Complex social behaviour can select for variability in visual features: a case study in Polistes wasps.

Abstract

The ability to recognize individuals is common in animals; however, we know little about why the phenotypic variability necessary for individual recognition has evolved in some animals but not others. One possibility is that natural selection favours variability in some social contexts but not in others. Polistes fuscatus wasps have variable facial and abdominal markings used for individual recognition within their complex societies. Here, I explore whether social behaviour can select for variability by examining the relationship between social behaviour and variability in visual features (marking variability) across social wasp taxa. Analysis using a concentrated changes test demonstrates that marking variability is significantly associated with nesting strategy. Species with flexible nest-founding strategies have highly variable markings, whereas species without flexible nest-founding strategies have low marking variability. These results suggest that: (i) individual recognition may be widespread in the social wasps, and (ii) natural selection may play a role in the origin and maintenance of the variable distinctive markings. Theoretical and empirical evidence suggests that species with flexible nesting strategies have reproductive transactions, a type of complex social behaviour predicted to require individual recognition. Therefore, the reproductive transactions of flexible species may select for highly variable individuals who are easy to identify as individuals. Further, selection for distinctiveness may provide an alternative explanation for the evolution of phenotypic diversity.

Full text

Page 1

doi: 10.1098/rspb.2004.2784
, 1955-1960
271
2004 Proc. R. Soc. Lond. B


wasps
Polistes
features: a case study in
Complex social behaviour can select for variability in visual


Email alerting service

here
right-hand corner of the article or click
Receive free email alerts when new articles cite this article - sign up in the box at the top

http://rspb.royalsocietypublishing.org/subscriptions
go to: Proc. R. Soc. Lond. B To subscribe to
This journal is © 2004 The Royal Society
on May 24, 2011rspb.royalsocietypublishing.orgDownloaded from

