Title: Florida Entomologist
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00098813/00098
 Material Information
Title: Florida Entomologist
Physical Description: Serial
Creator: Florida Entomological Society
Publisher: Florida Entomological Society
Place of Publication: Winter Haven, Fla.
Publication Date: 1983
Copyright Date: 1917
Subject: Florida Entomological Society
Entomology -- Periodicals
Insects -- Florida
Insects -- Florida -- Periodicals
Insects -- Periodicals
General Note: Eigenfactor: Florida Entomologist: http://www.bioone.org/doi/full/10.1653/024.092.0401
 Record Information
Bibliographic ID: UF00098813
Volume ID: VID00098
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: Open Access

Full Text

(ISSN 0015-4040)


(An International Journal for the Americas)

Volume 66, No. 1 March, 1983

The 66th Annual Meeting of the Florida Entomological Society; Second
Announcement and Call for Papers

LLOYD, J. E.-Insect Behavioral Ecology Symposium '82: Looking
Back and Going On ..-...........-- .....--------.....- -........- -............ 1
BURK, T. AND C. O. CALKINS,-Medfly Mating Behavior and Control
Strategies ...... .. ----- -----..... .......... ... 3
WAAGE, J. K.-Sexual Selection, ESS Theory and Insect Behavior:
Some Examples from Damselflies (Odonata) .............-.............-.... 19
LOUNIBOS, L. P.-Behavioral Convergences Among Fruit-Husk Mos-
quitoes ...-......---...--------.....--... ..----.... ... ...........--.... 32
FRANK, S. A.-A Hierarchical View of Sex-Ratio Patterns ...--......-. 42
SMITH, N. G.-Host Plant Toxicity and Migration in the Dayflying
Moth Urania ..----.......---------- ......------...... 76
WING, S., J. E. LLOYD AND T. HONGTRAKUL-Male Competition in
Pteroptyx Fireflies: Wing-Cover Clamps, Female Anatomy, and
Mating Plugs ...- ---~...... ------...----... -----..--.----.... 86

LOFGREN, C. S.-Introduction To The Symposium On Imported Fire
Ants, Southeastern Branch, Entomological Society of America,
January 26, 1982 ......... ......--------------------...... -------........ 92
BUREN, W. F.-Artificial Faunal Replacement for Imported Fire
A nt Control .-....................-- .............. .....-.....-....................................... 93
WOJCIK, D. P.-Comparison of the Ecology of Red Imported Fire Ants
in North and South America ......................-..... ............. ... 101
JOUVENAZ, D. P.-Natural Enemies of Fire Ants .....---......-......................... 111
APPERSON, C. S., AND C. T. ADAMS-Medical and Agricultural Impor-
tance of Red Imported Fire Ant .........................-........... .....-..... 121
VINSON, S. B.-The Physiology of the Imported Fire Ant Revisited .... 126

Continued on Back Cover

Published by The Florida Entomological Society


President ...----- -----------------........ ......... ...... -. A. C. (Abe) White
President-Elect ... ............................................. C. W McCoy
Vice-President .. ................................................ M. L. W right, Jr.
Secretary -...-- ------ .............---------- -.............. .- D. F. Williams
Treasurer ........ -......... ........................ ...... ................. ... D. P. W ojcik

J. R. Cassani
J. L. Knapp
D. C. Herzog
Other Members of the Executive Committee K. Lee
C. A. Morris
W. L. Peters
C. A. Musgrave Sutherland


Editor .. ..-----------.. -----------.............. ......... C. A. Musgrave Sutherland
Associate Editors ---------..- .. ... ........................ ....... ...... D. C. Herzog
F. A. Howard
M. D. Hubbard
J. R. McLaughlin
A. R. Soponis
H. V. Weems, Jr.
Business Manager ...........---.......---.... ...... ............... .............. .. D. P. W ojcik

FLORIDA ENTOMOLOGIST is issued quarterly-March, June, September,
and December. Subscription price to non-members is $15.00 per year in
advance, $5.00 per copy. Membership in the Florida Entomological Society,
including subscription to Florida Entomologist, is $15 per year for regular
membership and $5 per year for students. Inquiries regarding membership
and subscriptions should be addressed to the Business Manager, P. O. Box
12425, University Station, Gainesville, FL 32604. Florida Entomologist is
entered as second class matter at the Post Office in DeLeon Springs and
Gainesville, FL.
Authors should consult "Instructions to Authors" on the inside cover of
all recent issues while preparing manuscripts or notes. When submitting a
paper or note to the Editor, please send the original manuscript, original
figures and tables, and 3 copies of the entire paper. Include an abstract and
title in Spanish, if possible. Upon receipt, manuscripts and notes are ac-
knowledged by the Editor and assigned to an appropriate Associate Editor
who will make every effort to recruit peer reviewers not employed by the
same agency or institution as the authors(s). Reviews from individuals
working out-of-state or in nearby countries (e.g. Canada, Mexico, and
others) will be obtained where possible.
Manuscripts and other editorial matter should be sent to the Editor,
C. A. Musgrave Sutherland, 4849 Del Rey Blvd., Las Cruces, NM 88001.
Business matters for other Society officers can be sent to that individual at
the University Station address above.
This issue mailed May 15, 1983

The Florida Entomological Society will hold its 66th Annual Meeting on
9-12 August 1983 at the Sheraton Sand Key Hotel, Clearwater Beach, Flor-
ida. The location is 1160 Gulf Boulevard, Clearwater Beach, Florida 33515;
telephone-1-(813)-595-1611. Room rates will be $45.00 either single or
double; student rates will be a flat $15.00 per person, with a minimum of
3 students per room.
Questions concerning the local arrangements should be directed to:

DR. ROBERT W. METZ, Chairman
Local Arrangements Committee
Florida Entomological Society
1808 North 57th Street
Tampa, Florida 33619
Phone: 1-(813)-626-3184
Since many will present papers please tear out the sheet and submit be-
fore 1 JUNE 1983, to:
M. LEWIS WRIGHT, JR., Chairman
Program Committee, FES
P. O. Box 2185
Winter Haven, Florida 33880
Phone: 1- (813)-299-1131
Eight minutes will be allotted for presentation of oral papers, with 2
minutes for discussion. In addition, there will be a separate session for
members who may elect to present a Project (or Poster) Exhibit. The three
oral student papers judged to be the best on content and delivery will be
awarded monetary prizes during the meeting. Student authors must be
Florida Entomological Society Members and must be registered for the

Preregistration on Site
Full & Sustaining member $30.00 $35.00
Student not in student contest $12.00 $17.00
Student in contest $6.00 $11.00
Each extra banquet ticket 6.00 6.00
'Each fee includes one banquet ticket.


The following slide policy will govern slide presentations at the Annual
Meetings. Only Kodak Carousel projectors for 2 x 2 slides will be available.
However motion picture projectors will be available by special request to
the Local Arrangements Chairman prior to the date of the meeting.
Authors should keep slides simple, concise, and uncluttered with no more
than 7 lines of type on a rectangle 2 units high by 3 units wide. All printed
information should be readable to an audience of 300 persons.
A previewing room will be designated for author's use. A projectionist
will be available in the previewing room at least one hour before each session.
Authors are expected to give the projectionist their slides in the previewing
room prior to each session. Slides will be returned to the authors after each
session in the meeting room.
Authors are expected to organize their slides in proper order in their
personal standard Kodak Carousel slide tray (no substitution, please). Only
a few slide trays will be available in the previewing room from the projec-
tionist for hardship cases. Slides in the tray should be in correct order start-
ing with slot #1 of the tray and positioned correctly (position of slides to go
into tray: 1. upside down, and 2. lettering readable from this position upside
down and from right to left). A piece of masking tape should be placed on
the slide tray by the author and the following information should be written
on the tape: 1. author's name, 2. session date, and 3. presentation time.


-r .



Insect Behavioral Ecology-'82 Lloyd



This symposium on insect behavioral ecology is our fifth, counting the
original "Sociobiology of Sex" that Norm Leppla and I put together for
a branch meeting of the ESA (Fla. Ent. 62:1-34). They have been "well
received." Many authors report that they have had more requests for re-
prints of these articles than for any other they have written, and that the
requests come from scientists around the world. The paper-bound sym-
posium reprints are used for texts in graduate and undergraduate classes,
and refreshers to bring us all up to date in subject areas we have lost
track of or never have tracked at all. At least two articles have stimulated
adrenaline flow in researchers that needed the exercise. The papers con-
stantly remind us, on an increasing range of subjects, that we should al-
ways be tuning to adaptation in the organic world, and not get side-
tracked into the extremism of morphology for morphology's sake, physi-
ology for physiology's sake, or behavior for behavior's sake-though we
continue to organize instructions and syllabi into these artificial boxes.
They remind and hopefully show us that pragmatism, working both sides
of the practicality canal (Fla. Ent. 64:1), is to be preferred and presents
more of a challenge than either basic or applied science practiced alone.
The articles have reviewed and refocused subject areas, brought together
new combinations of information and ideas, cut through complexity to show
that some esoteric subjects may have something worth knowing in the
struggle to understand insect bionomics, and occasionally carried beyond the
confining range of Morgan's Canon to try to trap a problem in a pincer move-
ment; and always the authors have tried to make their papers understand-
able to those outside their speciality, to be interesting, to present the great-
est return for time spent by the reader, and yet retain a professional stand-
ard of quality appropriate to primary scientific literature. It ain't easy!
Reader mindset, attention span, basic understanding of biology, philosophy,
and English range broadly. Slang and colloquial expressions that add
spice and "turn some readers on," are a "put off," a "put down," or ob-
scure to another. (When I used the lightningbug "spermed on" as a com-
municatively vigorous malapropism to indicate both a spurring on and a sem-
inal contribution, some presumed it a typographical error.2) We shall con-
tinue, and I urge you to keep sending those cards and letters with sug-
gestions for topics, authors, and improvements. (If your idea is used we
can send you a photo of a "Real Insect" t-shirt.) Especially needed are
ideas for sources of financial support to pay travel and other expenses
for the speaker/authors, and some part of the cost of publishing the paper-
bound separates. Please send ideas to m'e and money to Treasurer Dan
Wojcik. Symposium-'82 was made possible by the contributions of Dale
Habeck, J. E. Lloyd, Jerry Stimac, Frank Slansky, Tom Walker, and IFAS,
through the efforts of Dan Shankland. The Society also provided money.
Thank you! And thanks also to the referees for this symposium, Jane

1Dept. of Entomology and Nematology, Univ. of Florida, Gainesville 32611.
2Espy, W. R. 1971. The game of words. Bramhall House, New York. (see p. 161)

2 Florida Entomologist 66 (1) March, 1983

Brockmann, Tim Forrest, Steve Frank, Reece Sailer, John Sivinski, John
Strayer, and Tom Walker; to Ngo Dong and Barbara Hollien for technical
assistance; to Laura Line Reep for artwork; and to the Executive and
Program Committees of FES, especially Abe White and Dan Wojcik.
Symposium-'83 will be held at the annual FES meeting at Clearwater
in August 1983. Florida Agricultural Experiment Station Journal Series
No. 4488.

?II 'iA *; ^- *
'" ''' '
iv 7.r OK

Symposium Participants. Back row, 1 to r: Steve Wing, Ted Burk, Bill
Walker, Jeanne Altmann, Jon Waage, Neal Smith, Carrol Calkins. Front
row: Jim Lloyd, Steve Frank, and Phil Lounibos. Photographed by Lionel
Stange, FADCS-DPI. 11 August 1982, Longboat Key, Sarasota, Florida.

Insect Behavioral Ecology-'82 Burk and Calkins 3



"There is an unfortunate division between pure and
applied science that is hurting everybody."
Paul Ehrlich

"We're in the middle of an eradication program and
can't afford the luxury of research."
a "high USDA official" quoted
by Ehrlich
(both from Walsh 1981)

"The agencies responsible for protection of public
interests in regard to agricultural pests must use the most
suitable methods available even though they are not ideal....
This sometimes results in controversy. Very specific control
measures such as biological, autocidal, and integrated tech-
niques require more intensive and specific research and
development than do broad spectrum techniques such as
Chambers et al. 1974


The "medfly" (Mediterranean fruit fly, Ceratitis capitata (Wied.)) is
one of the most destructive pests of fruits and vegetables worldwide. In
June 1980, medflies were found in detection traps in Los Angeles and San
Jose, California. Potential annual costs to the $14 billion California agri-
cultural industry were estimated at $59 million for chemical control, $38
million for quarantine and fumigation, and $260 million in crop losses
(minimum estimates, Hess 1981), not including such one-time costs as the
construction of extensive fumigation facilities, $497 million. Therefore, a
major medfly eradication effort was begun by the state of California in
conjunction with county and federal officials. Eventually the project ex-
tended over 1,300 square miles and cost more than $100 million. Success
in eradication was marked on October 21, 1982, by a champagne celebra-
tion involving project officials. Yet between initiation and completion, the
California project did not always run entirely smoothly. Throughout the
summer of 1981 nightly network news reports chronicled the ups and downs

*Theodore Burk is an Assistant Professor in the Department of Biology, Creighton Uni-
versity, Omaha, Nebraska 68178. He was formerly a postdoctoral Associate employed through
a cooperative agreement between the Insect Attractants, Behavior, and Basic Biology Re-
search Laboratory, Agric. Res. Serv., USDA, and the Department of Entomology and
Nematology, University of Florida, Gainesville. His research interests are in insect social
behavior, especially aggressive behavior and acoustic communication.
**Carrol O. Calkins is a Research Entomologist at the Insect Attractants, Behavior,
and Basic Biology Research Laboratory, Agric. Res. Serv., USDA, Gainesville, Florida. He
has been involved in biology, ecology, behavior, and host-plant resistance with the USDA
since 1960. From 1977 to 1980 he was head of the Seibersdorf Entomology Section, Inter-
national Atomic Energy Agency. Vienna, Austria. His research interests include mating
behavior, population dynamics and ecology.

Florida Entomologist 66 (1)

of the project and spotlighted many criticisms of it. Now that things have
settled down, it seems a good time for two behavioral ecologists to take a
retrospective look at medfly control strategies. Our intention is not to join
the criticisms of the California project. That project was in a pioneering
position, with decisions and improvisations necessarily being made on the
basis of information available. What we are attempting to do is to point
out some areas where the need for additional behavioral ecology research
was made apparent during the course of the California project. The need
for such behavioral ecology research will be encountered in future eradica-
tion projects which may involve other insects, especially fruit flies.
Some of the criticisms of the California project were technical: for ex-
ample, that original detection traps were placed at densities too low to
detect any but large established populations. Others, valid or not, were
more "political": all-out chemical spraying done too late (commercial grow-
ers) or too soon (suburban residents); management conflicts between state
and federal officials; differing emphases of "pure" and "applied" scientists
influencing pest control philosophies.' (For a summary of criticisms of
the California project see Jordan 1982a,b, Walsh 1981, Marshall 1981).2
One subject on which many doubts are centered is the use in California
of the sterile male release method (hereafter called SIT for "Sterile Insect
Technique") as one component of an integrated control program. As men-
tioned in the quote by Chambers et al. (1974), the control method selected
must be appropriate to the situation. In the heavily populated suburban
area of the San Jose infestation, SIT seemed an especially appropriate
method of control.3 Yet in mid-summer 1981, as the infested area increased,
use of SIT in the integrated control program of the California project
was abandoned. Prompted by the California experience, we consider in this
paper the merits of SIT for controlling medflies. We come to the conclusion
that the technique is sound, but its improvement relies to a great extent
not only on efficient tactics, but also on extensive collaboration between
theoretical and applied entomologists and on accurate knowledge of the
behavioral ecology of medflies (knowledge that was largely unavailable at
the beginning of the California project). Such information is, of course,
needed for most insect pests, not just medflies.

E. F. Knipling (1955) first proposed the idea of sterilizing and releasing
male insects to suppress insect populations. He has summarized the po-
tential advantages of SIT, concluding that population suppression can be
theoretically achieved, especially on incipient populations in new areas, at
a cost much lower than that of insecticide treatments alone and without
the ecological problems associated with insecticide use. The advantages of
this technique in residential areas are evident (Knipling 1979). There is
the additional advantage that sterile flies may be released, without eco-
logical concern, over an area much greater than that of the infestation
to act as a barrier against emigration. Since many of these advantages
met concerns felt by residents and officials in California, it was natural
for SIT to be adopted as part of the primary control program.4 Further-
more, officials could make use of a history of SIT programs going back to
the 1950s.

March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins

Sterile Insect Technique was used first in 1955 against screwworm
flies in the Caribbean and southeastern U. S. It was first used against fruit
flies in the Hawaiian and Marianas Islands (Steiner et al. 1962). The
melon fly (Dacus cucurbitae Coq.) was eradicated by the release of 257
million sterile flies on Rota Island (Steiner et al. 1965). SIT was used
against medflies on the Italian islands of Capri and Procida in 1966 and
1967 where populations were reduced by 99% but immigration from the
mainland prevented eradication (deMurtas et al. 1970). Other medfly SIT
eradication projects were conducted with varying degrees of success in
Tunisia in 1969 (Cheikh et al. 1975) and Central America in 1969 (Rhode
1970). Field programs of suppression have been conducted in Israel, Spain,
Cyprus, and the Canary Islands. Many of these projects have involved
collaboration between the U.S., U.N., I.A.E.A., F.A.O., and numerous host
countries. SIT was successfully used against medfly in Los Angeles in
1975-76 (Harris 1977) and again in 1980.
Currently, the state government of Western Australia is using SIT
against medflies (Fisher 1981). The largest ongoing medfly SIT program
began in Mexico and Guatemala in the late 1970s (Gibbs and Eerde 1981).
This program is achieving its goal of preventing the spread of medfly into
southern Mexico. An initial program of chemical spraying has been suc-
ceeded by release of billions of sterile males which act as a barrier at the
Mexican border.5 The rearing facility for this project is capable of produc-
ing more than 100 million flies per day, and had enough capacity to supply
over 100 million flies per week to the California project, over and above
needs in Mexico ( 1981 annual report, Programa Moscamed).
In California in 1981, however, the area of infestation seemed to spread
despite sterile male releases.6 Eventually it was decided to stop all further
releases and rely more heavily on chemical control, mainly by aerial spray-
ing of mixtures of malathion-protein hydrolysate baits.

The record provided by numerous SIT programs permits evaluations of
the factors responsible for success or failure. It was originally felt that
female monogamy was crucial, but the analyses of Zouros (1969) and Ito
and Kawamoto (1979) demonstrated that this was not the case. Provided
that sterile males are competitive (they need not be 100% competitive with
wild males), remating by females has only a marginal effect. Knipling (1979)
suggested the main requirements for SIT success against a new infestation
were early detection and production of high-quality males for release. Ito
and Kawamoto (1979) examined the effect of overflooding ratio (ratio of
released sterile to wild fertile males). They found a lack of correlation
between higher overflooding ratios and success of various projects: suc-
cess was obtained against some tephritids with ratios as low as 14:1, while
failure occurred with ratios as high as 1000:1 (summarized in Calkins et
al. 1982). Ito and Kawamoto (1979) concluded that two factors were
most important: immigration of additional flies into the control area and,
most important, adequate mating behavior by released males. They also
pointed out that, in regard to the importance of female monogamy, selec-
tion for resistance to SIT on wild females would probably not take the

Florida Entomologist 66(1)

direction of additional matings, but rather of increased discrimination in
making mate choices.

The theoretical analyses considered above suggest that SIT success
against a new introduction relies on quick detection, knowledge about med-
fly dispersal, and competent sexual behavior by released males. These are
all areas where knowledge from behavioral ecology can help in the forma-
tion and testing of hypotheses to generate information necessary for ap-
plied programs. In the balance of this paper, therefore, we briefly describe
basic medfly ecology, then consider the potential role of behavioral ecology
in the areas of dispersal, mating behavior, and trapping.
Although we know relatively little about medfly ecology, we do know
that the medfly is a versatile species. Native to East Africa, it is now
found in other areas including South Africa, the Mediterranean area,
Australia, Hawaii, and Central and South America. During this globe-
trotting, medfly populations must have experienced numerous population
bottlenecks followed by adaptation in various ways to local environments.
Thus we need to be alert to genetic variation in medfly populations.
Bateman (1976) divided tephritids into two ecological groups: temperate
species (such as the apple maggot Rhagoletis pomonella (Walsh)), which
are relatively K-selected, and tropical or sub-tropical species (such as the
medfly), which are more r-selected. Bateman's (1976) characterization of
tropical species applies almost without exception to the medfly. The tropical
species tend to be polyphagous. (Hagen et al. (1981) indicated that 253
hosts are known for the medfly. The quarantine list for California included
48 fruit and vegetable species.) Ripe hosts appear randomly dispersed in
time and space throughout the year, and are of short duration in a given
area. Thus, tropical tephritids must locate hosts, exploit them rapidly, then
disperse to find newly appearing hosts in other areas.
In keeping with its wide host range, the medfly is characterized by such
colonizing features as short development time, lack of diapause, and good
dispersal ability. We are only beginning to accumulate needed demographic
information (Shoukry and Hafez 1979), and predictions of the ability of
medflies to invade and survive in particular areas are notoriously unreliable.
Under ideal conditions adults can live up to 2 months, laying up to 40 eggs
per day. Eggs can hatch in 2-3 days, larvae can develop in 6-11 days,
pupation lasts as little as 6 days. Thus a generation may take as little as
3 weeks (there are 11-13 generations per year in Hawaii (Harris 1977),
but in northern California only 3 to 31/ generations are projected (Hagen
et al. 1981)). Obviously, development is slower and survival decreases under
less ideal conditions, but the range of tolerable conditions is not well known
and probably varies for different medfly populations.8 In general, it is
probably wise to assume that medfly can persist and thrive in any area
with a Mediterranean to tropical climate until proven otherwise. It can
also establish breeding populations during the summer in some temperate
regions. The California infestation proved to be somewhat tolerant of cold
temperatures and to have a wider host range than expected. New host
records occurred because this was the first time that medfly had come in
contact with these fruiting species.

March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins

In keeping with its r-selected ecology, medfly is a good disperser. Bate-
man (1972) suggests 3 general occasions for dispersal in tephritids: when
host fruit disappears, with the onset of more favorable seasonal conditions,
and when adults first emerge from pupae in the soil, with the first probably
being the most important.9 Several workers have associated mass dispersal
movements of medflies with disappearance of host fruit (Christenson and
Foote 1960, Steiner et al. 1961, 1962, Bateman 1972, Bateman 1976). The
California project included an extensive campaign of fruit stripping to
reduce resources available for medfly reproduction. An unintended result
of this fruit stripping may have been a strong stimulus to disperse. This
suggests that fruit stripping should be undertaken only when a barrier
of traps, sterile males, or chemical sprays surrounds the area. In California,
a 9 x 9-mile block around new detection points was sprayed for a projected
3 generations (at prevailing temperatures) following fruit stripping.
Medfly dispersal movements are characterized by frequent flights. Most
trapping studies catch more flies downwind from the release point (Wong
et al. 1982), although over short distances, flies orient upwind (Christen-
son and Foote 1960). Dispersal flights can take flies impressive distances.
Cirio and deMurtas (1974) trapped on an island flies that had been released
on the mainland of Italy 2.7 miles away. Steiner et al. (1962) recovered
marked flies in Hawaii 40 miles from the release site, perhaps due to dis-
placement by storm-related high-velocity winds.
There are many things we need to know about medfly dispersal. How
are dispersal tendencies in newly-emerged flies influenced by fruit availa-
bility? Is there a polymorphism in, dispersal tendency, even where resources
are present (as predicted by the theoretical dispersal models of Hamilton
and May (1977))? How do weather conditions influence dispersal propen-
sity? Is there a seasonal peak in dispersal in the medfly? What are the
details of orientation to wind direction and the role of winds in carrying
medflies over greater distances? An especially important area is that of
sex differences in dispersal.10 Do these exist in medflies? If so, how does
mating status influence this, and what are the implications for early de-
tection? (If mated females disperse more widely, but detection traps are
geared toward catching males, populations may be well established before
they are detected.) A large body of theoretical literature is developing on
the behavioral ecology of dispersal (see Baker 1978) which could be ap-
plied to studies of medfly movement.

One of the strongest achievements of behavioral ecology has been in
relating a species' ecology to its mating system (Emlen and Oring 1977).
Prokopy (1980) and Burk (1981) made this connection for tephritids. The
polyphagous, tropical species such as medfly have hosts that are distributed
in a random, patchy fashion. Males are unable to adopt the strategy of
defending a host fruit and swapping female access to it for a mating. The
mating system of medflies is therefore characterized by freer female choice
of mates from males aggregated into lekss" (male sexual display groups)
on the vegetation of host plants. In keeping with this female-choice, lekking

Florida Entomologist 66 (1)

mate system, male medflies produce a complex variety of sexual signals
(Burk 1981).
Only one study of medfly mating behavior under relatively natural
conditions has been published, the important field-cage study by Prokopy
and Hendrichs (1979) in Guatemala. They discovered the existence of two
mating modes by male medflies. Males search for, court and attempt to
force matings with females that are ovipositing in fruit in early morning
and late afternoon. Only 23% of male mounting of females on fruit re-
sulted in copulations, and only 15% of observed copulations took place on
fruit. In the middle of the day, males form leks, emit pheromone and
produce acoustic calls, and mate with arriving females. Seventy-four per-
cent of these female-initiated mounts were successful, and 85% of observed
matings took place in this way. These results are extremely valuable, but
the medfly mating system and its implications for SIT need to be con-
sidered in context of sexual selection theory. We have developed such an
explanatory model, based on observations of wild populations of the Car-
ibbean fruit fly (caribfly), Anastrepha suspense (Loew), a subtropical
species occurring in Florida which has biology and behavior remarkably
similar to the medfly (Burk 1983, Burk and Webb 1983).
The first thing a sexually mature male medfly must do (whether he is
a wild or released) is to find a mating area. One way to do this is to lo-
cate newly ripening fruit, a behavior that accounts for the occasional male
caught in color traps thought to mimic the visually attractive features of
fruit (Nakagawa et al. 1978). Perhaps an easier strategy for the ma-
jority of males is to go to areas where other males are already aggre-
gating and displaying, by responding to the female-attracting signals be-
ing produced by these males (Burk 1981, Alcock 1982). This perspective
accounts for the strong response of male tephritids of lekking species to
male-produced sex pheromones (reported for Dacus tryoni (Froggatt)
(Fletcher 1968) and A. suspense (Perdomo 1974), as well as for medfly
(Ohinata et al. 1977)). It also accounts for the pattern of responses of
tephritids to synthetic "parapheromones" (Chambers 1977a, see section be-
low on trapping). These synthetics (trimedlure, in the case of medfly) pri-
marily attract sexually mature males, but do attract females: (1) that are
unmated but sexually mature, (2) at times of day corresponding to the
normal mating period, and (3) when the number of wild males competing
with the synthetic-releasing trap is small (as when only females are re-
leased, when population density is low, or when most males have already
been trapped out) (Fitt 1981). It also explains why male response is so
strong. All males have to relocate a lek every day, while a female will only
respond once or a few times in her life. That sexual selection theory can
provide a convincing explanation of a previously puzzling pattern such
as this is a strong justification for using it in studies of other aspects of
medfly mating behavior.
Once a male medfly has found a lek, he establishes a single-leaf calling
station, from which he emits pheromone and produces calling sounds.
Prokopy and Hendrichs (1979) do not mention fights between males, only
attempted copulations of males with other males. In the caribfly, however,
males have to fight off other males to maintain their calling station. Losing
males are driven from the leaf, and sometimes are driven entirely out of

1March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins

the lek. The two major factors determining the outcome of fights are size
and prior residence (Burk 1983). Only those males that establish terri-
tories are able to produce female-attracting stimuli in the caribfly, and
therefore adopt the more successful of the two mating modes. The existence
of such territorial fighting is likely in the medfly and needs to be investi-
Male medflies may interact in the lek in another way, by mutually stim-
ulating each other to puff and call. Such "chorusing" is common in acoustic
insects (Greenfield and Shaw 1982), and is present in the caribfly (Burk,
unpubl. data) and in the Oriental fruit fly (Dacus dorsalis Hendel) D. L.
Chambers unpubl. data).
Male-male interactions, especially the aggressive interactions that re-
sult in some losing males being driven out of leks, may help to account for
the existence of the secondary "searching on fruit" mating mode. Cade
(1980) analyzed the existence of alternative mating strategies, concluding
that many strategies, such as conditional strategies consisted of "making
the best of a bad job" (Dawkins 1980). Searching on fruit is clearly less
successful for a male medfly than calling in a lek,11 and there is no indi-
cation of biased predation on lekking males that might even out long-term
mating success (Burk 1982). Searching may therefore be a tactic of males
unable to defend a calling territory. Hendrichs (1983) has found some
indications that this is the case in caribflies.
So far, we have only considered aspects of the first of Darwin's two
types of sexual selection, male-male competition. The second type, female
choice, is likely to be even more important in medflies. Female medflies
normally mate only once unless they deplete their stored sperm (Nakagawa
et al. 1971, Katiyar and Ramirez 1970). Females are therefore expected
to make very selective mate choices. In a species with no parental care,
females are expected to select vigorous males, male phenotype being indic-
ative of a superior overall genotype or of a genotype attractive to females
(giving a female either (1) offspring with "good genes" or (2) "sexy sons"
leading to a greater number of grandprogeny) (Trivers 1972). Work on
medflies should therefore be directed to discovering the features of a male's
phenotype that are attractive to such choosey females. Furthermore, we
should expect females to evaluate a number of different features. This would
reduce the likelihood of a female being "fooled" by the signals of a less
vigorous male, and would increase the likelihood that male variability is
heritable. Burk (1981) has considered these points for tephritids in general.
In the caribfly, male size seems to be an important characteristic. Large
males win more fights, have higher calling propensity, and their calling
song pulses are repeated more quickly (Burk 1983, Burk and Webb 1983).
Large size may reflect a superior larval genotype. Female carbflies prefer
large males over small males as mates (Burk and Webb 1983), and it is
likely that female medflies also prefer large mates.
Differences in courtship behavior may also be important to female med-
flies. In the caribfly, successfully mating males produce precopulatory songs
that are on average 10 dB louder than those of males that are rejected
by females (Burk and Webb 1983). Males that produce long precopulatory
songs mate 25% longer than males that produce little or no song (Burk

Florida Entomologist 66 (1)

unpubl. data). In the medfly, length of copulation affects the amount of
sperm transferred (Farias et al. 1972).
To summarize, we need to know at least the following things about
medfly mating behavior. What are the characteristics of the places where
leks form? How consistent is the response by males to male pheromones,
and what is the ontogenetic and diurnal periodicity of this response? What
is the extent of territorial aggression for calling sites in medflies, and what
factors influence the outcome of fights? Which males adopt the secondary
tactic of searching on fruit, is it part of a diurnal alternation practiced
by all males, or merely a last resort of males driven from leks? Do virgin
females go to host fruit, and do they ever mate there? Is there chorusing
by aggregated males? Do females orient to calling song as well as phero-
mone, or show increases in activity in its presence (as do female caribflies;
Burk unpubl. data)? Do large and small males or dominant and subordi-
nate males differ in signaling propensity or signal characteristics? Do fe-
males mate preferentially with large males or males in large aggregations?
How do females respond to between-male variation in courtship and pre-
copula songs? What predation pressures exist on signaling males or copu-
lating pairs? As you can see, given the absolute necessity that released
sterile males perform adequate mating behavior, we know very little about
mating in medflies.

