effect associated with an increase in larval infestation
levels or with the spatial or temporal spacing of clutches squares with the results of a
previous field study (Nufio and Papaj, APPENDIX C)
It is worth noting, however, that in our current laboratory experiments, the
proportion of variation in total survival and size that is explained by variation in egg
density was quite low in relation to field results (Nufio and Papaj, APPENDIX C)
(Figure I C & B). This implies that some factor correlated with egg density, and not egg
density itself, accounts for the striking effect of reuse on offspring survival observed in
the field. We believe that factor to be temporal patterning. In the field study, egg
density was confounded with the temporal spacing of successive clutches. A fruit
harboring relatively high densities of eggs was also a fruit in which the time between
first and last clutches was relatively long. We successfully estimated egg density in the
field, but had no way of assessing the temporal patterning of clutches. Yet, since this
study shows that clutches deposited at relatively later times suffered more in terms of
competition than clutches deposited earlier, temporal patterning could well account for
the observed patterns in reduced offspring fitness.
Why do females reuse hosts?
140

If reuse of larval hosts negatively impacts offspring survival and size at pupation, why
do females reuse hosts? One potential reason that females fail to reject previously
utilized hosts is that ovipositing females are unable to distinguish between previously
infested and uninfested hosts. However, several lines of evidence suggest that females
can distinguish between such hosts. First, in previous experiments, gravid female
walnut flies were not only found to selectively visit fruit containing artificial punctures,
but these females also frequently deposited clutches directly into the punctures (Papaj
1994, Lalonde and Mangel 1994). Secondly, like other tephritid fruit flies (Landolt and
Averill 1999), female R.juglandis deposit a MP on the fruit surface following clutch
deposition and this mark has a deterrent effect on oviposition (Nufio and
Papaj,APPENDIX E, also see Cirio 1972). In addition, R. juglandis females mark the
fhiit for a duration that is directly proportional to the size of their deposited clutch and
in turn, the marking pheromone appears to deter further oviposition in proportion to the
amount that is deposited on the fruit (Nuflo and Papaj, APPENDIX E, Papaj and
Alonso-Pimentel unpub. data). These results suggest that females behave as though
their larvae experience a level of competition that is proportional to larval density.
These results also imply that females behave as though such costs are acceptable to
bear, else why would they signal information about levels of competition in this way.
While costly in terms of offspring survival and weight at pupation, females
reuse hosts because such reuse provides them with direct beneflts, such as reduced time
to deposit clutches (Papaj and Alonso-Pimentel 1997), reduced ovipositor wear (Papaj
1993), and increased access to relatively impenetrable &uit (Lalonde and Mangel 1994)
141

(also see Nufio et al. 2000, Nufio and Papaj, APPENDIX C). In turn, these benefits may
allow females to deposit more clutches over time; presumably the increased numbers of
eggs that are laid by reusing fruit more than compensates for the losses experienced in
terms of reduced per capita offspring fitness. Finally, while reusing oviposition
punctures may provide females with benefits associated with the deposition of clutches,
the placement of multiple clutches into the same oviposition site may also increase the
survival of offspring relative to that of clutches that were deposited uniformly along a
fruit. Still, while these larval aggregations may increase relative levels offspring
survival, offspring still experience greater survival and increased pupal weights when
few clutches are deposited within a fruit. As such benefits experienced by larval
aggregations may be better taught of as a mechanisms that decrease or temper larval
competition for limited than a mechanism that enhances and Allee effect.
142

ACKNOWLEDGEMENTS
We thank Henar Alonso-Pimentel and Laurie Henneman for discussion and assistance
throughout. Sheridan Stone of the Fort Huachuca Wildlife Management office of the
US Army provided permission for collecting fhiit and logistical support in Garden
Canyon. This research was supported a National Science Foundation Minority Graduate
Research Fellowship, and NRICGP grant no. 93-37302-9126 to D.R.P and Sigma Xi.
We acknowledge Judie Bronstein, Reg Chapman, Bob Smith and Molly Hunter for
providing feedback on earlier drafts.
143

REFERENCES
Abrahamson WG, Weis AE, 1997. Evolutionary Ecology accross Three Trophic Levels:
Goldenrods, Gallmakers, and Natural Enemies. Princeton, NJ: Princeton
University Press.
Averill AL, Prokopy RJ, 1987. Intraspecific competition in the tephritid fruit fly
Rhagoletis pomonella. Ecology 68:878-886.
Bush GL, 1966. The taxonomy, cytology and evolution of the genus Rhagoletis in
North America (Diptera: Tephritidae). Bulletine of the Museum of Comparative
Zoology of Harvard University 134:431-562.
Byers JA, 1989. Chemical Ecology of Bark Beetles. Experientia 45:271-283.
Cirio U, 1972. Osservazion sul comportamento di ovideposizione deWdi Rhagoletis
completa in laboratorio. Proceedings, 9th Congress Italian Entomological
Society (Siena):99-117.
Corbet SA, 1973. Concentration effects and response of Nemeritis canescens to a
secretion of its host. Journal of Insect Physiology 19:2119-2128.
Courtney SP, KibotaTT, Singleton TA, 1990. Ecology of mushroom-feeding
Drosophilidae. Advances in Ecological Research 20:225-275.
Credland PF, Dick KM, Wright AW, 1986. Relationship between larval density, adult
size and egg production in the cowpea seed beetle, Callosobruchus maculatus.
Ecological Entomology 11:41-50.
144

Hausmann SM, Miller JR, 1989. Ovipositional preference and larval survival of the
onion maggot (Diptera: Anthomyiidae) as influenced by previous maggot
feeding. Journal of Economic Entomology 82:426-429.
Hedlund K, Bartelt RJ, Dicke M, Vet LEM, 1996. Aggregation pheromones of
Drosophila inmiigrans, D-phalerata, and D-subobscura. Journal of Chemical
Ecology 22:1835-1844.
Hoffmeister TS, Rohlfs M, 2001. Aggregative egg distributions may promote species
co-existence - but why do they exist? Evolutionary Ecology Research 3:37-50.
Howard DJ, Bush GL, 1989. Influence of bacteria on larval survival and development in
Rhagoletis (Diptera: Tephritidae). Annals of the Entomological Society of
America 82:633-640.
Howard DJ, Bush GL, Breznak JA, 1985. The evolutionary significance of bacteria
associated with Rhagoletis. Evolution 39:405-417.
Huang SL, Singh M, Kojima K, 1971. A study of frequency dependent selection
145
observed in the esterase-6 locus of Drosophila melanogaster using a conditioned
media method. Genetics 68:97-104.
Kirkendall LR, Kent DS, Raffa KF, 1997. Interactions among males, females and
offspring in bark and ambrosia beetles: the significance of living in mnnels for
the evolution of social behavior. In: The Evolution of Social Behavior in Insects
and Arachnids (Choe JC, Crespi BJ, eds). Cambridge, UK: Cambridge
University Press.
Lalonde RG, Mangel M, 1994. Seasonal effects on superparasitism by Rhagoletis
completa. Journal of Animal Ecology 63:583-588.

