Social conflict in ant larvae: Egg cannibalism occurs mainly in males and larvae prefer alien eggs

Abstract

In social organisms, cooperation is widespread. However, social groups also provide excellent opportunities for individuals to exploit the cooperative efforts of others, thus creating conflict. Social conflict is particularly common in ant societies but although much is known about conflicts between adult individuals, the selfish interests of developing offspring have been largely neglected. One taxonomically widespread way for offspring to engage in competition is through egg cannibalism. Here, we show that larvae of the ant Formica aquilonia compete by eating eggs, which increases their survival. Cannibalism behavior differs strongly between sexes, and males cannibalize 3 times as often as females. This points to sex-dependent benefits from cannibalism, for males possibly through removal of future competitors for mating opportunities. Larvae also preferentially eat eggs of distinct origin and odor profile, suggesting that they can detect and react to chemical cues. Thus, similar to adult ants, larvae possess the power to adjust their behavior to available information. We conclude that exploring the behavior of developing individuals can give new insight into social conflict in ants and social animals in general.

Full text

© The Author 2013. Published by Oxford University Press on behalf of
the International Society for Behavioral Ecology. All rights reserved. For
permissions, please e-mail: journals.permissions@oup.com
Behavioral
Ecology
The ocial journal of the
ISBE
International Society for Behavioral Ecology
Behavioral Ecology (2013), 24(6), 1306–1311. doi:10.1093/beheco/art067
Original Article
Social conflict in ant larvae: egg cannibalism
occurs mainly in males and larvae prefer
alieneggs
EvaSchultner,a Patriziad’Ettorre,b and HeikkiHelanteräa
aDepartment of Biosciences, Centre of Excellence in Biological Interactions, University of Helsinki,
Viikinkaari 9, 00014 Helsinki, Finland and bLaboratoire d’Ethologie Expérimentale Comparée, Université
Paris 13, 99 avenue Jean-Baptiste Clément, 93430 Villetaneuse, France
Received 14 March 2013; revised 2 June 2013; accepted 27 June 2013; Advance Access publication 2 August 2013
In social organisms, cooperation is widespread. However, social groups also provide excellent opportunities for individuals to exploit
the cooperative efforts of others, thus creating conflict. Social conflict is particularly common in ant societies but although much is
known about conflicts between adult individuals, the selfish interests of developing offspring have been largely neglected. One taxo-
nomically widespread way for offspring to engage in competition is through egg cannibalism. Here, we show that larvae of the ant
Formica aquilonia compete by eating eggs, which increases their survival. Cannibalism behavior differs strongly between sexes, and
males cannibalize 3 times as often as females. This points to sex-dependent benefits from cannibalism, for males possibly through
removal of future competitors for mating opportunities. Larvae also preferentially eat eggs of distinct origin and odor profile, suggest-
ing that they can detect and react to chemical cues. Thus, similar to adult ants, larvae possess the power to adjust their behavior to
available information. We conclude that exploring the behavior of developing individuals can give new insight into social conflict in
ants and social animals in general.
Key words: ant larvae, cannibalism, discrimination, kin selection, social conflict.
INTRODUCTION
In social organisms, ospring fitness is shaped by behavioral inter-
actions with other group members, such as competition for com-
mon resources (Mock and Parker 1997). Conflict reaches extreme
levels when ospring resort to killing each other, a common phe-
nomenon across a diversity of taxa including insects (Dickison
1992; Fincke 1994; Vijendravarma etal. 2013), spiders (Bilde and
Lubin 2001), snails (Baur 1992), and birds (Mock etal. 1990). Egg
cannibalism in particular is an eective way for developing indi-
viduals to gain resources (Elgar and Crespi 1992) and cannibal o-
spring typically benefit from improved survival (Roy etal. 2007),
higher growth and/or developmental rates (Osawa 1992; Roy etal.
2007; Crossland et al. 2011), increased adult body size (Osawa
2002; Michaud and Grant 2004), and competitor removal (Banks
1956; Wise 2006).
Although cannibals benefit from selfish behavior, cannibalism
is not without overall costs (Chapman et al. 1999), especially if
broods are composed of related individuals (Pfennig 1997). This
is because inclusive fitness costs of cannibalism are higher when
preying on close relatives instead of less related brood. Following
inclusive fitness theory (Hamilton 1964), the potential for conflict
should thus vary according to within-brood relatedness and reach
higher levels when broods are mixed, for example, when females
mate multiply or when many females breed together.
Because relatedness is an important determinant of potential
conflict (Ratnieks and Reeve 1992), the power to assess genetic
diversity is likely to be essential in shaping the outcome of
intrabrood conflict (Beekman etal. 2003). Specifically, selective
behavior to spare kin in mixed brood is only possible if indi-
viduals are able to recognize kin. In insects, information on kin-
ship can be encoded in complex chemical cues (Fletcher and
Michener 1987; Howard and Blomquist 2005) and developing
individuals have been shown to use this information to prefer-
entially choose nonrelatives over kin when cannibalizing (Dobler
and Kölliker 2009).