Input from behavioral ecology can be used in the design of traps that
increase the chance of early detection of medfly invasions. Traps work by
eliciting behavior that is normally adaptive to the stimuli provided by the
trap. An ideal trap would take into account the importance of different
resources to different age and sex classes of the medfly population, and
catch a demographic cross-section of the population. The most important
resources for mated females medflies are those related to feeding and
oviposition; for males, they are those related to the chance to encounter
impregnable females. For sexually receptive females the opportunity to
mate is an important resource also. Responses by individual medflies will
be strongest to cues characteristic of those important resources.
For many fruit flies, including the medfly, the main trap stimulus is
provided by a synthetic chemical ("parapheromone," Chambers 1977a)
which attracts mostly males.12 In the case of medfly, this synthetic is
trimedlure (tert-butyl-4 (or 5)-chloro-2-methyl-cyclohexanecarboxylic acid).
Other synthetics are used for Dacus species, and the attractancy of various
parapheromones follows taxonomic lines (Drew and Hooper 1981). In the
previous section we accounted for male response to these synthetics by
hypothesizing that they mimic species-specific sex pheromones (Fitt 1981),
and that males respond in order to enter mating aggregations. This raises
the possibility that the actual pheromone'may be used in traps. Ohinata et
al. (1977) compared response to traps containing trimedlure or pheromone,
and found that in general pheromone was at least as effective. Laboratory
assays elicited response from virgin females. Field tests resulted in re-
sponse by males. These authors offered no explanation for the apparent
contradiction. D. L. Chambers proposed in.a talk before the Entomological
Society of America (November, 1981) that precisely these results would

March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins

be expected if the materials were lekking pheromones. Virgin females would
respond (as in the laboratory tests where no males were present) as well
as males (as in field tests where few virgin females would be present). The
pheromone is a complex mixture of at least one ester, one alcohol, and
15 fatty acids, with males responding to various incomplete combinations
of the components, but females requiring all of the compounds. Again,
this is as expected on sexual selection grounds, with females being more
selective than males. Whether pheromone is used in large-scale projects
will probably depend on the economics of its synthesis and on detailed
comparisons of its relatively advantage (if any) over trimedlure in trap-
ping different age-sex classes.
Given that dispersal patterns are likely to vary between the sexes and
that infestations may begin through the activities of one or a few mated
females, it seems important to consider the incorporation into traps of
stimuli attractive to females as well as to males. Such stimuli would relate
to food or oviposition sites.
One successful trap incorporating such stimuli is the McPhail trap, an
invaginated glass trap containing fermenting yeast or hydrolyzed protein
in water (Chambers 1977a). For caribflies, with no male parapheromone
discovered, this remains the trap of choice. It is, however, bulky, vulnerable
and expensive. Trimedlure-baited medfly detection traps use the simple
Jackson trap (Harris et al. 1971), a plain white milk carton modified into
a triangular shaped container. Greany et al. (1982) and Davis et al. (1982),
studying laboratory and wild caribflies, respectively, demonstrated that
adding vertical orange stripes to the unbaited Jackson trap led to a 6-
fold increase in catch, by catching many females that would not respond
to a white trap. Prokopy (1976) has suggested that response to such bright
colors mimics adaptive response to the color of vegetation or ripe fruit.
If female medflies respond similarly, addition of color stripes to Jackson
traps might prove a significant step in improved detection. Work on this
question is in progress in Hawaii and Mexico.
Davis et al. (1982) identified another deficiency in the Jackson trap
design. The sticky trapping substance is on the inside of the trap. Yet both
medflies and caribflies land on the underside of objects, normally leaves
in a natural situation (Prokopy and Hendrichs 1979). Davis et al. (1982)
found that 90% of the caribflies attracted to a striped Jackson trap did
not go inside and eventually flew away. This led them to modify the Jackson
trap by removing the bottom surface and coating the roof-like ventral sur-
faces with the sticky substance. Trap catches in baited traps so modified
increased dramatically. This small study showed the importance of actually
sitting and watching flies respond to traps, rather than just making hourly
or daily counts.13

For the past 10 years, the importance of monitoring and improving the
quality of performance of mass-reared tephritids used in SIT programs
has been recognized (Huettel 1976, Chambers 1977b, Calkins et al. 1982).
A standard quality control system, RAPID, has been developed for medfly
(Boller et al. 1981, Calkins et al. 1982). It includes tests for (1) calibration
of pupal size, (2) flight ability, (3) irritability, (4) pheromone response,

Florida Entomologist 66 (1)

and (5) mating propensity. The role of behavioral ecologists in future
quality control development will be in suggesting additional tests or ways
of modifying existing tests to more adequately reveal the essential per-
formance characteristics required of the released males. Some suggestions
from our own work include: (1) use of the olfactometer to measure
pheromone response of males as well as females, given the importance of
this response in lek location; (2) use of leaf-mimicing patterns on the
top of mating propensity cages; (3) supplementing the standard mating
propensity test with choice tests involving one female and 2 males of
different phenotypes, to allow female mate choices to be fully expressed; (4)
measurement of the fighting ability of sterilized males; and (5) scrutiny
of precopulatory and copula song parameters as well as calling song param-
eters. In each case, we recognize that there is a trade-off between ease of
running a test and the importance of the information gained, but we must
guard against the assumption that we already know what qualities are
important in a released male.

In this paper we have suggested many areas of research that might
provide information to increase the likelihood of success for SIT against
medflies. This has not been intended to diminish the large amounts of work
already done, that have demonstrated the potential of large-scale, SIT-based
medfly programs. We do not accept the validity of a pure/applied distinction
in pursuing and applying the results of this research. By usual standards,
one of us would fall on each side of the pure/applied fence. Yet in work-
ing together we have found that both have benefited from considering be-
havioral ecology theory and real-world technical difficulties. In the final
analysis, both aspects will be equally important in the future medfly projects
that, given the nature of that insect, will surely be needed.


We evaluate the Sterile Insect Technique for control of the Mediterra-
nean fruit fly in light of the 1980-82 California medfly eradication program.
Use of SIT against new infestations relies on early detection, knowledge
of dispersal patterns, and adequate sexual behavior by released male medflies.
More information is needed in all of these areas, and the behavioral ecology
approach leads to the formation of many hypotheses than can be tested
to gain this information.


We would like to thank Dr. J. E. Lloyd for inviting us to participate
in the Insect Behavioral Ecology Symposium; Drs. D. L. Chambers and
J. C. Webb for their advice and help; Drs. M. D. Huettel and T. J. Walker
for commenting on the manuscript; Mr. Jorge Hendrichs for the many
ideas he provided during our collaborative research and for permission
to cite unpublished data; and Mrs. Elaine S. Turner for preparing the

March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins


'A report by the Congressional Office of Technology Assessment pointed
out that only 2% of USDA scientists were aged 30 or less, compared with
25% at NIH; for scientists over 50 the percentages were 39 (USDA) and
15 (NIH) (Wade 1982). Although organizational policy establishes emphasis
on basic or applied science, younger scientists tend to think more about basic
science because of more recent exposure to it at universities. The age
structure of USDA scientists, however, is a direct result of financial and
hiring constraints set by Congress.
2Jordan (1982a,b) has questioned whether eradication should have been
undertaken at all. Some of his arguments and our reactions:
(i) "It is hopeless to try to eradicate an entire species outdoors." Yet
no one in California was trying to eradicate an entire species; they were
trying to eliminate one local population that had recently colonized and
was probably not well adapted yet to the local environment.
(ii) Detection traps with synthetic chemical attractants don't work
because "medflies, unlike moths and other insects, seem to rely on vision
and host plant odors to lead them close to mating sites." In fact, there is
good evidence for an important role of pheromones in mating aggregation,
and for the likelihood that synthetic chemicals elicit such aggregation be-
havior. See our section on mating behavior.
(iii) "Sterile releases have scored only one major victory . Against
larger populations . these releases have not been effective." But see our
review below of SIT projects, and note the success of SIT in Mexico against
a much larger population and in more difficult terrain-see appendix note 5.
(iv) "Demographic studies show that only 4% of the medfly population
at a given time is adults; therefore it is useless to use aerial sprays." How-
ever, adults are the most vulnerable stage and all successfully reproducing
medflies must pass through this stage; it is merely necessary to repeat
sprays occasionally (in California that was done for 3 complete life cycles)
to kill newly-emerging adults.
(v) "Medflies lead sedentary lives, and probably would not even spread
from coastal areas into the San Joaquin Valley." But medflies are good
dispersers (see our section on dispersal); and they may invade southern
Germany from northern Italy! (Hagen et al. 1981).
(vi) "The medfly cannot survive San Jose winters or San Joaquin sum-
mers." This is based on laboratory studies in climatic chambers-but med-
flies survive worldwide in places colder than San Jose (in fact they over-
wintered in San Jose in 1979 and 1980) and as hot as the San Joaquin
Valley (the middle east, for example). This also neglects the existence
of favorable microhabitats to which medflies could retreat. On our lack
of knowledge of medfly environmental tolerance, see our section on medfly
ecology and appendix note 8.
(vii) "Even if medfly gets established, it could be kept at low densities
with parasitic wasps and nematodes." Yet the economic threshold for dam-
age to fruit is very low; parasites have not reduced medfly populations
to acceptable levels elsewhere (Hagen et al. 1981), Wharton et al. 1981);
it is probably easier to eliminate infestations before they become well estab-
lished than to wait until they are and then resort to keeping them at low
(viii) "The insect has colonized a coastal strip extending from Los
Angeles to San Francisco." No such strip exists; there was a small infesta-
tion in Los Angeles and a larger one near San Francisco, with nothing in
(ix) "The insect eye cannot perceive fine details." Insect ethologists, who

Florida Entomologist 66 (1)

hypothesize that species identify each other on the basis of details of
morphology and color patterns, might question this statement. Jordan may
fail to distinguish the varying selection pressure on females (to make fine
discrimination in mate choices) as opposed to males (who are selected to
make "quick and dirty" decisions so as not to miss mating opportunities).
(x) "Research has shown that (sperm from second-mating males)
will take precedence over those from previous matings." This is the general
rule for insects; however, in medflies the sperm from two matings seem
to mix rather evenly (Katiyar and Ramirez 1971, Hooper 1972).
Jordan does emphasize the need for further research. On this we agree
3Dr. M. D. Huettel (pers. comm.) points out that at present SIT is
used in conjunction with localized ground spraying or widespread aerial
spraying of chemical pesticides, and that a major benefit of behavioral and
ecological research on the medfly will be to determine whether SIT can be
made to work effectively without insecticides.
4Sterilized flies were supplied to the California project from Hawaii
(from USDA Agricultural Research Service and from a rearing facility
financed by the California authorities), Mexico (Programa Moscamed, Di-
reccion Gral. Sanidad Vegetal), and Peru (Peruvian Dept. of Agriculture).
5The crucial role of SIT in maintaining this barrier is shown by the
differing response of medfly and Anastrepha fruit fly populations when aerial
spraying ceased in 1981. Medflies are still absent, but Anastrepha num-
bers have bounced back (J. Hendrichs, pers. comm.).
6We will never know whether the infestation in California was made
worse by a release of unsterilized Peruvian flies. Some project officials feel
that as many as 100,000 fertile flies could have been released, and that the
project may well have been on its way to succeeding with the integrated
control methods until this accident happened (Marshall 1981). Some ob-
servers doubt that unsterilized flies were actually released. Problems with
identification (see appendix note 13) have added to the confusion.
7Dr. M. D. Huettel of the USDA Insect Attractants, Behavior, and
Basic Biology Research Laboratory, Gainesville, FL, has performed isozyme
analyses of medfly samples from different populations. His work provided
evidence that the California infestation originated in Central America
rather than Hawaii. (Huettel, pers. comm.). For the importance in SIT
of knowing the genetic characteristics of the target population, see Richard-
son et al. (1982).
s An example of our uncertainty about medfly environmental tolerance
is found in Hagen et al. (1981), p. 6. The following two statements occur
close together: "Pupae subjected to 280F are usually killed in 30 h, ex-
posure to 320F for 4 days and 420F for 10 days likewise results in very
high mortality" and ". . most larvae and pupae die after 3 or 4 h at 400F".
sBateman (1972) distinguishes between dispersive movements and non-
dispersive movements such as those involved in feeding, oviposition, and
mating activities in a local area. Medflies probably move from vegetation
to fruit to vegetation to fruit to vegetation during the course of a day
(Prokopy and Hendrichs 1979). Traps must be designed to catch medflies
performing both dispersal and local movenient.
lOMales normally disperse farther in mammals, females in birds (Green-
wood 1980). Johnson (1969) proposed that female insects were most often
the ones making long-range flights. Incest avoidance is one function of
such differences, but which sex disperses farther is influenced by resource
needs and mating systems.
"Prokopy and Hendrichs (1979) suggest that matings on vegetation
involve virgin females and matings on fruit involve already-mated females.

March, 1983

Insect Behavioral Ecology-'82 Burk and Calkins

If so, then even those matings that are achieved by males searching on fruit
are devalued by the degree of sperm mixing.
12The history of medfly attractants is interesting. A farmer's wife in
Australia, who put kerosene around a fence post to keep ants out of the
jam cooling on the post, found the kerosene attracted a swarm of male
medflies (Severin and Severin 1913); kerosene was widely used (along
with other petroleum distillates until the late 1950s. After World War II,
a USDA mass-screening program discovered that several esters of a fairly
good mosquito repellent attracted male medflies. This was called "siglure"
and was used in the Florida eradication program of the 1950s along with
kerosene and angelica seed oil (Steiner et al. 1957). Beroza et al. (1961)
found that adding HC1 to the double bond of siglure greatly increased at-
traction. This was called "medlure." The tert-butyl ester, called "trimedlure"
is twice as attractive to male medflies as medlure.
13In the SIT phase of the California project, Jackson traps were re-
placed with Steiner traps (Harris et al. 1971). These are clear, plastic
cylindrical traps with trimedlure and a knock-down chemical inside. They
have the advantage of being a dry trap that allows easy access to specimens.
(Note again, the absence of color and the requirement that flies enter
the cylinder.) However, flies would buzz around inside these traps before
expiring. This led to body surface contamination of unmarked flies with
the fluorescent powders used to mark released flies. This problem arose
because workers adopted, for a short period of time, the simple method
of merely looking at flies for powder, not crushing their heads to reveal
the true powder mark accumulated on the ptilinum when the fly emerged
from the fluorescent-marked pupa.

ALCOCK, J. 1982. Natural selection and communication among bark beetles.
Florida Ent. 65: 17-31.
BAKER, R. R. 1978. The evolutionary ecology of animal migration. Holmes
and Meier, New York.
BATEMAN, M. A. 1972. The ecology of fruit flies. Annu. Rev. Ent. 17: 493-
.1976. Fruit flies. Pages 11-49 in V. L. Delucchi ed. Studies in
biological control. Cambridge University Press, Cambridge, England.
1961. New attractants for the Mediterranean fruit fly. Agric. Food
Chem. 9: 361-5.
Measuring, monitoring and improving the quality of mass-reared
fruit flies, Ceratitis capitata Wied. 1. The RAPID quality control sys-
tem for early warning. Z. ang. Ent. 92: 67-83.
BURK, T. 1981. Signaling and sex in acalyptrate flies. Florida Ent. 64:
-- 1982. Evolutionary significance of predation on sexually signalling
males. Florida Ent. 65: 90-104.
1983. Behavioral ecology of mating in the Caribbean fruit fly,
Anastrepha suspense (Loew) (Diptera: Tephritidae). Florida Ent.
-- AND J. C. WEBB. 1983. Effect of male size on calling propensity,
song parameters, and mating success in Caribbean fruit flies,
Anastrepha suspense (Loew). Ann. Ent. Soc. America 76: 000-00.
CADE, W. 1980. Alternative male reproductive behaviors. Florida Ent. 63:

16 Florida Entomologist 66 (1) March, 1983

CALKINS, C. O., D. L. CHAMBERS, AND E. F. BOLLER. 1982. Quality control
of fruit flies in a sterile insect release program. Pages 341-55 in Ster-
ile insect technique and radiation in insect control. Int. Atomic En-
ergy Agency, Vienna.
CHAMBERS, D. L. 1977a. Attractants for fruit fly survey and control.
Pages 327-44 in H. H. Shorey and J. J. McKelvey, Jr. eds. Chemical
control of insect behavior: theory and applications. John Wiley &
Sons, New York.
1977b. Quality control in mass rearing. Annu. Rev. Ent. 22: 289-
Pest control by attractants: A case study demonstrating economy,
specificity, and environmental acceptability. BioScience 24: 150-2.
Suppression of the Mediterranean fruit fly in Tunisia with released
sterile insects. J. Econ. Ent. 68: 237-43.
CHRISTENSON, L. D., AND R. H. FOOTE. 1960. Biology of fruit flies. Annu.
Rev. Ent. 5: 171-92.
CIRIO, U., AND I. DEMURTAS. 1974. Status of Mediterranean fruit fly control
by the sterile-male technique on the island of Procida. Pages 5-16 in
Sterile insect technique and its field application. Proc. Panel Int.
Atomic Energy Agency, Vienna.
DAVIS, J. C., H. R. AGEE, AND D. L. CHAMBERS. 1982. Identification of trap
features that promote capture of the Caribbean fruit fly, Anastrepha
suspense (Loew). Florida Ent. 65: 000-00.
DAWKINS, R. 1980. Good strategy or evolutionarily stable strategy? Pages
331-67 in G. W. Barlow and J. Silverberg eds. Sociobiology: beyond
nature/nurture. Westview Press, Boulder, Colorado.
DEMURTAS, I. D., U. CIRIO, AND S. D. ENKERLIN. 1970. An experiment to
control the Mediterranean fruit fly on the island of Procida by the
sterile-insect technique. Pages 59-70 in Sterile-male technique for
control of fruit flies. Int. Atomic Energy Agency, Vienna.
DREW, R. A. I., AND G. H. S. HOOPER. 1981. The responses of fruit fly spe-
cies (Diptera: Tephritidae) in Australia to various attractants. J.
Aust. Ent. Soc. 20: 201-5.
EMLEN, S. T., AND L. W. ORING. 1977. Ecology, sexual selection, and the
evolution of mating systems. Science 179: 215-23.
FARIAS, G. J., R. T. CUNNINGHAM, AND S. NAKAGAWA. 1972. Reproduction
in the Mediterranean fruit fly: Abundance of stored sperm affected by
duration of copulation, and affecting egg hatch. J. Econ. Ent. 65:
FISHER, K. 1981. Fruit fly under attack . from the sterile-insect-tech-
nique. J. Agric. Western Australia 22 (2) : 51-2.
FITT G. P. 1981. Responses by female Dacinae to "male" lures and their
relationship to patterns of mating behaviour and pheromone re-
sponse. Ent. exp. & appl. 29: 87-97.
FLETCHER, B. S. 1968. Storage and release of a sex pheromone by the
Queensland fruit fly Dacus tryoni (IDiptera: Trypetidae). Nature 219:
GIBBS, R., AND E. EERDE. 1981. Irradiation proving successful in eliminat-
ing the medfly. Mexico 1(11) : 16-8.
GREANY, P. D., A. K. BURDITT, AND D. L. CHAMBERS. 1982. Effectiveness of
Jackson traps for fruit flies improved by addition of colored pat-
terns. Florida Ent. 65: 374-5.
GREENFIELD, M. D., AND K. C. SHAW. 1982. Adaptive significance of chorus-
ing in insects. Pages 00-00 in D. T. Gwynne and G. K. Morris eds.

Insect Behavioral Ecology-'82 Burk and Calkins

Orthopteran mating systems: Sexual competition in a diverse group
of insects. Westview Press, Boulder, Colorado.
GREENWOOD, P. J. 1980. Mating systems, philopatry and dispersal in birds
and mammals. Anim. Behav. 28: 1140-62.
HAGEN, K. W., W. W. ALLEN, AND R. L. TASSAN. 1981. Mediterranean
fruit fly: The worst may be yet to come. California Agric., March-
April: 5-7.
HAMILTON, W. D., AND R. M. MAY. 1977. Dispersal in stable habitats. Na-
ture 269: 578-81.
HARRIS, E. J. 1977. The threat of the Mediterranean fruit fly to American
agriculture and efforts being made to counter this threat. Proc. Ha-
waiian Ent. Soc. 22: 475-80.
S. NAKAGAWA, AND T. URAGO. 1971. Sticky traps for detection and
survey of three tephritids. J. Econ. Ent. 64: 62-5.
HENDRICHS, J. 1983. Sexual selection in wild and sterile Caribbean fruit
flies, Anastrepha suspense (Loew). M.Sc. Thesis, Univ. of Florida,
Gainesville. (In prep.).
HESS, C. E. 1981. Economic impact of the Mediterranean fruit fly. Report
presented to meeting of California St. Board of Food and Agric.
Aug. 6, 1981.
HOOPER, G. H. S. 1972. Sterilization of the Mediterranean fruit fly with
gamma radiation: Effect on male competitiveness and change in
fertility of females alternately mated with irridated and untreated
males. J. Econ. Ent. 65: 1-6.
HUETTEL, M. D. 1976. Monitoring the quality of laboratory-reared insects:
a biological and behavioral perspective. Environ. Ent. 4: 801-7.
ITO, Y., AND H. KAWAMOTO. 1979. Number of generations necessary to at-
tain eradication of an insect pest with sterile insect release method:
A model study, Res. Popul. Ecol. 20: 216-26.
JOHNSON, C. G. 1969. Migration and dispersal of insects by flight. Methuen
and Company, London.
JORDAN, W. H., JR. 1982a. A fruitless pursuit. Science 82 3(4): 62-8.
1982b. Invasion of the medfly. Nat. Hist. 91(5): 71-8.
KATIYAR, K. P., AND E. RAMIREZ. 1970. Mating frequency and fertility of
Mediterranean fruit fly females alternately mated with normal and
irradiated males. J. Econ. Ent. 63: 1247-50.
KNIPLING, E. F. 1955. Possibilities of insect control or eradication through
the use of sexually sterile males. J. Econ. Ent. 48: 459-62.
.1979. The basic principles of insect population suppression and
management. U. S. Dept. Agric. Handbook 512, 623 pp.
MARSHALL, E. 1981. Man versus medfly: Some tactical blunders. Science
CHAMBERS. 1971. Reproduction in the Mediterranean fruit fly: Fre-
quency of mating in the laboratory. Ann. Ent. Soc. America 64: 949-
URAGO, AND E. J. HARRIS. 1978. Visual orientation of Ceratitis capi-
tata flies to fruit models. Ent. Exp.' & Appl. 24: 193-8.
Mediterranean fruit fly: Laboratory and field evaluations of syn-
thetic sex pheromones. J. Environ. Sci. Health A12: 67-78.
PERDOMO, A. J. 1974. Sex and aggregation pheromone bioassays and mat-
ing observations of the Caribbean fruit fly, Anastrepha suspense
(Loew), under field conditions. Ph.D. Dissertation, Univ. of Florida,

Florida Entomologist 66 (1)

PROKOPY, R. J. 1976. Stimuli influencing tropic relations in Tephritidae.
Coll. Int. C.N.R.S. 265: 305-36.
1980. Mating behavior of frugivorous tephritidae in nature. Pages
37-45 in Proc. Symp. Fruit Fly Prob., XVI Int. Cong. of Ent., Kyoto.
AND J. HENDRICHS. 1979. Mating behavior of Ceratitis capitata on
a field-caged host tree. Ann. Ent. Soc. America 72: 642-8.
RHODE, R. H. 1970. Application of the sterile-male technique in Mediterra-
nean fruit fly suppression: A follow-up experiment in Nicaragua.
Pages 43-50 in Sterile male technique for control of fruit flies. Int.
Atomic Energy Agency, Vienna.
control of screwworms in North America. Science 215: 361-70.
SEVERIN, H. H. P., AND H. C. SEVERIN. 1913. A historical account on the
use of kerosene to trap the Mediterranean fruit fly (Ceratitis capi-
tata). J. Econ. Ent. 6: 347-51.
SHOUKRY, A., AND M. HAFEZ. 1979. Studies on the biology of the Mediter-
ranean fruit fly Ceratitis capitata. Ent. Exp. & Appl. 26: 33-9.
oils as Mediterranean fruit fly lures. J. Econ. Ent. 50: 505.
role of attractants in the recent Mediterranean fruit fly eradication
program in Florida. J. Econ. Ent. 54: 30-35.
W. C. MITCHELL, AND A. H. BAUMHOVER. 1962. Progress of fruit fly
control by irradiation sterilization in Hawaii and Marianas Islands.
Int. J. Appl. Rad. Isotopes 13: 427-34.
CHRISTENSON. 1965. Melon fly eradication by overflooding with
sterile males. J. Econ. Ent. 58: 519-22.
TRIVERS, R. L. 1972. Parental investment and sexual selection. Pages 136-
79 in B. Champbell ed. Sexual selection and the descent of man,
1871-1971. Aldine, Chicago.
WADE, N. 1982. Another look at agricultural research. Science 215: 483.
WALSH, J. 1981. Medfly continues to bug California. Science 214: 1221-3.
HART. 1981. Hymenopterous egg-pupal and larval-pupal parasitoids
of Ceratitis capitata and Anastrepha spp. [Dip.: Tephritidae] in
Costa Rica. Entomophaga 26: 285-90.
TANAKA, AND E. J. HARRIS. 1982. Mediterranean fruit fly: Dispersal
of wild and irradiated and untreated laboratory-reared males.
Environ. Ent. 11: 339-43.
ZOUROS, G. E. 1969. On the role of female monogamy in the sterile-male
technique of insect control. Ann. l'Inst. Phytopath. Benak:, New
Ser. 9(1) : 20-9.