Landolt PJ, Averill AL, 1999. Fruit Flies. In: Pheromones of Non-Lepidopteran Insects
Associated with Agricultural Plants (Hardie J, Minks AK, eds). New York:
CABI Publishing; 3-26.
Messina FJ, Renwick JAA, 1985. Ability of ovipositing seed beetles to discriminate
between seeds with differing egg loads. Ecological Entomology 10:225-230.
Nufio CR, Papaj DR, 2001. Host marking behavior in phytophagous insects and
parasitoids. Entomologia Experimentalis Et Applicata 99:273-293.
146
NuHo CR, Papaj DR, 2002a. Host marking behavior as a quantitative signal of infestation
levels in host use by the walnut fly, Rhagoletis juglandis. Appendix E.
Nufio CR, Papaj DR, 2002b. Reuse of larval hosts by the walnut fly, Rhagoletis
juglandis, and its implications for female and offspring performance. Appendix C.
Nufio CR, Papaj DR, Alonso-Pimentel H, 2000. Host utilization by the walnut fly,
Rhagoletis juglandis (Diptera: Tephritidae). Environmental Entomology 29:994-
1001.
Paine TD, Raffa KF, Harrington TC, 1997. Interactions among scolytid bark beetles,
their associated fiingi, and live host conifers. Annual Review of Entomology
42:179-206.
Papaj DR, 1993. Use and avoidance of occupied hosts as a dynamic process in tephritid
fimit flies. In: Insect-Plant Interactions (Bemays EA, ed). Boca Raton: CRC
E*ress; 25-46.
Papaj DR, 1994. Oviposition site guarding by male walnut flies and its possible
consequences for mating success. Behavioral Ecology and Sociobiology 34:187-
195.

Papaj DR, Alonso-Pimentel H, 1997. Why walnut flies superparasitize: time savings as
a possible explanation. Oecologia 109:166-174.
Price PW, 1988. Inversely density-dependent parasitism: the role of plant refuges for
hosts. Journal of Animal Ecology 57:89-96.
Prokopy RJ, 1981. Epideictic pheromones that influence spacing patterns of
phytophagous insects. In: Semiochemicals: Their Role in Pest Control (D.A.
Nordlund RLJ, and W. J. Lewis, ed). New York: Wiley Press; 181-213.
Prokopy RJ, Roitberg BD, 2001. Joining and avoidance behavior in nonsocial insects.
Annual Review of Entomology 46:631-665.
Pureswaran DS, Cries R, Borden JH, Pierce HD, 2000. Dynamics of pheromone
production and communication in the mountain pine beetle, Dendroctonus
ponderosae Hopkins, and the pine engraver, Ips pini (Sag) (Coleoptera:
Scolytidae). Chemoecology 10:153-168.
Raffa KF, Phillips TW, Salom SM, 1993. Strategies and mechanisms of host
colonization by bark beetles. In: Beetle-Pathogen Interactions in Conifer Forests
(Schowalter T, Filip G, eds). San Diego: Academic Press.
Reeve JD, Rhodes DJ, Turchin P, 1998. Scramble competition in the southern pine
beetle, Dendroctonus frontalis. Ecological Entomology 23:433-443.
Sang JH, 19S6. The quantitative nutritional requirements of Drosophila melanogaster.
Journal of Experimental Biology 33:45-72.
Telang A, Hammond SS, Opp SB, 1996. Effects of copulation firequency on egg-laying
and egg hatch in the walnut husk fly, Rhagoletis completa Cresson. Pan-Pacific
Entomologist 12:235-231.

van Alphen JJM, Galis F, 1983. Patch time allocation and parasitization efficiency of
Asobara tabida, a larval parasitoid of drosophila. Journal of Animal Ecology
52:937-952.
Wertheim B, 2001. Ecology of Dosophila aggregation pheromone: a multitrophic
approach: Wageningen University.
Wilson K, 1988. Egg laying decisions by the bean weevil Callosobruchus maculatus.
Ecological Entomology 13:107-118.
148

Table 1. The influence of various factors on total survivorship (from eggs deposited to
successful pupation) and the influence of various factors on survivorship of larvae from
successful egg hatch to emergence from a host fruit.
Total percent survival
(deposited egg to successful pupation)
149
MULTIPLE REGRESSION ANALYSES
DF ESTIMATE
Number of clutches deposited 5.83 -0.095 0.0689
Log number of eggs deposited
Fruit volume
Model error DF = 62
R'= 0.09
2.65
0.01
% Survival from egg hatch to larvae emerging from fruit
0.1088
0.9070
Number of clutches deposited 6.52 1 -0.1031 0.0345
Log number of eggs that hatched 2.31 1 0.1118 0.1339
Log number of eggs deposited
Fruit volume
Model error DF = 60
R'= 0.11
0.09
0.19
0.7637
0.6618

Table 2. The influence of various factors on egg hatch (either 100% hatch or not) and
survival from larvae emerging from a fruit and successfully pupating (either 100%
survival or not).
% Egg hatch
Number of clutches deposited
Number of eggs deposited
Fruit volume
150
MULTI-FACTORIAL LOGISTIC REGRESSION
Wald X' PF Estimate P
0.19
1.25
0.40
-0.14
-0.02

Table 3. The influence of various factors on median pupal weight within the density
manipulation experiment.
Median pupal weight
Log number of eggs hatched
Fruit volume
Log number of larvae emerging
from a fruit
Number of clutches deposited
Log number of eggs deposited
Model error DF = 61
R'= 0.59
151
MULTIPLE REGRESSION ANALYSES
F PF ESTIMATE P
20.21 1 -15.680

FIGURE CAPTIONS
Figure 1. Effects of offspring density within a walnut hoston o^spring survival and
weight. (A) Number of eggs deposited within each treatment. (B) Percent
survival of offspring within each treatment. (C) Median pupal weight of
offspring emerging from a fruit and successfully pupating. The number of
replicates for each treatment range from 11 to 14. Bars sharing same letters
are not significantly different (Tukey HSD, P < 0.05).
Figure 2. (A) Median pupal weight as a function of the number of eggs that hatched
within a fruit. (B) Median pupal weight as a function of the available fruit
volume per egg that hatched.
Figure 3. Effects of offspring density and spatial distribution of clutches within a
walnut host on offspring survival and weight. (A) Number of eggs deposited
within each treatment.(B) Percent survival of offspring within each treatment.
(C) Median pupal weight of offspring emerging from a fruit and successfully
pupating. Number of replicates for each treatment range from 11 to 14. Bars
sharing same letters are not signiflcantly different (Tukey HSD, P < 0.05).
152

FIGURE CAPTIONS continued
Figure 4. Effects of offspring density and temporally spacing clutches within
a walnut host on offspring survival and weight. (A) Number of eggs deposited
within each treatment (Tukey HSD, P < 0.05). (B) Percent survival of offspring
within each treatment. (C) Median pupal weight of offspring emerging from a
fruit and successfully pupating. The number of replicates for each treatment
range from 21 to 24. Bars sharing same letters (for B and C) are not significantly
different (Least squares contrasts removing density affects, p < .05).
153

FIGURE 1.
1 2 3 4 5
Number of Clutches Deposited within Fruit

Figure 2
MD
E
DC
n
Om
3
Qi
S
V
DA
E
ec
'S
a
e
es
W
s
Number of eggs hatciied
Fruit volume per hatched egg (cm^