Ants are important models in the study of social conflict (Bourke
and Franks 1995), but the selfish interests of larvae have been
widely neglected. However, the ecological and social environment
of ant societies provides many opportunities for competition
between larvae. First, ant brood is reared in batches of eggs and
larvae, providing larvae with easy access to highly nutritional eggs.
Address correspondence to E.Schultner. E-mail: eva.schultner@helsinki.fi.
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from

Schultner etal. • Ant larvae as players in social conict
Second, ants can spend more than half their lives as larvae, making
larval development an important phase of life history (Hölldobler
and Wilson 1990). Third, female caste fate, that is, developing into a
reproductive queen or nonreproductive worker, is often determined
by larval nutrition (Wheeler 1986, 1994). In addition, ant colonies
vary in their kin structure (Sundström etal. 2005) and thus in the
predicted intensity of competition among larvae. Finally, if ant
larvae use chemical information to discriminate among conspecifics
in a similar way as adults (Van Zweden and d’Ettorre), they may
also possess the power to adjust their behavior to dierent levels of
kinship.
In this study, we investigated brood conflict in the form of larval
egg cannibalism in the ant Formica aquilonia, a mound building spe-
cies that forms large networks of interconnected nests called super-
colonies (Rosengren and Pamilo 1983). Nests have hundreds of
reproductive queens and relatedness among nestmates approaches
zero (Pamilo 1982; Pamilo et al. 2005), creating potential for
extreme conflict among larvae. We measured levels of cannibalism
on dierent levels of experimentally manipulated brood kinship
and in larvae of dierent sexes. To test if eggs carry recognition
cues, we analyzed chemical profiles of eggs produced by queens of
distinct supercolonies. Finally, we estimated the eect of cannibal-
ism on the survival of larvae.
METHODS
Study colonies and collections
Queens and workers were collected from 2 supercolonies (LA:
n = 13 nests, 59.95000793°N/23.17383804°W; MY: n = 8 nests,
59.9859419°N/23.22908954°W) around Tvärminne Zoological
Station in southwestern Finland in April 2010 and 2011. The nest
fragments were transferred to plastic nest-boxes (40 × 30 × 20 cm)
coated with Fluon to prevent ants from escaping and peat and moss
provided as nest material. Boxes were kept in a greenhouse under
seminatural light and temperature conditions and watered and fed
daily with Bhatkar diet (Bhatkar and Whitcomb 1970).
Cannibalism experiments
We presented larvae with eggs of dierent origins to measure the
average level of cannibalism, the eect of egg diversity on cannibal-
ism, and the eect of cannibalism on larvae survival. Subsequent
sexing of a fraction of larvae used in experiments allowed testing
for dierences in cannibalism betweensexes.
Bioassays
Larvae and eggs were obtained by transferring queens from dif-
ferent nest fragments to the laboratory and isolating them in indi-
vidual Petri dishes that were kept in the dark at room temperature.
Eggs were removed once or twice daily from the queen dishes and
cleaned with strips of paper towel to minimize fungus growth. Eggs
were then moved to clean Petri dishes to track their maternal origin
and age. Strips of sponge cloth were placed around the eggs and
moistened daily to avoid desiccation. Eggs were left to hatch in the
dark at room temperature. After 9–11days, most eggs had hatched
(mean hatching success: 75%). Within a few hours after hatching,
each larva was moved onto a separate dish and placed ventrally on
top of a pile of 5 fresh eggs (1–3 days age), which were obtained
in the same way as larvae. Each larva was presented with a batch
of 5 eggs representing 4 levels of kinship—1) sibling: eggs from the
same mother queen (n=91 larvae), 2)nestmate: eggs from nestmate
queens (n=201), 3)colonymate: eggs from queens of a dierent nest
within the same supercolony (n=87), and 4)alien: eggs from queens
of a dierent supercolony (n = 80). For each of these treatments,
the number of eggs each larva had consumed was recorded once
a day for 2 consecutive days. When a larva was observed actively
feeding on an egg, this was also scored as cannibalized.
In the 2 treatments with the highest cannibalism levels (colony-
mate and alien), we measured the eect of cannibalism on larvae
survival by continuing assays for an additional 4days and recording
survival daily. Larvae were classified as dead when gentle probing
with forceps did not result in any response (i.e., moving, opening of
mandibles).
Statistical analysis of bioassays
Frequency of cannibalism incidents and the total number of eggs
eaten in each treatment were analyzed by fitting a mixed logistic
regression model (generalized linear mixed model, GLMM) with
presence of cannibalism or number of eaten eggs as response
variable, treatment as fixed explanatory variable, and colony as a
random factor (function glmer, package lme4 in R version 2.11.1;
Bates etal. 2012). Dierences between treatments were tested using
a priori treatment contrasts (Crawley 2007). The eect of canni-
balism on survival was estimated by comparing survival (days) of
cannibalistic and noncannibalistic larvae with a Wilcoxon–Mann–
Whitney test.