Page 2

Received 15 April 2004
Accepted 27 April 2004
Published online16 August2004
Complex social behaviour can select for variability in
visual features: a case study in Polistes wasps
Elizabeth A. Tibbettsy
Department of Neurobiology and Behavior, Cornell University, Ithaca,NY 14853, USA
The ability to recognize individuals is common in animals; however, we know little about why the pheno-
typic variability necessary for individual recognition has evolved in some animals but not others. One possi-
bility is that natural selection favours variability in some social contexts but not in others. Polistes fuscatus
wasps have variable facial and abdominal markings used for individual recognition within their complex
societies. Here, I explore whether social behaviour can select for variability by examining the relationship
between social behaviour and variability in visual features (marking variability) across social wasp taxa.
Analysis using a concentrated changes test demonstrates that marking variability is significantly associated
with nesting strategy. Species with flexible nest-founding strategies have highly variable markings, whereas
species without flexible nest-founding strategies have low marking variability. These results suggest that: (i)
individual recognition may be widespread in the social wasps, and (ii) natural selection may play a role in the
origin and maintenance of the variable distinctive markings. Theoretical and empirical evidence suggests
that species with flexible nesting strategies have reproductive transactions, a type of complex social behav-
iour predicted to require individual recognition. Therefore, the reproductive transactions of flexible species
may select for highly variable individuals who are easy to identify as individuals. Further, selection for dis-
tinctiveness may provide an alternative explanation for the evolution ofphenotypic diversity.
Keywords: individual recognition; skew theory; ornamentation; dominance hierarchies
1. INTRODUCTION
The phenotypic diversity within a single species can be
remarkable, from the polymorphic underwings of Catocala
moths (Kamil & Bond 2001), to the white and blue colour
phases of the lesser snow goose (Cooke et al. 1995), to the
markedly different sizes and shapes of leaf-cutting
ants (Ho ¨lldobler & Roces 2001). Many different selection
pressures are thought to cause the phenotypic diversity
within species, ranging from sexual selection to frequency-
dependent selection. Benefits derived from being individu-
ally distinctive and easily recognizable may offer an
important, but largely neglected, explanation for the origin
and maintenance of phenotypic diversity (Beecher 1989;
Johnstone 1997; Daleet al.2001).
Individualrecognitionoccurs
modalities,includingolfactory,
(reviewed in Dale et al. 2001). In every modality, the trait
used for recognition must be highly variable to allow accu-
rate discrimination of individuals (Dale et al. 2001). There
are two possible explanations for the evolution of the
phenotypic variability necessary for individual recognition:
(i) no selection favouring distinctiveness: the variability
could be neutral or functional in a non-recognition context
and have been co-opted for recognizing conspecifics; and
(ii) selection favouring distinctiveness: the variability could
benefit individuals by making them readily recognizable
and this benefit may have driven the evolution of pheno-
typic variability. One way to differentiate between these
alternatives is to test whether social behaviour predicted to
inmany
and
different
acousticvisual
involve individual recognition is associated with frequent
gains and/or infrequent losses of variability. This type of
association supports the hypothesis that variability has
evolved because distinctiveness benefits individuals of spe-
cies with this social behaviour.
Theoretical work suggests that several types of complex
social behaviour could select for the kind of variable mark-
ings necessary for individual recognition, including terri-
toriality (Ydenberg et al. 1988), reciprocal altruism
(Crowley et al. 1996), monogamous pairing (Dale et al.
2001), dominance (Barnard & Burk 1979; Van Rhijn &
Vodegel 1980; Dawkins & Guilford 1991) and repro-
ductive transactions (when group members yield repro-
duction to each other in exchange for benefits in a manner
predictedbytransactional
relationship between variability and social behaviour will
also provide valuable insight into which behaviours actually
select for distinctive, recognizable individuals, as little
empirical research hasaddressed this topic (Beecher 1989).
Social wasps in the genus Polistes are an attractive taxon
in which to examine whether social behaviour can select for
distinctiveness. First, it is known that individual recog-
nition occurs in at least one species in this genus. In Polistes
fuscatus, wasps use variation in facial and abdominal mark-
ings to recognize their nest-mates as individuals (Tibbetts
2002). Second, experimental results suggest that the social
behaviour of P. fuscatus may favour wasps with variable dis-
tinctive markings; easily recognizable wasps receive less
aggression from their nest-mates than unrecognizable
wasps (Tibbetts 2002). Third, Polistes is a large genus and
one finds diverse social behaviours of varying complexity
among its many species. Although specific information on
social behaviour is available for only a few Polistes species,
models). Examining the
yPresent address: ARL, Division of Neurobiology, University of Arizona,
Tucson, AZ85721,USA (eat11@email.arizona.edu).
Proc.R. Soc.Lond.B (2004)271,1955–1960
doi:10.1098/rspb.2004.2784
1955
#2004TheRoyalSociety
on May 24, 2011rspb.royalsocietypublishing.orgDownloaded from