Figure 3.
A.
B.
•2
u
c. ^
S 40
> .40
2-day 0 on Day 0 2 on Day 0 2 on Day 0
2 on Day 2 2 on Day 4
2-day 0 on Day 0 2 on Day 0 2 on Day 0
2 on Day 2 2 on Day 4
2-day 0 ^ on Day 0 2 on Day 0 2 on Day 0
2 on Day 2 2 on Day 4
Number and Temporal Pattern of Clutch Deposition
156

Figure 4.
B.
c.
-g 60
S a.
a 40 -
S

APPENDK E
HOST MARKING BEHAVIOR AS A QUANTITATIVE
SIGNAL OF INFESTATION LEVELS IN HOST USE BY
THE WALNUT FLY, RHAGOLETIS JUGLANDIS
158

Host marking beiiavior as a quantitative signal of infestation levels in host
use by the walnut fly, Rhagoletis juglandis
Cesar R. Nufio'-^
and
Daniel R. Papaj^
Department of Entomology*. Department of Ecology and Evolutionary Biology^ Center
for Insect Science^, and the Interdisciplinary Degree Program in Insect Science'^,
University of Arizona, Tucson, AZ 85721, USA
159

ABSTRACT
In phytophagous and entomophagous insects that develop in discrete hosts such as
seeds, small fruit or other insects, competition among offspring for available resources
can be intense. To minimize offspring competition within these discrete resources,
females have been found to reject hosts that harbor conspecific brood. Despite the costs
associated with reusing hosts and despite displaying the genus-typical host marking
behavior, walnut flies in the Rhagoletis suavis clade reuse hosts through the season. In
this study we examined the role of the putative marking behavior on the egg laying
decisions of the walnut fly R. juglandis. The amount of time females spent ovipositor
dragging following clutch deposition was correlated with the number of eggs a female
deposited within a fruit. Dragging time was not correlated with female size nor her egg
load prior to egg laying. The correlation between dragging time and deposited clutch
size suggests that the amount of putative marking pheromone deposited on the fhiit
surface may be a reliable indicator of the number of eggs previously deposited within a
fruit. In a second experiment, we therefore attempted to determine if responses to a
marked fruit depended on how long females had marked the fruit. Females were
allowed to ovipositor-drag on fruit for 0, 10,20, or 30 minutes and these fhiit were
placed within a field cage where other gravid females were allowed to forage. Whether
or not foraging females alighted on a particular host was not affected by the amount of
time that previous females had spent dragging on the fruit. However, the propensity to
probe the fruit and to deposit a clutch within the fruit was negatively correlated with
dragging time. In a follow-up experiment, females foraging within the freld cage
160

encountered either control fruit that contained clutches deposited by 25 females or fruit
previously exposed to 25 females that were also allowed to deposit marks on the fruit.
While females alighted on fruit containing marks or no marks equally, the presence of
marks reduced a foraging female's propensity to oviposit from 46% to nearly 10%. We
propose that R. juglandis deposits an active marking pheromone that deters egg laying
but that female physiology as well as fruit characteristics, temporal spacing of clutches
(inferred through patterns of fruit degradation), and other environmental variables affect
the degree to which females reuse hosts.
Key words marking pheromone reproductive trade-offs' oviposition deterrent'
oviposition-preference-offspring-perfonnance7?/tago/er/5 juglandis' superpatasitism'
Tephritidae walnut flies
161

INTRODUCTION
The decisions insects make about where their offspring will develop are especially
important for species with larval stages that are restricted to a particular environment or
host (Thompson 1983, Smith and Lessells 1985, Roitberg and Prokopy 1987, Bemays
and Chapman 1994). Because insect larvae that develop within discrete hosts (e.g.
flower buds, seeds, small fruit, stems, or other insects) may not be able to acquire more
resources if natal hosts become depleted, larval development can be affected very
strongly by level of competition for food resources. Given that brood competition can
reduce offspring survival, size and reproductive potential and increase the time required
to reach maturity (rev. Peters and Barbosa 1977, Fox et al. 1996, Blanckenhom 1998,
Sweeney and Quiring 1998, Allen 2001), it is not surprising that females in many
systems have evolved mechanisms for avoiding use of previously exploited hosts (rev.
Nufio and Papaj 2001).
In order to reduce the competition for larval resources their offspring may face,
females of a variety of species reject previously exploited hosts on the basis of visual or
tactile stimuli associated with eggs (Rausher 1979, Williams and Gilbert 1981, Shapiro
1981, Takasu and Hirose 1988) or larvae (Mappes and M^ela 1993, rev. Nufio and
Papaj 2001). Females in some phytophagous species use chemical cues associated with
larval frass (Mitchell and Heath 198S) or with plant damage caused by oviposition
(Cirio 1971) or larval feeding (Renwick and Radke 1981, Fitt 1984, Landolt 1993).
Similarly, in some entomophagous parasitoids, females may discriminate between
162

unparasitized and parasitized hosts on the basis of changes in a host's hemolymph
composition induced by parasitization (Fisher and Ganesalingam 1970).
In addition to cues produced incidentally by the presence of conspecifics, a
variety of entomophagous and phytophagous insects deploy signals of conspeciHc
presence that are actively produced by the sender female. In almost all cases, the signal
consists of a chemical termed a marking pheromone (MP) that has been deposited on or
in the host; this MP almost always has a deterrent effect on oviposition (rev. Roitberg
and Prokopy 1987, Hofsvang 1990, Landolt and Averill 1999, Nufio and Papaj 2001).
Tephritid fruit flies in the genus Rhagoletis, for example, deposit clutches within
developing fhiit, where larvae are constrained to feed and develop. In the apple maggot
fly, R. pomonella, larval competition is intense. Infestation of hawthorn berries typically
yields no more than one or a few pupae, even if considerably more eggs are deposited
into the fruit (Averill and Prokopy 1987, Feder et al. 1995). Following oviposition, a
female drags its ovipositor about the fruit surface, depositing a MP that deters other
females from reusing the occupied host. In R. pomonella. as well as other tephritids, MP
reduces larval competition by causing females to distribute their clutches uniformly
over host fruit (Bauer 1986, Averill and Prokopy 1989).
Members of the Rhagoletis suavis group, the so-called walnut flies, differ from
members of other Rhagoletis species groups with respect to their oviposition responses
to egg-infested fhiit. Not only do females deposit significant numbers of eggs into
previously egg-infested fruit, but they commonly deposit eggs directly into previously
established oviposition cavities formed within the walnut husks (Papaj 1993,1994,
Lalonde and Mangel 1994, Nufio et al. 2000). In Rhagoletis juglandis, the focus of the
163

present study, reuse of egg-infested fruit is common despite the fact that females engage
in what appears to be the genus-typical pattem of host-marking behavior (Papaj 1994,
Nufio et al. 2000).
Reuse of oviposition cavities, in particular, apparently provides females with
direct benefits such as reduced ovipositor wear (Papaj 1993), reduced time to deposit
clutches (Papaj and Alonso-Pimentel 1997). and increased access to less penetrable fruit
(Lalonde and Mangel 1994). However, the existence of such benefits begs the question
of why females appear to mark fruit in a manner almost identical to that observed in
Rhagoletis species that strongly avoid reuse of occupied fruit. We can think of three
hypotheses for the occurrence of this behavior in R. juglandis. First, host-marking
behavior may be an evolutionary relic. This would be somewhat surprising given
evidence that MP's have costs associated with their production and reception, as well
ascosts associated with eavesdropping by parasites and time spent engaged in ovipositor
dragging behavior (Hoffineister and Roitberg 1998, NuHo and Papaj 2001).
Nevertheless, this hypothesis is testable. If the behavior is relict, we would predict that
the behavior would result in no change in response to a fruit on the part of females
subsequently visiting the fruit.
Alternatively, the mark might serve to guide females to existing oviposition
cavities. If reuse of oviposition cavities confers direct benefits on females, then perhaps
a female might benefit from marking such sites, if the mark facilitated later use of that
site by her or her kin. In this case, we would predict that females would be attracted to
or arrested on fruit that bear the putative mark.
164