Sex bias in cannibalism
Sex is easily determined in Formica ants because males develop
from unfertilized, haploid eggs and females from fertilized, diploid
eggs. For sexing, a random sample of larvae from sibling (53 of
91 larvae) and nestmate treatments (141 of 201 larvae) was gen-
otyped at 8 polymorphic microsatellite loci designed for Formica
species and tested for successful cross-amplification in F.aquilonia:
FE13, FE19, FE21, and FE42 (Gyllenstrand etal. 2002); FL20 and
FL21 (Chapuisat 1996); FY4 and FY7 (Hasegawa and Imai 2004).
For DNA extraction, entire larvae were placed in individual wells
together with a 2.5:100µL Proteinase K–Chelex solution and left
to incubate overnight at 56°C. Polymerase chain reactions (PCRs)
were run in 10 µL reactions using 5-µL QIAGEN Type-It mic-
rosatellite multiplex buer, 3µL deionized water, 1µL optimized
primer mix, and 1µL DNA. PCR protocols were run according to
QIAGEN recommendations, products analyzed in 1:200 dilutions
in a 3730 ABI sequencer, and microsatellite peaks scored using
Genemapper software version 4.1. Larvae that were heterozygous
at 2 or more loci were classified as females, whereas individuals that
were homozygous at all successfully amplified loci were classified as
males. Based on allele frequencies from all genotypes, the probabil-
ity of wrongly assigning a diploid individual to the male category
was below 5% when at least 6 loci were used.
Statistical analyses of sexbias
Dierences between males and females in the number of cannibal-
ism incidents and the number of eaten eggs were tested with a chi-
square test for each treatment.
Analysis of chemical and genetic variation
In ants, individual odor profiles are both genetically and
environmentally determined (Van Zweden et al. 2010). To test
whether F. aquilonia larvae have access to information concerning
the origin of eggs, we analyzed genotypes of queens and the
surface hydrocarbon profiles of theireggs.
1307
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from

Behavioral Ecology
Chemical samples
One to 3 queens were isolated from each of 13 nests in supercol-
ony LA (n = 24) and 8 nests in supercolony MY (n = 14), placed
on individual Petri dishes and kept in the dark at room temperature
until they had laid 10 eggs. Queens were transferred to individual
tubes, killed by freezing, and stored at −20 °C for genetic analy-
sis (see below). Egg surface chemicals were extracted from a pooled
sample of 10 eggs per queen for 10 min in 10µL pentane (includ-
ing 5 ng/mL of n-C20 as internal standard). The extract was trans-
ferred to a clean vial and 3µL per sample was injected in an Agilent
6890N gas chromatograph (capillary column: Agilent HP-5MS,
30 m × 25 μm × 0.25 μm; split-splitless injector; carrying helium
gas at 1 mL/min) coupled to an Agilent 5375 mass spectrometer
(70 eV electron impact ionization). After an initial hold at 70 °C
for 1 min, the temperature was raised to 220 °C at 30°C/min and
then from 200 to 320°C at 5°C/min, where it was held for 5 min.
The areas of the 37 peaks common to all egg extracts were inte-
grated and compounds identified on the basis of their mass spectra
and retention times (Supplementary Figure S1) compared with the
internal standard using Agilent Chemstation software. For analyses
by compound group, compounds were separated into linear alkanes
(n= 9), alkenes (n =8), and methyl-branched compounds (n= 20).
Prior to analysis, absolute peak areas were normalized according to
Aitchison (1986).
Genetic samples
Two legs per queen were removed and genotyped according to the
same methods used for larvae sexing. We calculated mean pairwise
relatedness between queens (Queller and Goodnight 1989) using
GenAlEx version 6.5b4 (Peakall and Smouse 2012). To ensure
unbiased relatedness estimates, supercolony-specific allele frequen-
cies calculated from worker genotypes (16 workers from 21 nests in
each supercolony, data not shown) were used as population allele
frequencies.
Statistical analysis of chemical and genetic variation
Matriline origin To test whether egg profiles reflect maternal
genotype, we calculated Pearson correlations between pairwise
chemical distances (Manhattan distance) of eggs and pairwise
genetic distances based on relatedness values of their mother
queens. Correlations were calculated for all 37 compounds and
each compound group separately. All correlations were tested
for significance against a distribution obtained by calculating
correlations from 1000 random permutations of thedata.
Because discrimination between kin and nonkin eggs is primar-
ily relevant within the nest and the environmental component of
chemical profiles can confound analyses, only same-nest pairs of
queens and eggs were used in calculating correlations. Also, only
queens for whom at least 6 microsatellite loci amplified successfully
were included in the analysis. Final analysis was conducted with 8
queen pairs and their corresponding eggs from supercolony LA and
5 queen pairs and their eggs from supercolony MY (total n=13).
Supercolony origin We assessed the availability of cues on a
broader level by measuring supercolony-specific variation in all
egg profiles (n= 38). Principal component analysis (PCA) of all 37
compounds and each separate compound group (alkanes, alkenes,
and methyl-branched compounds), followed by linear discriminant
analysis (LDA) with leave-one-out cross-validation, were used to
measure classification success of eggs according to supercolony of
origin. LDA was based on the principal components that explained at
least 90% of the variation in the data set (Supplementary Table S2).