Page 3

models and empirical data indicate that nest-founding
strategy is a useful proxy for social behaviour (Reeve & Kel-
ler 2001), and nest-founding strategies are well docu-
mented. Species have one of three nest-founding strategies:
multiple foundress (all nests are started by multiple foun-
dresses) single foundress (all nests are started by a single
foundress), or flexible (nests can be started by a single
foundress or by multiple foundresses). All Polistes species
have workers for much of the colony cycle. Workers and
queens have identical marking variability (Tibbetts 2002),
as workers are fertile and may become queens (Reeve et al.
1998). Here, I test for a relationship between nest-found-
ing strategyand markingvariability across the Polistes.
(i) If variable markings are associated with all nesting
strategies, no conclusion can be drawn, but the pat-
tern suggests that worker interactions may be suf-
ficient toselect forvariability.
(ii) If there is no association between variable markings
and nesting strategies, nesting strategy may not be a
useful proxy for social behaviour or social behaviour
may not haveselected forvariability in the Polistes.
(iii) A significant association between nesting strategy and
variable markings (a) suggests a role for natural selec-
tion in the origin and maintenance of variable, dis-
tinctive markings, and (b) provides insight into the
type of social behaviour that may select for distinc-
tiveness.
2. METHODS
(a) Variability data
Information on the marking variability of each species was col-
lected by examining specimens from the Polistes collections at the
Museum of Natural History in New York, and Cornell University
in Ithaca, NY. All variability data were collected blind with respect
to species and social behaviour. Therefore, I was able to assess the
repeatability of variability scores by independently scoring the
variability of seven species in each museum. Variability was scored
only when at least 10 specimens of a species were available (the
mean number of specimens examined was 58; range 10–200). I
collected data on facial and abdominal variability, as both are used
for individual recognition in P. fuscatus (Tibbetts 2002). The
number of marking variants in a species was determined by count-
ing the number of facial or abdominal areas that varied and the
number of different marking variants for each of these areas.
Across the species examined, I found five discrete facial areas that
varied: inner eye, outer eye, clypeus edge, middle clypeus and
‘eyebrow’ (the area dorsal to the antennae; figure 1). There were
nine discrete abdominal areas that varied: the middle of each
abdominal segment (1–5) and the edge of abdominal segments 1–
4. I calculated the number of possible combinations of variable
areas and variants to get each species’ total variability. For
example, in figure 1, there are three areas that vary and two var-
iants for each of these areas. Therefore, this species has 23¼ 8
total variants. This method of assessing a species’ variability
assumes that marking areas vary independently of each other, an
appropriate assumption given the distribution of markings in P.
fuscatus (Tibbetts 2002). Further, this method of scoring a spe-
cies’s variability provides an estimate of the maximum variability
of each species rather than the exact number of variants in each
species. Species with less than five variants on either face or abdo-
men were categorized as low variability species. All species with
five or more variants in either face or abdomen were termed high
variability species. I chose five variants as the cut-off point
between low and high variability species as I was looking for the
exceptional variability expected in a species selected for distinc-
tiveness; accurate discrimination of individuals requires that all
individuals on a nest have different markings. This cut-off point
was determined after gathering all variability data, but was blind
with respect to the variability of particular species and their social
behaviour. Most low variability species had one variant and the
most high variability species had eight or more variants, so the
results are not sensitive to the specific cut-off point. If eight
variants were used as the cut-off, the mean probability of the
observed association between variability and social behaviour
occurring by random chance is 0.010. If three variants were used
as the cut-off, the mean probability is 0.066. Therefore, any cut-
off between three and eight yields similar results.
Variability data were collected on 69 species of Polistes; 14 spe-
cies of Mischocyttarus, the genus most closely related to Polistes
(Carpenter 1996), were also examined to provide an outgroup.
(b) Behaviouraldata
Searching the literature yielded data on nest-founding strategies
of 25 Polistes species. I defined multiple-foundress species as those
in which more than 98% of nests were started by multiple foun-
dresses, and single-foundress species as those in which more than
98% of nests were started by a single foundress. Species with
intermediate rates of singly and multiply founded nests were cate-
gorized ashaving ‘flexible’ nest-founding strategies.
(c) Phylogenetic analysis
Carpenter’s (1996) taxon-area phylogeny was used for the
analysis. His tree is based on a consensus tree that used 33 mor-
phological characters and species group distributions. Beha-
vioural data on 25 Polistes species and variability data on 69 Polistes
species plus an outgroup were mapped onto the tree using
MCCLADE (Maddison & Maddison 1992). The hypothesis that
eyebrow
inner eye
clypeus edge
middle clypeus
outer eye
Figure 1. Illustration of fourdifferent waspfaces. The five
general facial areas that varyacrosswasp taxa arelabelled.
1956E. A. Tibbetts
Complex social behaviour and visual featuresin Polistes wasps
Proc.R. Soc.Lond.B (2004)
on May 24, 2011rspb.royalsocietypublishing.orgDownloaded from