Finally, it is conceivable that the mark signals a cost associated with reuse,
specifically a cost in terms of losses in larval fitness associated with competition. In this
case, the mark would permit females to weigh the cost of competition to be suffered by
their progeny, were they to lay eggs in the occupied fruit, against any benefits that
accrue directly to females. Based on this final hypothesis, we might expect a female's
response to marked fruit to decline with increasing numbers of eggs in the fruit. This
expectation leads in turn to two predictions which are tested here. First, females should
deposit an amount of mark that is proportional to the size of their clutch. Second,
females should be deterred by the mark in proportion to the amount deposited.
165

MATERIALS AND METHODS
Natural history
Rhagoletis juglandis is a member of the walnut-infesting Rhagoletis suavis group (Bush
1966). In southern Arizona, this species is found on the Arizona walnut, Juglans major,
which occurs in montane canyons between 1200 and 2700 meters. These flies are
univoltine and females deposit clutches of ca. 16 eggs (Nufio et al. 2000) after piercing
the fruit surface with their ovipositor and hollowing out a small cavity in the walnut
husk. The larval stages feed on the husk of a single fruit, after which they emerge and
pupate in the soil beneath the natal tree. Pupae diapause through the winter and spring
and adults emerge during mid to late summer of a subsequent year.
General protocol
Adult female flies used in the laboratory and Held cage experiments originated as larvae
from fruit collected one to two years prior to their emergence in the lab from Garden
Canyon in the Huachuca mountains in southern Arizona. Upon emergence, flies were
reared in 3.79-liter plastic containers and provided with ad libitum water, sugar, and
slips of a yeast hydrolysate and sugar mixture. Flies were stored in a room with a
14: lOh lightrdark cycle and a day/night temperature of 32°C and 28 "C. Fruit used in all
tests were ripe Arizona walnut {Juglans major) fruit and were collected several days
prior to their use from several localities in Pima and Cochise counties.
166

Experiment 1. Relationship between clutch size and time spent engaged in a
putative host-marking behavior
In this experiment we attempted to determine if a female's putative host-marking
behavior potentially carries information about the size of the female's clutch. In
particular, we estimated the duration of ovipositor-dragging behavior in relation to
clutch size and in relation to two factors, egg load and female size, that are potentially
confounded with clutch size. Previous studies on congeners suggest that ovipositor-
dragging duration is a reasonable estimate of the amount of MP deposited on a fruit
(Averill and Prokopy 1980). Fruit used in this experiment had diameters of 26 to 34
nun. Flies used in laboratory experiments were 12 to 30 days old, post-eclosion. Ten to
fifteen mature females were placed into clear 16-fl oz (473-ml) plastic Solo cups, fitted
with petri dish lids, in which they were provided with ad libitum water, sugar, and a
yeast hydolysate and sugar mixture). A walnut fruit was hung from the top of the cage
and females were permitted to oviposite. After a female initiated oviposition, the firuit
upon which the fly oviposited were gently removed from the cup. Any other females
present on that fhiit were removed by aspiration. The fruit was then hung from the top
of an empty cup cage and the female allowed to complete oviposition. After oviposition
was completed, we recorded the length of time (in seconds) that a female dragged her
ovipositor over the walnut surface. Female walnut flies typically ovipositor-drag for
several seconds to minutes, then stop and often groom or walk about, and then resume
ovipositor-dragging behavior. So as to minimize the chances that we ended an
ovipositor-dragging session prematurely, in the experiments where we recorded
marking duration, we stopped a session only if a female flew from the host on which
167

she was ovipositor-dragging and spent 15 consecutive minutes off of the fhiit. Total
marking duration for a female did not include pauses that occurred between dragging
bouts, whether those pauses occurred on or off the fruit.
After an observation session for a particular female was terminated, the female
was placed into a labeled 1.5 ml plastic snap-cap vial, and frozen at -4°C. After the
experiment was completed, all females were dissected in saline under a stereoscope and
the number of mature eggs remaining in their ovaries counted. With the use of an ocular
micrometer, we estimated female size by measuring the length of the medial vein
bordering the anterior portion of the discal medial cell. This wing measure could be
made rapidly and previous laboratory experiments indicated that it was strongly
correlated with other indicators of female size such as thorax size, head width and
femur length (A. Lachman, unpublished data). A female's egg load at the time of
testing was estimated as the sum of the number of eggs deposited into the fruit and the
number of mature eggs remaining in her ovaries.
Experiment 2. Oviposition responses to fruit marlced for variable periods of time.
Generating fruit in dragging duration treatments
In this field cage experiment, we attempted to measure responses of gravid females that
encounter walnut hosts on which ovipositor-dragging behavior had occurred for 0, 10,
20, or 30 minutes. Fruit used in this field cage study, as well as those use in the
following field cage study, had diameters that ranged from 30 to 35 mm. Fruit in the
four dragging duration treatments were generated as follows. Four walnuts, similar in
size and ripeness, were randomly assigned to be marked for a given duration (0,10,20
168

or 30 minutes). Fruit were then hung from the tops of 473 ml clear, plastic cup cages
containing 10 to 15 gravid females (12-30 days post-eclosion). Females were permitted
to deposit clutches into the firuit and to engage in ovipositor-dragging behavior
afterwards. In this experiment, we were interested in assessing effects of variation in
ovipositor-dragging behavior, independent of the effects of variation in numbers of
clutches and numbers of oviposition cavities. We therefore controlled for variation in
clutches and cavities by having all fruit (even the fruit assigned to be in the unmarked
fruit treatment) receive 3 clutches deposited into the same oviposition cavity. In order to
ensure that clutches were placed into the same cavity, we gendy brushed gravid females
towards the previously-created cavities. After one or several attempts at new sites,
females usually encountered the previously made oviposition site and reused it.
The duration of ovipositor-dragging behavior on fruit in each of the four
treatments was achieved as follows. Fruit were placed into a cup cage and after a female
initiated oviposition, both fruit and fly were gently removed from the cup. Any other
females present on that fruit were removed by aspiration. The fruit was then hung from
the top of an empty cup cage and the female was allowed to complete oviposition. For
fruit assigned to the 0 minutes dragging treatment, females were removed inmiediately
after deposition of a clutch and before any ovipositor dragging behavior commenced.
Typically, female ovipositor dragging behavior follows the deposition of a clutch into a
fruit (see results below). Females in treatments where they were allowed to drag on the
host fruit nearly always did so after they spent several minutes with their ovipositors
inserted into the fruit. As such, we assumed that females not allowed to drag their
ovipositor nevertheless deposited clutches within fruit.
169