In addition, we verified supercolony-specific profile variation using
multivariate analysis (MANOVA, function adonis, package vegan;
Oksanen etal. 2012), again for all compounds and each compound
group separately.
RESULTS
Cannibalism experiments
Bioassays
Cannibalism levels rose gradually as diversity of eggs increased
from sibling to alien treatments. In the sibling treatment, 10% of
larvae (9/91) engaged in at least one act of cannibalism, whereas
in the nestmate treatment, 13% (27/201) engaged in cannibalism.
In total, larvae ate 2% and 3% of oered eggs in sibling and nest-
mate treatments, respectively. In the colonymate treatment, 21%
of larvae (18/87) were cannibals and consumed a mean of 4% of
eggs. In the alien egg treatment, 24% (19/80) of larvae engaged in
cannibalism and ate 5% of eggs. Both number of cannibalistic inci-
dents and the number of eaten eggs were significantly lower in the
sibling treatment compared with the alien treatment (response: inci-
dents, degrees of freedom [df]=3, P=0.039; response: number of
eaten eggs, df=3, P=0.025; Figure1 and Table1). There were no
significant dierences in the number of cannibalistic incidents or
eaten eggs among the other treatments.
Cannibalism levels were highest in colonymate and alien treat-
ments; therefore only these 2 treatments were continued for an
additional 4days in order to measure larvae survival. Because can-
nibalism levels did not dier between the two, we pooled data from
both treatments for survival analysis. Overall, cannibals survived
significantly longer than noncannibals (mean in days ± standard
error for cannibals: 2.94 ± 0.12 and noncannibals: 2.45 ± 0.07;
Wilcoxon–Mann–Whitney test, n=167, W=1642, P=0.002).
Sex bias in cannibalism
Of the 194 genotyped larvae, 88 were female and 106 were male
(sibling treatment: 21 females and 32 males; nestmate treatment:
0%
5%
10%
15%
20%
25%
30%
35%
Proportion
Sibling
(n = 91)
Nestmate
(n = 201)
Colonymate
(n = 87)
Alien
(n = 80)
*
Figure1
Mean proportion of cannibalism incidents (white) and number of eaten
eggs (black) across treatments. Bars indicate 95% confidence intervals.
Dierences were tested for each response separately with a GLMM for
binomially distributed data using treatment as fixed and colony as random
variables. *P<0.05.
1308
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from

Schultner etal. • Ant larvae as players in social conict
67 females and 74 males). In the sibling treatment, male larvae
engaged in cannibalism significantly more often than female lar-
vae (proportion of cannibalism incidents in females: 0% [0/21
larvae] and males: 21.8% [7/32 larvae]; Pearson’s chi-square test:
χ2= 4.27, df=1, P=0.038). The pattern was similar in the nest-
mate treatment with 5.9% (4/67 larvae) cannibalism in females and
20.2% (15/74 larvae) in males (Pearson’s chi-square test: χ2=4.75,
df=1, P=0.013; Figure 2). Males also ate significantly more eggs
than females in both treatments (Pearson’s chi-square test; sibling—
females: 0%, males: 4.4%; χ2=4.72, df=1, P=0.029; nestmate—
females: 1.5%, males: 4.6%, χ2=5.59, df=1, P=0.018).
Analysis of chemical and genetic variation
Matrilineorigin
Eggs laid by genetically distant queens were not more distinct in
their chemical profiles than eggs laid by closely related queens,
indicating that egg odors do not contain information about
matriline origin (Pearson correlation coecient, n = 13; all
compounds: r2= −0.20, P = 0.26; alkanes: r2 = −0.25, P = 0.21;
alkenes: r2 = 0.02, P = 0.55; methyl-branched compounds:
r2=−0.14, P=0.34).
Supercolonyorigin
Egg samples (92%) were classified correctly by supercolony origin
(PCA + LDA on PCs 1–10; Supplementary Table S2). Multivariate
analysis confirmed this result (MANOVA—factor: supercolony,
df = 1, F = 2.72, R2 = 0.07, P = 0.016). When compound groups
were analyzed separately, classification was similar using only methyl-
branched compounds (LDA on PCs 1–8: 84% correct classifica-
tion; MANOVA—factor: supercolony, df = 1, F =5.71, R2 = 0.17,
P=0.0019) but not when using linear alkanes (LDA on PCs 1–3: 52%
correct; MANOVA—factor: supercolony, df=1, F=0.25, R2=0.007,
P=0.82) or alkenes (LDA on PCs 1–4: 52% correct; MANOVA—fac-
tor: supercolony, df=1, F=1.62, R2=0.04, P=0.18).