March, 1983

Insect Behavioral Ecology-'82 Waage



Until the mid-1960's, much of the focus in animal behavior was on
identifying and describing species-specific, stereotyped behavior patterns
(e.g. courtship, oviposition) for comparison with other species and for
working out phylogenetic relationships (Klopfer & Hailman 1967). Natural
selection, when it was considered, was often assumed to favor behaviors
that were best for the survival and reproduction of the species. When
interactions within species were discussed it was generally assumed that
they were cooperative and governed by what was best for the species. These
general perspectives, while not universal, had considerable influence on
what behavior was studied and on how it was interpreted.
The late 1960's saw a resurgence of the Darwinian perspective, the
recognition that natural selection acts primarily at the level of individuals
(Williams 1966). Variations among individuals became important since
they provided clues to the dynamics of evolution. Behavioral interactions
were viewed in the context of reproductive selfishness, and previous expla-
nations of apparently cooperative behaviors were re-examined. It was
during this resurgence of Darwinian thinking that the disciplines of be-
havioral ecology and sociobiology emerged (Alcock 1979a; Alexander 1979;
Krebs & Davies 1978, 1981; Wilson 1975; Wittenberger 1981).
Sexual selection and ESS (Evolutionarily Stable Strategy) theories
have rapidly gained central positions in behavioral ecology theory (Blum &
Blum 1979; Campbell 1972; Caryl 1980; Dawkins 1980; Maynard Smith
1978, 1979, 1982; Parker, 1978; Smith 1983). The importance of the new
insights gained from these theoretical perspectives has been a major theme
of the Florida Entomological Society symposia on insect behavioral ecol-
ogy. In this paper I will focus on the sperm competition aspect of sexual
selection and the game theory models for animal contests of ESS theory.
Sexual selection and ESS theory are conceptually and mechanistically
linked. Both see behavioral interactions among members of the same spe-
cies as a major selective force and both emphasize how competition among
individuals can result in and maintain behavioral variants in populations.
I will briefly discuss some aspects of these theories and then illustrate
them with examples from the damselfly, Calopteryx maculata (De Beau-
vois). In the contexts of sperm competition and ESS models for animal
contests, these examples reveal otherwise un-noticed causal connections
among behaviors and produce testable hypotheses about the dynamics of
behavioral interactions.

*Jonathan K. Waage is an Associate Professor in the Program in Ecology and Evolu-
tionary Biology at Brown University. His research centers on sexual selection theory and
the evolution of odonate reproductive behavior. Recently this has focused on sperm com-
petition and its influence on odonate mating systems and postcopulatory behavior. He has
also published on reproductive isolation and character displacement. Current Address: Divi-
sion of Biology and Medicine. Box G. Brown University, Providence, RI 02912.

Florida Entomologist 66 (1)


Sexual selection generally involves competitive interactions within and
between sexes, and arises from the different evolutionary "interests" of
the sexes and the fact that one sex is usually limited in availability. It
occurs in two basic contexts: 1) mate choice (see Thornhill, 1980), and
2) intrasexual competition for mates (or more accurately, for fertiliza-
Historically, considerable attention has been devoted to what males
do to "obtain mates", but little work has focused on just how female mor-
phology, physiology, life history and behavior ultimately influence male re-
productive behavior. Sexual selection and ESS theories emphasize the fact
that much of the behavior of one sex must be largely influenced (in both a
proximate and ultimate sense) by the behavior of the other sex (e.g.
Parker 1979, Thornhill 1980, Waage 1983).
Intrasexual competition generally involves aggressive interactions, often
associated with territoriality and dominance, whereby some individuals
gain access to breeding sites or have first access to prospective mates. A
second kind of intrasexual competition, sperm competition, may be of
even greater significance, especially for insects (Parker 1970; Smith 1983).
It takes place after insemination but before fertilization of eggs, and
the competition involves males and their sperm. In the insects examined
so far the last male to mate appears to fertilize most of the eggs which
the female then oviposits (Boorman and Parker 1976; Smith 1983). Sperm
competition has resulted in a wide variety of mechanisms used by males
to remove, displace, or otherwise out-compete the sperm of rivals. The
existence of such mechanisms in turn selects for other devices (e.g. mating
plugs) and behaviors (e.g. postcopulatory guarding) which reduce a
male's risk of further sperm competition after he has mated (see Wing
et al, this symposium).
Although the emphasis is on male competition for female eggs, it is
essential to realize that females control this competition, making its very
existence possible. They do so because there are three prerequisites that
must be met for the evolution of sperm competition, all of which are attri-
butes of female morphology and behavior. These are: 1) a delay between
insemination and fertilization, 2) the accessibility of previous sperm to
males or their gametes, and 3) two or more matings per oviposition or
These conditions for the evolution of sperm competition are met by a
wide variety of organisms; and this in turn raises the fundamental ques-
tion-why should females mate more than once per clutch of eggs? Possi-
ble answers to this question range from obtaining increased genetic di-
versity among offspring to exchanging matings for protection during
oviposition or for access to oviposition sites controlled by males (Waage,
1983; Walker, 1980).


ESS theory utilizes game theory modeling and the concept of frequency-
dependent selection to explore why one or more behavioral variants in a
population are stable to invasion by others (Dawkins 1980; Maynard Smith

March, 1983

Insect Behavioral Ecology-'82 Waage

1982). It does not predict a "best" strategy in an absolute sense, but, rather,
seeks to explain stable equilibria among alternatives present or potentially
present in a population.
ESS theory involves a cost/benefit analysis of alternative behavioral
tactics in a given context (e.g. contests over resources). The units or
currency of this analysis are fitness increments. These are estimated by
time, energy, and risk spent versus survival and reproduction gained. In
the game theory metaphor, organisms "bid" fitness units in "games" against
other individuals (and their bids) in an "attempt" to make a fitness "profit."
A problem for applying ESS theory (and sexual selection theory) is that
one must find practical yet accurate measures of fitness units for indi-
viduals in natural populations. This is not an easy task (Howard 1979)
but its importance for understanding the evolution of behavior is em-
phasized by both ESS and sexual selection theory (e.g. Maynard Smith
1982; Wade and Arnold 1979).
Some general insights into animal behavior have emerged from the
ESS theory. First, the success (gained in fitness) of a particular strategy
will often depend largely on the nature and relative abundance of other
strategies in the population. Second, rather than there being one optimal
or best behavior in a given context, two or more alternative (equally fit)
behaviors may end up coexisting either as a polymorphism in the popula-
tion or as alternative tactics employed by an individual in a given situ-
ESS theory models were first developed (Maynard Smith 1979, 1982)
to provide a more rigorous answer to a general question: Why are most
fights settled by simple, ritualized displays or conventions and not by
escalated, physical struggles? The intuitive answer has previously been
that escalated fighting is too risky, even for the winner, and thus settling
contests by display or ritualized fighting would be favored. ESS theory,
by providing a formal, mathematical approach to analyzing simple con-
tests, has led to a better understanding of the "rules" and dynamics of
such contests.
A primary function of ESS theory is the generation of general per-
spectives and specific explanations for behaviors that would otherwise have
been overlooked. Maynard Smith (1982) provides a summary of the value
of ESS theory for a broad range of evolutionary questions. Several con-
tributors to the 1979 Florida Entomological Society Symposium have also
illustrated the usefulness of ESS theory for studying insect behavior in a
variety of contexts. These include: alternative reproductive tactics in crick-
ets (Cade 1980), alternative nesting behaviors of wasps (Brockmann 1980),
and parental care in Hemiptera (Smith 1980).



In damselflies and most dragonflies, males remain in tandem with or
otherwise protect ovipositing mates (Corbet 1962, 1980; Waage 1983). Prior
to Parker's (1970) introduction of sperm competition into sexual selection
theory, it was generally assumed that postcopulatory protection functioned
to lessen disturbance during oviposition (e.g. Corbet 1962). Parker's (1970)

Florida Entomologist 66 (1)

paper caused me to ask if postcopulatory guarding might also function
as a means of avoiding sperm competition. It now appears that sperm
displacement, and postcopulatory behavior to lessen its risk, are widespread
among odonates, especially zygopterans (Waage 1983). The following
discussion illustrates how pursuing such ideas in the context of sexual
selection (and ESS theory) has led to the discovery of a rather complex
set of behavioral interactions that illustrate how male and female perspec-
tives interact in shaping behavior.
Most Calopteryx maculata males are territorial at oviposition sites
(emergent vegetation) and, except at low population densities, most avail-
able sites are occupied by males. Because of repeated mating attempts by
these males, females have considerable difficulty in ovipositing without
mating and then being guarded by a male (Waage 1978). At low and
moderate densities territorial males have a mating advantage over non-
territorial ones by controlling female access to oviposition sites and ex-
changing matings for guarding. During copulation, C. maculata males
removed virtually all of the sperm of previous males and then replace
it with their own (Waage 1979a). Under these conditions nonterritorial
males, even if they do get to mate with an occasional female, are not
successful (fertilize no eggs) if that female must then mate with a terri-
torial male before ovipositing.
Females generally begin oviposition in their mate's territory, but may
change location before finishing. One might expect females to trade mat-
ings for guarding whenever they change oviposition sites. However, 1)
these females carry sufficient viable sperm for several bouts of oviposition,
2) mating involves time and may involve predation risk for females, and
3) odonate females cannot be directly forced to mate (females must actively
bring their abdomens forward to engage with the male's genitalia). There-
fore, it seems more reasonable that if females could avoid these mating
costs while still gaining access to new oviposition sites they would do so.
We now have a "conflict of interest" between males and females. The
dynamics are complex because the costs to males (lost fertilizations) vary
with frequency of females attempting oviposition without mating and it
is not certain that the cost to males is greater than that to females who
mate in exchange for guarding (time and risk). However, it would seem
reasonable to expect females to avoid the costs and risks of remating
whenever possible.1
In populations I have observed, males do occasionally guard females
with which they have not mated (up to 1/3 of males guarding mates also
guarded non-mates-Waage 1979b). This occurs largely because females
are able to exploit weaknesses in male behavior that allow them to sneak
in and be guarded by a male if he is already guarding a mate(s) (Waage
1979b). This "loophole" in the system allows the existence of alternative
female tactics. But the sneaky tactic can' only work if other females and
territorial males are "playing the game by the rules"-mating in exchange
for guarding.
There is one more step in this chain of behavioral interactions. If fe-
males who mate with nonterritorial males are able to oviposit and be
guarded without remating, then it is possible for nonterritorial males
to be successful (fertilize eggs), because their sperm will not be displaced.
The non-territorial "tactic" then becomes a viable option for reproduction.

March, 1983

Insect Behavioral Ecology-'82 Waage

In summary, female behavior (multiple mating per clutch) seems to
have led to male ability to displace sperm and thus to the selective pres-
sure on protecting mates during oviposition. Since oviposition sites and
receptive females are relatively scarce and defensible, males can generally
obtain matings in exchange for guarding which further selects for male
sperm removal ability (if one is going to pass up matings, with other
females, while guarding then it is best to be sure you are fertilizing all of
the eggs being deposited).
Once guarding becomes common some females can exploit male be-
havior for "free" access to oviposition sites and guarding, and in turn
other males can obtain fertilizations without the costs of obtaining and
defending territories. What remains to be determined is if and how fre-
quency-dependent selection may be influencing these interactions. Are there
two types of females in these populations (rule players and sneaks) with
equal fitness at certain relative abundances? Why have males not evolved
the ability to prevent females from gaining protection without mating?
Do territorial and nonterritorial males represent alternative tactics or
are non-territorial males just "making the best of a bad situation"? These
questions, generated by ESS and sexual selection theory, must still be
When looked for, the presence of multiple reproductive tactics within
populations appears to be fairly common (e.g. Blum & Blum 1979; Brock-
mann 1980; Cade 1980; Maier & Waldbauer 1979; Thornhill 1981). This
provides an important message for those studying insect behavior-focusing
on behavioral "norms' or species-specific behavior may often result in
overlooking some important behaviors and their causes.

I am currently studying the dynamics of territorial disputes in C.
maculata, and in particular, why some disputes escalate in duration and
intensity. Preliminary results allow me to test some predictions of ESS
models for animal contests.
The game theory models most relevant to my proposed work are those
for asymmetric contests (e.g. "Hawk-Dove-Bourgeois" or "War of Attri-
tion" models-Bishop and Cannings 1978; Caryl 1980; Hammerstein and
Parker 1982; Maynard Smith 1979, 1982; Maynard Smith and Parker 1976;
Parker and Thompson 1980).2 A basic asymmetry (there may be others)
for Calopteryx territorial contests is the fact that one male is a resident
and the other is not when a contest occurs. ESS game theory models predict
that such contests should be settled quickly and by conventions that involve
display and recognition of this asymmetry, which can reflect such factors
as size, energy or residency. Escalated contests should only occur if 1)
a second asymmetry outweighs the first (e.g. if the intruder is strong
enough to overcome a "resident wins" convention), or 2) there is confusion
by the contestants about the asymmetry or their roles in the convention
(e.g. both combatants "think" they are the resident).
Calopteryx disputes involve chasing and displaying by males and are
of variable duration and intensity, ranging from a short, one-sided chase
of an intruder out of the territory to an intense and rapid spiraling,
often covering an area several meters long (Waage, 1978). Most fights

Florida Entomologist 66 (1)

March, 1983

(86%) last less than 15 sec and are of low intensity. Others occur at
escalated levels and may last up to several hours (broken into a series of
extended fights between the same males). Why then should some contests
take so long to settle? One possible answer is the game theory prediction
that escalations would occur if habitat or behavioral factors resulted in
"'confusion over residency."
Table 1 shows the duration and level of intensity (on a scale of 1 to
3) of 1032 territorial disputes recorded in 1980. More than 3/4 of all ob-
served fights were between a resident and a nonterritorial intruder or be-
tween neighboring territory residents. The remaining fights ("confusion")
involved interlopers or situations in which two males accidentally both
became residents in the same territory (see Fig. 1A & B). The mean
durations and intensity levels of fights for the intruder and neighbor con-
texts did not differ significantly. Over 98% of the disputes were won by
the resident (or if neighbors fought, each retained his territory). Con-
tests between neighbors often decreased in duration and intensity as they
appeared to work out the boundaries between their territories (Table 2A
and Fig. 1A). Such boundary disputes accounted for many of the escalated
fights early in the day. The proportion of level 2 & 3 (high intensity)
fights was lowest for the intruder context (31%), higher in the neighbor
context (53%), and highest in the context I have called "confusion" (86%).
The distribution of fights among intensity levels differs significantly for the
three contexts. Fights in the "confusion" context were significantly longer
than intruder and neighbor disputes.
As Fig. 1 illustrates some of this "confusion" was largely habitat-
mediated. Certain distributions of perch and oviposition sites seemed to
make escalated contests more likely. A second apparent cause of "confu-
sion" appeared to result from situations in which residents left their terri-
tories for one to several minutes while mating or during long chases and
fights. If an interloper took over the territory during this period, an
escalated contest with the resident generally resulted. A small number of
intruders seemed to persist and cause escalation for no obvious reason.
They may have been particularly aggressive males or their behavior might
be influenced by unobserved previous experience. Fights involving these
males were included in the neighbor or intruder contexts in Table 1. Table


1 2 3

INTRUDER 479 (46.4%) 13.4 (1.5) 69 24 7
NEIGHBOR 313 (30.3%) 11.8 (1.1) 47 44 9
"CONFUSION"* 240 (23.3%) 43.9 (6.8) 14 62 24

*"Confusion" = shared perch & oviposition sites and interloners (qp tpxt nnr vlio 1I

Insect Behavioral Ecology-'82 Waage 25

S0 /

--- l) _


0 0 0w

zJ "\ o |
> \ ,0 O 0 a

SI I W bn Q) Flo.

E .E
aC < F; .1 i ?

CO U. d -

Q 0
-Q -
'E a
0 z p 0

og h --<

I< C .; 1
A0 (1
o, C -

0. >c >

D 0 < 3 t 2
h)CL 0..9

S0 ,. < '-
U) 0U 0 C)) (1 d

LU LU L Ud '0
0 4 / 0 *0

C- T0
0 (D0


Florida Entomologist 66 (1)

2 provides some examples of escalations involving the "confusion" context
(Fig. 1A & B) and intruding males that persisted for unknown reasons
(Fig. 1C). The "vs others" data in Table 2 are included to show that the
context of the interaction and not the aggressiveness of the resident was
responsible for any escalation.
The fact that most escalations occurred in contexts where some degree
of confusion over territory residency (Fig. 1) is likely is consistent with
the predictions of game theory models. However, to avoid the circularity
of calling escalated fights "confusion" and then claiming confusion escalates
fights, it is necessary to do two things. First, as I have done in Table 1,
fights must be organized by context and then examined for the degree of
escalation. Second, I must demonstrate that by creating the situation I
hypothesize to result in "confusion" I can escalate fights between any two
males. Experiments to do the latter are in progress and I shall briefly
present some of their results.
One experiment, which takes advantage of the fact that C. maculata
territories are centered around oviposition sites (Waage 1973), involves the
manipulation of males and their territories using movable oviposition sites
(emergent vegetation tied to fishing line). Two such sites are set up about
3 m apart. When each is occupied by a male, the durations of fights be-
tween the new neighbors are recorded. After this background data is ob-
tained ("apart" in Table 3), the sites are moved together thereby putting
both males in the same territory. Both should act as if they are resident
and escalated fights are predicted because of the resulting residency con-
fusion. Moving the oviposition sites apart again restores the separate ter-
ritories and should lead to de-escalated fights. This is exactly what happens
(Table 3). A variation of this experiment is to leave both oviposition sites
together until one male eventually wins (only one territory exists). This
manipulation should reveal whether other asymmetries exist that would
allow the eventual winner to be predicted from previous observations on
his size, fighting or mating success.


A. Neighbors settling boundaries-first 50 min of day:
FIRST 30 MIN N = 8 MEAN = 69.8 sec
NEXT 20 MIN N = 9 MEAN = 6.6 sec
B. Two males (X and Y) defending the same oviposition site:
X VS OTHERS N = 14 MEAN = 15.6 sec
X VS Y N= 9 MEAN = 169.3 sec
C. Interloper takes territory while resident away fighting:
RES. VS OTHERS N = 8 MEAN = 23.7 sec
RES. VS INTERLOPER N = 4 MEAN = 103.3 sec
D. Intruder persists-intruder = previous resident:
RES. VS OTHERS N = 22 MEAN = 16.9 sec
RES. VS INTRUDER N = 43 MEAN = 45.3 sec

March, 1983

Insect Behavioral Ecology-'82 Waage



A. APART 9 7.7 (1.6)
TOGETHER 2 176.0 (-)
APART 7 7.0 (2.0)
TOGETHER 6 45.2 (26.7)
B. APART 6 4.6 (1.2)
TOGETHER 3 110.7 (76.5)
APART 6 4.8 (1.3)
TOGETHER 6 319.3 (234.6)
C. APART 4 7.7 (2.1)
TOGETHER 8 145.3 (64.9)
APART 9 16.4 (3.9)
TOGETHER 7 91.3 (18.1)

In order to test the interloper aspect of what I am calling "confusion",
I am using a variation of the resident removal and re-release experiments
which Davies (1978) did with the butterfly Pararge aegeria. I remove a
resident, hold him in a cooler, and re-introduce him once an interloper has
become established in the territory. When the former resident is released
both males should then consider the territory theirs and escalations are
expected. Preliminary experiments with C. maculata have shown that
residents can be removed and held in coolers for as long as 45 min with
no obvious effects on fighting ability. As expected, the resulting fights
with interlopers were escalated, and only rarely was the interloper able
to keep the territory when the former resident was released. Future ex-
periments will test if it is possible to influence how difficult it is for a resi-
dent to displace an interloper by varying the time an interloper is present
or the number of matings he has before releasing the resident.
This brief description of some preliminary results from my work on
Calopteryx fighting behavior shows some of the predictions of game theory
models do hold. These are: 1) residents usually win and do so in short
contests, and 2) escalation occurs primarily when confusion exists over
who is resident. In order to fully understand the dynamics of these terri-
torial disputes, ESS models require answers to the following questions:
1. What are the actual costs and gains for the contestants?
-Time, energy spent or matings missed
2. Why should the intruder give up?
-Physical or experience asymmetry
-Cost/gain asymmetry
-Simple "resident wins" convention
3. What is the role of proximate factors in these contests?
-Oviposition site density and distribution
-Male density
-Female presence or arrival rate

Florida Entomologist 66 (1)

4. Do variations in male or territory quality affect contest dynamics
and results?
5. How do contest duration and intensity interact?
-Intruders determine duration, who changes the intensity
The identification of such questions and their relationship to understanding
the dynamics of animal contests have been a major contribution of ESS

Sexual selection and evolutionarily stable strategy (ESS) theories pro-
vide valuable insights into the evolution and dynamics of insect behavior.
Both theories stress the importance of selection in the context of inter-
actions among conspecifics and its frequency dependent nature. Sperm
competition is a particularly significant aspect of sexual selection theory
for insect behavior. It points out the influence of differing male and female
evolutionary perspectives on mating and postcopulatory behavior.
A major importance of ESS theory to studying behavior is its identifica-
tion of and explanations for the coexistence of alternative behaviors in a pop-
ulation. Previous approaches to behavior assumed only a single, best be-
havior should exist. ESS game theory models also provide a formal, mathe-
matical approach to determining the rules for and dynamics of animal
The relevance of sexual selection (sperm competition) and ESS to
damselfly behavior is illustrated with examples of male/female interactions
and territorial contests.

I thank Joy Bergelson, Jane Brockmann, Ola Fincke, Jim Lloyd and
Doug Morse for comments on earlier drafts on this paper. Joy Bergelson,
Mark Camara, Donna Fernandes and Martin Platt assisted in collection of
some of the data presented on Calopteryx fighting behavior. Donna Fer-
nandes did most of the compilation and analysis of this data and provided
many helpful suggestions. The research reported on has been generously
supported by grants from the National Science Foundation (DEB 77-
15904 and DEB 80-04282) to the author. I thank Jim Lloyd for inviting
me to participate in his ongoing series of symposia on insect behavioral

1. The situation is actually more complex (see Alcock 1979b and Waage
1979b). Males attempting matings with new females while guarding mates
are 1) not assured of mating the new female (50% chance) and 2) are
likely to lose previous mates to other males while courting or mating a
new one. The resultant balance between guarding non-mates and attempt-
ing matings with all females probably depends on: 1) the actual costs and
benefits involved for both sexes, 2) the population density (which affects
the rate of female arrivals and the chance a male will lose an ovipositing
mate if he attempts to court a new female), and 3) the number of times
a male can effectively mate in a short period of time. If such factors can

March, 1983

Insect Behavioral Ecology-'82 Waage

be measured and adjusted by density, then an ESS analysis of the condi-
tions for guarding non-mates and the relative frequency of non-mates
guarded can be done. Such an analysis is in progress.
2. I do not mean to imply that my approach is to directly test ESS
models. Most of those in the literature are over-simplified and intended for
heuristic analyses and not direct testing (Maynard Smith 1982). However,
the general predictions of these models are testible and the models are
modifiable. Calopteryx contests do not readily fit the "Hawks and Doves"
models (little or no risk of injury exists and fights are of variable duration)
and may be more appropriately approximated by "War of Attrition" models.
The contests are probably asymmetric, at least to the extent that they in-
volve residents versus intruders, and thus the most general predictions
of all asymmetric models (which are similar-Maynard Smith 1979, 1982)
can be tested. More importantly, testing the predictions leads to the meas-
urement of costs and benefits involved in the contest dynamics and to the
experimental determination of the "rules" for determining winners.


ALCOCK, J. 1979a. Animal Behavior: An Evolutionary Approach. Sinauer
Assoc., Inc. Sunderland, Mass.
ALCOCK, J. 1979b. Multiple mating in Calopteryx maculata (Odonata:
Calopterygidae) and the advantage of non-contact guarding by males.
J. Nat. Hist. 13: 439-46.
ALEXANDER, R. D. 1979. Darwinism and Human Affairs. University of
Washington Press, Seattle.
BISHOP, D. T. AND C. CANNINGS. 1978. A generalized war of attrition. J.
Theor. Biol. 70: 85-124.
BLUM, M. S. AND N. A. BLUM (eds.) 1979. Sexual Selection and Repro-
ductive Competition in Insects. Academic Press, New York.
BOORMAN, E. AND G. A. PARKER. 1976. Sperm (ejaculate) competition in
Drosophila melanogaster, and the reproductive value of females to
males in relation to female age and mating status. Ecol. Ent. 1:
BROCKMANN, H. J. 1980. Diversity in the nesting behavior of Muddaubers
(Trypoxylon politum Say: Sphecidae). Florida Ent. 63: 53-64.
CADE, W. 1980. Alternative male reproductive behaviors. Florida Ent. 63:
CAMPBELL, B. G. (ed.). 1972. Sexual Selection and the Descent of Man
1871-1971. Aldine Publ. Co., Chicago.
CARYL, P. G. 1980. Escalated fighting and the war of nerves: games theory
and animal combat. In, P. P. G. Bateson & P. H. Klopfer (eds.). Per-
spectives in Ethology. Ch. 10, pp. 199-224. Plenum Press, N.Y.
CORBET, P. S. 1962. A Biology of Dragonflies. M. F. & G. Witherby, Ltd.,
1980. Biology of Odonata. Ann. Rev. Ent. 25: 189-217.
DAVIES, N. B. 1978. Territorial defense, in the speckled Wood Butterfly
(Pararge aegeria): the resident always wins. Anim. Behav. 26:
DAWKINS, R. 1980. Good strategy or evolutionary stable strategy? In G.
W. Barlow & J. Silverberg (eds.) Sociobiology: Beyond Nature/
Nurture? AAAS Selected Symposium 35. pp. 331-67. Westview Press,
HAMMERSTEIN, P. AND G. A. PARKER. 1982. The asymmetric war of attri-
tion. J. Theor. Biol. 96: 647-82.

Florida Entomologist 66 (1)

HOWARD, R. D. 1979. Estimating reproductive success in natural popula-
tions. American Nat. 114: 221-31.
KLOPFER, P. H. AND J. P. HAILMAN. 1967. An Introduction to Animal Be-
havior. Ethology's First Century. Prentice-Hall, Engelwood Cliffs,
KREBS, J. R. AND N. B. DAVIES. 1978. Behavioural Ecology: an Evolution-
ary Approach. Sinauer Assoc., Inc., Sunderland, Mass.
1981. An Introduction to Behavioral Ecology. Sinauer Assoc., Inc.,
Sunderland, Mass.
MAIER, C. T. AND G. P. WALDBAUER. 1979. Dual mate-seeking strategies in
male syrphid flies (Diptera: Syrphidae). Ann. Ent. Soc. America 72:
MAYNARD SMITH, J. 1978. The evolution of behavior. Sci. American 239:
1979. Game theory and the evolution of behavior. Proc. R. Soc.
London B 205: 475-88.
1982. Evolution and the Theory of Games. Cambridge University
Press, Cambridge.
MAYNARD SMITH, J. AND G. A. PARKER. 1976. The logic of asymetric con-
tests. Anim. Behav. 24: 159-75.
PARKER, G. A. 1970. Sperm competition and its evolutionary consequences
in the insects. Biol. Rev. 45: 525-67.
1978. Selfish genes, evolutionary games, and the adaptiveness of
behaviour. Nature 274: 849-55.
.1979. Sexual selection and sexual conflict. In M. S. Blum and N. A.
Blum (eds.). Sexual Selection and Reproductive Competition in In-
sects. pp. 123-66. Academic Press, New York.
PARKER, G. A. AND E. A. THOMPSON. 1980. Dung fly struggles: a test of
the war of attrition. Behav. Ecol. Sociobiol. 7: 37-44.
SMITH, R. L. 1980. Evolution of exclusive postcopulatory parental care in
the insects. Florida Ent. 63: 65-78.
(ed.). 1983. Sperm Competition and the Evolution of Animal Mat-
ing Systems. Academic Press, New York (in press).
THORNHILL, R. 1980. Competitive, charming males and choosy females:
was Darwin correct? Florida Ent. 63: 5-30.
1981. Panorpa (Mecoptera: Panorpidae) Scorpionflies: Systems
for understanding resource-defense polygyny and alternative male
reproductive efforts. Ann. Rev. Ecol. & Syst. 12: 355-86.
WAAGE, J. K. 1973. Reproductive behavior and its relation to territoriality
in Calopteryx maculata (Beauvois) (Odon.; Calopterygidae). Be-
haviour 47: 240-56.
1978. Oviposition duration and egg deposition rates in Calopteryx
maculata (Beauvois) (Zygoptera; Calopterygidae). Odonatologica
7: 77-88.
1979a. Dual function of the damselfly penis: Sperm removal and
transfer. Science 203: 916-18.
.1979b. Adaptive significance of postcopulatory guarding of mates
and non-mates by male Calopteryx maculata (Odonata). Behav. Ecol.
Sociobiol. 6: 147-54.
.1983. Sperm competition and the evolution of odonate mating sys-
tems. In R. L. Smith (ed) Sperm Competition and the Evolution of
Animal Mating Systems. Academic Press, New York (in press).
WADE, M. J. AND S. J. ARNOLD. 1980. The intensity of sexual selection in
relation to male sexual behavior, female choice, and sperm prece-
dence. Anim. Behav. 28: 446-61.