Page 4

changes in marking variability (dependent variable) are associated
with nest-founding strategy (independent variable) was tested
using Maddison’s (1990) concentrated changes test. This test
examines whether changes in marking variability occur more fre-
quently on branches where species have the predicted nest-found-
ing strategy or an unknown nest-founding strategy than expected
by chance. The null hypothesis is that gains and losses of variable
markings are equally probable on every branch. Specifically, the
test addresses whether a flexible nesting strategy facilitates the
evolution of variability (if gains of variability are associated with
flexible nesting strategies) and/or the maintenance of variability
(losses of variability are not associated with flexible nesting strat-
egy). The number of gains and losses of variability was estimated
by MCCLADE, using the most parsimonious reconstruction of
character evolution (figure 2). The concentrated changes test is
sensitive to the number of gains and losses, so the data were also
analysed with the most conservative numbers of gains and losses
to ensure that the results were robust. Carpenter’s (1996) phy-
logeny contains polytomies. As the concentrated changes test can-
not be performed with polytomies, all polytomies were randomly
resolved by MCCLADE. Then the concentrated changes test was
run for each of five randomly selected resolutions. The exact prob-
ability of the observed relationship occurring by chance was calcu-
lated by MCCLADE in most analyses. In two cases, exact values
could not be calculated, so probabilities were estimated with a
1000-iteration simulation.
(d) Correlation analysis
The relationship between marking variability and nesting strat-
egy was also analysed independently of phylogeny by using a v2
analysis.
3. RESULTS
There was no relationship between the number of speci-
mens examined and the classification of a species with
respect to marking variability (high or low) (p ¼ 0:32,
t ¼ 1, d:f: ¼ 47). Therefore, the number of specimens
examined didnot bias the variability categorizations.
Specimens of seven species were examined in both
museums to determine the repeatability of variability cate-
gorizations. Three species were categorized as highly vari-
able in both museums, and four species were categorized as
low variability in both museums (Fisher’s exact test:
p ¼ 0:028). Therefore, each sample was an accurate rep-
resentation of that species’ true variability and my method
of scoring variability yielded consistent results, even
between museums.
The distribution of marking variability in 69 species of
Polistes is shown in figure 2. The ancestral state is low varia-
bility. Highly variable markings evolved a minimum of five
times. The distribution of marking variability and nest-
founding strategy in 25 Polistes species is shown in figure 3.
All eight species with highly variable markings have a flex-
ible nesting strategy, but there are four species with a
Figure 2. The distribution of marking variability among69
Polistes species using Carpenter’s (1996)phylogeny. Black
indicates lowvariability. Grey indicates high variability. The
most parsimonious reconstruction of variability is shown.
Analysis wasalso performed using less parsimonious
reconstructions of variability (markedwith Aand B)to ensure
theresults arerobust.
Mischocyttarus
sulcatus
gigas
jadwigae
olivaceus
rothneyi
schach
tepidus
fastidiosus
humilis
japonicus
sagittarius
saussurei
smithii
snelleni
stigma
africanus
albincinctus
chinenesis
gallicus
marginalis
riparius
tenellus
annularis
aterrimus
buyssoni
camanchus
canadensis
cavapyta
dominicus
erythrocephalus
infuscatus
kaibabensis
lanio
mexicanus
actaeon
bicolor
billardieri
cinerascens
pacificus
occipitalis
subsericeus
testaceicolor
carnifex
arizonensis
bahamensis
consobrinus
crinitus
cubensis
exclamans
goeldii
incertus
instabilis
minor
peruvianus
simillimus
stabilinus
versicolor
weyrauchorum
xanthogaster
major
apachus
bellicosus
carolina
dorsalis
flavis
fuscatus
metricus
perplexus
poeyi
B
A
Complex social behaviour and visualfeatures inPolistes wasps
E. A.Tibbetts 1957
Proc.R. Soc.Lond.B (2004)
on May 24, 2011 rspb.royalsocietypublishing.orgDownloaded from

Page 5

flexible nesting strategy that do not have highly variable
markings (P. olivaceous, P. lanio, P. metricus and P. car-
olina).
Table 1 gives the probabilities that the observed associ-
ation between changes in variability and nest-founding
strategy would occur through random chance. (Prob-
abilities are given for three reconstructions of ancestral
states and five random resolutions of the polytomies seen in
figure 3.) The probability of the observed association
occurring through random chance is below 0.05 for every
analysis.
Similar results are found when the relationship between
nest-founding strategy and variability is analysed using a v2
analysis. There are 13 species with non-variable markings
and non-flexible nesting strategies, eight species with vari-
able markings and flexible nesting strategies, four species
with non-variable markings and flexible nesting strategies,
and zero species with variable markings and a non-flexible
nesting strategy. This distribution is significantly different
from random (v2¼ 12:7, p ¼ 0:0004). If the non-flexible
species are more specifically categorized into single-
foundress and multiple-foundress species, the statistical
difference between nesting strategy and variability is also
significantly different from
p ¼ 0:0017; 10 non-variable single-foundress species,
three non-variable multiple-foundress species, zero vari-
able single-foundress species, and zero variable multiple-
foundress species). Therefore, flexibly nesting species are
significantly more likely to have variable markings, regard-
less of whetherone controls forphylogeny.
random(v2¼ 12:7,
4. DISCUSSION
There are eight Polistes species with the kind of highly vari-
able markings necessary for individual recognition, sug-
gesting that individual recognition is widespread in the
Polistes. Further, gains of marking variability are signifi-
cantly clustered in areas of the phylogeny where species
have an unknown or flexible nest-founding strategy (foun-
dresses have the option of starting a nest alone or in a
group). In areas of the phylogeny where foundresses always
nest alone or always nest in a group, all species have
invariant markings. There is also a significant association
between nesting strategy and marking variability inde-
pendent of phylogenetic relationships; flexible species have
variable markings and non-flexible species have non-vari-
ablemarkings.
The robust association between marking variability and
flexible nesting strategies suggests that variable markings
have evolved because some aspect of flexible nest-founding
selects for distinctive individuals. The complex social
behaviour of flexible species is the most plausible selection
pressure, as markings are known to be a social signal, unob-
trusive without close interaction, and not associated with
physiological or behavioural differences (Tibbetts 2002).
Further, P. fuscatus who are easily recognizable received
less aggression from their nest-mates than P. fuscatus who
are unrecognizable, suggesting that being recognizable may
provide social benefits (Tibbetts 2002). Finally, theory and
empirical results indicate that flexible species (such as P.
fuscatus) have reproductive transactions (reviewed in Reeve
& Keller 2001), precisely the type of complex social behav-
iour predicted to favour distinctiveness (Beecher 1989;
Johnstone 1997; Daleetal. 2001).
Distinctiveness probably provides individual and colony-
level benefits in species with reproductive transactions. At
the individual level, distinctive wasps probably receive less
aggression than non-distinctive wasps. Theoretical and
empirical evidence indicates that, in flexible species, the
dominant foundress concedes a fraction of the group’s
reproduction to the subordinate in exchange for the sub-
ordinate’s help (reproductive transactions (Reeve & Rat-
nieks 1993; Reeve et al. 2000; Reeve & Keller 2001)). The
dominant makes such concessions because each subordi-
nate has the option of starting her own nest (Reeve 2000).
Precise reproductive shares vary with rank, so it is not sur-
prising that there is constant aggression between wasps of
adjacent ranks. Wasps of other ranks receive little
aggression, probably because they are not perceived as an
gigas
jadwigae
olivaceus
__________
rothneyi
humilis
________
japonicus
snelleni
stigma
chinenesis
riparius
annularis
canadensis
erythrocephalus
lanio
______
cinerascens
_____________
subsericeus
crinitus
exclamans
___________
versicolor
__________
apachus
_________
bellicosus
___________
carolina
_________
dorsalis
_________
fuscatus
_________
metricus
_________
Figure 3. The distribution of marking variability andnest-
founding strategy in 25Polistes species. Black indicates low
variability. Grey indicates high variability. Underlined species
haveflexible nest-founding strategies. Other species are
multiple foundress or singlefoundress.
1958E. A. Tibbetts
Complex social behaviour and visual featuresin Polistes wasps
Proc.R. Soc.Lond.B (2004)
on May 24, 2011 rspb.royalsocietypublishing.orgDownloaded from