DISCUSSION
Our experiments show that potential conflict among ant larvae is
reflected in their behavior in the form of egg cannibalism. Around
10% of larvae cannibalized sibling eggs, and incidence of cannibal-
ism rose to 24% when larvae were presented eggs from an alien
supercolony. This is the first evidence that ant larvae can adjust
their behavior to brood composition and available sensory infor-
mation. Cannibalism increased larvae survival and male larvae
engaged in cannibalism 3 times as often as females. These results
emphasize the importance of including developing individuals in
studies of conflict in ants and social insects in general and oer a
new perspective on the role of males in ant societies.
In ants, egg eating can play an important role during early colony
life when founding queens lay nonviable, trophic eggs to feed the first
generation of larvae (Crespi 1992). In contrast, cannibalism of viable
eggs by adults and feeding of eggs to older larvae are thought to occur
as a result of resource shortages (Wilson 1971; Sorensen et al. 1983;
Heinze etal. 1999) and help increase colony resistance to starvation
(Rueppell and Kirkman 2005). Our experiments confirm that eggs are
highly nutritious and that their consumption improves larvae survival
in the absence of other food sources and thus can provide a means for
larvae to regulate their own food intake. In addition, egg eating may
allow larvae to grow faster than their nestmates and possibly avoid
being eaten themselves, a benefit of cannibalism in other taxa (Polis
1981). Especially female larvae, whose adult caste depends on larval
nutrition, would gain from the increased growth and/or develop-
mental rates commonly associated with cannibalism (Fox 1975; Polis
1981; Roy etal. 2007; Crossland etal. 2011). In stingless bees, some
Table1
GLMM analysis of incidents of cannibalism (a) and number of eaten eggs (b)
Fixed eects Random eect, colony (n=18)
Treatment Estimate Standard error z-Value PVariance Standard deviation
(a)
Intercept (sibling) −2.248 0.373 −6.025 <0.001
Nestmate 0.364 0.429 0.849 0.396 0.134 0.366
Colonymate 0.807 0.482 1.673 0.094
Alien 0.987 0.480 2.054 0.040
(b)
Intercept (sibling) −3.932 0.352 −11.169 <0.001
Nestmate 0.441 0.399 1.104 0.269 0.084 0.289
Colonymate 0.734 0.444 1.651 0.099
Alien 0.973 0.435 2.237 0.025
0%
10%
20%
30%
40%
Proportion of cannibals
Sibling
(n = 53)
(4/67)
Nestmate
(n = 141)
(15/74) (0/21) (7/32)
* *
Figure2
Proportion of cannibalism incidents performed by females (white) and
males (black) in sibling and nestmate treatments. Numbers above bars
indicate the number of cannibalistic larvae/total number of larvae from
each sex. Dierences between sexes were analyzed with a chi-square test.
*P<0.05.
1309
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from

Behavioral Ecology
females selfishly determine their caste, which may be linked to under-
lying physiological dierences between larvae (Wenseleers etal. 2005).
Conversely, cannibalism could be a behavioral mechanism of selfish
caste determination in female ant larvae, although this remains to be
shown (Rüger etal. 2007).
The strong male bias in our experiments furthermore raises the
question of sex-specific benefits from cannibalism (Osawa 2002).
In F. aquilonia, male and female reproductives are reared from the
first batch of eggs laid in spring (Rosengren etal. 1987; Gösswald
1989). The larvae used in our experiments were obtained from this
brood cohort; thus, it does not seem likely that the observed male
bias was due to dierences in cannibalism between worker-destined
and sexual brood. Instead, males may engage in cannibalism more
often if the associated gain in nutrition increases sperm quantity
or quality (Delisle and Hardy 1997). This could be especially
relevant in ants, where spermatogenesis typically occurs only during
larval development (Hölldobler and Bartz 1985). In addition,
developmental consequences of cannibalism such as increased adult
body size (Osawa 2002) may select for higher levels of cannibalism
in males. For instance, leg and wing length of adult males have been
positively correlated with mating success in the ant Pogonomyrmex
occidentalis (Wiernasz etal. 1995). Finally, if competition for mating
opportunities among males is stronger than in females, removal of
competitors through egg cannibalism may be more advantageous to
males. This could be important in F.aquilonia, where mating often
occurs in or near the nest (Fortelius etal. 1993), but further studies
of population structure would be needed to confirmthis.
Intrabrood relatedness and the ability of individuals to assess
kinship are known to mediate the occurrence and intensity of can-
nibalism in animals (Pfennig 1997). However, in our experiments,
levels of cannibalism did not dier significantly between sibling,
nestmate, and colonymate treatments. Together with the lack of
evidence for matriline-specific cues in egg odors, this suggests that
lack of information constrains discrimination within colonies. To
our knowledge, the only other study of discriminative egg canni-
balism in ant larvae also failed to find evidence for nestmate rec-
ognition (Urbani-Baroni 1991) although chemical cues were not
investigated here. In social insects, nest-level costs of individual
selfishness are thought to select against the evolution of kin-infor-
mative cues (Ratnieks 1991; Keller 1997) and only a few studies
have found such cues in ants (Nehring etal. 2011; Helanterä et al.