March, 1983

Insect Behavioral Ecology-'82 Waage 31

WALKER, W. F. 1980. Sperm utilization strategies in nonsocial insects.
American Nat. 115: 780-99.
WILLIAMS, G. C. 1966. Adaptation and Natural Selection. Princeton Univ.
Press, Princeton.
WILSON, E. 0. 1975. Sociobiology. Harvard Univ. Press, Cambridge, Mass.
WITTENBERGER, J. F. 1981. Animal Social Behavior. Duxbury Press, Boston.

Florida Entomologist 66 (1)



Animals have often independently evolved analogous traits in response
to similar selection pressures in different localities. Examination of the
habitats of these species may reveal common features of their environments
that prompted evolution of the shared traits (Alcock 1979). Although the
evolutionary convergence of distinct taxa is a widely accepted principle,
most examples in the literature concern morphological similarities among
birds or mammals of differing phylogenetic ancestries (e.g., Mayr 1963,
MacArthur 1972). The convergence of behavioral traits has received less
attention because behaviors are more tedious to inventory and comparative
studies are fewer.
Because of their limited and stereotyped repertoires, insects ought to
provide excellent material for studying behavioral convergences. Obvious
convergences have appeared in unrelated social insects such as ants and
termites. Within single families, similar behavioral traits may have evolved
independently, such as the stutter-trill acoustic signal which is found in
three distinct subfamilies of Gryllidae (Alexander 1962, T. Walker, pers.
comm.). For the family Culicidae, medical entomology has fostered the
accumulation of sufficient natural history information to permit a contrast
of behaviors which occur in divergent genera.
In the course of studies in Africa and South America (Lounibos 1978,
1980, Lounibos and Machado-Allison 1983), my attention was drawn to
certain behavioral similarities among mosquitoes which had evolved, ap-
parently independently, to occupy water-containing fruit husks. Here I will
compare ovipositional and larval behaviors which have been observed among
two or more unrelated mosquitoes inhabiting fruit husks in the Neotropical,
Afrotropical, and Oriental Regions, and I will suggest that these speciali-
zations have evolved independently in different biogeographic zones in re-
sponse to similar selective forces of the husk microhabitat.


The mosquito genera Trichoprosopon, Eretmapodites, and Armigeres
are confined, respectively, to the Neotropical, Afrotropical, and Oriental
Regions and presently have no geographical overlap (Fig. 1). Most species'
oviposit in plant-held waters such as treeholes, leaf axils, fallen leaves, or
fruits. Members of these three genera are the most abundant, albeit not
the exclusive, mosquito inhabitants of water-containing fruit husks.2 From
his observations in the Old World tropics, Mattingly (1969) suggested that
only husks and shells containing unpolluted fluid are used by mosquitoes
other than Eretmapodites or Armigeres. In husks containing Trichoprosopon,
Eretmapodites, or Armigeres, individuals of these genera far outnumbered

*Phil Lounibos is an Entomologist at the Florida Medical Entomology Laboratory in
Vero Beach. His current research focuses on mosquito ecology and behavior. Address: Uni-
versity of Florida, Florida Medical Entomology Laboratory, 200 9th Street SE, Vero Beach,
FL 32960.
Florida Agricultural Experiment Station Journal Series No. 4369.

March, 1983

Insect Behavioral Ecology-'82 Lounibos

Fig. 1. The geographic distributions of the three mosquito genera special-
ized for the occupancy of water-containing fruit husks in their immature
stages. Numbers of recognized species in each genus and records of countries
of occurrence were derived from Knight and Stone (1977), Knight (1978),
and Zavortink (1979).
the fruit occupants from other mosquito genera (Machado-Allison and
Alvarado unpublished, Raymond et al. 1976, Bick 1951).

To provide a habitat for larval mosquitoes, fruits must be gnawed, par-
tially eaten, cut, or broken and dropped in a position to catch and hold
rain water. In East and West Africa, monkeys and baboons are important
consumers of the fruits with fleshy endocarps which subsequently offer a
habitat for species of the Eretmapodites chrysogaster (s.l.) group (Ray-
mond et al. 1976, Lounibos 1978) (Fig. 2A). Husks harboring Eretmapodites
held 65 ml or less of fluid (Lounibos 1980), and their availability in native
African forests was both seasonal and patchy (Raymond et al. 1976). The
irregular occurrence of husks may explain why fruit "specialists" such as
E. subsimplicipes also oviposit in alternative containers such as fungus
cups, leaves, and snail shells on the forest floor (Lounibos 1980).
Fruits that are cultivated and/or used by man provide important habi-
tats for species of Trichoprosopon, Eretmapodites, and Armigeres. Coconut
shells and cacao husks are the commonest examples (Figs. 2B, C). In a
wet cacao husk, but less often in a coconut shell, the medium contains de-
caying fruit. The culicid fauna of husks may be regulated, in part, by the
contribution of fruit exudate to the medium. Species of Trichoprosopon,
Eretmapodites, and Armigeres appear to be specially adapted to the viscous
ooze videe Lounibos 1978) produced by the breakdown of fleshy endocarp.

34 Florida Entomologist 66(1) March, 1983


-A 4.-j*
"*>*S r ^'i, f^ ^ ''" .

S ," .IMB i


t -4 k D-
Fig. 2. A-C. Discarded fruit husks inhabited by larval mosquitoes. A.
Strychnos spinosa husk which contained Eretmapodites subsimplicipes at
the Kenya coast. B. Theobroma cacao husk which contained Trichoprosopon
digitatum on a plantation in eastern Venezuela. C. Cocos nucifera shell
which contained E. subsimplicipes and Aedes aegypti at the Kenya coast.
D. Female T. digitatum brooding eggs in a cacao husk. Scale bars (approx.) :
A-C = 2.0 cm; D = 5.0 mm.

Fruits cultivated by man on plantations are more regularly spaced
that the patchily distributed fruits of tropical forests. Presumably, the
specialist mosquitoes of cacao plantations, such as E. chrysogaster in West
Africa (Mattingly, cited in Hopkins 1952) or T. digitatum in the Caribbean
(Aitken et al. 1968, Lane 1953), in evolutionary time first inhabited native,
uncultivated fruits prior to exploiting the habitat bonanzas of plantations.
In the Neotropics, Theobroma cacao, the source of commercial chocolate,
grew wild and co-occurred with T. digitatum long before it was sown on
plantations (Van Hall 1932).

Among mosquito genera which ordinarily do not inhabit fruits, chemical
factors which cue oviposition have been identified for species of Culex (e.g.
Ikeshoji 1966), Aedes (e.g. Kalpage and Brust 1973), and Wyeomyia
(Istock et al. 1983). The stimulatory chemicals may be bacterial metabolites
in the oviposition medium (Ikeshoji et al. 1967), may be derived from
mosquito larvae or pupae (Kalpage and Brust 1973), or from the plant
parts that create the microhabitat (Istock et al. 1983). The available in-
formation suggests that mosquitoes that oviposit in small, discrete water
bodies, such as fruit husks, are more likely to be cued by a chemical.

Insect Behavioral Ecology-'82 Lounibos

Members of all three fruit-inhabiting genera respond to oviposition
chemicals. In oviposition choice experiments, T. digitatus preferred fluid
from cacao husks to fluid derived from Alocasia axils, an available habitat
not used by this species (Lounibos and Machado-Allison 1983). On the Kenya
coast, one or more chemicals derived from wild fruits stimulated oviposition
by E. subsimplicipes, but Eretmapodites quinquevittatus, which eschews
fruit husks in nature, preferred pure spring water for oviposition (Lounibos
1978). Although species of Armigeres have not been tested directly for
responsiveness to oviposition stimulants from fruits, in the laboratory A.
kuchingensis showed a strong preference for oviposition in polluted water
(Thomson 1941). To date, there is no evidence whether oviposition stimu-
lants. from fruits are unique to this microhabitat.

Post-ovipositional, maternal care has been recorded in six species of
Armigeres and one species of Trichoprosopon. Strickland (1917) was the
first to observe fertile eggs attached to the hind legs of adult Armigeres
flavus, a Malaysian species whose larvae inhabit coconut shells and bamboo
internodes. He suggested that egg carrying enabled females to deposit their
progeny as newly hatched larvae into inaccessible habitats or where eggs
might be subject to "some danger". Egg carrying by A. flavus was con-
firmed by Barraud (1934) who presumed that females introduced eggs or
incompletely hatched larvae by poking their hind legs through small holes
in bamboo nodes.
Macdonald (1960) indicated four additional Malaysian Armigeres, A.
annulitarsis, A. balteatus, A. inchoatus, and A. magnus, that also carry eggs
on their hind legs. Mattingly (1971b) added A. traubi to the list of
Armigeres known to exhibit this maternal behavior. Because two egg-
carrying Malaysian species occur frequently in open containers, Macdonald
(1960) contested the notion that egg carrying had evolved for exploiting
inaccessible habitats, but he offered no alternative hypothesis to account
for this behavior. Egg batches attached to the hind legs of Armigeres have
been discovered on females attacking man as well as on pinned museum
specimens (Macdonald 1957, Mattingly 1971a).
Females of T. digitatum from Venezuela brood their egg rafts in cacao
husks (Fig. 2D) from oviposition until egg hatch (Lounibos and Machado-
Allison 1983), a period requiring 26-30 hours at 26-29C (Aitken et al.
1968). Although oviposition behavior of this species was described by
Pawan (1922), post-ovipositional egg attendance was not recognized until
Aitken et al. (1968) noted its occurrence in a laboratory colony. Lounibos
and Machado-Allison (1983), after demonstrating that rafts in nature
were almost always guarded by females, conjectured that brooding behavior
evolved to protect eggs from predation or desiccation. Rafts that were not
guarded were observed to break up, and eggs floated to the sides of husks
where they could be stranded and more subject to desiccation or predation.
The eggs of neither T. digitatum nor egg-carrying Armigeres are resistant
to desiccation (Lounibos, unpublished, Macdonald 1960).
It may simply be a coincidence that the only records of post-ovipositional
egg attendance by mosquitoes occur in two fruit-inhabiting genera,
Armigeres and Trichoprosopon. More likely, some property of this micro-

Florida Entomologist 66 (1)

March, 1983

habitat or containers in general has promoted the independent evolution
of egg carrying among Armigeres and brooding by T. digitatum.

Three general modes of feeding behavior: filtering, browsing, and pre-
dation are recognized for larval mosquitoes (Surtees 1959). Larvae of all
three genera of fruit specialists are adapted for both browsing and faculta-
tive predation.
Knight (1971) compared the larval mandible of all but three of the
known mosquito genera. Among these, members of 18 genera have mandi-
bular teeth less prominent than the fruit-husk species (Fig. 3); these
mosquitoes are generally characterized as filter-feeders (Bates 1949). In the
evolutionary transition from filter feeding to predation,3 the mandibular
teeth are known to increase in size and sclerotization (Harbach 1977). The
mandibular teeth of species from ten other genera,4 known as facultative or
obligate predators (Bates 1949), are equally well-developed as in species
of fruit specialists. (Fig. 3).
Armigeres are as well suited for browsing the inner walls of fruits as for
carnivory. Lounibos (1978) suggested that frugivory by larvae permitted


Fig. 3. Mandibles of fourth instar larvae of A. Armigeres subalbatus
from southern Japan, B. Eretmapodites s'ubsimplicipes from eastern Kenya,
C. Trichoprosopon digitatum from eastern Venezuela, D. Culiseta melanura
from Florida, USA. A. and D. are redrawn from Knight (1971) and B.
and C. are drawn from larvae collected from fruit husks by this author.
Certain structural details have been omitted to emphasize the relative sizes
of the mandibular teeth. Abbreviations (after Knight 1971): DT = dorsal
teeth; MdB = mandibular brush; Mp membranous process; SpA = spinose
area; VT = ventral teeth. Scale bar = 0.05 mm.

Insect Behavioral Ecology-'82 Lounibos

the maintenance of the extraordinary densities of E. subsimplicipes found
in African Saba and Strychnos fruits. Venezuelan cacao husks support den-
sities of T. digitatum comparable to E. subsimplicipes in native African
fruits and, like E. subsimplicipes, T. digitatum larvae rasp the fruit endocarp
with their mandibles (Lounibos, unpublished).
Predation has been confirmed among larvae of various species of
Armigeres (Tanaka et al. 1979, Bates 1949), Eretmapodites (Haddow 1946,
Lounibos 1980), and Trichoprosopon (Arnett 1950, Zavortink 1979, Seifert
and Barrera 1981). While laboratory studies have concentrated on preda-
tion upon other species of culicid prey (Haddow 1946, Lounibos 1980), in
nature fruit-husk Armigeres (Bick 1951), Eretmapodites (Lounibos 1978),
and Trichoprosopon (Seifert and Barrera 1981) occur most commonly only
with conspecifics, and most predation may be assumed to be intraspecific
(although Haddow (1946) regarded E. chrysogaster as resistant to attack
by members of its own species). In bracts of Heliconia area, uniform-aged
cohorts of larvae are maintained by larger T. digitatwm preying upon
smaller conspecifics (Seifert and Barrera 1981). The common occurrence
of T. digitatum belonging to a single instar in cacao husks (Machado-Allison
and Alvarado, unpublished) may be the product of stage-specific canni-

The capacity to grow and metamorphose in a viscous, fetid, and an-
aerobic medium is shared by fruit husk inhabitants of all three genera.
In New Guinea, Bick (1951) noted great concentrations of Armigeres breinli
and Armigeres milnensis in minute amounts of "putrid, semi-liquid ma-
terial" in coconut shells. In parts of China, favored breeding sites of
Armigeres obturans (= subalbatus) are tubs of fermenting urine used by
farmers for fertilizer (Feng 1937-38). Even when the fruit medium occu-
pied by E. subsimplicipes evaporates to a mucilaginous sludge, larvae con-
tinue to thrive (Lounibos 1978). In Panama, Galindo et al. (1951) ob-
served the capacity of T. digitatum larvae to endure desiccation and fluid
too polluted for other species.
Facilitating movement through their semi-aquatic medium, members of
all three fruit-husk genera are capable of sinuous crawling uncharacter-
istic of larval mosquitoes. Mattingly (1969) commented on the convergence
of this locomotory behavior among Eretmapodites and Armigeres: "In both
cases the larvae swim with a remarkable vibratory 'shimmying' motion,
possibly adaptive to progress through a viscous medium." Among Armigeres,
specific citations of crawling by larvae include A. theobaldi on Curcuma
flowers (Chari 1940), and A. breinli whose larvae in coconut husks have
"an eel-like movement whereby they are able rapidly to burrow into the
semi-liquid filth on which they appear tofeed" (Paine and Edwards 1929).
Hopkins (1952) documented crawling of larvae of a Ugandan species of
Eretmapodites and West African E. quinquevittatus. T. digitatum larvae
are capable of crawling in cacao husks (Lounibos, unpublished). Eretma-
podites larvae also tend to remain appressed to the bottom of a container
even after all fluid has been poured out (Hopkins 1952). I have observed
the tenacious adherence of larvae of E. subsimplicipes inhabiting fruit
husks in Kenya as well as T. digitatum in Venezuelan cacao husks.

Florida Entomologist 66(1)

The 'shimmying' swimming motion used by larvae was stated by Mat-
tingly (1969) to extend to Eretmapodites and Armigeres pupae. Further,
pupae of fruit-husk genera are unusually sedentary, remaining submerged
on their sides for long periods on the bottom of a container; this has been
observed both among Ugandan Eretmapodites (Haddow 1946) and T.
digitatum from Venezuela (Lounibos, unpublished).

Fruit-husk inhabiting Trichoprosopon, Eretmapodites, and Armigeres
are presumed to represent "grades", sets of species that share a level of
evolutionary organization attained repeatedly by diverged lines (Futuyma
1979). The evidence for independent evolution of adaptations to fruit husks
is circumstantial yet persuasive. Unlike cosmopolitan genera such as Aedes,
Culex, or Anopheles, there is no present-day geographical overlap among
Trichoprosopon, Eretmapodites, or Armigeres. The tribe Sabethini, which
includes Trichoprosopon, is almost exclusively neotropical (Knight and
Stone 1977). Although Eretmapodites and Armigeres are both placed in the
tribe Aedini, the culicid fauna of the Afrotropical and Oriental Regions
are very distinct with few species occurring in both regions (Mattingly
1962), supporting the hypothesis of independent evolution of these faunas.
Independent, parallel evolution of various traits within one insect family
is well documented for the Drosophilidae where chromosomal arrangements
have allowed independent reconstruction of phylogenies (e.g., Throckmorton
An objective of studying behavioral convergences is resolution of the
selective pressures that promoted parallel and independent evolution. Be-
haviors described in this paper are not necessarily direct adaptations to
fruit husks, but may have existed as pre-adaptations that allowed species
to exploit the habitat. Regardless, two properties of the husk microhabitat
were probably important in either shaping common courses of adaptation
or influencing pre-adapted colonists. Firstly, fruit husks, like many small
vessels occupied by mosquitoes, provide a transient and patchy habitat,
and these properties of containers may have contributed to the evolution
of egg attendance to protect against desiccation, and responsiveness to
chemical stimuli to locate the irregularly distributed microhabitat. Secondly,
and more specific to fruit husks, is the odoriferous and viscous medium
which probably fostered the invasion by pre-adapted mosquitoes or the
evolution per se of responsiveness to ovipositional chemicals, larval frugi-
vory, crawling larvae, and recumbent pupae.
Some groundwork in mosquito behavior and evolution has been laid
because of the relevance of these insects to human health. Hopefully the
comparisons and conjectures of this paper made possible by medical ento-
mology will provide ideas for and be tested by future research.

Three genera of mosquitoes, Trichoprosopon, Eretmapodites, and Armi-
geres, contain species which are specialized for colonizing water-containing
fruit husks. Behavioral specializations include: responsiveness to a chemical
oviposition stimulant in fruits; post-ovipositional maternal care of eggs;

March, 1983

Insect Behavioral Ecology-'82 Lounibos 39

larval frugivory; facultative cannibalism; crawling larvae and recumbent
pupae in their fetid, viscous, and anaerobic media. Trichoprosopon, Eretma-
podites, and Armigeres are confined, respectively, to the Neotropical, Afro-
tropical, and Oriental Regions where, it is argued, behaviors common to
fruit husk inhabitancy evolved independently.


Comments from and discussions with J. Day, H. Frank, J. Kitzmiller, J.
Lloyd, C. Machado-Allison, J. Rey, J. Sivinski, and T. Walker helped clarify
topics and their presentation- in this paper. I am also grateful to C. Baker
for preparing Fig. 3.


1The sabethine genus Trichoprosopon currently is composed of 21 species
(Zavortink 1979). Mention of the genus in this paper refers solely to
T. digitatum, the best-known species (Zavortink 1981). Eretmapodites, the
only culicid genus restricted to the Afrotropical Region (Edwards 1941),
now embraces 38 species (Knight and Stone 1977, Knight 1978). Most rec-
ords of fruit inhabitancy refer to members of the E. chrysogaster complex
of 12 species, which includes E. subsimplicipes (Gillett 1972). At last count,
46 species were included in Armigeres (Knight and Stone 1977). Species
from both subgenera, Armigeres and Leicesteria, have been found in fruits.
2In Africa, species of Aedes, Culex, and Toxorhynchites also have been
recorded as occupying fruits of the Loganiaceae and Apocynaceae (Ray-
mond et al. 1976, Lounibos 1980), and on islands in the Pacific, Tripteroides
and Aedes species may be found in coconut shells and rat-gnawed cacao
husks (Baisas and Ubaldo-Bagayon 1952, Laird 1956). In eastern Venezuela
one species of Culex occasionally occurs in the cacao husks ordinarily oc-
cupied by T. digitatum (Machado-Allison and Alvarado, unpublished).
3Filter (plankton) feeding is not necessarily the primitive mode among
larval Culicidae. Harbach (1977) considers it equally likely that browsers
(scavengers) are primitive and gave rise to both filter feeders and pred-
4The genera whose larval mandibles as depicted by Knight (1971) are
comparable in size and form to fruit-husk mosquitoes: Aedes (Mucidus),
Culex (Lutzia), Heizmannia, Limatus, Malaya, Phonomyia, Sabethes,
Topomyia, Tripteroides, and Zeugnomyia.

oratory colonization of Trichoprosopon digitatum (Rondani) (Dip-
tera: Culicidae). Mosquito News 28: 445-54.
ALCOCK, J. 1979. Animal behavior: an evolutionary approach. Sinauer,
Sunderland, Mass.
ALEXANDER, R. D. 1962. Evolutionary change in cricket acoustical com-
munication. Evolution 16: 443-67.
ARNETT, R. H. 1950. Notes on the distribution, habits, and habitats of
some Panama culicines (Diptera: Culicidae). J. N. Y. Ent. Soc. 58:
BAISAS, F. E. AND A. UBALDO-BAGAYON. 1952. Notes on Philippine mos-
quitoes, XVI. Genus Tripteroides. Manila Bureau of Printing.
BARRAUD, P. J. 1934. Fauna of British India, Diptera, Vol. V: Family

40 Florida Entomologist 66(1) March, 1983

Culicidae, Tribes Megarhinini and Culicini. London: Taylor and
BATES, M. 1949. The natural history of mosquitoes. MacMillan, N. Y.
BICK, G. H. 1951. The ecology of the mosquito larvae of New Guinea. Pac.
Sci. 5: 392-431.
CHARI, M. O. T. 1940. Armigeres theobaldi breeding in Curcuma flowers.
J. Malar. Inst., India 3: 263-64.
EDWARDS, F. W. 1941. Mosquitoes of the Ethiopian region. III. Culicine
adults and pupae. British Museum, London.
FENG, L.-C. 1937-38. A critical review of the literature regarding the
records of mosquitoes in China. II. Subfamily Culicinae, tribes
Megarhinini and Culicini. Pek. Nat. Hist. Bull. 12: 285-318.
FUTUYMA, D. J. 1979. Evolutionary biology. Sinauer, Sunderland, Mass.
GALINDO, P., S. J. CARPENTER, AND H. TRAPIDO. 1951. Ecological observa-
tions on forest mosquitoes of an endemic yellow fever area in Pan-
ama. Amer. J. Trop. Med. 31: 98-137.
GILLETT, J. D. 1972. Common African mosquitoes and their medical im-
portance. Heinemann Ltd., London.
HADDOW, A. J. 1946. The mosquitoes of Bwamba County, Uganda. IV.
Studies on the genus Eretmapodites, Theobald. Bull. Ent. Res. 37:
HARBACH, R. E. 1977. Comparative and functional morphology of the
mandibles of some fourth stage mosquito larvae. Zoomorphologie 87:
HOPKINS, G. H. E. 1952. Mosquitoes of the Ethiopian Region I.-Larval
bionomics of mosquitoes and taxonomy of culicine larvae. British
Museum (NH), London.
IKESHOJI, T. 1966. Studies on mosquito attractants and stimulants. III.
SThe presence in mosquito breeding waters of a factor which stimu-
lates oviposition. Jap. J. Exp. Med. 36: 67-72.
IKESHOJI, T., T. UMINO, AND S. HIRAKOSO. 1967. Studies on mosquito at-
tractants and stimulants. Part IV. An agent producing stimulative
effects for oviposition of Culex pipiens fatigans in field water and the
stimulative effects of various chemicals. Jap. J. Exp. Med. 37: 61-9.
ISTOCK, C. A., K. TANNER, AND H. ZIMMER. 1983. Habitat selection by the
pitcher-plant mosquito, Wyeomyia smithii: behavioral and genetic
aspects. IN J. H. Frank and L. P. Lounibos (eds.) Phytotelmata:
terrestrial plants as hosts of aquatic insect communities. Plexus,
Medford, N. J.
KALPAGE, K. S. P. AND R. A. BRUST. 1973. Oviposition attractant produced
by immature Aedes atropalpus. Environ. Ent. 2: 729-30.
KNIGHT, K. L. 1971. Comparative anatomy of the mandible of the fourth
instar mosquito larva (Diptera: Culicidae). J. Med. Ent. 8: 189-205.
KNIGHT, K. L. 1978. Supplement to a catalog of the mosquitoes of the
world (Diptera: Culicidae). Ent. Soc. Amer., College Park, Mary-
KNIGHT, K. L. AND A. STONE. 1977. A catalog of the mosquitoes of the
world (Diptera: Culicidae). Ent. Soc. Amer., College Park, Mary-
LAIRD, M. 1956. Studies of mosquitoes and freshwater ecology in the South
Pacific. Bull. R. Soc. N. Z. 6: 1-213.
LANE, J. 1953. Neotropical Culicidae. Univ. of Sao Paulo, Brazil.
LOUNIBOS, L. P. 1978. Mosquito breeding and oviposition stimulant in fruit
husks. Ecol. Ent. 3: 299-304.
LOUNIBOS, L. P. 1980. The bionomics of three sympatric Eretmapodites
(Diptera: Culicidae) at the Kenya coast. Bull. Ent. Res. 70: 309-320.
LOUNIBOS, L. P. AND C. E. MACHADO-ALLISON. 1983. Oviposition and egg

Insect Behavioral Ecology-'82 Lounibos

brooding by the mosquito, Trichoprosopon digitatum, in cacao husks.
Ecol. Ent. 8 (In press.)
MACARTHUR, R. H. 1972. Geographical ecology. Harper and Row, N.Y.
MACDONALD, W. W. 1957. Malaysian Parasites. XVI. An interim review of
the non-anopheline mosquitoes of Malaya. Stud. Inst. Med. Res.
Malaya. 28: 1-34.
MACDONALD, W. W. 1960. Malaysian parasites. XXXVIII. On the syste-
matics and ecology of Armigeres subgenus Leicesteria (Diptera,
Culicidae). Stud. Inst. Med. Res. Malaya 29: 110-53.
MATTINGLY, P. F. 1962. Towards a zoogeography of the mosquitoes. Syst.
Assoc. Publ. 4: 17-36.
MATTINGLY, P. F. 1969. The biology of mosquito-borne disease. American
Elsevier, N. Y.
MATTINGLY, P. F. 1971a. Mosquito eggs. XIII. Genus Armigeres Theobald
Mosq. Syst. News. 3: 122-29.
MATTINGLY, P. F. 1971b. Mosquito eggs. XIV. Genus Armigeres Theobald
(continued) and Aedes subgenus Alanstonea Mattingly. Mosq. Syst.
News. 3: 130-37.
MAYR, E. 1963. Animal species and evolution. Belknap Press, Harvard
Univ., Cambridge, Mass.
PAINE, R. W. AND F. W. EDWARDS. 1929. Mosquitos from the Solomon
Islands. Bull. Ent. Res. 20: 303-16.
PAWAN, J. L. 1922. On the eggs and oviposition of Psorophora (Janthino-
sona) posticata, Wied. (Culicidae). Bull. Ent. Res. 12: 481-2.
RAYMOND, H. L., M. CORNET, AND P. Y. DIENG. 1976. Etudes sur les
vecteurs sylvatiques du virus amaril. Inventaire provisoire des habi-
tats larvaires d'une foret-galerie dans le foyer enddmique du Senegal
oriental. Cah. ORSTOM series Ent. Med. Parasitol. 14: 301-6.
SEIFERT, R. P. AND BARRERA, R. 1981. Cohort studies on mosquito (Diptera:
Culicidae) larvae living in the water-filled floral bracts of Heliconia
aurea (Zingiberales: Musaceae). Ecol. Ent. 6: 191-7.
STRICKLAND, C. 1917. A curious adaptation of habit to its environment of
a Malayan mosquito. J. Straits Br. Asiatic Soc. 75: 39.
SURTEES, G. 1959. Functional and morphological adaptations of the larval
mouthparts in the subfamily Culicinae (Diptera) with a review of
some related studies by Montchadsky. Proc. R. Ent. Soc. London (A)
34: 7-16.
TANAKA, K., K. MIZUSAWA, AND E. S. SAUGSTAD. 1979. A revision of the
adult and larval mosquitoes of Japan (including the Ryukyu Archi-
pelago and the Ogasawara Islands) and Korea (Diptera: Culicidae).
Contrib. Ameri. Ent. Inst. 16: 1-987.
THOMSON, R. C. M. 1941. Studies on the behaviour of Anopheles minimus,
Part IV. The composition of the water and the influence of organic
pollution and silt. J. Malar. Inst., India 4: 63-102.
THROCKMORTON, L. H. 1975. The phylogeny, ecology, and geography of
Drosophila. IN R. King (ed.) Handbook of genetics, Vol. 3: Inverte-
brates of genetic interest. Plenum, N. Y.
VAN HALL, C. J. J. 1932. Cacao. MacMillan, London.
ZAVORTINK, T. J. 1979. Mosquito studies. XXXV. The new sabethine genus
Johnbelkinia and a preliminary reclassification of the composite genus
Trichoprosopon. Contrib. American Ent. Inst. 17: 1-61.
ZAVORTINK, T. J. 1981. Species complexes in the genus Trichoprosopon.
Mosq. Syst. 13: 82-5.