For firuit assigned to the remaining treatments, females were permitted to drag
their ovipositors over the fhiit surface until they stopped or until 10, 20, or 30 minutes
of aggregate duration on the fhiit had been achieved. If the total duration of ovipositor-
dragging behavior by the first female did not reach criterion for the treatment to which
the fruit was assigned, the fruit was returned to a cup cage of females and the above
cycle repeated. This procedure continued either until the dragging duration criterion
was reached or until a fruit had received 3 clutches within the same cavity. Sometimes,
the aggregate dragging duration on a fruit reached criterion before three clutches were
deposited. In that case, females were permitted to add clutches to the existing cavity
but were not permitted to ovipositor-drag after completion of the clutch.
170
Usually, fruit that required 10, 20 and especially 30 minutes of marking required
more than three females (and hence more than 3 clutches) to mark the fruit. So as to
reach ovipositor-dragging duration criterion, we arranged for females to deposit
clutches into non-treatment fruit but to engage in ovipostor-dragging behavior on
treatment fiiiit. This was done as follows. After these females finished ovipositing into
non-treatment fruit, they were gently coaxed onto a thin paper tip (made of filter paper),
placed at the end of a 15.2 cm long wooden dowel, which had been previously dipped
into a sucrose solution. These females were then placed onto fhiit that had reach their
three clutch maximums but which still required females to drag ovipositors on the fiuit
for a greater amount of time. Females so transferred readily engaged in ovipositor-
dragging behavior on the new host (i.e. E*rokopy et al. 1982) Once the criterion dragging
duration was reached on a given fruit, females were aspirated from the fruit.

After two replicates of the four treatments were generated, fruit were stored
overnight in a refrigerator and used in experiments the following morning.
Behavioral observations
Females used in both field cage experiments were treated as follows. Once a cohort of
females, reared as in the general protocol above, was 12-25 days old, they were chilled
and individually marked by placing two dots of non-toxic tempura paint on their thorax.
By using four colors in any combination, we were able to assign unique marks to all test
females. After females were marked, they were placed into a separate 3.79-liter plastic
container containing ad libitum water, sugar, and slips of a yeast hydrolysate and sugar
mixture, as well as two ripe walnuts. Females were allowed to deposit clutches over a
24 hr period, so as to reduce egg load and provide them with egg-laying experience.
This procedure was designed to heighten their sensitivity to host marking pheromone
since, in other species, egg load (van Randen and Roitberg 1996) and previous
experience (Potting et al. 1997, Roitberg et al. 1993) have been found to affect a
female's responses to marked hosts.
After being exposed to a walnut host for 24 hrs, 30 - 40 females were released
into a 3 m-wide by 3 m tall cylindrical, nylon-screen field cage erected within an
evaporator-cooled greenhouse at the University of Arizona Agriculture Experimental
Station in Tucson, and allowed to acclimate for 48 hrs before the experiment
commenced. During this time, females had access to slips of a yeast hydrolysate and
sugar mixture, and small water bottles with extended cotton wicks distributed among
the branches of two 2-3m high potted walnut trees placed inside the cage.
171

On the morning of the test, treatment fruit were removed from the refrigerator
and allowed to warm for roughly an hour and a half before use in field cage tests. Fruit
were then distributed uniformly among several branches. To control for position effects,
fruit were rotated over positions within the trees every 45 minutes. Behavioral
observations typically began at 0900 and ended anywhere between 1200 and 1530,
depending on female activity levels.
After distributing treatment fruit, we recorded the first fruit on which a
particular female alighted and whether the female probed the fruit with her ovipositor or
initiated oviposition into the fruit. A probing consisted of a female curling her abdomen
and firmly poking the fruit with the end of her ovipositor. The initiation of an egg-
laying or oviposition event was noted when a female placed her ovipositor into an
oviposition cavity constructed previously by females or less often when a female
initiated the construction of her own oviposition cavity into the fruit. Creation of an
oviposition cavity occurs when a female inserts her ovipositor into the walnut husk and
then pivots repeatedly around the ovipositor. At the point at which a test female
stopped pivoting and became quiescent, she was immediately removed from the fruit in
order to prevent her from depositing eggs into the fruit. Diuing the experiment,
172
greenhouse temperature and humidity ranged from 22-2TC and 40%-50%, respectively.
Because of the amount of time required to set up each of the dragging duration
treatments, replicate fruit were typically used for two consecutive observation periods.

Experiment 3. Female responses to heavily marked fruit
In the ^rst Held cage experiment a fruit assigned to the 30 minute treatment typically
required the efforts of 9 to 12 females. In the following experiment, we were interested
in increasing the level of ovipositor-dragging duration still further and used a different
protocol to establish fruit treatments.
173
Fruit (30 to 35 mm in diameter) were paired for size and then randomly assigned
to a 'dragged on' or 'not dragged on' treatment. Fruit in the 'not dragged on' treatment
were wrapped in parafilm, whereas those in the 'dragged on' treatment were not.
Females could readily oviposit into the parafilm-wrapped fruit; however, during
ovipositor dragging after oviposition, the ovipositor could not make contact with the
fruit surface. Experiments on related fruit fly species indicated that such treatment
prevented deposition of MP on a fruit surface (Papaj et al. 1989). Both members of a
replicate pair contained 5 punctures made with a 00 insect pin and spaced haphazardly
over the fruit surface. The punctures were intended to stimulate and facilitate egg
deposition especially for the fruit wrapped in paraHlm.
Fruit in both treatments were hung within separate 3.79-liter plastic containers
that contained ad libitum water, sugar, and slips of a yeast hydrolysate and sugar
mixture. Twenty-five medium to large flies were placed into each cage for 48 hrs and
allowed to deposit clutches. After 48hr elapsed, fruit were removed from the cages and
'not dragged on' treatment fruit were unwrapped. Six fruit (three of each treatment)
were distributed uniformly among the branches of the test trees in the field cage. Details
of the field cage setup were identical to that described for Experiment 2. We recorded

probing attempts and initiation of oviposition by individually marked females on fruit of
each treatment.
174

RESULTS
Experiment 1. Effects of clutch size on time spent engaged in a putative marking
iiehavior
In this experiment we attempted to determine if time spent ovipositor dragging was
related to a female's clutch size, and two factors, body size (estimated from wing vein
length) and egg load (calculated as the sum of the size of her clutch and the number of
mature eggs remaining in her ovaries), potentially confounded with clutch size. Linear
regression analysis indicated that deposited clutch size was significantly related to egg
load (R' = 0.26, df = I, 37; P = 0.001, slope different from zero, t = 3.59) but not to
body size (R" = 0.004, df = 1, 38; P = 0.70, slope not different from zero). In this
experiment, larger females had marginally higher egg loads than smaller females (R' =
0.10, df = 1,37; P = 0.052, slope different from zero, t = 2.00). On average, female egg
loads consisted of 31 mature eggs (n=38) and females deposited 33 percent of the eggs
they carried.
All females in this experiment dragged their ovipositors along the fruit surface
following clutch deposition. Host marking duration conunonly exceeded 5 minutes (X=:
285.5 sec, +.s.e. 25.3, n = 54; range 22-734 sec). In a forward-iterated stepwise multiple
regression analysis including clutch size, egg load and female size as independent
variables found that host-marking duration was significantly positively related to clutch
size (t = 3.57, df = 35, P = 0.001) (Figure 1), significantly negatively related to egg load
(t = -2.19, df= 36, P = 0.04), but not significantly related to our estimate of female size
(t s 0.26, df= 35, P > 0.5). The model retaining clutch size and egg load explained
175