2013). In line with this, evidence for kin-preferential behavior is
limited (Van Zweden etal. 2010, but see Hannonen and Sundström
2003 for an example in Formica fusca), even though patterns resem-
bling nepotism may arise through passive processes (for example in
Formica exsecta, Holzer, Kümmerli, etal. 2006). Our results suggest
similar constraints act on kin discrimination in larvae.
On the supercolony level, eggs diered in their chemical pro-
files. Methyl-branched compounds in particular showed a strong
supercolony signal, confirming that substance classes dier in their
importance as recognition cues (Van Zweden etal. 2010). In ants,
ecient discrimination between colony members and intruders
is essential in maintaining the integrity and organization of colo-
nies (Hölldobler and Wilson 1990) and adult workers use colony-
specific odors to discriminate nestmates from non–nestmates (Van
Zweden and d’Ettorre). In supercolonial ants, exchange of work-
ers, brood, and queens between nests of the same supercolony
can lead to blending of odors on a large spatial scale, and our
results indicate that supercolony-level cue mixing in the maternal
environment plays an important role in determining egg odors in
F.aquilonia.
Indeed, when larvae were presented with eggs carrying odors dis-
tinct from those they typically encounter within their own supercol-
ony, cannibalism levels increased significantly. Thus when dierences
are extreme, ant larvae react to chemical cues. In nature, larval dis-
crimination of alien eggs may play a role when queens enter foreign
nests. Ant supercolonies often originate from a single founding col-
ony and can have strict boundaries (Holzer, Chapuisat, etal. 2006;
Helanterä 2009). Still, queens may move across supercolonies and
attempt adoption in foreign nests (Holzer etal. 2008). Several spe-
cies of wood ants including F.aquilonia dominate our study area and
exchange of queens between supercolonies or even intrusion by het-
erospecific queens is thus not improbable (Czechowski 1996).
Overall, we found that egg cannibalism in F.aquilonia is shaped
by the ability of larvae to assess the chemical diversity of eggs and
intrinsic dierences between sexes regarding its intensity. Notably,
sex-specific dierences regarding the benefits of egg eating may
play a role in mediating larval cannibalism. Ant larvae are thus not
the powerless individuals they have been seen as so far. Instead, our
study shows that developing ospring and male larvae in particu-
lar may be a neglected force in ant societies. Moreover, this work
demonstrates that ants, whose kin structure variation and chemical
information processing are already well understood, can be fruitful
models for studies of intrabrood conflict.
SUPPLEMENTARY MATERIAL
Supplementary material can be found at http://www.beheco.
oxfordjournals.org/
FUNDING
This study was funded by the University of Helsinki (490057 to
H.H.) and the Academy of Finland (140990, 135970 to H.H.).
We thank Jan Oettler, Stuart West, and an anonymous referee for construc-
tive comments on an earlier version of the manuscript and Unni Pulliainen,
Siiri Fuchs, and Chloé Leroy for help with data collection.
Handling editor: Anna Dornhaus
REFERENCES
Aitchison J. 1986. The statistical analysis of compositional data. London:
Chapman and Hall.
Banks C. 1956. Observations on the behaviour and early mortality of cocci-
nellid larvae before dispersal from egg shells. Proc R Entomol Soc Lond.
31:56–61.
Bates D, Maechler M, Bolker BM. 2012. lme4: linear mixed-eects mod-
els using S4 classes. R package version 0.999999-0. [cited 2013 February
25]. Available from: http://cran.r-project.org/package=lme4.
Baur B. 1992. Cannibalism in gastropods. In: Elgar M, Crespi B, editors.
Cannibalism: ecology and evolution among diverse taxa. New York:
Oxford University Press. p. 102–127.
Beekman M, Komdeur J, Ratnieks F. 2003. Reproductive conflicts in social
animals: who has power? Trends Ecol Evol. 18:277–282.
Bhatkar A, Whitcomb W. 1970. Artificial diet for rearing various species of
ants. Fla Entomol. 53:229–232.
Bilde T, Lubin Y. 2001. Kin recognition and cannibalism in a subsocial spi-
der. J Evol Biol. 14:959–966.
Bourke A, Franks N. 1995. Social evolution in ants. Princeton (NJ):
Princeton University Press.
Chapman J, Williams T, Escribano A. 1999. Fitness consequences of
cannibalism in the fall armyworm, Spodoptera frugiperda. Behav Ecol.
10:298–303.
Chapuisat M. 1996. Characterization of microsatellite loci in Formica lugu-
bris B and their variability in other ant species. Mol Ecol. 5:599–601.
1310
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from

Schultner etal. • Ant larvae as players in social conict
Crawley M. 2007. Statistical modelling. The R book. Chichester (UK): John
Wiley & Sons. p. 388–448.
Crespi B. 1992. Cannibalism and trophic eggs in subsocial and eusocial
insects. In: Elgar M, Crespi B, editors. Cannibalism: ecology and evolu-
tion among diverse taxa. New York: Oxford University Press. p. 176–213.
Crossland MR, Hearnden MN, Pizzatto L, Alford RA, Shine R. 2011. Why
be a cannibal? The benefits to cane toad, Rhinella marina [=Bufo marinus],
tadpoles of consuming conspecific eggs. Anim Behav. 82:775–782.