Florida Entomologist 66 (1)



"Is it a boy or a girl?" (Todo el mundo)

Most organisms have two sexes; thus it is reasonable to ask what is
the ratio of the sexes, and does this ratio fluctuate in a predictable manner.
Basic researchers have long recognized that the sex ratio is a phenotypic
trait of an organism, and therefore may be subject to natural selection.
In addition, it has been widely noted that (i) the relative fitness of a
male versus a female ranges widely according to a variety of ecological
factors (e.g., size of individual, local sex ratio, inbreeding), implying that
the selection of sex-ratio traits should be strong; and (ii) relative to
other traits studied by behavioral ecologists, sex ratios can often be meas-
ured with a high degree of precision. Hence, the subject of sex ratios is
ideal for testing the efficacy of natural selection, and for testing our under-
standing of the selective process. Applied researchers also have a long
history of interest in sex ratios. Efficiency of live-stock breeding would
be greatly enhanced by control over the sex ratio; mass-rearing of para-
sites as biological control agents is often improved by a female-biased sex
ratio; and causing a male-biased sex ratio in a population of pests would
aid in reducing the amount of damage.
This paper is an introduction to a way of viewing sex-ratio puzzles
that has had some success-both theoretically and empirically-and points
out some of the difficulties of sex-ratio problems. After a methodology has
been developed, the following questions will be addressed: What sex-ratio
patterns are observed in nature? What are the causes and what are the
ecological correlates of the patterns? What sort of logic is required to ex-
plain each pattern? What general themes underlie the patterns? Some
assumptions and extensions of the theory are discussed in the appendix.1


Before introducing the categories of sex-ratio patterns, we must have
(i) definitions to describe the possible relationships between genotypes and
their expressed phenotypes, (ii) a common understanding of proximate and
ultimate causes, (iii) an appreciation of the interactions among the bio-
logical levels of life (e.g., genes, individuals, populations), and (iv) a
language to describe how organisms allocate their energy and resources
with respect to sex ratios (the language of economics). These four topics
are discussed in turn.

The relationship between an individual's genotype and the sex ratio it
produces (phenotype) is an important descriptor of sex-ratio patterns.
Three possible relationships exist. (i) The sex ratio produced by an indi-
vidual is genetically determined; i.e., a particular genotype expresses a

S. A. Frank is a graduate student in the Department of Zoology, University of Florida.
His research focuses on the natural history and sex ratios of fig wasps. Current Address:
Department of Zoology, University of Florida, Gainesville, FL 32611.

1March, 1983

Insect Behavioral Ecology-'82 Frank

fixed sex ratio and the sex ratio differs (except for random variation) for
different genotypes. (ii) The sex ratio produced by an individual is inde-
pendent of its genotype, and becomes fixed before the individual becomes
an adult. The sex ratio may be determined by the environment during
pre-adult development, or it may be independent of the environment and
therefore constant for all individuals in the population. A fixed response
independent of the genotype is referred to as a phenotypically canalized re-
sponse. (iii) The sex ratio produced by an individual is independent of
its genotype, and may be variable during the adult stage. This is the situa-
tion in which individuals respond to ecological correlates of fitness (see
next section), and is referred to as a phenotypically plastic response.

Since organisms are, for the most part, products of evolution by natural
selection, most biologists agree that the ultimate cause of many behavioral
and physiological attributes of an organism is the maximization of fitness.
For example, the ultimate cause (or adaptive significance) of food acquisi-
tion is obtaining energy in order to increase the number of progeny suc-
cessfully reared. We also know from our everyday experience that the be-
havior and physiology of an organism is often a response to its environ-
ment. Ecological cues (such as daylength, population density) that trigger
particular genetic, physiological, developmental, or psychological mechanisms
are referred to as proximate cues, and the mechanisms are referred to as
proximate causes. These mechanisms, and the response to particular cues,
usually evolve by the process of natural selection, and therefore often have
an adaptive significance. Thus, in an evolutionary sense, the particular
cues that are used by an organism to trigger behavioral and physiological
responses ought to be those that have historically been correlates of success-
ful (in terms of fitness) responses. Therefore, I often refer to proximate
cues as ecological correlates of fitness. This usage stresses the key postu-
late-the proximate cues that organisms respond to only have meaning
(in an evolutionary sense) because they are correlates of fitness (but see
Williams 1966 and Gould and Vrba 1982 for relevant caveats).2
The specific proximate cues which are correlates of the sex ratio (and
fitness, since sex ratio affects fitness) in different species follow no dis-
cernible pattern. For example, different species of mass-reared insects yield
different sex ratios under similar ecological conditions (e.g., Hoelscher and
Vinson 1971, Bouletreau 1976).3 In any species the particular proximate
cues (e.g., host size, daylength) used are likely to be those that have his-
torically correlated with a successful (in terms of fitness) sex-ratio pattern,
and which the organism can assess. In other words, the correlates must con-
tain information about which sex-ratio response confers the greatest fit-
ness and the individual organism must be able to assess and respond to this
information. Since any particular phylogeny is the result both of chance
events and of the process of natural selection, just which proximate cues
an organism uses are not predictable with certainty. Phenotypic plasticity
of the sex ratio often does not exist (e.g., Williams 1979). This may be
because (i) no ecological correlates of fitness with respect to sex ratios
exist, or (ii) the correlates exist, but the organism has not evolved a
response to these correlates.4

44 Florida Entomologist 66(1) March, 1983


A variety of sex-ratio patterns occurs in nature (see Charnov 1982
for a review). Presenting these patterns one at a time leaves the impres-
sion that they are only vaguely related and due to a potpourri of causes
and effects. In order to capture all of the patterns within a single cast, we
must weave a broader net. The approach I will use is a hierarchical one
with respect to genetic organization. Collections of genes make up chromo-
somes, a particular set of chromosomes forms an individual genome, an
isolated group of individuals forms a subpopulation, and the collection
of all subpopulations is the population gene pool. The similarities and dif-
ferences among several sex-ratio patterns can be understood within such
a hierarchical framework. For example, a 1:1 population sex ratio may
be observed, while the sex ratio of the progeny of particular individuals
may differ from 1:1 in a predictable manner (e.g., Trivers and Willard
1973). At what levels are the causes of these individual- and population-
level patterns to be found? The patterns will be classified according to the
levels) at which natural selection is analyzed, and the level at which
the sex-ratio effect is observed.
A common view, which will be extended here, is that the effects of na-
tural selection can be observed as changes in gene frequencies within the
population (e.g., Crow and Kimura 1970). Alleles are the lowest unit,
or atom, of the hierarchy that we are interested in, and the population
the highest level. The fitness of an allele within the population describes
whether the frequency of that allele is increasing, decreasing, or remain-
ing stable. A means of quantifying the fitness of an allele within the pop-
ulation is the number of replicates an allele produces divided by the
average number of replicates produced by all alleles within the population,
over one or more generations. The number of replicates produced can be
called reproductive success (=RS), thus fitness of an allele within the
population is

fitness (allele/pop) =RS (allele/pop) /E[RS (allele/pop)].
This phrase is read as 'the fitness of a particular allele within the popula-
tion is equal to the RS of that allele within the population divided by
the expected (or average) RS of all alleles within the population gene
pool.' This notation is used throughout the paper.5
The key to the hierarchical approach is realizing that the fitness of an
allele within the population can be partitioned into components. For ex-
ample, analyzing the success of an allele within the population can be
carried out by first analyzing the success of that allele within a subpopula-
tion, and then analyzing the success of the subpopulation within the entire
population (Price 1970, 1972, Hamilton 1975; see fig. 1). In general, the
fitness of a unit at a given level (e.g., chromosome) within a higher level
(e.g., individual) is the number of alleles the unit contributes (RS) within
a higher level divided by the average number of alleles contributed by
all similar units [E(RS)] within the higher level. Fitness is measured over
some time period, usually one or more generations. The interpretation of
fitness values is simple. A fitness (allele/subpop) less than one occurs when
a particular allele is decreasing in frequency within that subpopulation. A

Insect Behavioral Ecology-'82 Frank



fr 4/,6 fr-3/6
FITNESS (SUBPOP/POP)= 0.25= 15 46


frP SE


POP TOTAL: fr 28/80

fr= 39/B

FITNESS (ALLELE/POP) 39=28 = 1.39
Fig. 1. A simplified example analyzing the components fitness(allele/
subpop) and fitness(subpop/pop) for a population made up of two subpopu-
lations (I and II). fr is the frequency of the allele, and N is total population
size. The frequency of the allele decreases in both subpopulations, however
for this example it is assumed that the subpopulation decreases in size when
the frequency is less than 0.5, and the subpopulation increases in size when
the frequency is greater than 0.5. Therefore, although the frequency of
the allele decreases in both subpopulations, in this example the allele in-
creases in frequency within the population. Although this example is arti-
ficial, the analysis of fitness components is a useful tool for understanding
certain types of sex-ratio patterns discussed in this paper.
fitness greater than one means an increase in frequency over time, and a
fitness equal to one occurs when the frequency is remaining stable. Further
properties of the hierarchy will be developed while discussing particular
patterns. The utility of the approach will be apparent when the patterns
are viewed as a whole.

Over long periods of time natural selection favors behavioral and physio-


Florida Entomologist 66 (1)

logical patterns of individual organisms (and entities at other levels of
the hierarchy) that are consistent with the maximization of the indi-
vidual's (or entity's) fitness within the population. The sort of behavior
or physiology one is often interested in concerns the allocation of resources.
For example, producing a particular ratio of sons and daughters, or a
particular ratio of "male" and "female" flowers, is an allocation of
resources into two alternatives-males and females. A useful language
for discussing the allocation (=investment) of resources already exists-
the language of economics. A particular investment pattern is said to have
costs and benefits in terms of fitness. When resources are limiting, natural
selection favors investment patterns that maximize the benefit/cost ratio
(see Staddon 1980 for a review of this approach).

Observed sex ratios often form patterns at several biological levels
(e.g., individuals, populations), and these patterns are the result of both
selection acting at various genetic levels and the relationship between geno-
types and phenotypes. Each pattern is named according to an interesting
phenomenon underlying the pattern. Table 1 lists the patterns in the order
they will be presented, and shows how the patterns are classified according
to the level (s) at which fitness components are analyzed, the level of the ob-
served sex-ratio effect, and the genotype-phenotype relationship. These five
patterns are chosen to illustrate the recurrent themes underlying most sex-
ratio patterns. These themes are discussed in the conclusion. Sex ratios
are given as 'males :females' or 'males/total'.3

The first question is why are the population sex ratios of many species
about 1:1. To analyze this question the expected fitness (ind/pop) through
a single son is compared to expected fitness(ind/pop) through a single
daughter as the population sex ratio changes.6,7 Fitness of an individual
in the population is RS/E(RS). I will analyze fitness over 2 generations,
where RS is the number of grandprogeny through a single male or female
offspring divided by 4 (in diploid organisms each parent contributes 1/2
of the genome to an offspring, and thus a grandprogeny represents 1/4 of
each grandparent). Consider the situation depicted in table 2. The expected
fitness of an individual through a daughter added to this population is 2/3,
through a son 4/3. In general, when males are rarer than females, expected
fitness to an individual through a son is greater than through a daughter,
and thus selection at the level of the individual within the population favors
the production of males. Similarly, when females are rarer than males,
expected fitness is greater through daughters, and the production of fe-
males is favored. Thus, the population sex ratio tends to stabilize at 1:1.
However, this conclusion is only approximately correct, and a closer look
reveals a more subtle pattern.
The situation where either one male or one female offspring was added
to the population was just considered. However, producing a male or a
female may require different amounts of resources. Perhaps females are
provisioned with more food than males, as in many solitary bees and wasps

March, 1983

Insect Behavioral Ecology-'82 Frank

0 0

4 )


*a T

41 E



^- ai '
N *r

a) a*





(U *^<



2 'rb





4 .5
^ ai fQ

_* a S 3 1
lll jR


4- 4. 2

h p4

,-' Q4

0 0

0 0



u a)

-' >







Florida Entomologist 66 (1)


A DAUGHTER. Sex ratio of the generation being added to is 1
male per 2 females, the population size is assumed to be large,
mating random, and the average clutch per female is assumed
to be 10. Thus, a diploidd) parent expects 10 grandchildren
through a daughter for a RS of 10/4 = 2.5. Since the sex ratio
is 1:2, each male, on average, sires the progeny of 2 females,
and thus is expected to sire 20 progeny. Expected RS through
a son is therefore 20/4 = 5. Expected fitness is average RS
by a certain strategy (male v. female) divided by E(RS) over
all strategies, (5+2.5)/2 = 3.75. Thus, under these conditions,
expected fitness through a son is greater than through a

Population sex ratio of generation being added to: 1:2 (males:females)
Average clutch size/female = 10
expected RS of parent
per individual fitness =
offspring by sex RS/E(RS)

1 male 20/4 = 5 5/3.75 = 4/3
if parent produces
1 female 10/4 = 2.5 2.5/3.75 = 2/3

(Krombein 1967). In this case different amounts of resources are invested
in male and female progeny.8 As an example suppose that the cost in terms
of resources to produce one female is equal to the cost of producing 2
males (see table 3). The situation can be analyzed as before, but asking
instead what are the expected witnesses through sons and daughters per
unit investment (rather than per offspring, as above). In this case, what
is the fitness through 1 daughter versus 2 sons? The example in table 3
shows expected witnesses when the population sex ratio is 1:1, males being


of producing 2 males and the cost of producing 1 female each
represents one unit of investment. Since for a single unit of
investment a parent may produce 2 males or 1 female, the
fitness through 2 males must be compared to the fitness through
1 female. The remainder of the table is obtained as in table 2.
With the population sex ratio 1:1, production of the cheaper
sex, males, is favored.

Population sex ratio of generation being added to: 1:1
Cost of 2 males = 1 unit of investment = cost of 1 female
Average clutch size/female=10
expected RS of parent fitness=
per unit investment RS/E(RS)

2 males 20/4=5 5/3.75=4/3
parent produces
1 female 10/4=2.5 2.5/3.75=2/3

March, 1983

Insect Behavioral Ecology-'82 Frank 49

favored over females. In general, the sex that is cheaper to produce will
be more abundant than the more costly sex [this is Fisher's (1958) theory,
reviewed by Charnov (1982)].
At what point does the sex ratio stabilize in this situation? Or, at what
population sex ratio are the expected witnesses through the 2 alternatives,
1 daughter or 2 sons, equal? Table 4 demonstrates that the witnesses are
equal when the population sex ratio equals the ratio of the number of
males to females per unit investment. This is generally true, and since
producing 1 female represents 1 unit of investment, and producing 2
males also represents 1 unit of investment, the population sex ratio sta-
bilizes such that the investment ratio in male and female progeny is 1:1-
i.e., where the sex ratio is 2:1 (fig 2).
Equal investment is a population-level attribute, therefore the problem
of genotype-phenotype interaction is usually irrelevant. Only when
individuals respond to a fluctuating population sex ratio is phenotypic
plasticity important. An example of such a response is given in pattern
(2) below.

An Example
Data consistent with the equal investment hypothesis have been collected
for a few species of Hymenoptera (see Charnov 1982 for other examples).
Solitary bees and wasps often seem to have sex ratios skewed toward the
sex that is cheaper to produce (Trivers and Hare 1976, but see Alexander
and Sherman 1977 for problems in interpretation). Two studies have been
conducted that were designed to test the equal investment hypothesis on
natural populations of the social wasp Polistes. Noonan (1978) studied P.
fuscatus, and Metcalf (1980) examined P. metricus and P. variatus (table
5). These data represent several nests from each population. Wet weight
of individuals was measured at eclosion, and was used to quantify invest-

3, EXCEPT THE POPULATION SEX RATIO IS 2:1. Expected number of
grandchildren through one daughter is the average clutch, or
10. Since the population sex ratio is 2:1, each male, on average,
sires one-half of a female's total progeny. Thus expected number
of grandchildren through each male is 5, and through 2 males
10. Fitnesses per unit investment for the 2 alternatives, 2 males
versus 1 female, are equal. Therefore the sex ratio stabilizes at
2 males per female, which is the ratio of the number of males
to females produced per unit investment (see fig 2).

Population sex ratio of generation being added to: 2:1
Cost of 2 males = 1 unit of investment= cost of 1 female
Average clutch size/female=10
expected RS of parent fitness=
per unit investment RS/E(RS

2 males 10/4=2.5 2.5/2.5=1
parent produces
1 female 10/4=2.5 2.5/2.5=1

Florida Entomologist 66 (1)


0.25 0.5

0.75 1.0

INVESTMENT RATIO (male:female)
Fig. 2. Graphic representation of equal investment in the sexes theory.
The solid curves represent the expected fitness of a single male and a
single female added to a population as a function of the overall population
investment ratio, where the population investment ratio is the amount of
resources allocated to male progeny (or male function in hermaphrodites)
to the amount allocated to female progeny (or function). For example,
when the overall investment ratio is 0.25, males have a higher expected
fitness than females, thus the investment ratio will increase. The expected
witnesses of males and females are equal only when the population invest-
ment ratio is 0.5. General proofs of this result can be found in Charnov
(1982) and papers cited by him. The stable sex ratio will equal the number
of males produced per unit investment to the number of females produced
per unit investment. An intuitive argument can be constructed as follows.
Let the number of males produced per unit investment be M, and the num-
ber of females per unit investment be F. Define the proportion of males
(sex ratio) as x, the proportion of females as 1-x, and the population size
as N. Since the sex ratio is stable when the total investment in males equals
the total investment in females, Nx*/M = N(1-x*)/F, where x* is the
stable sex ratio. Hence, x* = M/(M+F) = M:F. So, when 3M = F, x* =
0.25 (represented by the dashed curves).

ment in males and females (Noonan 1978). In P. fuscatus the observed
sex ratio is male biased and the investment ratio per individual is female
biased, yielding a population investment ratio very close to 1:1. In P.
metricus the opposite is observed, a female-biased sex ratio and a male-
biased investment ratio per individual, also yielding a population invest-

March, 1983

Insect Behavioral Ecology-'82 Frank

SPECIES OF Polistes. P. fuscatus DATA FROM NOONAN (1978), P.
metricus AND P. variatus FROM METCALF (1980). ALL RATIOS

P. fuscatus P. metricus P. variatus

Population sex ratio 1.07:1.00 0.82:1.00 0.94:1.00
at eclosion
Sample size 819 17,701 3630
Wet-weight ratio at
eclosion 0.94:1.00 1.23:1.00 1.01:1.00
Population investment ratio
(weight ratio) (sex ratio) 1.01:1.00 1.01:1.00 0.95:1.00

ment very nearly 1:1. See Noonan (1978) and Metcalf (1980) for statis-
tical analyses and discussion of these data.

Equal investment is a population-level pattern; however, empirical
studies suggest that the investment ratio of individuals sometimes differs
from 1:1 (see Charnov 1982). Why might such individual variation exist?
If expected fitness (ind/pop) per unit investment through one sex is greater
than through the other sex in a particular ecological setting, then the in-
vestment ratio in that setting is predicted to be biased toward the more
successful sex (Trivers and Willard 1973, Charnov et. al. 1981). The hypoth-
esis can be stated as follows: the ecological correlates of relative fitness
of the sexes per unit investment are also correlates of individual-level in-
vestment ratios in the sexes (fig. 3). In other words, individuals can adjust
their investment ratio according to local conditions in a phenotypically
canalized or plastic manner. Notice that pattern (1) can be a special case of
this hypothesis, where the population sex ratio is an ecological correlate
of relative fitness of the sexes (Werren and Charnov 1978). As an example,
in the mite Marcrocheles offspring sex ratio varies inversely with adult
population sex ratio (Filipponi et. al. 1972, Filipponi and Petrelli 1975).

Example 1

Lariophagus distinguendus (Pteromalidae) is a hymenopteran parasite
that lays one egg per host (Charnov et. al. 1981). In Hymenoptera unfer-
tilized eggs give rise to males and fertilized eggs to females (haplodiploidy).
Females are known to have some control over the sex of each egg laid, so
the necessary mechanism for variation of the sex ratio exists at the level
of individual phenotypes (i.e., plasticity) (Flanders 1956). Host sizes that
a parasite encounters in nature often vary, and a positive correlation be-
tween host size and offspring size has been demonstrated for Lariophagus
(Charnov et. al. 1981). The sex ratio is male biased on small hosts and
female biased on large hosts (fig. 4, see also Clausen 1939). If it were
known that small males have a greater fitness than small females, and
large females have a greater fitness than large males (e.g., that size affects

52 Florida Entomologist 66(1) March, 1983








Fig. 3. Hypothetical relationship between expected fitness(ind/pop) for
the alternatives male and female as a function of offspring size. In this situ-
ation, if an individual is small it has a higher expected fitness as a male,
and if large a higher expected fitness as a female. Offspring size often has
correlates; in example 1 (see text) host size is a positive correlate of off-
spring size; in example 2 parasite density is a negative correlate of offspring

number of eggs laid by females more drastically than number of successful
matings by males), then the hypothesis that correlates of relative fitness
of the sexes (e.g., host size) are also correlates of the sex ratio would be
supported by this study. Empirical evidence of relative fitness shifts of
the sexes with size has not been obtained for Lariophagus (measuring fit-
ness is difficult), thus direct support of the hypothesis cannot be claimed.
However, likely correlates of individual fitness (fecundity for females,
age at death for males) as related to size were measured for Lariophagus
and Heterospilis prosopoidis (Braconidae), and the data suggest that
changes in fitness with size are greater for females than for males (Charnov
et. al. 1981). This is consistent with the observed sex-ratio response.
What ecological cues might Lariophagus use to determine whether a
host is large or small? A logical hypothesis is that the insect assesses the
size of the host (or a correlate of size, such as tunnel diameter of a burrow-
ing host) relative to other hosts encountered, and adjusts the sex ratio of
its offspring according to the above prediction. Charnov et. al. (1981) con-
ducted an experiment with Lariophagus to test this hypothesis (fig. 4). A
particular host size was presented alone, with smaller hosts, and with larger
hosts. The sex-ratio shifts observed support the hypothesis that relative
host size is assessed by these wasps (see fig. 4 for details).

Insect Behavioral Ecology-'82 Frank 53

1.0" Alo

(9 \
Z 14/ 32

LU .8 66 \

V)65 23
-] .6
S2120 \
Z .38 ". \29
a. b. c.

o \ e \17 23

S. 9 ae62
Q -\ 39

14*- ,24 45 -*43 \ / 3- 27

6 .8 1.0 1,2 1.4 16 1.8 2.0

Fig. 4. Correlation of relative host size and sex ratio in Lariophagus.
Curve b represents the sex ratio produced by female wasps presented se-
quentially with 20 hosts of a single size. Curves a and c represent the sex
ratio where the wasps were presented with an alternating sequence of two
host sizes, the hosts differing by 0.4 mm. Curve c represents the sex ratio
when the host was the smaller of two alternating sizes, and a where the
host was the larger of two alternating sizes. For example, hosts of 1.4 mm
were offered alone and gave a sex ratio 15% males (curve b). When 1.4 mm
hosts were offered alternately with 1.8 mm hosts, the 1.4 mm hosts yielded
a sex ratio of 30% males (curve c). When 1.4 mm hosts were offered al-
ternately with 1.0 mm hosts, the 1.4 mm hosts yielded a sex ratio of 2%
males (curve a). Host size was determined by measuring tunnel diameter
of the burrowing host. The number of emerging wasps is given for each
datum point. From Charnov et. al. (1981).

Example 2

Some parasites lay several eggs on a single host. Often as the number
of developing parasites on a host increases, the size of each emerging para-
site decreases due to competition for food among the larvae. In other words,
initial parasite density is a negative correlate of eventual parasite size
[e.g., Hydromeris contorta (Mermithidae), Charnov (1982) citing Johnson
(1955, not seen)]. The sex that has the greater increase in expected fitness
with increasing size should be favored in hosts with few parasites (large
individual parasites emerging), the other sex in crowded hosts (small in-
dividuals) (see fig. 3). The nematode Hydromeris provides a good example.
The sex of mermithid nematodes is determined after birth, while the nema-
tode is still immature, and environmental conditions have been shown to
be correlates of sex (i.e., sex is a phenotypically canalized trait) (Christie

54 Florida Entomologist 66(1) March, 1983

1929, Anya 1976). If a greater change in expected fitness with size for
females is assumed, as in example 1, the prediction is that females will be
favored in hosts with few parasites, males in hosts with many parasites.
Figure 5 shows the number of nematodes on a single host and the sex
ratio. The data are consistent with the prediction that density is a correlate
of sex. Given the assumption that relative fitness of the sexes differs as a
function of density (or size) as in figure 3, then correlates of relative fit-
ness of the sexes are correlates of the sex ratio. Figure 6 shows a hypo-
thetical relationship between an ecological correlate of fitness and the
expression of a sex-ratio pattern.


When the population is divided into small, isolated groups (=demes),
the expected fitness of an individual within a deme (ind/deme) is greatest
according to equal investment in the sexes (due to selection within demes,
or intrademic selection), and the fitness of the deme within the population
(deme/pop) is greatest when the investment ratio of the deme is extremely
female biased (due to selection between demes, or interdemic selection)
(paternal care assumed unimportant). The fitness of the individual within
the population as a function of its investment ratio in the sexes is a com-
promise of the conflict between the two fitness components, (ind/deme)
and (deme/pop). The resulting sex-ratio pattern due to conflicts among



0 o.6-

z 0.5-
I- 0.4

0. 0.3-
a.. 0.3
0. 0.2


1 2 3 4 5 6 7 8
Fig. 5. Correlation of number of nematodes per host and sex ratio. Few
nematodes per host corresponded to relatively large emerging individuals,
and many nematodes per host corresponded to relatively small individuals.
Data of Johnson (1955, not seen) for the mermithid nematode Hydromeris,
redrawn from Charnov (1982).

Insect Behavioral Ecology-'82 Frank 55





9 RS > dCRS cORs > $ RS


Fig. 6. Link between proximate cues and observed sex-ratio patterns.
In order to determine the size of a host, parasites often use ecological cor-
relates of host size such as daylength, ambient temperature, chemical prop-
erties of the host, and as in this example, tunnel diameter of a burrowing
larva. The same proximate cue, large tunnel diameter which implies a
large host, may signal high fitness of a male-biased or female-biased in-
dividual-level sex ratio to different parasite species according to their
particular natural histories. The two pathways are similar to examples 1
and 2 of pattern (2) in the text, which assume the fitness-size relationship
shown in figure 3.

levels of the hierarchy is surprising, and this pattern allows us a view of
a type of hierarchical organization rarely observed.
It is generally thought that the conditions under which interdemic
selection will occur are very rare, and this has influenced many biologists'
perceptions of nature. Indeed, in the vast majority of cases interdemic selec-
tion is likely to be trivial (Williams 1966, but see Wilson 1980). But the
situations where such selection may be relevant do exist and include ex-
traordinary life styles (Hamilton 1967, 1978, 1979). In order to illustrate
this difficult problem I will discuss a particular population breeding struc-
ture that is common among many groups of parasitic Hymenoptera and
mites (although probably rare among organisms in general). Consider an
isolated resource patch that is ephemeral in time and space, such as a larval
host. One to a few mated females arrive foundresseses) and lay eggs.
The eggs develop in this patch, and eventually the adults emerge. All mat-
ings occur just after eclosion within the isolated patch. In addition, for
this example we assume that the resource patch is depleted after only one

Florida Entomologist 66 (1)

generation. Therefore the already-mated females must disperse to find and
colonize new patches, groups of these dispersing females settling on a new
patch in a random fashion.