28.7% of overall variation in host marking duration (marking time in seconds = 31.6 •
(clutch size)- 5.3 • (egg-load) + 116). By itself, egg load did not explain marking
duration (R^ = 0.001, df = 1, 38; P = 0.79, slope not different from zero).
Experiment 2. Oviposition responses to fruit on which females ovipositor-dragged
for 0,10,20 or 30 minutes.
In this field cage experiment, females alighted on all models with more or less equal
frequency (Figure 2). The tendency of a female to probe the fruit with her ovipositor
decreased as the aggregate duration of ovipositor-dragging behavior increased (Figure
2). A logistic regression relating propensity to probe a fhiit (yes or no) to the amount of
time that females ovipositor-dragged on a host (0, 10, 20 or 30 minutes) yielded a
significant effect of dragging duration on probing propensity (X" = 5.35, df = I, p =
0.02). The resulting logistic regression model estimated that every minute of ovipositor
dragging reduced a female's propensity to probe a fruit by 3.5% (+1.5%).
176
The effect of ovipositor-dragging duration on oviposition was due at least in part
to an effect on the propensity of a female to initiate probing events before leaving the
fruit (Figure 2). Actual oviposition events declined as host-marking duration increased.
A logistic regression relating propensity to attempt oviposition (yes or no) to the
amount of time females were allowed to mark a particular host (0, 10, 20 or 30 minutes)
showed the effect of dragging duration on oviposition to be significant (X^ = 7.82, df =
1, p = 0.005). The resulting logistic regression model estimated that every minute of
ovipositor dragging on a given fruit reduced a female's propensity to attempt
oviposition subsequently into that fruit by 4.1% (+1.5%).

Experiment 3. Female responses to fruit exposed to 25 females over a 48 hr period.
In this second field cage experiment, females again alighted on all models with more or
less equal frequency (Figure 3). The tendency of a female that landed on a fruit to probe
the fruit with her ovipositor was numerically less on fruit potentially bearing a mark
than on fruit that could not have borne a mark. However, the difference was not
significant (G" = 2.67, df = I, p = 0.10). By contrast, females attempted oviposition
significantly less often into fruit potentially bearing a mark than on fruit that could not
have borne a mark (X* = 11.47, df = I, p < 0.001). In this experiment, the presence of
females reduced a test female's propensity to oviposit from nearly 46% to nearly 10%.
177

DISCUSSION
Evidence for host-marking and oviposition deterrence
Of the three hypotheses presented above (see Introduction), our results provide support
only for the third, namely that ovipositor-dragging behavior in Rhagoletis juglandis is a
bonaHde host-marking behavior and that the mark deters oviposition (Figures 2 and 3).
In this respect, both the form of the host-marking behavior and the deterrent
consequences of that behavior are similar to those described for numerous other
Rhagoletis species (rev. Prokopy and Papaj 1999, Landolt and Averill 1999).
Noteworthy about the current finding is the fact that this pattern of host-marking
behavior and response to the mark occurs despite the fact that walnut flies actively reuse
fruit and even reuse existing oviposition cavities (Papaj 1993, 1994, Lalonde and
Mangel 1993).
178
Nufio and Papaj (2001) identified a number of lines of evidence commonly used
to document the existence of a marking pheromone (MP). Our study gathered some but
not all of these lines of evidence. We have, for example, identified a behavior pattern
that is logically consistent with the deposition of a chemical mark. We attempted to
rule out the effects of other potential stimuli, such as the eggs themselves or recent fruit
damage associated with oviposition, by having each treatment contain an equal number
of clutches and punctures. However, we have not strictly shown that a chemical is
deposited on the fruit during ovipositor-dragging behavior. For example, we have not
made extracts of the MP and shown that, when applied to uninfested fruit, such extracts
deter oviposition. Such information is important to obtain; it would rule out, for

example, the possibility that ovipositor-dragging behavior does not involve deposition
cf a chemical but rather generation of some kind of damage to the fruit surface.
In the case of a species such as Rhagoletis juglandis, such an alternative seems
highly unlikely. Abundant evidence gathered for other, relatively closely related
Rhagoletis species (rev. Landolt and Averill 1999) indicates that deposition of a mark is
involved during a behavior very similar to the one observed in R. juglandis.
Nevertheless, we hope to generate and assay MP extracts for R. juglandis in future
work.
Marking pheromone as an indicator of brood numbers
Walnut flies differ from other Rhagoletis species where host marking has been studied
in that eggs are laid in batches, rather than singly. The tendency for females to lay eggs
in batches suggests a possible function of an MP, namely to convey information about
the number of eggs deposited into fruit. In this study, we obtained evidence that a
female's mark potentially encodes information about clutch size. Specifically, we
found that the amount of time that a R. juglandis female spends dragging her ovipositor
on the fruit surface is proportional to the size of the clutch that she has just laid (Figure
1). Studies on related Rhagoletis species suggest that the duration of ovipositor-
dragging behavior is a reliable indication of the amount of MP deposited on the fhiit
surface (AveriU and Prokopy 1980). If this also applies to host marking in R, juglandis,
it follows that the amount of MP deposited by a /?. juglandis female is positively
correlated with the size of her clutch. Moreover, results of the field-cage experiment in
which females were presented with fruit varying in the aggregate duration of host-
179

marking suggest that females are indeed sensitive to variation in the amount of MP
found on a fruit. In short, R. juglandis appears potentially able to signal not only the
presence of eggs in a fruit but their number as well.
Host-marking duration is an imperfect indication of the amount of marking
pheromone on a fhiit. Congruence between marking duration and the amount of MP
placed on a host may be affected by variation in the speed at which females dragged
their ovipositors along the fruit surface, the width of the deposited trail, the consistency
in the amount of MP deposited along a trail, and female diet (Hendry 1976, Averill and
Prokopy 1988). These factors could account in part for the considerable scatter in the
relationship between host-marking duration and clutch size (see Figure 1).
Is the potential to encode and transmit information about clutch size realized
during signaling under field conditions? We have not conducted field observations of
oviposition behavior in relation to levels of infestation. What we can say is that reuse in
the field occurs to such an extent that such information would be useful to females. In
the field, walnut flies typically deposit clutches of ca. 16 eggs into fruit (Nufio et al.
2(XX)). Because females actively reuse hosts in the field, after 4-S days a fruit may
180
contain ca. 45 eggs, and it is not unusual to find fruit into which females deposited 80 or
more eggs (Nufio et al. 2(XX), Nufio and Papaj, APPENDIX C). Reuse has negative
consequences from the perspective of any given larva, decreasing larval survival and
reducing the weight at which offspring will pupate. In laboratory studies, a reduction in
pupal weight translates to a reduction in lifetime female fecundity (Nufio and Papaj,
APPENDIX Q. Females would thus seem to benefit from a sensitivity to the
accumulating costs associated with reusing a host, a sensitivity that we detected in our