Czechowski W. 1996. Colonies of hybrids and mixed colonies; interspecific
nest takeover in wood ants (Hymenoptera, Formicidae). Memorabilia
Zool. 50:1–116.
Delisle J, Hardy M. 1997. Male larval nutrition influences the reproduc-
tive success of both sexes of the spruce budworm, Choristoneura fumiferana
(Lepidoptera: Tortricidae). Funct Ecol. 11:451–463.
Dickison JL. 1992. Egg cannibalism by larvae and adults of the milk-
weed leaf beetle (Labidomera clivicollis, Coleoptera: Chrysomelidae). Ecol
Entomol. 17:209–218.
Dobler R, Kölliker M. 2009. Kin-selected siblicide and cannibalism in the
European earwig. Behav Ecol. 21:257–263.
Elgar M, Crespi B, editors. 1992. Cannibalism: ecology and evolution
among diverse taxa. New York: Oxford University Press.
Fincke O. 1994. Population regulation of a tropical damselfly in the larval
stage by food limitation, cannibalism, intraguild predation and habitat
drying. Oecologia. 100:118–127.
Fletcher D, Michener C. 1987. Kin recognition in animals. Chichester
(UK): John Wiley & Sons.
Fortelius W, Rosengren R, Cherix D, Chautems D. 1993. Queen recruitment
in a highly polygynous supercolony of Formica lugubris (Hymenoptera:
Formicidae). Oikos. 67:193–200.
Fox L. 1975. Cannibalism in natural populations. Annu Rev Ecol Syst.
6:87–106.
Gösswald K. 1989. Die Waldameise. Wiesbaden (Germany): AULA-Verlag.
Gyllenstrand N, Gertsch P, Pamilo P. 2002. Polymorphic microsatellite
DNA markers in the ant Formica exsecta. Mol Ecol Notes. 2:67–69.
Hamilton W. 1964. The genetical evolution of social behavior I. J Theor
Biol. 7:1–16.
Hannonen M, Sundström L. 2003. Worker nepotism among polygynous
ants. Nature. 421:910.
Hasegawa E, Imai S. 2004. Characterization of microsatellite loci in red
wood ants Formica (s. str.) spp. and the related genus Polyergus. Mol Ecol
Notes. 4:200–203.
Heinze J, Foitzik S, Oberstadt B. 1999. A female caste specialized for
the production of unfertilized eggs in the ant Crematogaster smithi.
Naturwissenschaften. 95:93–95.
Helanterä H. 2009. Do unicolonial wood ants favor kin? J Biol. 8:56.
Helanterä H, Aehle O, Roux M, Heinze J, d’Ettorre P. 2013. Family-based
guilds in the ant Pachycondyla inversa. Biol Lett. 9:20130125.
Hölldobler B, Bartz S. 1985. Sociobiology of reproduction in ants. In:
Hölldobler B, Lindauer M, editors. Experimental behavioral ecology and
sociobiology. Sunderland (MA): Sinauer Associates. p. 237–257.
Hölldobler B, Wilson E. 1990. The ants. Cambridge (MA): Belknap-
Harvard University Press.
Holzer B, Chapuisat M, Keller L. 2008. Foreign ant queens are accepted
but produce fewer ospring. Oecologia. 157:717–723.
Holzer B, Chapuisat M, Kremer N, Finet C, Keller L. 2006. Unicoloniality,
recognition and genetic dierentiation in a native Formica ant. J Evol
Biol. 19:2031–2039.
Holzer B, Kümmerli R, Keller L, Chapuisat M. 2006. Sham nepotism as a
result of intrinsic dierences in brood viability in ants. Proc R Soc B Biol
Sci. 273:2049–2052.
Howard RW, Blomquist GJ. 2005. Ecological, behavioral, and biochemical
aspects of insect hydrocarbons. Annu Rev Entomol. 50:371–393.
Keller L. 1997. Indiscriminate altruism: unduly nice parents and siblings.
Trends Ecol Evol. 12:99–103.
Michaud J, Grant A. 2004. Adaptive significance of sibling egg cannibalism
in Coccinellidae: comparative evidence from three species. Ann Entomol
Soc Am. 97:710–719.
Mock D, Drummond H, Stinson C. 1990. Avian siblicide. Am Sci.
78:438–449.
Mock D, Parker G. 1997. The evolution of sibling rivalry. New York:
Oxford University Press.
Nehring V, Evison SEF, Santorelli LA, d’Ettorre P, Hughes WOH. 2011.
Kin-informative recognition cues in ants. Proc R Soc B Biol Sci.
278:1942–1948.
Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’Hara R,
Simpson GL, Solymos P, Stevens M, Wagner H. 2012. vegan: commu-
nity ecology package. R package version 2.0-5. [cited 2013 February 27].
Available from: http://CRAN.R-project.org/package=vegan.
Osawa N. 1992. Sibling cannibalism in the ladybird beetle Harmonia axyridis:
fitness consequences for mother and ospring. Res Popul Ecol. 34:45–55.