Numerical Example

In this example individual fitness within the population is partitioned
into individual fitness within a deme and deme fitness within the popula-
tion.A Individual fitness within the deme is analyzed as in pattern (1),
taking into account the very small size of the deme. [An assumption of
pattern (1) is that mating occurs randomly within the population. Here
mating occurs randomly within the deme, and no mating occurs between
members of different demes, so fitness(ind/deme) is analyzed instead of
fitness (ind/pop)]. As in pattern (1), RS is analyzed over 2 generations.
The method of determining witnesses of individuals within a deme is demon-
strated in table 6, for 2 foundresses forming a deme. From table 6 note
that foundress I has the same sex ratio in 2 cases, 1:1, but her fitness
changes according to the sex ratio adopted by foundress II. Since the
fitness of each foundress depends on the sex ratio of the other foundress,
the situation has a game-like quality (Hamilton 1967). Such an approach
can be expanded to include all combinations of sex ratios by foundresses
I and II. As Hamilton (1967) showed for 2 foundresses per deme, and Col-
well (1981) demonstrated for arbitrary number of foundresses per deme
when analyzing only the fitness component (ind/deme), the overall in-
vestment ratio within a deme stabilizes at 1:1. Generally, the overall in-
vestment ratio within the next higher stage of the hierarchy in which the
mating occurs at random stabilizes at 1:1. [In pattern (1), the population is
the next stage above individuals in which mating is random.] As we pro-
ceed to the next component, demes within a population, a conflict with the
individual level emerges (Hamilton 1967, 1979, Colwell 1981, Wilson and
Colwell 1981).
For the particular form of breeding structure being discussed, the fit-
ness of a deme within the population is positively correlated with number
of colonizing females produced. This conclusion follows from the fact that a
deme's contribution to the next generation depends on the number of colon-
izers it produces. Therefore, selection among demes within the population
favors demes with the most female-biased sex ratio possible such that
there are just enough males to mate with all of the females. The result
of conflicting selection pressures between individuals within a deme and
demes within the population is a stable sex-ratio pattern intermediate
between the extremes.
Hamilton (1967, 1979), and later Colwell (1981) and Wilson and Col-
well (1981) arrived at a general solution for the stable sex ratio in diploid
organisms, where clutch sizes of the fouridresses are equal. The solution is

r*= 1/2[(n-1)/n] n=2,3,4,...

where r* is the sex ratio (males/total) of a deme with n foundresses.9 The
effect is a predicted sex-ratio pattern at the level of the deme. A similar
solution for the haplodiploid case has also been found (Hamilton 1979, see
fig. 7).

March, 1983

Insect Behavioral Ecology-'82 Frank 57

4) ) 0.

0 0 T

;4 to mm0 C

1 a a
U~~ 's '"

ed Fd

) a, 0. H 0

a) 00 C NO C0
C/I^ -S 0'3 2 -

F- gq CdUS < "V 00 '1 .O 0f
> +N

(V C/2C1 C C
0 0' U C

w E-1r
^Ffl ^a n ^S

H ..o C so -
SCd 4.1

S00 0 0

Cdi wn: tH ?- d :tr -41 00 o 00
7SI ~F

41 0,.
4/i W -U A

Cd 00 Cd Cd

E- 1)'- C
C)^ ^ S -< -e
~jaHa II E

fc~<~~- C) ^ ^- ~
C) c

N II f s Hrx Eu- ^ Ec

-igH C)s o-

E C3 )
iii ul fl ci o>

CIoH S^ S>.i 5A

C)~c C) I

- ^ Onm
FI00 a,"o

58 Florida Entomologist 66(1) March, 1983

Empirical Support
Females of the hymenopteran parasite Nasonia vitripennis (Ptero-
malidae) lay up to 50 eggs on each dipteran host puparium encountered,
and 2 or more females often lay eggs on a single host (natural history
reviewed by Whiting 1967). The adult males usually emerge first, and
stay on or by the host waiting for the females to emerge. The males lack
the ability to fly, and all mating occurs near the host. The mated females
then fly off to find a fresh host to parasitize. Since these are hymenopteran
parasites, they are haplodiploid, and the females have the potential to
control the sex ratio of their clutches (phenotypically plastic sex ratio).
These parasites have the sort of breeding structure outlined above, and
Werren (1980a, 1980b, in press) examined the sex ratio to see if it con-
formed to Hamilton's (1979) prediction for haplodiploid organisms. The
data he obtained seem to agree with the predicted sex ratio (fig. 7).



O) 0.50 *

,. l .............. d .........................-
0 S
o S......t ......\

S 0.25 *'

2 4 6 8 10 12 14 16

Fig. 7. Correlation of sex ratio and number of foundresses per host for
Nasonia reared on Sarcophaga bullata in the laboratory. Mean +/- 1.96
s. e. are shown for each level. The dotted line represents the predicted sex
ratio for haplodiploid organisms with the breeding structure as described
in the text, where r*=1/2[(n-1)/n] [(4n-2)/(4n-1)] (Hamilton 1979,
Werren in press). Notice that this formula only differs from the one in the
text for diploid organisms by the factor [ (4n-2) / (4n-1)]. Haplodiploidy
can be treated as a case of sex-linked sex-determination control, where fe-
males are XX and males are XO. This sex-linked sex determination creates
an additional female-biasing selection pressure as discussed in pattern (5).
Data and experimental details in Werren (1980b).

Insect Behavioral Ecology-'82 Frank

More Levels in the Hierarchy?
Although the Nasonia data fit the prediction fairly well, other parasitic
wasps have a sex ratio distinctly more female biased than predicted [e.g.,
Melittobia (Eulophidae), fig wasps (Agaonidae), see Hamilton 1967, 1979
for a review]. The highly skewed sex ratios among these species remain a
mystery. I suggest that adding further levels to a hierarchical analysis be-
tween the level of demes and the population may provide a clue.
Consider the natural history of fig wasps as an example (see Wiebes
1979, Janzen 1979, and Hamilton 1979 for reviews of fig-wasp biology).
Fig trees bear inflorescences (=figs) that are hollow receptacles with hun-
dreds of staminate and pistillate flowers opening into a sealed, central
cavity. A few female wasps foundresseses) enter the fig through a nar-
row passage, and once inside can never leave. They pollinate some of the
pistillate flowers with pollen carried from the fig from which they emerged,
lay an egg in each ovary of the remaining pistillate flowers, and then die.
Each larva develops by eating the tissue of a single flower. About one
month later the next generation of wingless males emerges and mates with
the not-yet-eclosed females within the still-sealed fig. Therefore, each in-
dividual fig constitutes a deme (as described in the previous discussion)
that forms from a pool of colonizing females and disintegrates after one
generation. Hamilton's (1979) prediction about the sex ratio is relevant
(given in fig 7), since the fig wasps appear to have the sort of breeding
structure assumed in that model. Data indicate that the sex ratio of some
fig-wasp species does become less female biased with increasing number of
foundresses, however at a rate much slower than predicted (Frank in prep.,
fig. 8).
Let us return to the natural history of fig wasps for a moment. The
fig-wasp species for which data are reported in figure 8 occupy fig trees that
occur in the Caribbean, throughout the Florida Keys archipelago, and in the
extreme southern portion of the Florida peninsula. The host-tree's distri-
bution on the peninsula is quite patchy, and occasional frosts cause local
extinctions followed by recolonization (pers. obs.). Each fig tree bears
thousands of figs per crop, and several hundred wasps are reared from
each fig. Thus a single tree produces between 10 thousand and 10 million
fig wasps per crop. The emerging wasps fly and/or are blown to a new
tree bearing figs at a stage receptive to wasps. Since the trees occur in
patches isolated by long distances over land or water, the emerging wasps
will probably most often enter receptive figs within their own patch. Occa-
sionally a few wasps will be blown from one patch to a neighboring patch.
Thus each patch of trees represents a distinct subpopulation of wasps.
Fitness (ind/pop) can be partitioned into individuals within a deme (= a
fig), demes within a subpopulation (= an isolated patch of trees), and
subpopulations within the population. The sex ratio that maximizes fitness
(ind/subpop) is now represented by the formula stated in figure 7, since
the subpopulation now satisfies the assumptions of that model [i.e., a two-
level analysis of fitness (ind/deme) and fitness (deme/subpop)]. The con-
tribution of a subpopulation to the whole population is dependent on the
number of colonizing females produced by that subpopulation, creating an
additional female-biasing selection pressure. This population structure may
cause the stable sex ratio at the level of the deme to be more female biased

60 Florida Entomologist 66 (1) March, 1983



27 -

2x ;

9 1

I 2 3 4 5 6 7 8
Fig. 8. Correlation of sex ratio and number of foundresses per fig for
the fig wasp Pegoscapus sp. reared from figs of Ficus citrifolia. Each
datum point represents the sex ratio of the wasps from a single fig; the
average number of wasps per fig is 207. The solid curve is the predicted sex
ratio for haplodiploid organisms as in figure 7, which assumes random mat-
ing within a deme (=fig) and no differentiation among subpopulations. The
dashed curve is the predicted sex ratio assuming some assortative mating
and/or differentiation among subpopulations [Pind/fig K Psubpop/pop = 0.52;
see Wright 1969, Frank in prep for notation]. Data, statistical analyses,
and experimental details in Frank (in prep).

than in a population structure such as Nasonia's, where a significant par-
tition of demes/pop may not exist.'0
This explanation of fig-wasp sex ratios is hypothetical, and will re-
quire further testing on several fig-wasp species over a variety of host-
tree distributions (i.e., continuous distributions with poorly defined sub-
populations to isolated distributions with well-defined subpopulations). An
advantage of this approach is that it is amenable to simple modelling,
which will reveal exactly which measurements of isolation are important.
The hypothesis is that as the distribution of a species becomes relatively
(among all species of fig wasps) more isolated, the sex ratio at a given
number of foundresses per fig will be relAtively low for that species. Also,
within any species the sex ratio will increase with increasing number of
foundresses per fig (i.e., phenotypic plasticity is expected, see next section).

Ecological Correlates of Breeding Structure and Sex Ratio
As discussed in pattern (2), the best sex ratio for an individual mother
may change with ecological parameters-in this case it changes with the
9 1

0 -

I 2 3 4 5 6 7 8
Fig. 8. Correlation of sex ratio and number of foundresses per fig for
the fig wasp Pegoscapus sp. reared from figs of Ficus citrifolia. Each
datum point represents the sex ratio of the wasps from a single fig; the
average number of wasps per fig is 207. The solid curve is the predicted sex
ratio for haplodiploid organisms as in figure 7, which assumes random mat-
ing within a deme (=fig) and no differentiation among subpopulations. The
dashed curve is the predicted sex ratio assuming some assortative mating
and/or differentiation among subpopulations [Pind/fig Psubpop/pop = 0.52;
see Wright 1969, Frank in prep for notation]. Data, statistical analyses,
and experimental details in Frank (in prep).

than in a population structure such as Nasonia's, where a significant par-
tition of demes/pop may not exist.10
This explanation of fig-wasp sex ratios is hypothetical, and will re-
quire further testing on several fig-wasp species over a variety of host-
tree distributions (i.e., continuous distributions with poorly defined sub-
populations to isolated distributions with well-defined subpopulations). An
advantage of this approach is that it is amenable to simple modelling,
which will reveal exactly which measurements of isolation are important.
The hypothesis is that as the distribution of a species becomes relatively
(among all species of fig wasps) more isolated, the sex ratio at a given
number of foundresses per fig will be relatively low for that species. Also,
within any species the sex ratio will increase with increasing number of
foundresses per fig (i.e., phenotypic plasticity is expected, see next section).

Ecological Correlates of Breeding Structure and Sex Ratio
As discussed in pattern (2), the best sex ratio for an individual mother
may change with ecological parameters-in this case it changes with the

Insect Behavioral Ecology-'82 Frank

number of foundresses in a patch. Therefore, a phenotypically plastic re-
sponse to ecological correlates of breeding structure is possible. Consider
the natural history of Nasonia outlined above. The females search for
puparia and lay up to 50 eggs per host. Each female probably parasitizes
several hosts. When a female lays eggs on a host she may be the first
parasite, or the host may already have been parasitized by one or more
foundresses. The sex ratio that maximizes her fitness (ind/pop) depends on
how many other foundresses lay eggs on that host (Werren in press). What
are the ecological correlates of foundress number that a female can assess?
It has been demonstrated that a female can differentiate between para-
sitized and unparasitized hosts (Wylie 1965), and adjust her sex ratio ac-
cordingly (Holmes 1972, Werren 1980a). Such discrimination may be a
chemical cue (Holmes 1972). Because two females parasitizing the same
host may never see each other, chemical cues are likely candidates for being
assessed as ecological correlates of fitness. Females may also respond to
high densities of hosts or conspecifics assessed visually. What about fig
wasps? The foundresses must enter and oviposit within a tiny, sealed cavity
from which they never leave, and it seems likely that the foundresses adjust
their sex ratio according to an assessment of number of other foundresses.
This may be accomplished either tactilely, or by reacting to chemical cues
that are correlated with number of foundresses.11

In the patterns examined so far, fitness of levels below the individual
(=subgenomic) within the population have been assumed to be equivalent
to fitness of individuals (=genomes) within the population. Such an as-
sumption is legitimate when a phenotypic attribute benefits all subgenomic
elements equally.6,10 For example, changes in expected fitness(chromosome/
ind) as a function of sex ratio are the same for all autosomal chromosomes
of a diploid individual. When, depending on some phenotypic attribute,
the fitness of a subgenomic element within the population differs from the
fitness of individuals within the population that carry these elements, a
conflict arises (see Birky 1978, Cosmides and Tooby 1981 for reviews).
These subgenomic elements may include bacterial and viral infections, cellu-
lar organelles such as mitochondria that contain genetic material, and a
variety of other extranuclear cellular symbionts (reviewed by Margulis
1981). The fitness(element/pop) often differs from the fitness(ind/pop) in
which the element resides as a function of the sex ratio of the individual's
progeny (i.e., the sex ratio is a phenotypic attribute of the individual).
Figure 9a represents a case where such a conflict is likely to exist. In this
case a bacterial infection is passed from generation to generation through
the eggs of a female. The infection is not passed through sperm. Therefore,
the mode of inheritance of the infection is from mother to daughter, or
matrilineal. The expected fitness of the infection within an individual host's
progeny clearly increases as the sex ratio produced by its host becomes
more female biased. The fitness of the individual host within the population
with respect to sex ratio is often greatest according to patterns (1), (2),
and (3). A conflict exists between the infection and much of the genome
(= intragenomic conflict).12 Several cases of extranuclear genetic elements
affecting the sex ratio have been documented (i.e., the sex ratio of the
host is genetically determined by the element; for example, Sakaguchi and

Florida Entomologist 66 (1)

March, 1983

Poulson 1963, Johnson 1977, Werren et. al. 1981, Skinner 1982, Bryant et.
al. 1982, reviewed by Uyenoyama and Feldman 1978). Male sterility in
maize is probably controlled by mitochondrial genes, and therefore is an-
other case of this pattern (Rhoades 1933, Levings and Pring 1976).
Several questions about sex-ratio biases controlled by subgenomic ele-
ments need to be addressed. When they exist, do they spread through the
entire population? If not, what is their frequency, and why is this frequency
observed? Since the interests (in terms of fitness) of the sex-ratio biasing
elements conflict with the interests of other parts of the genome, do forms
of suppression exist (i.e. are there alleles that can reduce or remove the
effect of the sex-ratio bias even when the element is present)? The effect
of the sex-ratio bias due to these elements is expressed at the level of the
individual that has such an element, but what about the population sex
ratio? To determine how common a subgenomic element affecting the sex
ratio will be within a population, fitness (element/pop) must be analyzed.
Applying the hierarchical method of analyzing fitness, fitness (element/pop)
is a function of both fitness (element/ind) and fitness (ind/pop) harboring
these elements. This partitioning of fitness into components highlights bio-
logically important phenomena such as suppression (intragenomic conflict),



Y o

9 A


Cf (X Y)



Fig. 9. Uniparental inheritance pedigrees. (a) Matrilineal inheritance
of an extranuclear (cytoplasmic) genetic element. The element is passed
from mother daughter through the egg-the element may be passed to a
son through the egg, but is not passed through sperm, so males are a dead
end in the pedigree. Cytoplasmic elements, including mitochondria, chloro-
plasts, and bacterial and viral symbionts are almost always inherited
matrilineally. (b) Patrilineal inheritance of a Y chromosome. Males con-
tribute a Y chromosome to sons and an X chromosome to daughters. Since
females are XX, daughters represent a dead end in the pedigree of a Y

Insect Behavioral Ecology-'82 Frank

and leads to predictions of observed sex-ratio patterns at the individual
and population level, as we shall see below.

Empirical Evidence
The existence of extranuclear particles causing a sex-ratio bias is well
documented (see above references); however, clear answers to the above
questions do not exist for any single population. Here I briefly summarize
some observations on the terrestrial isopod Armadillidium vulgare (Howard
1942, 1958), and a few other species of isopods. (i) Individual females con-
sistently produce either female-biased broods (thelygeny), 1:1 broods
(amphogeny), or male-biased broods (arrhenogeny). (ii) The brood sizes
are the same for all sex ratios, suggesting there is no reduction in fecundity
associated with sex-ratio traits, and ruling out sex-biased mortality as an
explanation. (iii) A male parent has no effect on the brood's sex ratio. (iv)
Other loci (e. g., eye color) segregate in a Mendelian fashion, ruling out
parthenogenesis. (v) Thelygenic females have a particle in the ovarial tis-
sue absent from amphogenic females. Ovarial tissue grafted from thelygenic
females to amphogenic females converted the amphogenic females to thely-
geny. The nature of arrhenogeny is not entirely understood. Arrhenogeny
may be a different response to the same particle by genetically different
females, different particles, or some other mechanism (Juchault and Le-
grand 1970, Legrand and Juchault 1972). (vi) The mode of sex determina-
tion is unknown, although suggestions have been made (Bacci 1965). (vii)
The inheritance of sex-ratio traits is somewhat confusing. Howard (1958)
reviewed data of his own and other workers, and concluded that (a) in
Trichoniscus provisorius and A. vulgare the typical pattern is amphogenic
females giving rise to amphogenic daughters, and thelygenic females giv-
ing rise to both thelygenic and arrhenogenic daughters. The effect of a
male parent's history (i.e., from an amphogenic or arrhenogenic brood)
has not been studied, and (b) in Cylisticus convexus it has been demon-
strated that the male parent's history has an effect on the sex-ratio charac-
teristics of his daughters. Both Howard (1958) and Johnson (1977) have
suggested that the inheritance pattern may be due to an interaction be-
tween extranuclear elements and nuclear genes. This hypothesis seems
reasonable with respect to both the observations mentioned above and the
theory of intragenomic conflict over control of the sex ratio as outlined
above. (viii) The sex-ratio theory outlined in this paper predicts equal
investment in the sexes over the entire population, in spite of individual-
level biases [since when a sex is rare, it has a higher expected fitness than
the more abundant sex, see pattern (1)]. Nuclear genes interacting with
extranuclear elements could provide a genetic mechanism for variability
of population-investment ratios, allowing equal investment in spite of sex-
ratio distortion in infected individuals (Johnson 1977). The observed pop-
ulation sex ratios are difficult to interpret due to small sample size, method
of sampling, and the fact that the particle may spread in an infectious
manner under laboratory rearing conditions. Due to the nature of the sex-
ratio trait, the accuracy of an estimate probably depends more on the
number of broods reared than on total number of offspring counted, and
the variance of the population sex ratio is likely to be high. Johnson (1977),
studying Venezillo evergladensis, reports the largest random sample that
I could find in the literature. For the life-time broods of 87 females, the

Florida Entomologist 66 (1)

overall sex ratio was 1311:1310 (males:females), with what appear to
be 6 thelygenic and 6 arrhenogenic females. Johnson cites Vandel as hav-
ing found 42 arrhenogenics and 32 thelygenics in a sample of 88 T. pro-
visorius, and 4 of each type in A. vulgare in a sample of 15. Johnson
(unpub. data) found 4 thelygenics and 5 arrhenogenics in a random sample
of 30 gravid A. vulgare females collected from the wild, with an overall
sex ratio of 1085:1089 (males:females). Thus, it seems likely that popula-
tion sex ratios at birth do not differ greatly from 1:1 (but see Howard

When considering the effects of selection, the level of the individual
seems the most natural to contemplate; it is individuals that we can see
actively reproducing and competing. It is more difficult to imagine sub-
genomic elements competing for their own reproductive success, as in the
case of the matrilineal infections discussed in pattern (4). Consider the
case of a diploid organism. Each chromosome is paired with a homologue,
and only one of the two paired chromosomes is found in any single gamete.
It is often assumed that each chromosome of the pair has an equal chance
of appearing in any gamete due to random assortment. But, a chromosome
that can bias assortment such that it is in more than half of the successful
gametes (referred to "as a driving chromosome", or as "a case of meiotic
drive", reviewed by Zimmering et. al. 1970, White 1973, Crow 1979, see also
Bryant et. al. 1982) will increase in relative frequency within the individual's
progeny that it resides in, since the expected RS of a chromosome/ind is
one-half the number of progeny by that individual. Therefore whether
a driving chromosome will spread in the population depends on (i) the
fitness of (ind/pop) with such a chromosome, and (ii) the ability of other
elements in the genome to suppress a driving chromosome.13
The sex chromosomes of Drosophila are designated XY for males and
XX for females. The male's contribution of an X or Y chromosome deter-
mines the sex of his progeny (see fig. 9b).14 Thus, a driving X chromosome
in a male causes a female-biased sex ratio of his progeny; similarly, a driv-
ing Y chromosome causes a male bias. What about the fitness of an indi-
vidual within the population bearing a driving sex chromosome? For ex-
ample, consider a male with a driving X chromosome. Most of his progeny
will be females. The sex ratio of his progeny that would maximize his fit-
ness within the population often differs from the female-biased sex ratio
caused by the driving X chromosome [according to patterns (1), (2), and
(3)]. Thus, although a driving chromosome may increase its own fitness
(chrom/ind) within an individual above one, it may decrease the fitness of
the entire genome within the population (including itself), and may or
may not have a fitness (chrom/pop) greater than one.12 If a driving chromo-
some decreases fitness(ind/pop), selection favors other elements of the
genome that can suppress the driving chromosome. Thus the apparently
random assortment often observed in meiosis may actually be a competi-
tive stalemate of subgenomic elements (Alexander and Borgia 1978); for
example, a Y chromosome with the ability to suppress the driving behavior
of an X may yield a 1:1 sex ratio. Driving X chromosomes are widespread
in wild populations of some Drosophila species (reviewed by Zimmering et.
al. 1970, White 1973); an interesting sex-ratio-distortion condition in D.

March, 1983

Insect Behavioral Ecology-'82 Frank

paramelancia was studied by Stalker (1961). There are two types of driv-
ing X's, called 'Northern' and 'Southern', corresponding to locations of the
populations where they are most commonly found, although some overlap
exists. Northern X's have their sex-ratio-distortion effect suppressed by
Southern Y's (i.e., males with Northern X, Southern Y do not produce ex-
cessively female-biased broods). Southern X's are not suppressed by any
type of Y's. In this case, no reduction in fecundity was reported for males
with a driving X chromosome and a female-biased progeny sex ratio. Other
cases of drive with reduced fitness(ind/pop), or suppression due to other
chromosomes, have been documented (White 1973, Crow 1979).
Hartl (1977) reviews potential applications of meiotic drive for breed-
ing and for population control of insect pests. Hamilton (1967) pointed
out that the introduction of a driving Y chromosome might be an effective
form of control in species with XY (heterogametic) males, but also noted
a possible complication due to suppression of driving Y's by other elements
of the genome, and suggested methods for overcoming this difficulty. Thus
meiotic drive and extranuclear elements may be of practical use, and will
certainly help explain some of the variability of sex ratios observed in

Five sex-ratio patterns were discussed in this paper (table 1).15 Two
general themes of sex-ratio patterns emerge-(i) ecological correlates of
fitness and (ii) conflicts within and among the levels of selection. These two
themes are not mutually exclusive, and both affect and are affected by
the mode of sex determination. (i) When relative fitness of the sexes per
unit investment varies according to some ecological correlate, then local
sex ratios are expected to vary according to that ecological correlate. There
are at least two complications. First, not only must ecological correlates
of relative fitness exist, but whatever controls the sex ratio must also,
in some sense, be able to respond to the levels of that correlate (by a proxi-
mate mechanism causing phenotypic plasticity or canalization which re-
sponds to proximate cues such as temperature, daylength, etc.). This leads
to the second complication. There are many types of sex-determination
systems, and the sex-determination system may constrain the nature of the
genotype-phenotype relationship. The sex-determination system may also
constrain which genetic entities can control the sex ratio. (ii) The second
general theme is conflict within and among the levels of selection. A par-
ticular subset of the genome (e.g., a chromosome, or an extranuclear ele-
ment) may maximize its own fitness according to a sex ratio different
from the sex ratio that maximizes the fitness of other subsets of the
genome [and hence different from fitness (ind/pop)]. The phenotype of an
individual depends on the resolution of this inherent conflict (which may
be within or among levels of the hierarchy), and this resolution will de-
pend on whatever elements) can control the expressed sex ratio. Thus,
there is much overlap among genotype-phenotype interactions, ecological
correlates of fitness, the mode of sex determination, and the levels of
A hierarchical approach to sex ratios has at least two advantages. First,
it provides insight into the fitness components which must be analyzed

Florida, Entomologist 66(1)

when considering particular problems, and this highlights biologically
interesting phenomena such as intragenomic conflict and population breed-
ing structure. Second, a hierarchical approach provides a framework for
specifying the level at which a sex-ratio pattern will be observed, such as
individuals infected with a sex-ratio distorting bacteria, or the population
as a whole when considering overall investment in males and females. A
hierarchical approach also provides insight into a much broader class of
problems than sex ratios. Current research has revealed that what we
usually refer to as an individual's genome is actually a rich mosaic of
genetic elements of diverse taxonomic origins (Margulis 1981). These
genetic elements within an individual have both common and conflicting
interests with respect to their own witnesses, leading naturally to the ex-
pression of both cooperative and antagonistic traits. Such cooperative-
antagonistic symbioses are perhaps more common than is immediately ap-
parent (Axelrod and Hamilton 1981); hints that knowledge of such sym-
bioses may be of great practical value have recently appeared in the litera-
ture (e.g. Stoltz and Vinson 1979, Edson et. al. 1981, Clark 1982). For
example, a termite hindgut is know to be a veritable ecosystem unto itself
(reviewed by Margulis 1981). The success of any microbe within a termite
depends mostly on the success of the termite, thus the association is mainly
a cooperative one. However, a microbe's success also depends on its suc-
cess within the subpopulation of similar microbes in the gut, and a microbe
may be transmitted horizontally (i. e., infectiously among termites) as
well as vertically (i.e., inherited from generation to generation). Thus, it
is likely that on some occasions the interests of the microbe and the ter-
mite will conflict, and that the microbe may alter the phenotype of its
host accordingly [e.g., a prediction is that a microbe is more likely to be-
come virulent as its host becomes old or sick, thus increasing its ability
for horizontal (infectious) transmission when the probability for vertical
(inheritable) transmission is low; for an elegant example of this, see Lewin
(1977) and Ptashne et. al. (1982) on the biology of the temperate bacterio-
phage lambda, see also Axelrod and Hamilton (1981)]. With specific knowl-
edge about a termite gut, or any genetic mosaic, testable predictions about
subtle patterns of conflict flow naturally from a hierarchical view. Recog-
nizing these subtle patterns will occasionally prove critical to both the
identification and unraveling of elusive puzzles.