field cage assays. What remains to be demonstrated is a sensitivity under field
conditions.
To our knowledge, there exists only one other insect for which a possible
relationship between clutch size and marking duration has been suggested. A correlation
between marking time and clutch size deposited within a host was inferred for the
gregarious egg parasitoid, Telenomus fariai (Bosque and Rabinovich 1979). In this
study, however, clutch size was not measured directly but inferred from the number of
progeny emerging from a host. Such a result could reflect a marking effort- clutch size
correlation, but alternative explanations are possible (for example, females might mark
in relation to the size of their eggs which might in turn influence an offspring's
probability of survival). These authors also did not examine how the amount of time a
female spent dragging on the egg surface affected the behavior of other females that
were exposed to such hosts. In this case, further study is required to both determine
whether marking time in this gregarious parasitoid is related to the number of eggs
females initially deposit within their hosts, though such a relationship appears likely,
and the potential graded affects it might have on the behavior of newly arrived females.
In a few other cases in which females make a graded assessment of level of infestation
within a host, the underlying mechanismhas not been conclusively established (Bakker
et al. 1972, 1990, van Lenteren and Debach 1981, van Dijken and Waage 1987).
In our study, we determined that host-marking decreases a female's propensity
to deposit a clutch within a fruit. While not investigated in this study, MP might also
work to decrease the size of the clutch a female that chooses to reuse a host might
deposit within the host. A reduction in the number of eggs allocated to a previously
181

utilized host has been found in the tephritid fly Ceratitis capitata (Papaj et al. 1990) and
several other insects (Bakkeret al. 1972, Dcawa and Okabe 1985, van Dijken and
Waage 1987).
182

ACKNOWLEDGEMENTS
We thank Henar Alonso-Pimentel and Laurie Henneman for discussion and assistance
throughout. Sheridan Stone of the Fort Huachuca Wildlife Management office of the
US Army provided permission for collecting fruit and logistical support in Garden
Canyon. This research was supported a National Science Foundation Minority Graduate
Research Fellowship, and NRICGP grant no. 93-37302-9126 to D.R.P. We
acknowledge Judie Bronstein, Reg Chapman, Bob Smith and Molly Hunter for
providing feedback on earlier drafts.
183

REFERENCES
Allen GR, Hunt J, 2001. Larval competition, adult fimess, and reproductive strategies in
the acoustically orienting Ormiine Homotrixa alleni (Diptera: Tachinidae).
Journal of Insect Behavior 14:283-297.
Averill AL, Prokopy RJ, 1980. Release of oviposition-deterring pheromone by apple
maggot flies, Rhagoletis pomonella (Walsh) (Diptera: Tephritidae). Journal of
the New York Entomological Society 88:34.
Averill AL, Prokopy RJ, 1987. Intraspecific competition in the tephritid firuit fly
Rhagoletis pomonella. Ecology 68:878-886.
Averill AL, Prokopy RJ, 1988. Factors influencing release of host marking pheromone
by Rhagoletis pomonella Flies. Journal of Chemical Ecology 14:95-111.
Averill AL, Prokopy RJ, 1989. Host marking pheromones. In: World Crop Pests: Fruit
Flies. Their Boiology, Natural Enemies and Control (Robinson AS, Hooper G,
eds). Amsterdam: Elsevier; 207-219.
Bakker K, Eijsackers HJP, van Lenteren JC, Meelis E, 1972. Some models describing
distribution of eggs of parasite Pseudeucoila bochei (Hym., Cynip.) over its
hosts, larvae of Drosophila melanogaster. Oecologia 10:29-57.
Bakker K, Peulet P, Visser ME, 1990. The ability to distinguish between hosts
containing different numbers of parasitoid eggs by the solitary parasitoid
Leptopilina heterotoma (Thomson) (Hym, Cynip). Netherlands Journal of
Zoology 40:514-520.
184

Bauer G, 1986. Life-history strategy of Rhagoletis altemata (Diptera: Trypetidae), a
fruit fly operating in a 'non-interactive' system. Journal of Animal Ecology
55:785-794.
Bemays EA, Chapman RF, 1994. Host Plant Selection by Phytophagous Insects. New
York: Chapman and Hall.
Bemays EA, Gonzalez N, Angel J, Bright KL, 1995. Food Mixing by Generalist
Grasshoppers - Plant Secondary Compounds Structure the Pattern of Feeding.
Journal of Insect Behavior 8:161-180.
Blanckenhom WU, 1998. Adaptive phenotypic plasticity in growth, development, and
body size in the yellow dung fly. Evolution 52:1394-1407.
Bosque C, 1973. Host marking behavior by parasitic wasp Telenomus fariai
(Hymenoptera: Scelionidae). Acta Cientifica Venezolana 24:88-88.
Bradbury JW, Vehrencamp SL, 1998. Principles of Animal Communication.
Sunderland: Sinauer Associates, Inc.
Bush GL, 1966. The taxonomy, cytology and evolution of the genus Rhagoletis in
North America (Diptera: Tephritidae). Bulletine of the Museum of Comparative
Zoology of Harvard University 134:431-562.
Cirio U, 1971. Reperti sul meccanismo stimolo-risposta nell' ovideposizione del Dacus
oleae Gmelin (Diptera: Tryptidae). Redia 52:577-600.
Denno RF, McClure MS, Ott JR, 1995. Interspecific interactions in phytophagous
insects - competition reexamined and resurrected. Annual Review of
Entomology 40:297-331.
185

Feder JL, Reynolds K, Go W, Wang EC, 1995. Intraspecific and interspeci^c
competition and host race formation in the apple maggot fly, Rhagoletis
pomonella (Diptera: Tephritidae). Oecologia 101:416-425.
Fisher RC, Ganesalingam VK, 1970. Changes in composition of host haemolymph after
attack by an insect parasitoid. Nature 227:191-193.
Fitt GP, 1984. Oviposition behavior of two tephritid fruit flies, Dacus tryoni and Dacus
jarvisi, as influenced by the presence of larvae in the host fruit. Oecologia
62:37-46.
Fox CW, Martin JD, Thakar MS, Mousseau TA, 1996. Clutch size manipulations in two
seed beetles: Consequences for progeny fitness. Oecologia 108:88-94.
Hendry LB, 1976. Insect phermones: diet related. Science 192:143-145.
Hofftneister TS, Roitberg BD, 1998. Evolution of signal persistence under predator
exploitation. Ecoscience 5:312-320.
Hofsvang T, 1990. Discrimination between unparasitized and parasitized hosts in
hymenopterous parasitoids. Acta Entomologica Bohemoslovaca 87:161-175.
Dcawa T, Okabe H, 1985. Regulation of egg number per host to maximize the
reproductive success in the gregarious parasitoid, Apanteles glomeratus L
(Hymenoptera: Braconidae). Applied Entomology and Zoology 20:331-339.
Lalonde RG, Mangel M, 1994. Seasonal effects on superparasitism by Rhagoletis
completa. Journal of Animal Ecology 63:583-588.
Landolt PJ, 1993. Effects of host plant leaf damage on cabbage looper moth attraction
and oviposition. Entomologia Experimentalis Et Applicata 67:79-85.

Landolt PJ, Averill AL, 1999. Fruit Flies. In: Pheromones of Non-Lepidopteran Insects
Associated with Agricultural Plants (Hardie J, Minks AK, eds). New York:
CABI Publishing; 3-26.
Mappes J, M^ela I, 1993. Egg and larval load assessment and its influence on
oviposition behaviour of the leaf beetle Galerucella nymphaeae. Oecologia
93:38-41.
Mayhew PJ, 1997. Adaptive patterns of host-plant selection by phytophagous insects.
Oikos 79:417-428.
Messina FJ, Renwick JAA, 1985. Ability of ovipositing seed beetles to discriminate
between seeds with differing egg loads. Ecological Entomology 10:225-230.
Mitchell ER, Heath RR, 1985. Influence of Amaranthus hybridus L. allelochemics on
oviposition behavior of Spodoptera exigua and Spodoptera eridania
(Lepidoptera: Noctuidae). Journal of Chemical Ecology 11:609-618.
NuHo CR, Papaj DR, 2001. Host marking behavior in phytophagous insects and
parasitoids. Entomologia Experimentalis Et Applicata 99:273-293.
187
Nuflo CR, Papaj DR, 2002. Reuse of larval hosts by the walnut fly, Rhagoletis juglandis,
and its implications for female and offspring performance. Appendix C.
Nufio CR, Papaj DR, Alonso-Pimentel H, 2000. Host utilization by the walnut fly,
Rhagoletis juglandis (Diptera: Tephritidae). Environmental Entomology 29:994-
1001.
Papaj DR, 1990. Fruit size and clutch size in wild Ceratitis capitata. Entomologia
Experimentalis Et Applicata 54:195-198.

Papaj DR, 1993. Use and avoidance of occupied hosts as a dynanuc process in tephritid
fruit flies. In: Insect-Plant Interactions (Bemays EA, ed). Boca Raton: CRC
Press; 25-46.
Papaj DR, 1994. Oviposition site guarding by male walnut flies and its possible
consequences for mating success. Behavioral Ecology and Sociobiology 34:187-
195.
Papaj DR, Alonso-Pimentel H, 1997. Why walnut flies superparasitize: time savings as
a possible explanation. Oecologia 109:166-174.
Papaj DR, Roitberg BD, Opp SB, 1989. Serial effects of host infestation on egg
allocation by the Mediterranean fruit fly: a rule of thumb and its fimctional
significance. Journal of Animal Ecology 58:955-970.
Peters TM, Barbosa P, 1977. Influence of population-density on size, fecundity, and
developmental rate of insects in culture. Annual Review of Entomology 22:431 -
450.
Potting RPJ, Snellen HM, Vet LEM, 1997. Fitness consequences of superparasitism and
mechanism of host discrimination in the stemborer parasitoid Cotesia flavipes.
Entomologia Experimentalis Et Applicata 82:341-348.
Prokopy RJ, Malavasi A, Morgante JS, 1982. Oviposition deterring pheromone in
Anastrepha fraterculus flies. Journal of Chemical Ecology 8:763-771.
Prokopy RJ, Papaj DR, 1999. Behavior of flies of the genera Rhagoletis, Zonosemata,
and Carpomya (Trypetinae: Caipomyina). In: Fruit Flies (Tephritidae):

Quiring DT, McNeil JN, 1984. Intraspecific larval competition reduces efficacy of
oviposition-deterring pheromone in the alfalfa blotch leafminer, Agromyza
frontella (Diptera: Agromyzidae). Environmental Entomology 13:675-678.
Rausher MD, 1979. Egg recognition: Its advantage to a butterfly. Animal Behaviour
27:1034-1040.
Renwick JAA, Radke CD, 1981. Host plant constituents as oviposition deterrents for
the cabbage looper, Trichoplusia ni. Entomologia Experimentalis Et Applicata
30:201-204.
Roitberg BD, Prokopy Itl, 1987. Insects that mark host plants. Bioscience 37:400-406.
Roitberg BD, Sircom J, Roitberg CA, Vanalphen JJM, Mangel M, 1993. Life
Expectancy and Reproduction. Nature 364:108-108.
SAS, 2000. JMP, Version 4. Statistics and Graphics Guide. SAS Institute. Gary, North
Carolina.
Shapiro AM, 1981. The pierid red-egg syndrome. American Naturalist 117:276-294.
Smith RH, Lessells CM, 1985. Oviposition, ovicide, and larval competition in
granivorous insects. In: Behavioral ecology: ecological consequences of
adaptive behaviour (Sibly RM, Smith RH, eds). Oxford: Blackwell Scientific
Publication; 423-448.
Sweeney J, Quiring DT, 1998. Oviposition site selection and intraspecific competition
influence larval survival and pupal weight of Strobilomyia neanthracina
(Diptera: Anthomyiidae) in white spruce. Ecoscience 5:454-462.

Takasu K, Hirose Y, 1988. Host discrimination in the parasitoid Ooencyrtus nezarae:
the role of the egg stalk as an external marker. Entomologia Experimentalis Et
Applicata 47:45-48.
Thompson JN, 1983. Selection pressure on phytophagous insects feeding on small host
plants. Oikos 40:438-444.
van Dijken MJ, Waage JK, 1987. Self and conspecific superparasitism by the egg
parasitoid Trichogramma evanescens. Entomologia Experimentalis Et Applicata
43:183-192.
van Lenteren JC, Debach P, 1981. Host discrimination in three ectoparasites {Aphytis
coheni, A. lingnanensis and A. melinus) of the oleander scale (Aspidiotus nerii).
Netherlands Journal of Zoology 31:504-532.
van Randen EJ, Roitberg BD, 1996. The effect of egg load on superparasitism by the
snowberry fly. Entomologia Experimentalis Et Applicata 79:241-245.
Williams KS, Gilbert LE, 1981. Insects as selective agents on plant vegetative
morphology: egg mimicry reduces egg laying by butterflies. Science 212:467-469.
190

FIGURE CAPTIONS
Figure 1. The relationship between clutch size and duration of ovipositor-dragging
behavior.
Figure 2. The proportion of females that probed the fruit surface with their ovipositor;
and the proportion of these female's that initiated egg-laying into a fruit. N's
indicate the number of females landing on a fruit.
Figure 3. The proportion of females that probed the fruit surface with their ovipositor;
and the proportion of these female's that initiated egg-laying into a fitiit. N's
indicate the number of females landing on a fruit.
191

Figure 1.
QD
s
D£
« xt
E I
•a.!
•5 ^
s u
e o
k
e
a
s
ft.
0.8
0.6
0.4
0.2
n=39
Probed
Control 10 20 30
192
Oviposited
witiiin the host
Minutes Previous Females were Allowed to Host Mark

Figure 2.
w
OA
8
k
98
s
o
a.
QH
s
800 n
600 -
400 -
200 - I :
8
• a
-r-
12
Number of eggs deposited
"T—
16
193
—I
20

Figure 3.
MD
s
eu
J
75
1.1
o I
S b
e e
••i
e
Qi
s
A.
0.8
•S i 0.6
E ®
0.4
0.2
n =28
Punctures +
Clutches
Fruit treatments
11 = 26
Punctures +
Clutches + MP
Probed
194
Oviposited
witiiin tile host