Osawa N. 2002. Sex-dependent eects of sibling cannibalism on life history
traits of the ladybird beetle Harmonia axyridis (Coleoptera: Coccinellidae).
Biol J Linn Soc. 76:349–360.
Pamilo P. 1982. Genetic population structure in polygynous Formica ants.
Heredity. 48:95–106.
Pamilo P, Zhu D, Fortelius W, Rosengren R, Seppä P, Sundström L. 2005.
Genetic patchwork of network-building wood ant populations. Ann Zool
Fennici. 42:179–187.
Peakall R, Smouse P. 2012. GenAlEx 6.5: genetic analysis in Excel.
Population genetic software for teaching and research—an update.
Bioinformatics. 28:2537–2539.
Pfennig D. 1997. Kinship and cannibalism. BioScience. 47:667–675.
Polis G. 1981. The evolution and dynamics of intraspecific predation. Annu
Rev Ecol Syst. 12:225–251.
Queller D, Goodnight K. 1989. Estimating relatedness using genetic mark-
ers. Evolution. 43:258–275.
Ratnieks F. 1991. The evolution of genetic odor-cue diversity in social
Hymenoptera. Am Nat. 137:202–226.
Ratnieks F, Reeve H. 1992. Conflict in single-queen Hymenopteran soci-
eties: the structure of conflict and processes that reduce conflict in
advanced eusocial species. J Theor Biol. 158:33–65.
Rosengren R, Fortelius W, Lindström K, Luther A. 1987. Phenology and
causation of nest heating and thermoregulation in red wood ants of the
Formica rufa group studied in coniferous forest habitats in southern. Ann
Zool Fennici. 24:147–155.
Rosengren R, Pamilo P. 1983. The evolution of polygyny and polydomy in
mound-building Formica ants. Acta Entomol Fenn. 42:65–77.
Roy HE, Rudge H, Goldrick L, Hawkins D. 2007. Eat or be eaten: preva-
lence and impact of egg cannibalism on two-spot ladybirds, Adalia bipunc-
tata. Entomol Exp Appl. 125:33–38.
Rueppell O, Kirkman RW. 2005. Extraordinary starvation resistance in
Temnothorax rugatulus (Hymenoptera, Formicidae) colonies: demography
and adaptive behavior. Insect Soc. 52:282–290.
Rüger M, Fröba J, Foitzik S. 2007. Larval cannibalism and worker-induced
separation of larvae in Hypoponera ants: a case of conflict over caste deter-
mination? Insect Soc. 55:12–21.
Sorensen A, Busch T, Vinson S. 1983. Factors aecting brood cannibalism
in laboratory colonies of the imported fire ant, Solenopsis invicta Buren
(Hymenoptera: Formicidae). J Kans Entomol Soc. 56:140–150.
Sundström L, Seppä P, Pamilo P. 2005. Genetic population structure
and dispersal patterns in Formica ants—a review. Ann Zool Fennici.
42:163–177.
Urbani-Baroni C. 1991. Indiscriminate oophagy by ant larvae: an explana-
tion for brood serial organization? Insect Soc. 38:229–239.
Vijendravarma R, Narasimha S, Kawecki T. 2013. Predatory cannibalism
in Drosophila melanogaster larvae. Nat Commun. 4:1789.
Wenseleers T, Ratnieks F, Ribeiro M, Alves D, Imperatriz-Fonseca V. 2005.
Working-class royalty: bees beat the caste system. Biol Lett. 1:125–128.
Wheeler D. 1986. Developmental and physiological determinants of caste
in social Hymenoptera: evolutionary implications. Am Nat. 128:13–34.
Wheeler D. 1994. Nourishment in ants: patterns in individuals and societ-
ies. In: Hunt J, Nalepa C, editors. Nourishment and evolution in insect
societies. Boulder (CO): Westview Press. p. 245–278.
Wiernasz D, Yencharis J, Cole B. 1995. Size and mating success in males
of the western harvester ant, Pogonomyrmex occidentalis (Hymenoptera:
Formicidae). J Insect Behav. 8:523–531.
Wilson E. 1971. The insect societies. Cambridge (MA): Harvard University
Press.
Wise DH. 2006. Cannibalism, food limitation, intraspecific competition, and
the regulation of spider populations. Annu Rev Entomol. 51:441–465.
Van Zweden J, Brask J, Christensen J, Boomsma J, Linksvayer T, d’Ettorre
P. 2010. Blending of heritable recognition cues among ant nestmates cre-
ates distinct colony gestalt odours but prevents within-colony nepotism.
JEvol Biol. 23:1498–1508.
Van Zweden J, d’Ettorre P. 2010. The role of hydrocarbons in nestmate
recognition. In: Blomquist GC, Bagnères A-G, editors. Insect hydrocar-
bons: biology, biochemistry and chemical ecology. Cambridge (MA):
Cambridge University Press. p. 222–243.
1311
by guest on October 20, 2015http://beheco.oxfordjournals.org/Downloaded from