Sex ratios are complex phenomena. In order to formulate a general
theory that spans this complex issue, four concepts were first established:
(1) the possible relationships between genotypes and their expressed pheno-
types were stated; (2) the significance in terms of fitness (ultimate cause)
of behavior was contrasted with environmental (proximate) cues associated
with behavior; (3) the levels of a genetic hierarchy (genes, chromosomes,
individuals, population) at which natural selection acts effectively, and the
level at which sex-ratio patterns are observed were described; and (4)
the language used to analyze investment of resources into males and fe-
males by examining benefit/cost ratios was defined. After these four con-
cepts were discussed, five general sex-ratio patterns were described. In
each case the theory was presented briefly, followed by a supporting em-

March, 1983

Insect Behavioral Ecology-'82 Frank 67

pirical study. Two general themes emerged; (i) an individual organism's
expressed sex ratio may be consistent with ecological correlates of the rela-
tive fitness of the sexes, if the organism can assess these correlates, and
(ii) conflicts in the direction of selection on the sex ratio at different bio-
logical levels (e.g., chromosomes, individuals) may result in a complex
sex-ratio pattern.

1There is a large body of theory on sex ratios, mostly in the mathemati-
cal style of theoretical population genetics (Charnov 1982 and refs). Such
a rigorous approach is helpful in specifying assumptions and testable pre-
dictions. However, in a broad description of sex-ratio patterns a rigorous
population-genetics approach may obscure interesting biological phenomena
(e.g., phenotypically plastic responses to the environment versus genetically
determined responses). Therefore, I have adopted a heuristic approach
that highlights biologically important concepts and often parallels the
mathematical theory.
2The mechanisms involved in the expression of traits can usually be
thought of as having a genetic component (in organisms such as insects).
However, the phenotypic trait actually expressed may be a response to
an environmental (proximate) cue mediated by that mechanism such that
the response is independent of the genotype (=phenotypic plasticity). For
example, if a parasitic insect has the ability (genetically determined mech-
anism) to vary its sex ratio according to the size of its host, the expressed,
or observed sex ratio (phenotypic trait) may not have a genetic component
(i.e., may be phenotypically plastic). In other words, the expressed pheno-
typic trait may have a direct genetic component (genetically determined),
or may be independent of (or uncoupled from) the genotype by being medi-
ated through a mechanism that yields a variable response. This observation
will be important for deciding how to analyze particular sex-ratio patterns.
3It must be stressed that the patterns I discuss do not refer to the sex
ratio among adults, and do not account for differential mortality and dif-
ferent behavioral patterns between the sexes that may lead to observed
sex-ratio biases. In general, the patterns refer to the total allocation ratio
of resources into males and females (Charnov 1982), e.g., the sex ratio of a
clutch of eggs by an insect that does not give parental care to its offspring.
These concepts are discussed later in this paper.
4If no correlates of fitness exist, then a genetically determined or pheno-
typically canalized response is likely, since nothing is gained by phenotypic
plasticity. A difficult but very important problem is why, when correlates
do exist, a mechanism that responds to these correlates sometimes evolves
and sometimes does not. Part of the answer must be whether the organism
can evolve the ability to assess the level of the correlate and a mechanism
to mediate the response (a phylogenetic constraint). Another part of the
answer lies in an implicit assumption about the levels of selection; the
individual is considered the unit of selection in the problem stated above.
For sex ratios, this assumption is not always met, and thus the problem as
stated is misleading. See notes 6 and 9 for further discussion.
5Usually one refers to a change in the frequency of an allele (fitness)
only with respect to alternative alleles found at the same locus. The ac-
cumulation of evidence in the past few years suggesting that gene con-
version, gene duplication, mobile genetic elements, and viruses are not rare
phenomena (Cold Spring Harbor Symposium on Quantitative Biology, v.
45 1980) makes the concepts represented by 'locus'I 'alternative allele', and

Florida Entomologist 66 (1)

'population gene pool' somewhat fuzzy. Thus, a broader view is taken here
that does not restrict alleles to a particular location within the genome
and allows the analysis of fitness to be conducted within a class broader
than the population gene pool (e.g., the population in the usual sense plus
symbiotic viruses).
6I am analyzing fitness (ind/pop), thus ignoring the component (allele/
ind) necessary to complete the analysis of (allele/pop). A difficult and criti-
cal problem of evolutionary theory is to determine the level (s) at which
selection can be meaningfully analyzed (Williams 1966, Lewontin 1970,
Hamilton 1975, Alexander and Borgia 1978). Theoretical models of be-
havior often posit alternative alleles that have associated phenotypic
characters expressed in their host, and then analyze fitness (allele/pop)
(see Dawkins 1980 for discussion of models). However, in many situations
(e.g., dominant autosomal allele, no allele-genotype interaction, Mendelian
segregation, these being common assumptions of many population-genetic
models) the benefits of a phenotypic character associated with a particular
allele, or set of alleles, are shared equally by all alleles residing in that
individual. For the majority of population-genetic models the phenotypic
trait associated with the allele that obtains the greatest fitness is the same
phenotypic trait that maximizes the number of offspring of the individual
possessing that allele. Thus, in many cases, the component fitness (allele/
ind) adds nothing to the analysis in a heuristic sense, and requires the
strong assumption of an allele with a direct phenotypic effect. Often, a
phenotypically plastic response (i.e., variability of the sex ratio at the
level of individual phenotypes) yields the greatest individual RS within the
I am not suggesting that the individual is the unit of selection, and the
gene is not, or vice versa. Rather, as a heuristic device, I am considering
whether the underlying mechanism of sex-ratio variability is phenotypic
or genetic (or a combination of the two). If sex-ratio control is genetic,
then whether or not there are conflicts of interest among subgenomic ele-
ments over the sex ratio needs to be considered in order to determine the
lowest level analyzed for the arguments presented here. The reasons for
this will be clearer after patterns (4) and (5) are discussed.
7A second assumption here is that parents control the sex-ratio of their
progeny, based either on the parents' genotype and/or phenotype. This cer-
tainly is not true of many plants and other organisms which are her-
maphrodites, and for other sex-determination systems (Bacci 1965, Ber-
gard 1972, Charnov and Bull 1977). Charnov (1982) handles this problem
with a general approach: He asks what allocation ratio of resources, into
males versus females, maximizes the benefit/cost ratio with respect to
whatever entity controls this allocation ratio (e.g., a parent controls its off-
spring's sex, a hermaphrodite controls the allocation ratio into its sexes,
etc.). Using this approach Charnov shows that equal investment holds for
a much wider class of situations than parental control of progeny sex
sThe concept of "investment of resources" is widely used, but rarely ex-
plicit. Organisms invest time, calories, ,rare nutrients, etc., and determin-
ing which of these is critical to measure empirically is difficult. The theory
of fitness sets may be applied (see Levins 1968). In practice it is often
useful to ask the question, which resource is most limiting (e.g., calories),
and thus places the greatest constraint on investment patterns.
9The contrast in approach between Hamilton (1967, 1979) and Colwell
(1981) highlights a major theme of this paper. Hamilton assumed the
breeding structure outlined in the text, and additionally (i) that deme sizes
are constant (=n) throughout the population, and (ii) the sex ratio is

March, 1983

Insect Behavioral Ecology-'82 Frank

genetically determined. When sex-ratio control is autosomal, the optimum
is given by the formula in the text, and when the genetic control of the
sex ratio is sex-linked or extranuclear, the optimum is reported in Hamilton
(1979, fig 6). The optimum sex ratio varies widely according to the type
of genetic control [see also patterns (4) and (5) in this paper], highlight-
ing the influence of intragenomic conflict on the sex ratio. A disadvantage
of this approach is that deme sizes are likely to vary in nature, and thus
the optimum sex ratio will fluctuate according to ecological conditions (i.e.,
deme size). Therefore, phenotypically plastic sex ratios may be favored;
but intragenomic conflict cannot be ignored, since a given phenotype favors
different subsets of the genome unequally. Hamilton (1967) showed that
for n=2, the sex ratio that yields the greatest number of grandprogeny
to a foundress is the same as predicted by the formula in the text, 1/4.
This agreement between an autosomal model and a phenotypic model is
not surprising (see note 6). Colwell (1981) demonstrated the formula
in the text is the unbeatable sex ratio in a phenotypic sense (number of
grandprogeny), and it also seems to be the phenotypic solution for variable
deme sizes (I know of no explicit statement of this in the literature). This
suggests that when analyzing what I call fitness (ind/pop), the best sex
ratio for an individual foundress is given by the formula in the text, and
hence phenotypic plasticity is advantageous when deme sizes vary. Colwell
(1981) suggested that this phenotypic result is independent of the under-
lying genetic basis of sex-ratio variability. He implicitly assumed, however,
that fitness (ind/pop) is equivalent with autosomal fitness within the
population gene pool; and in light of Hamilton's (1979) demonstration of
the importance of intragenomic conflict, this suggestion seems erroneous.
Returning to Hamilton's (1979) population-genetic approach, some bio-
logically interesting insights emerge. First, recall that when one takes
a random sample from a population, the expected sampling variance is
less than the population variance-E (s2) = [ (n-- ) /n]0-2<-2. This implies
that when a few individuals randomly settle in a patch dememe), they are
genetically more homogeneous than the population as a whole. This leads
to a biologically meaningful interpretation of the formula in the text. r* =
1/2[V(wg)/V(t)], where V(wg) is the within-group genetic variance and
V(t) is the population variance (Hamilton 1979). With random assort-
ment of colonizing females between generations, E[V(wg)] = E(s2) =
[(n-l)/n]0-2, and V(t) = 0-2, hence r*=1/2[(n-1)/n]. Notice that the
lower the within-group variance at the loci controlling the sex ratio, the
more "related" ind/group are, and the more female biased the sex ratio
"OThis is a logical extension of note 9. The driving force behind female-
biased sex ratios is the ratio V(wg)/V(t). Each fig represents a random
selection of wasps mostly from the local subpopulation (Frank, in prep.),
and therefore genetic variance within a fig(=wg) is a sampling variance
of the total variance in the subpopulation. In addition, it is likely that the
subpopulations differ in genetic composition due to sampling error, selection,
and drift (see Wright 1969), and possible that assortative mating occurs
within a fig (Frank, in prep.), which effectively reduces V(wg)/V(t).
Note that V (wg)/V(t) = [V (wg)/V(subpop) ] [V(subpop)/V(t)], which
identifies the contribution of differentiation among subpopulations (Wright
1969) to female-biased sex ratios (Hamilton 1979). Figure 8 suggests that
the sex ratio is phenotypically plastic, foundresses responding to the num-
ber of other foundresses (or some correlate of foundress number). It is
not clear what the quantitative relationship is between the autosomal genetic
variance model and a phenotypic (number of grandprogeny) model, although
the predicted trends are very likely the same.

Florida Entomologist 66 (1)

11If the sex ratio is controlled by a subset of the genome, and these
alleles can produce a phenotypically plastic response, then it is not num-
ber of foundresses, but genetic variance within a fig at the loci controlling
the sex ratio that is important (see note 10). Hence, if these loci have a
mechanism that can assess cues correlated to genetic relatedness and pro-
duce a sex-ratio response, such a response is expected. Genetic recognition
would be implicated if, for a given number of foundresses, relatedness is
a correlate of sex ratio.
12Fitness (element/pop) is a function of both fitness (element/ind) and
fitness(ind/pop), Individuals harboring these elements (or a driving chro-
mosome) may have a reduced fitness(ind/pop), possibly explaining why
such elements do not dominate many populations, or when they are wide-
spread, why they are often associated with complex suppressor systems
(this will be discussed in the text). Note that fitness(element/pop) can be
less than one, even though fitness (element/ind) is greater than one, due to
the component (ind/pop).
"3Similar to note 6, benefits due to a particular allele on a chromosome
are usually shared equally by all alleles on that chromosome, assuming no
recombination. As a caveat, it is important to note that with recombination
chromosomes are not stable entities, and thus analysis usually focuses on
alleles (Dawkins 1976). In the text I refer to chromosomes because it is
easy to observe the effect of a driving sex chromosome-a sex-ratio bias.
Identifying alleles underlying sex-ratio distortion may be quite difficult
(but see Crow 1979).
14Notice that a Y chromosome which is inherited patrilineally is not
different from an extranuclear genetic element that is inherited patrilin-
eally, except that a chromosome is located in the nucleus and extranuclear
elements are not. Thus patterns (4) and (5) can be grouped into a more
general class, where the defining characteristic is a conflict between fit-
ness(subgenomic element/ind) and fitness(ind/pop), when examined with
respect to the investment ratio in the sexes.
"There are many other patterns discussed in the literature that were
not mentioned in this paper. Most are included in Charnov (1982). For
example, much has recently been written about plant sex ratios (e.g., Bar-
rett and Helenurm 1981 and refs), the ability of adults to change sex in
some fish (sequential hermaphroditism, e.g., Warner and Hoffman 1980,
Warner 1982 and refs), parent-offspring conflict and the sex ratio of
eusocial colonies (e.g., Trivers and Hare 1976, Alexander and Sherman
1977), competition among siblings of the same sex for resources which may
lead to a population investment ratio different from 1:1 (local resource
competition, Clark 1978; this pattern may be relevant to vertebrates, e.g.,
Clutton-Brock et. al. 1981, Clutton-Brock 1982), competition among sib-
lings between the sexes (Pickering 1980), and bivoltinism in insects with
partial overlap of generations, a pattern which may yield an investment
ratio that changes from generation to generation within a single season
(Seger in press, see also Longair 1981). In general, these patterns are ex-
tensions of the basic notions reviewed in this paper.
16Sam Skinner has pointed out to me some work that is more recent
than that cited in the text. Legrand et. al (1980), studying Porcellio dila-
tatus, present evidence for three types of sex chromosomes (X, Y, and Z),
of which any individual has two; and they discuss the occurrence of thely-
geny, arrhenogeny, and amphogeny, in light of these sex chromosomes and
the inheritable extranuclear particles that affect the expression of gender.
Juchault and Legrand (1981 and refs) report that Armadillidium vul-
gare males are homogametic (ZZ) and females heterogametic (ZW), and
that two feminizing factors act to make genetic males functionally a re-

March, 1983

Insect Behavioral Ecology-'82 Frank

productively successful female; and one masculinizing factor that, when
present simultaneously with a feminizing factor, causes genetic males to
be intersexual. The population dynamics and interaction of these factors
with sex chromosomes to yield a global sex ratio of approximately 1:1 in
nature are discussed in these papers.
Bulnheim (1978), synthesizing his own work on the amphipod Gammarus
duebeni, reports that the sex of an individual has environmental, genetic
(nuclear), and parasitic (extranuclear elements) components. The sex
of an uninfected individual depends on an interaction between a polygenic
sex-determination system and photoperiod during ontogeny (the photo-
sensitive phenocritical period of sex differentiation is delimited by the sec-
ond to fourth molts). In addition to environmental and genetic components,
this amphipod harbors two transovarially transmitted matrilineall) micro-
sporidians that have a feminizing influence on a developing individual, but
these parasites interact with environmental and probably nuclear genetic
factors in determining gender.

My thoughts about sex ratios have been deeply influenced by my under-
graduate courses at the University of Michigan with R. D. Alexander, W.
D. Hamilton, and B. S. Low, and by my graduate advisor H. J. Brockmann
and her students at the University of Florida. All errors in interpretation
are my own. The following have provided helpful discussion and comments,
J. Cohen, C. Johnson, B. S. Low, J. R. Lucas, P. Murphy, J. A. Pounds, S.
H. Richey, J. Sivinski, T. J. Walker, and particularly H. J. Brockmann,
T. G. Forrest, A. Grafen, J. E. Lloyd, and M. Obin. E. L. Charnov pro-
vided a preprint of his book, which was an invaluable resource. I especially
wish to thank my mentors, H. J. Brockmann and J. E. Lloyd, for their
wonderful patience during the writing of this paper. Fig wasp research
was supported by a grant from the Alexander Bache Fund and the Theo-
dore Roosevelt Memorial Fund, two grants from Sigma Xi, and by funds
from the Dept. of Zoology at the University of Florida and I. N. and M. E.
Frank. The Everglades National Park, and Van Waddill and the Agricul-
tural Research and Education Center, Homestead, provided permits and
facilities. W. D. Hamilton introduced me to fig wasps, and has generously
shared his ideas. This research was conducted in partial fulfillment of
the requirements for the M. S. degree at the University of Florida. Figure
4 was reprinted by permission from Nature, Vol. 289, p. 29, copyright (c)
1981 Macmillan Journals Limited, and by permission from E. L. Charnov.
Figure 5 was modified and reprinted by permission from The Theory of
Sex Allocation by E. L. Charnov, p. 44, copyright (c) 1982 Princeton Uni-
versity Press. Figure 7 was modified and printed by permission from Stud-
ies in the Evolution of Sex Ratios by J. H. Werren, copyright (c) 1980
J. H. Werren.

ALEXANDER, R. D., AND G. BORGIA. 1978. Group selection, altruism, and the
levels of organization of life. Annu. Rev. Ecol. Syst. 9: 449-74.
AND P. W. SHERMAN. 1977. Local mate competition and parental
investment in social insects. Science 196: 494-500.
ANYO, A. 0. 1976. Physiological aspects of reproduction in nematodes.
Adv. Parasit. 14: 267-351.
AXELROD, R., AND W. D. HAMILTON. 1981. The evolution of cooperation.
Science 211: 1390-96.

Florida Entomologist 66 (1)

BACCI, G. 1965. Sex Determination. Pergamon, Oxford.
BARRETT, S. C. H., AND K. HELENURM. 1981. Floral sex ratios and life his-
tory in Aralia nudicaulis (Arliaceae). Evolution 35: 752-62.
BERGARD, J. 1972. Environmental and physiological control of sex deter-
mination and differentiation. Annu. Rev. Ent. 17: 57-74.
BIRKY, C. W. JR. 1978. Transmission genetics of mitochondria and chloro-
plasts. Annu. Rev. Genet. 12: 471-512.
BOULETREAU, M. 1976. Effect of photoperiod experienced by the adults on
the sex ratio of the progeny in Pteromalus puparum (Hymenoptera:
Chalcidae). Ent. Exp. Appl. 19: 197-204.
BRYANT, S. H., A. T. BECKENBACH, AND G. A. COBBS. 1982. "Sex-ratio"
trait, sex composition, and relative abundance in Drosophila pseudo-
obscura. Evolution 36: 27-34.
BULNHEIM, H. P. 1978. Interaction between genetic, external and para-
sitic factors in sex determination of the crustacean amphipod Gam-
marus duebeni. Helgolander wiss. Meeresunters. 31: 1-33.
CHARNOV, E. L. 1982. The Theory of Sex Allocation. Monographs in Popu-
lation Biology No. 18. Princeton Univ. Press, Princeton.
-- AND J. J. BULL. 1977. When is sex environmentally determined?
Nature 266: 828-30.
Sex ratio evolution in a variable environment. Nature 289: 27-33.
CHRISTIE, J. R. 1929. Some observations of sex in the Mermithidae. J.
Exp. Zool. 53: 59-76.
CLARK, A. B. 1978. Sex ratio and local resource competition in a prosimian
primate. Science 201: 163-65.
CLARK, T. B. 1982. Spiroplasmas: diversity of arthropod reservoirs and
host-parasite relationships. Science 217: 57-59.
CLAUSEN, C. P. J. 1939. The effect of host size upon the sex ratio of
hymenopoterous parasites and its relation to methods of rearing
and colonization. J. New York Ent. Soc. 47: 1-9.
CLUTTON-BROCK, T. H. 1982. The red deer of Rhum. Nat. Hist. 91(11):
S. D. ALBON, AND F. E. GUINNESS. 1981. Parental investment in
male and female offspring in polygynous animals. Nature 289: 487-
Movable Genetic Elements. Cold Spring Harbor Laboratory, New
COLWELL, R. K. 1981. Group selection is implicated in the evolution of
female-biased sex ratios. Nature 290: 401-04.
COSMIDES, L. M., AND J. TOOBY. 1981. Cytoplasmic inheritance and intra-
genomic conflict. J. Theor. Biol. 89: 83-129.
CROW, J. F. 1979. Genes that violate Mendel's rules. Sci. Amer. 240(2):
AND M. KIMURA. 1970. An Introduction to Population Genetics
Theory. Burgess Publishing Co., Minneapolis.
DAWKINS, R. 1976. The Selfish Gene. Oxford Univ. Press, N.Y.
1980. Good strategy or evolutionarily stable strategy? Pages 331-67
in G. W. Barlow and J. Silverberg, eds. Sociobiology: Beyond Na-
ture/Nurture? AAAS, Westview Press, Boulder, CO.
Virus in a parasitoid wasp: suppression of the cellular immune re-
sponse in the parasitoid's host. Science 211: 582-84.
FILIPPONI, A., B. MOSNA, AND B. PETRELLI. 1972. The optimum temperature
of Macrocheles muscaedomesticae as a population attribute. Riv.

March, 1983

Insect Behavioral Ecology-'82 Frank 73

Parasitt. 32: 193-218.
-- AND G. PETRELLI. 1975. Laboratory studies on the autoecology of
three species of the iGlaber group (Acarina, Mesostigmata). 3. Sex
ratio and population dynamic. Riv. Parasitt. 36: 295-308.
FISHER, R. A. 1958. The Genetical Theory of Natural Selection, 2nd ed.
Dover Publications, N. Y.
FLANDERS, S. E. 1956. The mechanism of sex-ratio regulation in the (para-
sitic) Hymenoptera. Insectes Soc. 3: 325-34.
FRANK, S. A. in preparation. The Natural History and Sex Ratio of the
Florida Fig Wasps. M. S. thesis, University of Florida.
GOULD, S. J., AND E. S. VRBA. 1982. Exaptation-a missing term in the
science of form. Paleobiology 8: 4-15.
HAMILTON, W. D. 1967. Extraordinary sex ratios. Science 156: 477-88.
1975. Innate social aptitudes of man: an approach from evolu-
tionary genetics. Pages 133-155 in R. Fox, ed. Biosocial Anthropology.
Wiley, N. Y.
1978. Evolution and diversity under bark. Pages 154-175 in L. A.
Mound and N. Waloff, eds. Diversity of Insect Faunas. Blackwell
Scientific Publications, Oxford.
1979. Wingless and fighting males in fig wasps and other insects.
Pages 167-220 in M. S. Blum and N. A. Blum, eds. Sexual Selection
and Reproductive Competition in Insects. Academic Press, N. Y.
HARTL, D. L. 1977. Applications of meiotic drive in animal breeding and
population control. Pages 63-88 in E. Pollak, O. Kempthorne, and
T. B. Bailey, Jr., eds. Proceedings of the International Conference
on Quantitative Genetics, 1976. Iowa State Univ. Press, Ames.
HOELSCHER, C. E., AND S. B. VINSON. 1971. The sex ratio of a hymenop-
terous parasitoid, Campoletis perdistinctus, as affected by photo-
period, mating, and temperature. Ann. Ent. Soc. America 64: 1373-
HOLMES, H. B. 1972. Genetic evidence for fewer progeny and a higher
percent males when Nasonia vitripennis oviposits in previously para-
sitized hosts. Entomophaga 17: 79-88.
HOWARD, H. W. 1942. The genetics of Armadillidium vulgare Latr. II.
Studies on the inheritance of monogeny and amphogeny. J. Genet.
44: 143-59.
1958. The genetics of Armadillidium vulgare Latr. IV. Lines
breeding true for amphogeny and thelygeny. J. Genet. 56: 1-10.
JANZEN, D. H. 1979. How to be a fig. Ann. Rev. Ecol. Syst. 10: 13-51.
JOHNSON, A. A. 1955. Life history studies on Hydromeris contorta, a
nematode parasite of Chironomus plumosus. Ph. D. Thesis, Univer-
sity of Illinois.
JOHNSON, C. 1977. Evolution of sex ratios in the isopod, Venezillo ever-
gladensis. Evolution 31: 603-10.
JUCHAULT, P., AND J. J. LEGRAND. 1970. Intersexualit6 et monogenie chez
les Crustac6s isopodes terrestres: induction de la th6lygenie chez
Armadillidium vulgare; facteurs contrSlant le pseudo-hermaphro-
disme masculin externe chez Porcellip dilatatus. Ann d' Endocrinologie
31: 525-30.
-- AND 1981. Contribution a l'6tude qualitative et quantitative
des facteurs contr6lant le sexe dans les populations du crustac6
Isopode terrestre Armadillidium vulgare latreille. III. Populations
n'h6bergeant pas le facteur f6minisant F (Bacteroide intracyto-
plasmique). Arch. Zool. exp. Gen. 122: 117-31.
KROMBEIN, K. V. 1967. Trap-nesting Wasps and Bees: Life Histories,
Nests, and Associates. Smithsonian Press, Washington, D. C.

Florida Entomologist 66 (1)

LEGRAND, J. J., AND P. JUCHAULT. 1972. Mise en evidence dans une popu-
lation d'Armadillidium vulgare Latr. (Crustac6 Isopode Oniscoide)
de deux types de lign6es arrh6nog6nes en relation avec ces facteurs
6pig6n6tiques A effet respectivment masculinisant et feminisant. C.
R. Acad. Sci. Paris 274: 1554-57.
J. P. MOCQUARD, AND G. MARTIN. 1980. Polymorphisme
d'origine geographique portant sur la ((valence)) des chromosomes
sexuels et phenomenes de monog6nie resultant du croisement de
differentes populations de Porcellio dilatatus Brandt. (Crustac6 Iso-
pode terrestre). Reprod. Nutr. Develop. 20: 23-59.
LEVINGS, C. S. III, AND D. R. PRING. 1976. Restriction endonuclease analy-
sis of mitochondrial DNA from normal and Texas cytoplasmic
male-sterile maize. Science 193: 158-60.
LEVINS, R. 1968. Evolution in Changing Environments. Princeton Univ.
Press, Princeton.
LEWIN, B. 1977. Gene Expression, Vol. 3. Wiley, N. Y.
LEWONTIN, R. C. 1970. The units of selection. Annu. Rev. Ecol. Syst. 1:
LONGAIR, R. W. 1981. Sex ratio variations in xylophilous aculeate Hy-
menoptera. Evolution 35: 597-600.
MARGULIS, L. 1981. Symbiosis in Cell Evolution. Freeman, San Francisco.
METCALF, R. A. 1980. Sex ratios, parent-offspring conflict, and local com-
petition for mates in the social wasps Polistes metricus and Polistes
variatus. Am. Nat. 116: 642-54.
NOONAN, K. M. 1978. Sex ratio of parental investment in colonies of the
social wasp Polistes fuscatus. Science 199: 1354-56.
PICKERING, J. 1980. Larval competition and brood sex ratios in the gre-
garious parasitoid Pachysomoides stupidus. Nature 283: 291-92.
PRICE, G. R. 1970. Selection and covariance. Nature 227: 520-1.
1972. Extension of covariance selection mathematics. Ann. Hum.
Genet. 35: 485-90.
PTASHNE, M., A. D. JOHNSON, AND C. O. PABO. 1982. A genetic switch in
a bacterial virus. Sci. Amer. 247 (5) : 128-40.
RHOADES, M. M. 1933. The cytoplasmic inheritance of male sterility in
Zea mays. J. Genet. 27: 71-93.
SAKAGUCHI, B. AND D. F. POULSON. 1963. Interspecific transfer of the
"sex-ratio" condition from Drosophila willistoni to D. melanogaster.
Genetics 48: 841-61.
SEGER, J. in press. Partial bivoltinism may cause alternating sex-ratio
biases that favour eusociality. Nature.
SKINNER, S. W. 1982. Maternally inherited sex ratio in the parasitoid
wasp Nasonia vitripennis. Science 215: 1133-34.
STADDON, J. E. R., ed. 1980. Limits to Action: The Allocation of Indi-
vidual Behavior. Academic Press, N. Y.
STALKER, H. D. 1961. The genetic systems modifying meiotic drive in
Drosophila paramelancia. Genetics 46: 177-202.
STOLTZ, D. B., AND VINSON, S. B. 1979. Viruses and parasitism in insects.
Adv. Virus Research 24: 125-71.
TRIVERS, R. L., AND H. HARE. 1976. Haplodiploidy and the evolution of so-
cial insects. Science 191: 249-63.
-- AND D. E. WILLARD. 1973. Natural selection of parental ability
to vary the sex ratio of offspring. Science 179: 90-92.
UYENOYAMA, M., AND M. W. FELDMAN. 1978. The genetics of sex ratio
distortion by cytoplasmic infection under maternal and contagious
transmission: an epidemiological study. Theor. Pop. Biol. 14: 471-97.
WARNER, R. R. 1982. Mating systems, sex change and sexual demography

March, 1983

Insect Behavioral Ecology-'82 Frank 75

in the Rainbow Wrasse, Thalassoma lucasanum. Copeia 1982: 653-61.
AND S. G. HOFFMAN. 1980. Local population size as a determinant
of mating system and sexual composition in two tropical marine
fishes (Thalassoma spp.). Evolution 34: 508-18.
WERREN, J. H. 1980a. Sex ratio adaptations to local mate competition
in a parasitic wasp. Science 208: 1157-59.
1980b. Studies in the evolution of sex ratios. Ph. D. Thesis, Uni-
versity of Utah, Salt Lake City.
in press. Sex ratio evolution under local mate competition. Evolu-
S. W. SKINNER, AND E. L. CHARNOV. 1981. Paternal inheritance of
a daughterless sex ratio factor. Nature 293: 467-68.
WHITE, M. J. D. 1973. Animal Cytology and Evolution, 3rd ed. Cambridge
Univ. Press, Cambridge.
WHITING, A. R. 1967. The biology of the parasitic wasp, Mormoniella vitri-
pennis. Quart. Rev. Biol. 42: 333-406.
WIEBES, J. T. 1979. Co-evolution of figs and their insect pollinators. Annu.
Rev. Ecol. Syst. 10: 1-12.
WILSON, D. S. 1980. The Natural Selection of Populations and Communi-
ties. Benjamin/Cummings Publishing Co., Menlo Park, California.
AND R. K. COLWELL. 1981. Evolution of sex ratio in structured
demes. Evolution 35: 882-97.
WILLIAMS, G. C. 1966. Adaptation and Natural Selection. Princeton Univ.
Press,j Princeton.
1979. The question of adaptive sex ratio in outcrossed vertebrates.
Proc. Royal Soc. London B. 205: 567-80.
WRIGHT, S. 1969. Evolution and the Genetics of Populations, Vol. 2. Uni-
versity of Chicago Press, Chicago.
WYLIE, H. G. 1965. Discrimination between parasitized and unparasitized
house fly pupae by females of Nasonia vitripennis (Walker) (Hy-
menoptera: Pteromalidae). Canadian Ent. 97: 279-86.
ZIMMERING, S., L. SANDLER, AND B. NICOLETTI. 1970. Mechanisms of meiotic
drive. Annu. Rev. Genet. 4: 409-36.

Florida Entomologist 66 (1)



"October 15th, 1936. Referring to that point about high death
rate in captive broods, it is of interest to note that when this took
place there was an almost equal death rate in wild larvae left on
trees-death not due to parasitism."
"I believe that this is one method of natural control by means of
a change in the chemical composition of the foliage-retarding larval
growth and even causing death."
Quoted from letter to Sir Edward Poulton from V.G.L. van Someren in
Nairobi. (van Someren, 1937). His insight has been overlooked.

The diurnal members of the moth family Uraniidae are familiar insects
because most "coffee table" books on the Lepidoptera figure the spectacu-
lar Chrysiridia madagascariensis, the uraniid of that island, and almost
all depict one of the forms of Urania described in this paper (Fig. 1). These
handsome black and green Papilio-like moths sometimes have spectacular
annual flights throughout the Neotropics (Williams, 1958; Skutch, 1970;
Smith, 1972 and 1982). The numbers involved varied greatly from year to
year (Fig. 2 and Smith, 1982). Urania are not always nomadic, and pop-
ulations with all life stages may be found in seasonal wet forest during
every month of the year. They burst out of such areas, usually in July or
August in Central America and northern South America (Fig. 3). "Return"
flights that can be detected by casual observation are not annual nor wide-
spread. Thus the obvious conclusion-not all moths participate in the an-
nual "emigration" flight.1 Why do some, often many, emigrate each year?
Why do some not emigrate? I think I have discovered part of the answer.
The sole larval food plants of Urania moths (including the Chrysiridia
of Madagascar, Catala, 1940) are species of Omphalea (Euphorbiaceae)
which are mainly big, woody lianas, though some are trees. The distribu-
tion of the Uraniidae coincides exactly with that of Omphalea spp., and is
peculiar: the lowland Neotropics including Cuba and Jamaica, extreme
eastern Tanzania, Madagascar, Papua, and the southern Philippines. No
other insect to my knowledge regularly eats the leaves of Omphalea, which
are apparently full of defensive compounds, although some beetle nips
holes in the youngest leaves.2 In the wild adult Urania apparently feed
only on nectar of certain species with fluffy-white mimosoid-like flowers
like Inga or Eupatorium spp. (Smith, 1982). If they do not feed as adults
they do not breed and they die.3
Female Urania lay eggs on the undersides of the leaves usually in

*Neal Smith is a biologist with the Smithsonian Tropical Research Institute in Panama.
His research interests are in the evolutionary biology of birds and insects.
Current address: Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Re-
public of Panama or Smithsonian Tropical Research Institute, APO Miami 34002.

March, 1983

Insect Behavioral Ecology-'82 Smith


Fig. 1. The Central American form, U. fulgens, of the day-flying moth
clutches of about 80 but single eggs and small clutches are also found.
There are usually five instars; the early ones graze the surface of the
leaves while the later ones are capable of eating an entire leaf. The typical
generation time from egg to adult is about 46 days (Smith, 1982). Up to
seven generations a year may be produced in a big flight year, with three

Florida Entomologist 66 (1)



FLIGHTS II11111111111111T

01' 07' 15 24' 32' 40' 46' 5455' 61'64' 69' 73' 77' 80'81
I --X- 7.5 YEARS-- |
--X13.5 YEARS---


Fig. 2. Summary of population fluctuations and emigratory flights of
Urania moths in Central America and northern South America.
or four of these coming out during the nomadic period. Adults have lived
up to 66 days in the lab and I would not be surprised if they lived at least
as long as this in the wild.
0. diandra is the only species of Omphalea in lower Central America
and South America. This often huge woody liana sends its tendrilled
branches above the supporting canopy in swamp forests behind mangroves,
in forests at the edge of sandy beaches, and in deep monsoon-type forests
far removed from the ocean. I suspect that it occurs in patches separated
from other such pockets by many kilometers. The ranges of Urania and
Omphalea do not coincide at all times. In the dry season Urania (as adults
or larvae) are found only in those OQnphalea "patches" located in wet
relatively seasonal areas (eg. certain areas on the Atlantic side of Central
America). And then not all "patches" in those areas will contain Urania.
It is only during the nomadic phase that Urania are capable of nearly
filling up the entire range of Omphalea. There are "patches" of low-lying
Omphalea on both sides of Panama that I have repeatedly censused for
Urania and found none for 3-6 years. But, in all such areas, Urania even-
tually moved in, sometimes persisting for several years and in others dis-
appearing within months. There is yet no clear picture of how Omphalea
is distributed.
The results of Urania rearing experiments from 1973 suggested that
some aspect of the Urania caterpillars' consumption of Omphalea leaves
determined whether they would emerge as emigratory or sedentary adults.4
A rearing experiment in which Jeff Waage and I put field-collected larvae
and adults on clones of Omphalea was responsible for focusing attention
on a key to the solution. The adults laid eggs and these hatching larvae
amid ones from the wild were placed on potted clones of Omphalea "A"
(collected in 1970). They did well (too well in fact, for we feared that

March, 1983

Insect Behavioral Ecology-'82 Smith



Fig. 3. Distribution of Urania moths and their general flight directions
when in nomadic phase. The former population in Jamaica is now appar-
ently extinct. The Cuban form differs only slightly in appearance from
the others. It is said to also undergo population fluctuations and north-
south flights.

0a* ,

Florida Entomologist 66 (1)

they would consume all potted clones!). In the field Omphalea "A" had
been heavily hit by the 1973 flight but looked in good shape in 1974.5 We
brought back leaves from it to feed the larvae in the lab. There were no
larvae on it.6 Some lab larvae refused to cat it, others at it but grew slowly,
and others died. Assuming that the plant had been sprayed with insecticide
we put the remaining larvae back on the potted clones where they did
well. We noted that when larvae were moved from one Omphalea to an-
other (for more convenient observation) they often left the plant and
wondered why it was adaptive for larvae to "imprint" on the plant on
which they were hatched.
At this point by chance I had a conversation with a plant pathologist
to whom I described my Urania-Omphalea dilemma. "Sounds like a phyto-
alexin response (Cruickshank, 1963) or probably more like that wound re-
sponse' that Green and Ryan (1972) have described." I had expected that
plants would respond in evolutionary time to pest pressure and evolve chem-
ical defensive and other mechanisms. But the thought that a plant could
respond to grazing by the induced production of "anti-herbivore com-
pounds" in a matter of days or even hours was completely new and gave
me an entirely different view of the Omphalea problem. My records showed
that the larvae of 1974 were either on vines that had escaped attack (rea-
son?) for the previous three years, or were vines that I had missed in
my censuses.
The obvious experiment was to treat genetically identical clones to
different grazing regimes and to bioassay their acceptability on Urania
larvae. No matter where the eggs were laid, Urania larvae always climb
up to the most actively growing part of the vine and begin surface grazing
the youngest leaves. I grazed with a razor the leaves of 3 meter long clones.
At about day 10 larvae usually moult to 3rd instar and begin to disperse.
At this stage they can bite through a mature leaf. I mimicked this with a
paper punch (Fig. 4) and caused over the next 12-15 days as much damage
as 80 larvae would do before they pupate. I then waited the appropriate
time-a "Urania generation"-and repeated the process with some of the
clones again, and finally some clones received three grazings.
The new leaves of each of the four groups were fed to sib larvae after
the new leaves had appeared on the thrice grazed clones. Larvae grown
on leaves from thrice grazed clones grew significantly slower and had much
higher mortality rates than did larvae fed leaves from the other groups
(Smith, 1982, Fig. 5). Grazing could have resulted in an increase in
toxic compounds or a decrease in the nutritive quality of the clone. Since
the clones grew at about the same rate and there seemed to be no difference
in the color or size of the new leaves, I favor the toxicity hypothesis. How-
ever this experiment did not adequately control for the "recovery" of the
once or twice grazed plants since I tested the leaves only after the new
leaves came out on the thrice-grazed plants. The other clones could have
"recovered" during that time (about 220 days). My present doubts are
further reinforced by reviewing the mixed results that came from the
1973 experiments4 (Smith, 1982). Clearly leaf quality (grazing history)
was one of the major uncontrolled factors. It is "biologically reasonable"
that a plant ought to refrain from expending valuable energy on toxic
compounds unless provoked by repeated attacks. In addition, it is usually
the third generation after the last flight of the previous nomadic phase

March, 1983

Insect Behavioral Ecology-'82 Smith

Fig. 4. Experimental simulation of Urania grazing using a paper punch.

that becomes nomadic. If toxicity goes up slowly and comes down slowly,
this would suggest that the cycles in Urania (Fig. 2) reflect the chemical
cycles in Omphalea (Smith, 1982).7
It is interesting to note that the consequences of the grazing history
of a plant seem not only to act as the ultimate factor for emigration but
the apparent increase in level of toxic compounds may also be the prowi-
mate clue producing emigrating or non-emigrating individuals.4 Figure 6
suggests a model of Urania nomadic flights.

Florida Entomologist 66(1)

100 -







IN 20


0 I 2 3
Fig. 5. The effects of the number of previous grazings of genetically
identical clones of Omphalea diandra on the development and mortality of
Urania larvae which were eating new leaves.

Day-flying Neotropical Urania moths undergo cyclical population fluctu-
ations and massive one-way flights. There may be "return" flight several
months later, but these are scattered, not always annual, and composed of
relatively few recently emerged individuals that are usually in reproductive
diapause. Presumably not all moths emigrate in the southward flight. Why
do some leave and others remain? The sole larval food plant are species
of the woody, often huge lianas Omphalea spp (Euphorbiaceae). The dis-
tribution of Omphalea and Urania coincide exactly but not all patches of
Omphalea are occupied by Urania at any given time. Few insects and
no other Lepidopterian eat Omphalea. Imitating feeding of Urania larvae,

March, 1983

Insect Behavioral Ecology-'82 Smith 83




Fig. 6. A suggested model of Urania nomadic flights. Not all individuals
in a population emigrate, and not all populations contribute to the emi-
gration flight. Return flights may consist of few individuals and thus are
difficult to detect. These individuals may also be many generations removed
from the original immigrants.

I experimentally grazed with razor and later paper punched the leaves of
genetically identical clones of Omphalea. Larvae grew normally on the
controls, and on leaves from clones earlier "grazed" once or twice. But
when presented with leaves from clones "grazed" the third time, they
either refused to eat them, ate them and died, or grew slowly and had
higher pupal mortality. A hypothesis is proposed that the chemical re-
sponse of the plants also provides not only the ultimate cause but also
that the level of toxicity provides the proximate clue as to whether the
adults emigrate or not. They may be seeking vines which have received less
grazing and or have escaped grazing entirely for several years and thus
have lowered toxicity. I also suggest that it is the plant's response that
determines the four or eight year cycle in Urania.

'A one way migration poses some problems most of which I feel are
more apparent than real. If all the "migrating" genes moved out of an area,
say Veracruz, Mexico, how then does this area continue to give rise to suc-
cessive generations of migrants? One explanation is that one does not de-
tect the few individuals that move back from other locales to the south.
This need not occur very often, perhaps .every few years, but the females
are quite fecund and population build up could be very rapid. We simply
do not know how far an individual migrates, nor do we have any idea of
the genetics as they directly relate to the nomadic tendency.
2Leaf-cutting ants (Atta spp.) also occasionally attack. I have examined
specimens of all the described species and all possess extra floral nectaries
in one place or another on their leaves. In 0. diandra, pugnacious ants
(not Atta spp.) move up and down the liana seeking out secreting nec-
taries-yet another defensive line against herbivores.

Florida Entomologist 66 (1)

3The absence of an adult food source in an area could certainly be a
factor causing adults to leave. In 1978 I released several hundred recently
emerged adults in an area with much Omphalea but apparently lacking
flowers suitable to the adults. Approximately one half of the individuals
were fed before release on sugar water, honey and desolved buillon. All
of the adults disappeared within a few hours and were not seen again
during the following week.
4In those experiments I assumed that there was a migratory type and
non-migratory type, and that something during the larval period deter-
mined which emerged. Two sets of hypotheses were tested, first, abundant
larval food would give rise to non-migratory adults; scarcity would pro-
duce migratory adults. Second, low larval density would produce non-
migratory adults; high density would give rise to migratory adults. In
both cases the experimental groups behaved one way or the other, but
the results were neither consistent nor predictable.
5The flight of 1973 was a big one (Fig. 2). As noted by several authors
(e.g., Smith, 1972) the year following big flights is often (note exception
1954-55, Fig. 2) a poor one for Urania emigration. The obvious explana-
tion-that they ate all the Omphalec-I believe, is not, the correct one.
Plants that had been badly chewed up by October 1973 were completely
leafed out again by January 1974.
6Urania populations at nadir in both wild and lab. Gregarine parasites
were thought to be implicated, but later were judged not to be a serious
source of mortality. Larvae found to "imprint" on particular plants but
significance of such behavior not understood.
'Experiments would have been repeated in 1981 except that there was
no Urania flight in 1981, a unique occurrence with reliable data back to
1954, and anecdotal data back to 1900. The year 1981 was the wettest year
in recorded meteorogical history in Panama, but how this might have af-
fected Urania is not known.

CATALA, R. 1940. Variations Experimentales de Chrysiridia madagascarien-
sis Less. [Lep. Uraniidae].Archives Mus. National d'Histoire Na-
turelle. Sixieme Serie. Tome Dix-SeptiBme.
CRUICKSHANK, I. A. M. 1963. Phytoalexins. Annu. Review of Phytopathol-
ogy Vol. 1: 351-73.
GREEN, T. R. AND C. A. RYAN. 1972. Wound-induced Proteinase Inhibitor
in Plant Leaves: a possible defense mechanism against insects. Sci-
ence 175: 776-77.
PRESTON, F. W. 1969. Diversity and stability in the biological world, in
Diversity and stability in ecological systems. Brookhaven Symposia
in Biology. No. 22: 1-2.
SKUTCH, A. F. 1970. Migrations of the American moth Urania fulgens.
The Entomologist 103: 192-97.
SMITH, N. G. 1972. Migrations of the day-flying moth Urania in Central
and South America. Carib. J. Sci., 12,(1-2) : 45-58.
1982. Periodic migrations and population fluctuations by the Neo-
tropical day-flying moth Urania fulgens through the Isthmus of Pan-
ama. pp. 331-44 in The Ecology of a tropical forest: seasonal rhythms
and long-term changes. ed. by E. Leigh, A. S. Rand and D. Windsor.
Smithsonian Institution Press and Oxford Univ. Press.
VAN SOMEREN, V. G. L. 1937. Chemical changes in the food-plant: a cause
of failure in rearing larvae. Proc. R. Ent. Soc. Lond (A) 12. pt. 1-2.
WILLIAMS, C. B. 1958. Insect migration. Collins. London: 235 p.

March, 1983

Insect Behavioral Ecology-'82 Smith 85

URQUHART, F. A. 1960. The monarch butterfly. University of Toronto
Press: 361 p.

86 Florida Entomologist 66 (1) March, 1983



A variety of male insect structures are used to hold females during
mating including mandibles (Sivinski 1981), tergal gin traps (pinching
organs) (Morris 1979), genital claspers (Wing 1982), antennae and modi-
fied legs (Parker 1970). Such structures may prevent other males from tak-
ing over a receptive female (Parker 1970), and they may enable the male
to manipulate the female to his advantage-e.g. for forceful insemination
(Thornhill 1982; see also, Lloyd 1979). The hooked elytral tips of Pteroptyx
fireflies, long a puzzle to biologists who observed them and also the primary
criterion for separating the genus Pteroptyx from Luciola, are used to
clamp the female during mating in Pteroptyx valida, and probably other
species. Here we present the details of this clamping as it occurs in P.


At sites near Bangkok, Thailand, we collected fireflies by hand and by
sweeping tree foliage with an aerial net. Fresh specimens were dissected
and others were preserved in alcoholic Bouin's solution for sectioning. Pairs
in copula (duration unknown) were collected by gently clipping their leaf
perch from the tree and placing it in a cyanide jar. Dead pairs that re-
mained joined were transferred to alcoholic Bouin's solution. Four fixed
pairs were embedded in paraffin and sectioned (10 microns), two longi-
tudinally and two in cross section. Slides were stained with Delafield's
hematoxylin and eosin. Four individual males and four females were pre-
pared the same way.


Copulations occur in certain trees where P. valida fireflies congregate.
Males perch on foliage of a tree and flash signals, and a female, having
joined a particular male on his perch, is courted by the male (Lloyd et al.,
in press; see also, Case 1980; and Lloyd 1973). Once the female allows
intromission, control over the congress passes to the male because of the use
of his clamp.
In valida males, and probably males of other Pteroptyx species, a "geni-
tal pocket" opens at the tip of the abdomen (Fig 1; Lloyd et al., in press).
The pocket is formed by the terminal abdominal sternite and tergite which

*Steve Wing is a graduate student in the Dept. of Entomology and Nematology at the
University of Florida. His research interests include phenomena associated with sperm
transfer in insects.
James E. Lloyd is a Professor of Entomology at the University of Florida, Gainesville,
Fla. 32611
Tawatchai Hongtrakul works at the Pesticide Research Laboratory, Dept. of Agriculture,
Bangkhen, Bangkok 9, Thailand, where he supervises scouting programs in field crops. His
research is concerned with developing guidelines for the use of pesticides.
Florida Agriculture Experiment Station Journal Series No. 4469.

Insect Behavioral Ecology-'82 Wing, Lloyd, Hongtrakul 87

Fig. 1. Pteroptyx valida male. Note the hooked elytral tip and "genital
pocket" in the terminal abdominal segment.

are not joined distally. The male genitalia are situated within the pocket,
and are covered by a two-piece sheath when not in use. The location of the
aedeagus requires that the female terminalia enter the male's pocket for
intromission. When the female terminalia are in the male's genital pocket
and the two are turned tail-to-tail (Fig. 2; Lloyd and Wing 1981), the
female is gripped in the jaws of the male's copulatory clamp. The lower jaw
of the clamp is the terminal abdominal sternite of the male (Fig. 3). The
hind margin of this sternite curves upward, forming a lobed ridge (Bal-
lentyne and McLean 1970), and this heavily sclerotized ridge presses up-
ward on the venter of the female's abdomen during copulation. The oppos-
ing (upper) jaw of the clamp is the male's strongly deflexed elytral tips
(Fig. 1 and Fig. 3). (The deflexed elytral tips, found only on males, are the
characteristic from which the genus takes its name: Ptero=wing, ptyx=
fold.) During mating the male elytra are positioned under those of the
female (Fig. 2), and the male hooks his elytral tips around the anterior
margin of the female's sixth abdominal tergite. In the grip of male's elytral
hooks dorsally and his sternite ventrally, the female abdomen is doubly
bent (Fig. 3). Thus, the female vagina and abdominal tip are swallowed
and clamped by male structures.
In valid females a bursa copulatrix (Fig. 4), similar to those found
in certain other beetles (Becker 1956a,b; Surtees 1961) receives the male
ejaculatory products. This chamber is reinforced by sclerotized plates and

A dA

Fig. 2. A copulating P. valida pair: orientation for Fig. 3. Area "A"
(delimited by arrows) shows the region shown in longitudinal section in
Fig. 3.

Florida Entomologist 66 (1)

.... ......" elytron

bursa terminal tergite
Santerior --

plug (dorsal)

f terminal ster te-- .... ..
S aedeagus
spermatheca ". 9erminalia

S9 anterior
Fig. 3. Longitudinal section of a mating P. valida male and female. See
Fig. 2 for orientation. Note that the female's elytra, which normally rest
above the male's elytra as in Fig. 2, are not shown.

fishbone-like rods that probably prevent damage to the female by the male
clamp, for it is in this region of the abdomen where the lower jaw of the
male clamp is applied. The elytral hooks of the male press at the region
of the abdomen where the anterior end of the bursa is situated (Fig. 3).
The spermathecal duct and a large female accessory gland join the
bursa at its anterior end, and the common oviduct joins posteriorly (Fig. 4).
Dissections of fresh specimens showed that the bursae of mated females con-
tain a rubbery "plug" that is probably the remains of a spermatophore
uood to carry the sperm from the aedeagus to the spermatheca (Fig. 4).
The plug consists of a rubbery layer enclosing a fluid core. The male repro-
ductive system includes three pairs of accessory glands, products of which
form the plug. No mated female, including those kept alive for up to five
days before dissecting, was found to be without a plug. This suggests that
a plug remains in place at least semi-permanently, and perhaps for the
life of the female. A plug is carried in the anterior, expanded end of the
bursa (Fig. 4) where it does not interfere with egg passage. The plug
may prevent subsequent insemination of the female by other males due
to blocking access to the spermatheca (see Gregory 1965; Parker 1970).
The inner surface of the bursa has sclerotized teeth that may hold the
plug away from the wall of the bursa, creating a space through which
sperm can migrate from the spermatheca to the eggs at the opposite (pos-
terior) end of the bursa (see Hinton 1964, for a discussion of such teeth).
When female insects mate more than once, male competition results in
tactics of many forms. Males reduce competition from sperm of subsequent
mates by using mating plugs, mate guarding, and prolonged copulation
(Parker 1970, Sivinski 1983a). Males overcome such measures by pre-
vious males by removing mating plugs, etc. (Parker 1970). Males reduce

March, 1983

Insect Behavioral Ecology-'82 Wing, Lloyd, Hongtrakul 89


position of plug


sclerotized plate


common oviduct-

Fig. 4. Internal reproductive structures of a P. valid female.

competition from previous males by removing (Waage 1979) or prevent-
ing the use of their sperm (Parker 1970, Wing 1982). Females have a
reproductive interest in the matter and may act to overcome male tactics
that keep females from realizing benefits of multiple mating (see Walker
1980, Sivinski 1983b, and Sakaluk and Cade 1983 for discussion of such
benefits). Males may counter these female actions, and so on (see Lloyd
P. valida copulations occur amidst high densities of intensely compet-
ing males. The clamp used by valida males may function in the manipula-
tion of the female during placement of the plug, resulting in a reduced
chance of subsequent insemination of the female. The clamp may also pre-
vent take-over of the female by other males during copulation. Whatever
its function, the valida clamp is the first report of a male insect using wing
covers to control female genitalia during copulation.

Pteroptyx valida males use their hooked wing covers to clamp females
during mating. Female internal reproductive anatomy includes reinforced
structures in the region where the clamp presses. The male clamp may
have a role in positioning the spermatophore within the female, and/or
the clamp may prevent other males from taking over the female.

Florida Entomologist 66 (1)


This work was made possible by financial support from the National
Geographic Society (Grant #221680). Assistance from Prayoon Deema,
Nuan Chan, James L. Nation, Kathy Dennis, Thelma Carlysle, and Donald
W. Hall is gratefully acknowledged. The fireflies were drawn by Laura
Reep. Internal structures were drawn by Susan Wineriter. Helpful reviews
of the manuscript were made by Tim Forrest, John Sivinski, and Tom


BALLENTYNE, L. A., AND M. R. MCLEAN. 1970. Revisional studies on the
firefly genus Pteroptyx Olivier (Coleoptera: Lampyridae: Lucio-
linae: Luciolini). Trans. American Ent. Soc. 96: 223-305.
BECKER, E. C. 1956a. The phyletic significance of the female internal organs
of reproduction in the Elateridae. Proc. 10th Int. Cong. Ent. 1:
BECKER, E. C. 1956b. Revision of the Nearctic species of Agriotes. Canadian
Ent. 88: 1-101.
CASE, J. F. 1980. Courting behavior in a synchronously flashing, aggrega-
tive firefly, Pteroptyx tener. Biol. Bull. 159: 613-625.
GREGORY, G. E. 1965. The formation and fate of the spermatophore in the
African migratory locust Locusta migratoria migratorioides Reiche
and Fairmaire. Trans. R. Ent. Soc. London 117: 33-66.
HINTON, H. E. 1964. Sperm transfer in insects and the evolution of
haemocoelic insemination. Pages 95-107 in K. C. Highnam, ed. Insect
reproduction. R. Ent. Soc. London Symp. 2.
LLOYD, J. E. 1973. Fireflies of Melanesia: bioluminescence, mating behav-
ior, and synchronous flashing (Coleoptera: Lampyridae). Environ.
Ent. 2: 991-1008.
LLOYD, J. E. 1979. Mating behavior and natural selection. Florida Ent.
62: 17-34.
LLOYD, J. E. 1981. Sexual selection: individuality, identification, and rec-
ognition in a bumblebee and other insects. Florida Ent. 64: 89-118.
LLOYD, J. E., AND S. R. WING. 1981. Photo story: copulating Pteroptyx
valida fireflies. Florida Ent. 64: 459.
LLOYD, J. E., S. R. WING, AND T. HONGTRAKUL. 1980. The behavioral ecol-
ogy of Thailand fireflies. Nat. Geog. Rep. (in press).
MORRIS, G. K. 1979. Mating systems, paternal investment, and aggres-
sive behavior of acoustic Orthoptera. Florida Ent. 62: 9-16.
PARKER, G. A. 1970. Sperm competition and its evolutionary consequence
in the insects. Biol. Rev. 45: 525-567.
SAKALUK, S. K., AND W. H. CADE. 1983. The adaptive significance of fe-
male multiple mating in house and field crickets. In D. T. Gwynne
and G. K. Morris, eds. Orthopteran mating systems: sexual competi-
tion in a diverse group of insects. ,Westview Press, Boulder, Colorado.
SIVINSKI, J. 1981. "Love bites" in a Lycid beetle. Florida Ent. 64: 541.
SIVINSKI, J. 1983a. Predation and sperm competition in the evolution of
coupling durations, particularly in the walkingstick Diapheromera
veliei. In D. T. Gwynne and G. K. Morris, eds. Orthopteran mating
systems: sexual competition in a diverse group of insects. Westview
Press, Boulder, Colorado. (in press)
SIVINSKI, J. 1983b. Sperm in competition. In R. L. Smith, ed. Sperm com-
petition in the evolution of animal mating systems. Academic Press,

March, 1983

Insect Behavioral Ecology-'82 Wing, Lloyd, Hongtrakul 91

New York. (in press)
SURTEES, G. 1961. Spermathecal structures in some Coleoptera associated
with stored products. Proc. R. Ent. Soc. London (A) 36: 144-152.
THORNHILL, R. 1983. Rape in Panorpa scorpionflies and a general rape hy-
pothesis. (in press)
WAAGE, J. K. 1979. Dual function of the damselfly penis: sperm removal and
transfer. Science 203: 916-918.
WALKER, W. F. 1980. Sperm utilization strategies in non-social insects.
American Nat. 115: 780-799.
WING, S. R. 1982. The reproductive ecologies of three species of fireflies.
M. S. Thesis. Univ. of Florida, Gainesville. 68 pp.

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs