me and we: the interplay between individual and group

COIS-49; NO. OF PAGES 9

Me and we: the interplay between individual and groupbehavioral variation in social collectivesAdria C LeBoeuf1 and Christina M Grozinger2

Available online at www.sciencedirect.com

ScienceDirect

In social insects, substantial behavioral variation exists

among individuals and across colonies. Here, we discuss the

role of individual variation in shaping behavioral tendencies

of social groups, and highlight gaps in our knowledge about

the role of the social group in modulating individual

behavioral tendencies. We summarize our knowledge of the

genetic mechanisms underpinning these processes, and

describe the use of genomic tools to better understand the

influence of social context on individuals. We discuss rapid

collective phasic transitions, in which a group of individuals

engages in a common novel behavior together, as a

potentially highly informative model system in which to

comprehensively investigate the interplay between individual

and group variation.

Addresses1 Department of Ecology and Evolution, Center for Integrative Genomics,

University of Lausanne, UNIL-Sorge, Batiment Biophore, CH-1015

Lausanne, Switzerland2 Department of Entomology, Center for Pollinator Research, Center for

Chemical Ecology, The Pennsylvania State University, 1 Chemical

Ecology Lab, Orchard Road, University Park, PA 16802, USA

Corresponding author: Grozinger, Christina M ([email protected])

Current Opinion in Insect Science 2014, 3:xx–yy

This review comes from a themed issue on Social Insects

Edited by Laurent Keller and Nathalie Stroeymeyt

http://dx.doi.org/10.1016/j.cois.2014.09.010

2214-5745/# 2014 Elsevier Ltd. All right reserved.

IntroductionIndividuals in social insect colonies can vary significantly

in their behavioral tendencies — their propensity to

engage in particular behaviors — and there is mounting

evidence that a great deal of behavioral variation exists

across colonies as well [1��,2��]. While there have been

many studies examining how group behavior emerges

from individual behavior through self-organization [3],

how individual variation and group variation influence

each other has not been comprehensively considered in

social insects. For example, is the group’s behavior simply

an average of the behavioral tendencies of the individual

group members, or do some individuals disproportio-

nately affect group behavioral tendencies? To what

Please cite this article in press as: LeBoeuf AC, Grozinger CM: Me and we: the interplay between in

dx.doi.org/10.1016/j.cois.2014.09.010

www.sciencedirect.com

extent can and do individuals change their tendencies

depending on their social environment (e.g., the beha-

vioral tendencies of their group)? In this review we will

consider variation on the individual scale, variation on the

group scale, and the interplay between them in social

insects. Furthermore, we discuss the application of mol-

ecular and genomic approaches to understand the mech-

anisms mediating these processes.

Discussions of the effects of individual behavioral vari-

ation on group function in social insects have largely

focused on mature colonies and established social groups,

in which individual variation leads to improved division of

labor of colony tasks [4,5��,6]. However, the behavior of

social insect colonies changes over time as colonies form,

mature, and eventually die, and the behavioral tendencies

of the individuals may change correspondingly [7�].Furthermore, colonies can undergo rapid collective pha-

sic transitions, such as migrating or swarming, wherein a

large number of individuals, regardless of their current

behavioral or physiological state, engage in a common

behavior simultaneously [8–10]. These rapid collective

phasic transitions thus appear to be a special case of social

behavior, in which the social signals generally override

individual variation. We discuss examples and possible

roles of individual variation in shaping these processes.

Variation in behavioral traits of individuals andgroupsIn all animal species, including in social insects, individ-

uals can vary profoundly in their behavioral tendencies

(for excellent recent reviews, see Refs. [1��,2��]). Beha-

vioral variation often results from differences in response

thresholds to key sensory stimuli, where some individuals

respond rapidly to cues presented at low intensities, while

others require prolonged exposure to high intensities of

these cues before a response is elicited [11,12]. Simple

shifts in sensory sensitivity can result in global shifts in

an organism’s social behavior: for example, honey bee

workers with high response thresholds to queen phero-

mone are more likely to activate their ovaries and com-

pete with the queen over reproduction [13], less likely to

rear new queens if the queen is lost [14], and exhibit

distinct genome-wide brain gene expression patterns [15]

compared to workers with low response thresholds.

Individual variation in response thresholds can be rela-

tively stable over the lifetime of an individual (if it is due

to genetic variation or early developmental factors) or

change as a result of condition (age, reproductive status,

dividual and group behavioral variation in social collectives, Curr Opin Insect Sci (2014), http://

Current Opinion in Insect Science 2014, 3:1–9



mailto:[email protected]


http://www.sciencedirect.com/science/journal/aip/22145745

2 Social Insects


immune status), experience (including learning), social

interactions, or environmental context (reviewed in Refs.

[1��,2��]). Few studies have examined whether the same

regulatory neural and genetic mechanisms underlie beha-

vioral differences that are stable between individuals or

plastic throughout an individual’s lifetime, though recent

genomic studies suggest this is likely to be the case [16];

for example, a common set of genes is differentially

expressed between workers from aggressive Africanized

and more docile European honey bee strains (genetic

variation), between aggressive older bees and more docile

younger bees (age-dependent variation), and between

honey bees that are exposed to alarm pheromone or

solvent controls (environmental/social context variation).

Interestingly, expression of a specific subset of these

genes is associated with aggressive behavior in honey

bees, paper wasps, and mice, suggesting that these genes

may mediate aggressive tendencies across species [17�].

Many recent studies of individual variation have been

devoted to the analysis and documentation of ‘keystone’

individuals, individuals who have a disproportionately

large, and in some cases, irreplaceable effect on group

dynamics in relation to generic individuals (as defined

and reviewed over many animal species in Ref. [18�]). As

noted by Modlmeier et al. [18�], in reference to social

insects, the only true fixed, irreplaceable keystone indi-

viduals are queens. Workers can occupy keystone roles,

though this is typically temporary and these workers can

be readily replaced by their nestmates [19]. Keystone

workers may be those that have especially low response

thresholds for a specific task-associated signal and thus

respond very rapidly. While there have been several

studies examining the genomic mechanisms distinguish-

ing queens and workers in social insect species (for

examples see Refs. [20–23]), there have been few studies

of the genomic mechanisms that distinguish keystone

workers from their nestmates. A recent study of scout

honey bees (which seek out novel food resources in their

environment and subsequently recruit their nestmates)

demonstrated that these bees have distinct brain gene

expression profiles compared to their non-scout foraging

nestmates [24]. Expression of several genes involved in

biogenic amine signaling differed between the two

groups. Treating colonies with octopamine and gluta-

mate resulted in increased overall scouting behavior,

while treatments with dopamine antagonists inhibited

it. Comparative studies in other model systems will be

necessary to determine if common genes underlie scout-

ing behavior across social insect species. Additionally, it

would be of great interest to determine if common genes

underlie the behavior of different types of keystone

individuals, even if the behaviors of those individuals

are quite distinct. For example, keystone behaviors

could arise from the tuning of molecular activators such

as hormones (discussed in Ref. [13]), biogenic amines

[25], or neuropeptides [26,27].


dx.doi.org/10.1016/j.cois.2014.09.010


Social insect collectives also exhibit variation in their

group behavioral traits, with three predominant modes

of variation: stable — inter-colony differences that remain

intact over the lifetime of the colony, developmental —

shifting over the course of colony maturation, and pha-sic — behaviors that vary over short timescales (Figure 1).

We will focus on stable group behavioral variation and

discuss rapid collective phasic transitions in a later

section. Social insect colonies exhibit stable variation

in aggressiveness, foraging choices, queen number

(reviewed in Refs. [1��,2��]). In many cases, these differ-

ences are due to genotypic differences, though the under-

lying genetic mechanisms have not always been

characterized. Pogonomyrmex barbatus colonies surveyed

over 30 years show heritable behavioral variation in their

foraging choices, with parent and offspring colonies

selecting similar days to forage in terms of the desiccation

threat vs. foraging pay-off [28]. In populations of Apismellifera honey bees selected for pollen versus nectar

foraging at the colony level, there are genomic differences

associated with the insulin signaling pathway [29�,30]. A

bimodal colony-wide type of variation occurs in the fire

ant Solenopsis invicta, associated with a non-recombining

region encompassing the Gp-9 locus. Whether the

workers accept one or multiple queens in their colony

depends on which Gp-9 alleles are present in the workers

and in the queen [31�,32] (Figure 2). Other physiological

traits are also linked with the Gp9 allele: monogyne forms

typically produce larger and more fecund queens, larger

workers, and more alates than do same-sized polygyne

colonies [33].

What is the effect of individual variation on thesocial group?In general, stable, genetic variability across individuals

plays a largely positive role in the performance and fitness

of social groups [4,5��], likely through facilitation of

division of labor. Based on differing response thresholds

and spatial location, individuals segregate themselves

amongst different tasks [6,34,35,36,37��], with increased

task specialization often leading to increased efficiency

[37��,38,39,40]. However, an individual’s behavioral

tendencies can also alter the overall behavioral tendencies

of the group, though effects that can be variable and non-

additive, reflecting the interactions between the individual

and its nestmates.

The type of effect (linear vs. non-linear) that individual

variation has on the group’s behavioral tendencies

depends on the type of behavior, mode of communication

or information transfer, and responsiveness of the group

members (Figure 3). In simple stimulus-response beha-

viors, individual members shape the group’s behavior in

an additive, linear, manner; for example, several individ-

uals encounter a stimulus (i.e., a corpse) but only a subset

of high-responding individuals responds to that stimulus

(and remove the corpse [41]). In such a case, for a given





Individual-level and group-level behavioral variation LeBoeuf and Grozinger 3


Figure 1

Population Density

Gre

gario

us

Phasic Variation: Locusts

Biv

ouac

Aggr

essiv

enes

s

Age of Individual

% T

ime

out o

f Nes

t

large/oldcolony

small/youngcolony

Phasic Variation:Swarming Honey Bees

Age of Individual

Stable Variation

Developmental Variation

colony A

colony B

Population Density

(a)

(b)

(c)

(d)

Sol

itary

% T

ime

out o

f Nes

t

Hos

tN

est

New

Nes

t

Current Opinion in Insect Science

Schematics demonstrating modes of group-scale variation. (a) A colony can vary in the stable behavioral traits of its component members, here with

thicker lines indicating a larger number of colony members. Timescale: colony lifetime [1��,29�,70]. (b) Developmental variation can vary the strength of

division of labor over the lifetime of a colony: small colonies require workers to multi-task more, while large colonies can afford greater task

specialization. Timescale: Months to years [37��,38,39,40]. (c) Rapid collective phasic transitions between different behavioral states in locusts. Solid

lines correspond to stable states, while dotted lines correspond to unstable transition states, with the gap between the dotted lines indicating the

extent of hysteresis [64]. (d) Rapid collective phasic transitions across stable states in swarming honey bees. Solid lines and point correspond to stable

states, while dotted lines correspond to unstable transition states. When a honey bee colony becomes large and dense (threshold density signified by

the dashed gray line), the honey bees swarm: a subset of the honey bees leave the home nest together, migrate to a temporary location, and form a

bivouac; when a new nest site has been decided upon (decision threshold not pictured in the schematic), the bivouac group then migrates en masse to

the new site. [43,66].

number of corpses, the rate of corpse removal should scale

with the number of low-response-threshold individuals in

the group. In cases of positive feedback loops where a

signal is amplified, such as recruitment, effects of indi-

vidual variation on the group are by definition non-linear

[42��]. The degree of response amplification in recruit-

ment depends on the means of recruitment. Active

recruitment that alters the response thresholds of many

other individuals yields a highly non-linear effect. When a

honey bee recruits other bees to a new food resource with

a waggle dance, she changes the behavior of many others

simultaneously, who after visiting the resource them-

selves, may recruit more bees with their own waggle

dances [43]. Recruitment methods such as tandem run-

ning in ants, where a single ant recruits others one by one,

exert a more mildly non-linear effect on the colony [44��].Non-linear responses can also come about by passive

means through a process called stigmergy, where for

example, bystanders may simply join others that are enga-

ging in a certain behavior [34,45�]. Thus, in the absence of a

positive feedback loop, the group’s behavior should reflect


dx.doi.org/10.1016/j.cois.2014.09.010


the average behavior of the colony members, but if positive

feedback is used, the group’s behavior is predicted to be

driven by that of the most responsive member — which

could correspond to an individual occupying a keystone

role (Figure 3). Note, however, that colony-level behaviors

generally involve a combination of many individual-level

behaviors and are often brought about by different types of

positive and negative feedback.

There are several examples in which social groups of

individuals with mixed behavioral tendencies have been

created and the impacts on the groups’ behavioral

tendencies have been examined. The effect of the Gp-9non-recombining region on polygyny and monogyny in

fire ants represents a non-linear effect (Figure 2): the

queen-acceptance behavior of an entire monogyne

Gp-9B/Gp-9B colony can be shifted from single-queen

acceptance to multiple-queen acceptance if only 5–10%

of the colony’s workers have the Gp-9b allele [32]. These

experiments suggest that this small proportion of workers

is actively influencing the behavior of the other 90% of





4 Social Insects


Figure 2

Gp-9B/Gp-9Bindividual

Gp-9B/Gp-9bindividual

wo

rker

s

accepted queen(s)

Monogyne Monogyne Polygyne PolygyneGp-9B/Gp-9B Gp-9B/Gp-9B Gp-9B/Gp-9b Gp-9B/Gp-9b

proportion of Gp-9B/Gp-9b workers

thre

shol

d pe

rcen

tage

for

colo

ny-w

ide

beha

vior

al c

hang

e


Polygyny and monogyny in fire ants. Fire ants exhibit a colony level dimorphism controlled by a large non-recombining region encompassing the Gp-9

allele: when the colony is composed of only Gp-9B/Gp-9B (dark red) workers, they will accept only a single queen of their same genotype; if as few as

5% of the colony’s workers are of the Gp-9B/Gp-9b genotype (light red), the colony will accept multiple queens specifically of the Gp-9B/Gp-9b

genotype and will execute any queens with a Gp-9B/Gp-9B genotype [31�,32,33,62]. Note that the Gp-9b/Gp-9b genotype is lethal.

workers likely through some highly propagated or volatile

signal. In social spiders, the foraging success of the group

corresponds to the foraging abilities of its most aggressive

members [46]. However, while a group’s ‘boldness’ (the

latency to investigate a possible prey item) correlates most

strongly to the behavior of the boldest individual, it also is

significantly correlated with the average boldness score

across the group [47��]. Studies of honey bee colonies

consisting of a mix of defensive and gentle bees indicate

that the guards (which are arguably functioning as keystone

individuals) responded as rapidly as in defensive colonies,

but the number of bees subsequently recruited was inter-

mediate between gentle and defensive colonies [48]. In

mixed colonies of docile European and more aggressive

Africanized honey bees, the Africanized guards attacked a

stimulus first, and subsequent responders were mixed

genotypes [49]. Thus, in this case, the group behavior

reflects both the response thresholds of keystone individ-

uals and recruited individuals.

The signals produced by individuals to influence their

nestmates are regulated by genomic mechanisms, but

these have not been studied comprehensively. In honey

bees, pheromones strongly influence worker behavior,

and there is ample evidence that exposure to specific

pheromones can trigger gene expression changes


dx.doi.org/10.1016/j.cois.2014.09.010


corresponding to the associated behavioral changes in

recipient bees [16,50,51]. Few studies, however, have

examined which genes actually regulate the production

of pheromones in social insects. A recent study examining

the gene expression patterns in the pheromone-produ-

cing mandibular glands of honey bee queens, queenright

workers, queenless reproductive workers and queenless

sterile workers demonstrated that while caste differences

have the strongest influence on gene expression patterns,

social context had a significantly stronger effect than the

reproductive state [52�]. Thus, the social context can

influence the expression of genes involved in the pro-

duction of the social signals that create social context.

What is the effect of the group on theindividual’s behavior?An individual’s behavior is clearly strongly influenced by

the social cues and signals (such as pheromones) it receives

from other members of its social group [1��,53]. On a larger

scale, the overall behavioral tendencies of the group can

impact the behavior of the individual. For example, in

honey bee colonies of 50% mixed aggressive Africanized

and 50% docile European bees, Africanized bees were

more likely to guard than European bees, but there was

no difference in guarding behavior between these two

genotypes in colonies of 25% Africanized, 75% European







Figure 3

Individual Variation

ColonyPhenotype

Linear Non-Linear

without interaction, average personality dominates

with interaction, extreme personalities can dominate

recruitment

foragingbehavior

CORPSE

undertakingbehavior

Linear Effect

Non-linear Effect

(a) (b)


The effect of individual variation on colony behavioral tendencies. Individual variation can alter the group phenotype either linearly or non-linearly

(a). When the colony behavioral phenotype is an average of the component individual behavioral phenotypes, the mechanism can be assumed to be

linear. Alternately, when the colony behavioral phenotype reflects the most extreme individual(s), the mechanism is likely to be non-linear. This linearity

or non-linearity comes about through the strength and dynamics of the interactions between individuals in the group (b). Namely, when a positive-

feedback loop is employed to recruit or amplify a response, more extreme individuals who initiate these responses will have a disproportionately large

effect on the group. Note, however, that colony-level behaviors may reflect a combination of linear and non-linear behaviors.

bees [54]. In cross-fostering experiments with anarchistic

honey bee workers (which activate their ovaries even in

the presence of a queen and brood) and wild-type

workers, anarchistic bees always activated their ovaries

more readily than wild-type workers under any con-

dition, but both types of bees were more likely to activate

their ovaries when housed in an anarchistic colony,

suggesting reduced production of queen and brood

pheromones (which typically inhibit worker reproduc-

tion [55]) in anarchistic colonies [56]. In many cases, it is

possible that the behavior of the individual simply

reflects its response thresholds in the context of the

group: for example, a moderately aggressive individual

in the presence of highly aggressive individuals may not

exhibit aggression.

However, several recent studies suggest that the effect of

the social group can go beyond simply ranking individuals

according to their response thresholds. When honey bees

from genotypes displaying low levels of hygienic behavior

(removal of diseased larvae) were caged with bees dis-

playing high levels of hygienic behavior, the low-line bees

actually increased their hygienic activity [57]. This

increased task performance did not appear to be due to

recruitment by the high-line bees, but rather to a more

general effect of the social context. Studies of fanning

behavior in small groups of caged honey bees demon-

strated that group size strongly influenced the likelihood

that individual bees would fan (a greater percentage of

bees fan in larger groups) as well as the temperature

threshold at which fanning would be initiated (larger

groups fan at lower thresholds) [58]. Group size also


dx.doi.org/10.1016/j.cois.2014.09.010


influences information flow through colonies: larger colo-

nies of honey bees with intact dance language were more

efficient at recruiting foragers to new food resources

than smaller colonies [59]. Thus, individual response

thresholds and information flow, both of which shape

group behavior, can change with group size and likely

with other components of group dynamics (for review of

effects of group size see Ref. [60]).

A number of studies on global gene expression patterns

across many social insect species have demonstrated

that social environment often has a greater impact on

global gene expression than an individual’s own repro-

ductive state (in fire ants [61�,62], papers wasps [17�],bumble bees [63�] and honey bees [52�]). Furthermore, in

the case of paired fire ant foundresses which formed

hierarchies with a dominant and subordinate foundress,

switching the dominance rank by switching partners had a

greater impact on gene expression patterns than the rank

itself [61�]. In the case of the Gp-9 non-recombining

region in fire ants (Figure 2), when comparing workers

of the same genotype, Gp-9B/Gp-9B, in either the social

context of monogyne colonies (with Gp-9B/Gp-9B queens

and workers) or polygyne colonies (which contain both

Gp-9B/Gp-9b and Gp-9B/Gp-9B workers and multiple

Gp-9B/Gp-9b queens), worker gene expression profiles

reveal a greater effect of the colony’s behavioral pheno-

type than of the worker’s own genotype [62]. Thus,

genomic expression patterns can represent a sensitive

bioassay for the effects of social context, but the function

of these expression differences and the cues that regulate

them remain to be determined.





6 Social Insects


The role of individual variation in rapidcollective phasic transitionsSocial insect colonies can also exhibit rapid collective

phasic transitions during their lifecycle. In these beha-

viors, multiple individuals, regardless of their current

activities, are induced to perform a coordinated behavior,

such locusts transitioning from solitary to gregarious

phase [64], ants engaging in mating flights [65], as well

as reproductive colony fission and colony migration in ant,

bee and wasp species [9,10,65,66]. For example, when

honey bee colonies undergo reproductive fission —

‘swarming’ (reviewed in Refs. [43,66]) — thousands of

bees forego their normal behavioral diversity and tran-

sition through a series of stereotyped behaviors: exodus of

the swarm from the mother colony, temporary relocation

to a bivouac while scouts search for a new nest site, lift off

of the swarm and finally migration to the new nest site to

establish the new colony. These unified transitions

through a specific series of stereotyped behaviors are

analogous to a dynamical system transitioning through

stable fixed points (Figure 1d).

As in other social behaviors, rapid collective phasic tran-

sitions can be brought about by active signaling from

keystone individuals, as in swarming in honey bees

[43,66] or colony migration in some ants [9]. In the

case of honey bees, when colony exodus or swarm liftoff

is imminent, a small subset of individuals (the scouts,

representing �5% of the swarm) use vibration signals,

piping, and/or buzz-running to activate the other individ-

uals in the group and trigger swarm movement. These

scouts also search for and select the new nest site, and

lead the airborne swarm to the new nest by performing

‘streaking’ behavior, in which they fly rapidly in the

direction of the new nest. Thus, a small handful of key-

stone individuals are responsible for directing the swarm’s

behavior and movement. Interestingly, the same bees that

serve as scouts during swarming serve as scouts for new

foraging resources in established colonies, and thus com-

mon genomic and neurophysiological processes likely

drive these behaviors [24]. It is unknown if there is differ-

entiation among the non-scout bees in the swarm: typically,

the young bees in the colony join the swarm and thus they

are likely not particularly variable since they are in the very

early stages of behavioral maturation, but there does

appear to be genetic variation in the likelihood to join a

swarm (reviewed in Ref. [66]). In Temnothorax ants, a group

of individuals initiate scouting and evaluation of new nests,

and then recruit others through tandem running. The

scouts and recruits then physically carry the remaining

ants to the new nest site. Thus, the recruits may exhibit

a phasic transition in their behavioral state, but the

individuals who are simply carried to the new site likely

do not experience any behavioral change. Interestingly,

among the primary scouts there can be significant variation

in their effectiveness to find nest sites and recruit others

to those sites, due to their previous experiences [67].


dx.doi.org/10.1016/j.cois.2014.09.010


Rapid collective phasic transitions can also be brought

about passively, through an integration of group status by

each individual, which in turn translates to behavioral

change. Colony fission in both honey bees [43,66] and in

Temnothorax ants [68] is correlated with a threshold in-

nest density, a passive measure. In honey bees, increasing

population size and density is thought to limit the distri-

bution of queen pheromones throughout the colony,

triggering the production of new queen rearing which

is the earliest indicator that a colony is preparing to swarm

[43,66]. Within-nest population density in Temnothoraxants is correlated with polydomy, such that ant colonies

undergoes fission and separate into multiple nests when

population density inside the nest is high [68]. In locusts,

an increase in population density promotes the individual

to transition from the solitary to the gregarious phase [64].

Increasing population density leads to changes in levels of

microRNA-133 in the brain of individual locusts [69��].This microRNA triggers changes in expression of various

genes involved in the dopamine synthesis pathway,

resulting in the dramatic behavioral, physiological, and

morphological changes associated with this behavioral

phase change [64,69��]. Shifts between these two phases

are even accompanied by hysteresis (Figure 1c), such that

population densities that triggered the shift to gregarious

phase needed to be greatly reduced to initiate the tran-

sition back to solitary phase [64].

Rapid collective phasic transitions represent a special

type of social behavior, in which the group social behavior

temporarily overrides the variation present in individuals,

and thus provides a fascinating system in which to

examine the interplay between individual and group-

level behavioral variation. For example, unlike stable

social groups, there is low behavioral variation amongst

individuals engaging in these phasic transitions. In these

cases, behavioral diversity might negatively affect group

function, though this remains to be tested. The individ-

uals participating in a rapid collective phasic transition

appear to enter a common behavioral state, suggesting

that pre-existing preferences and tendencies may be

erased, and it is unclear if and how these re-emerge after

the phasic transition is completed, and how this process is

regulated at the genomic and neurophysiological level.

Finally, it remains to be determined what mechanisms

allow the keystone individuals (i.e., the scout bees) to

remain immune to the effects of the social group and able

to maintain their own distinct behavioral repertoires.

ConclusionSocial insects exhibit considerable inter-individual and

inter-colonial variation in behavior, as well as behavioral

shifts that take place over different timescales in both

the individual-scale and colony-scale. Furthermore, the

behavioral tendencies of the individual influence the

tendencies of the group, and vice versa (‘me’ affects

‘we’ and ‘we’ affects ‘me’). However, the relative strength







of these effects has not been broadly considered (for

example, to what extent can the group override the

tendencies of the individual? To what extent are group

behaviors the result of additive or synergistic interactions

among individual group members?). Just as network

motifs such as feedback loops have been extremely help-

ful in the understanding of genetic and neural circuits

[42��], these structures can also inform our experimental

approaches to social networks, where qualitatively differ-

ent behaviors emerge based on the strength and dynamics

of the component social interactions. While there has

been ample research on the proximate mechanisms med-

iating individual variation, we have less information about

the genes that underpin the production of social signals

which in turn influence group dynamics, and the genes

that are responsive to social context. Finally, studying the

interactions between the individual and the group during

rapid collective phasic transitions may be particularly

informative. As in stable groups, individuals can influence

the group’s behavior through active or passive mechan-

isms, but because these individuals will eventually per-

form the same behavior, the resultant changes are quite

dramatic. While the majority of the studies examining the

interplay between individual and group behavioral

tendencies have used stable groups, social groups them-

selves are dynamic: the role that individual variation plays

in the functioning of the group under diverse social and

environment conditions is an open area for investigation.

AcknowledgementsWe would like to thank Clare Rittschof, Pavan Ramdya, NathalieStroeymeyt and Laurent Keller for helpful discussion and comments on themanuscript. This work was supported by an NSF-CAREER grant to CMG,and ACL is supported by the University of Lausanne, the Swiss NationalScience Foundation, ERC Starting Independent Grant to Richard Bentonand an ERC Advanced Grant to Laurent Keller.

References

1.��

Jandt JM, Bengston S, Pinter-Wollman N, Pruitt JN, Raine NE,Dornhaus A, Sih A: Behavioural syndromes and social insects:personality at multiple levels. Biol Rev Camb Philos Soc 2014,89:48-67.

A comprehensive review of the application of ‘behavioral syndromes andpersonalities’ to study of social insects.

2.��

Jeanson R, Weidenmuller A: Interindividual variability in socialinsects — proximate causes and ultimate consequences. BiolRev 2013.

A comprehensive review of the proximate factors regulating individualbehavioral variation in social insects and effects of this variation on socialgroups.

3. Detrain C, Deneubourg JL: Self-organized structures in asuperorganism: do ants ‘behave’ like molecules? Phys Life Rev2006, 3:162-187.

4. Mattila HR, Seeley TD: Genetic diversity in honey bee coloniesenhances productivity and fitness. Science 2007, 317:362-364.

5.��

Modlmeier AP, Liebmann JE, Foitzik S: Diverse societies aremore productive: a lesson from ants. Proc Biol Sci 2012,279:2142-2150.

This paper provides further evidence that increased behavioral variationamong individual workers leads to improved colony performance. Whileprevious studies in honey bees manipulated genetic diversity (e.g.,inseminated the queen with semen from one versus ten males) in order


dx.doi.org/10.1016/j.cois.2014.09.010


to manipulate behavioral variation within the colony, this study usednaturally occurring colonies of Temnothorax longispinosus ants andquantified behavioral variation among individuals and correlated thisparameter to colony level traits.

6. Oldroyd BP, Fewell JH: Genetic diversity promoteshomeostasis in insect colonies. Trends Ecol Evol 2007,22:408-413.

7.�

Evans LJ, Raine NE: Changes in learning and foragingbehaviour within developing bumble bee (Bombus terrestris)colonies. PLOS ONE 2014:9.

This manuscript explores how the foraging strategies of individual colonymembers change during colony growth and maturation.

8. Sumpter DJT: The principles of collective animal behaviour.Philos Trans R Soc B Biol Sci 2006, 361:5-22.

9. Visscher PK: Group decision making in nest-site selectionamong social insects. Annu Rev Entomol 2007, 52:255-275.

10. McGlynn TP: The ecology of nest movement in social insects.Annu Rev Entomol 2012, 57:291-308.

11. Scheiner R, Page RE, Erber J: Review article Sucroseresponsiveness and behavioral plasticity in honey bees(Apis mellifera). Apidologie 2004, 35:133-142.

12. Arenas A, Giurfa M, Sandoz JC, Hourcade B, Devaud JM,Farina WM: Early olfactory experience induces structuralchanges in the primary olfactory center of an insect brain. Eur JNeurosci 2012, 35:682-690.

13. Hoover SER, Winston ML, Oldroyd BP: Retinue attraction andovary activation: responses of wild type and anarchistic honeybees (Apis mellifera) to queen and brood pheromones. BehavEcol Sociobiol 2005, 59:278-284.

14. Pankiw T: Queen rearing by high and low queen mandibularpheromone responding worker honey bees (Apis mellifera L.).Can Entomol 1997, 129:679-690.

15. Kocher SD, Ayroles JF, Stone EA, Grozinger CM: Individualvariation in pheromone response correlates with reproductivetraits and brain gene expression in worker honey bees. PLoSONE 2010, 5:e9116.

16. Alaux C, Sinha S, Hasadsri L, Hunt GJ, Guzman-Novoa E,DeGrandi-Hoffman G, Uribe-Rubio JL, Southey BR, Rodriguez-Zas S, Robinson GE: Honey bee aggression supports a linkbetween gene regulation and behavioral evolution. Proc NatlAcad Sci U S A 2009, 106:15400-15405.

17.�

Toth AL, Tooker JF, Radhakrishnan S, Minard R, Henshaw MT,Grozinger CM: Shared genes related to aggression, rather thanchemical communication, are associated with reproductivedominance in paper wasps (Polistes metricus). BMC Genomics2014, 15:75.

In a genome-wide analysis of gene expression patterns in paper wasps,the authors demonstrate that environment (either season or group com-position) is a greater contributor to brain gene expression patterns thancaste or physiological state, indicating that environmental factors canpotentially override individual variation.

18.�

Modlmeier AP, Keiser CN, Watters JV, Sih A, Pruitt JN: Thekeystone individual concept: an ecological and evolutionaryoverview. Anim Behav 2014, 89:53-62.

Comprehensive and clarifying review on the keystone individual concept.

19. Tenczar P, Lutz CC, Rao VD, Goldenfeld N, Robinson GE:Automated monitoring reveals extreme interindividualvariation and plasticity in honeybee foraging activity levels.Anim Behav 2014, 95:41-48.

20. Ferreira PG, Patalano S, Chauhan R, Ffrench-Constant R,Gabaldon T, Guigo R, Sumner S, Gabaldon T, Guigo R:Transcriptome analyses of primitively eusocial waspsreveal novel insights into the evolution of sociality andthe origin of alternative phenotypes. Genome Biol 2013,14:R20.

21. Grozinger CM, Fan Y, Hoover SER, Winston ML: Genome-wideanalysis reveals differences in brain gene expression patternsassociated with caste and reproductive status in honey bees(Apis mellifera). Mol Ecol 2007, 16:4837-4848.



http://refhub.elsevier.com/S2214-5745(14)00093-5/sbref0005












































































8 Social Insects


22. Graff J, Jemielity S, Parker JD, Parker KM, Keller L: Differentialgene expression between adult queens and workers in the antLasius niger. Mol Ecol 2007, 16:675-683.

23. Bonasio R, Li Q, Lian J, Mutti NS, Jin L, Zhao H, Zhang P, Wen P,Xiang H, Ding Y et al.: Genome-wide and caste-specific DNAmethylomes of the ants camponotus floridanus andharpegnathos saltator. Curr Biol 2012, 22:1755-1764.

24. Liang ZS, Nguyen T, Mattila HR, Rodriguez-Zas SL, Seeley TD,Robinson GE: Molecular determinants of scouting behavior inhoney bees. Science 2012, 335:1225-1228.

25. Muscedere ML, Johnson N, Gillis BC, Kamhi JF, Traniello JF:Serotonin modulates worker responsiveness to trailpheromone in the ant Pheidole dentata. J Comp Physiol ANeuroethol Sens Neural Behav Physiol 2012, 198:219-227.

26. Brockmann A, Annangudi SP, Richmond TA, Ament SA, Xie F,Southey BR, Rodriguez-Zas SR, Robinson GE, Sweedler JV:Quantitative peptidomics reveal brain peptide signatures ofbehavior. Proc Natl Acad Sci U S A 2009, 106:2383-2388.

27. Galizia CG, Kreissl S: Honeybee neurobiology and behavior. InHoneybee neurobiology and behavior. Edited by Galizia CG,Eisenhardt D, Giurfa M. Netherlands: Springer; 2012:211-226.

28. Gordon DM: The rewards of restraint in the collectiveregulation of foraging by harvester ant colonies. Nature 2013,498:91-93.

29.�

Page RE, Rueppell O, Amdam GV: Genetics of reproduction andregulation of honeybee (Apis mellifera L.) social behavior.Annu Rev Genet 2012, 46:97-119.

Provides excellent overview of the mechanisms regulating worker fora-ging strategies in honey bees and how these shape colony-level traits.Furthermore, the review highlights the role of common mechanismsunderlying behavioral variation across multiple timescales.

30. Linksvayer TA, Fondrk MK, Page RE: Honeybee social regulatorynetworks are shaped by colony-level selection. Am Nat 2009,173:E99-E107.

31.�

Wang J, Wurm Y, Nipitwattanaphon M, Riba-Grognuz O, Huang Y-C, Shoemaker D, Keller L: A Y-like social chromosome causesalternative colony organization in fire ants. Nature 2013,493:664-668.

The authors describe the large genetic polymorphism underlying a majordifference in fire ant colony phenotype. It is worth considering what othercolony-wide differences might have such a genetic element underpinningthem.

32. Ross KG, Keller L: Experimental conversion of colony socialorganization by manipulation of worker genotype compositionin fire ants (Solenopsis invicta). Behav Ecol Sociobiol 2002,51:287-295.

33. Ross K, Keller L: Ecology and evolution of social organization:insights from fire ants and other highly eusocial insects. AnnuRev Ecol Syst 1995, 26:631-656.

34. Richardson TO, Christensen K, Franks NR, Jensen HJ, Sendova-Franks AB: Ants in a labyrinth: a statistical mechanicsapproach to the division of labour. PLoS ONE 2011, 6:e18416.

35. Mersch DP, Crespi A, Keller L: Tracking individuals showsspatial fidelity is a key regulator of ant social organization.Science 2013, 340:1090-1093.

36. Sendova-Franks AB, Hayward RK, Wulf B, Klimek T, James R,Planque R, Britton NF, Franks NR: Emergency networking:famine relief in ant colonies. Anim Behav 2010, 79:473-485.

37.��

Goldsby HJ, Dornhaus A, Kerr B, Ofria C: Task-switching costspromote the evolution of division of labor and shifts inindividuality. Proc Natl Acad Sci U S A 2012, 109:13686-13691.

A computational modeling and in silico evolution approach to determiningthe costs of task switching in division of labor.

38. Jeanson R, Fewell JH, Gorelick R, Bertram SM: Emergence ofincreased division of labor as a function of group size. BehavEcol Sociobiol 2007, 62:289-298.

39. Flanagan TP, Letendre K, Burnside WR, Fricke GM, Moses ME:Quantifying the effect of colony size and food distribution onharvester. Ant Foraging 2012, 7:1-9.


dx.doi.org/10.1016/j.cois.2014.09.010


40. Holbrook CT, Barden PM, Fewell JH: Division of labor increaseswith colony size in the harvester ant Pogonomyrmexcalifornicus. Behav Ecol 2011, 22:960-966.

41. Sun Q, Zhou X: Corpse management in social insects. Int J BiolSci 2013, 9:313-321.

42.��

Alon U: An introduction to systems biology: design principles ofbiological circuits. 2007.

An excellent introduction to biological circuit and network analysis.

43. Seeley TD: Honeybee democracy. Princeton University Press;2010.

44.��

Shaffer Z, Sasaki T, Pratt SC: Linear recruitment leads toallocation and flexibility in collective foraging by ants. AnimBehav 2013, 86:967-975.

This study combines theory and experiment to analyze different types ofpositive feedback through recruitment.

45.�

Arganda S, Perez-Escudero A, de Polavieja GG: A common rulefor decision making in animal collectives across species. ProcNatl Acad Sci U S A 2012, 109:20508-20513.

A modeling study working with decision-making data from a range ofbiological organisms.

46. Pruitt JN, Riechert SE: How within-group behavioural variationand task efficiency enhance fitness in a social group. Proc BiolSci 2011, 278:1209-1215.

47.��

Pruitt JN, Grinsted L, Settepani V: Linking levels of personality:personalities of the ‘average’ and ‘most extreme’ groupmembers predict colony-level personality. Anim Behav 2013,86:391-399.

This study demonstrates how the personality traits of the individualsshape the personality of the group, and specifically examines the relativeeffects of extreme vs average individuals.

48. Paleolog J: Behavioural characteristics of honey bee(Apis mellifera) colonies containing mix of workers ofdivergent behavioural traits. Anim Sci Pap Rep 2009,27:237-248.

49. Guzman-Novoa et al.: Genotypic effects of honey bee (Apismellifera) defensive behavior at the individual and colonylevels: the relationship of guarding, pursuing and stinging.Apidologie 2004, 35:15-24.

50. Grozinger CM, Sharabash NM, Whitfield CW, Robinson GE:Pheromone-mediated gene expression in the honey bee brain.Proc Natl Acad Sci U S A 2003, 100 Suppl.:14519-14525.

51. Alaux C, Le Conte Y, Adams HA, Rodriguez-Zas S, Grozinger CM,Sinha S, Robinson GE: Regulation of brain gene expression inhoney bees by brood pheromone. Genes Brain Behav 2009,8:309-319.

52.�

Malka O, Nino EL, Grozinger CM, Hefetz A: Genomic analysis ofthe interactions between social environment and socialcommunication systems in honey bees (Apis mellifera). InsectBiochem Mol Biol 2014, 47:36-45.

This analysis of global gene expression patterns in pheromone-producingglands in honey bees which demonstrate that social status (presence ofabsence of a queen) plays a much greater role in regulating geneexpression patterns than ovary activation state, indicating that environ-mental factors can potentially override individual variation.

53. Janis I: Victims of groupthink: a psychological study of foreign-policy decisions and fiascoes. Mifflin: Houghton; 1972, .

54. Hunt GJ, Guzman-Novoa E, Uribe-Rubio JL, Prieto-Merlos D:Genotype–environment interactions in honeybee guardingbehaviour. Anim Behav 2003, 66:459-467.

55. Le Conte Y, Hefetz A: Primer pheromones in socialhymenoptera. Annu Rev Entomol 2008, 53:523-542.

56. Barron AB, Oldroyd BP: Social regulation of ovary activation in‘anarchistic’ honey-bees (Apis mellifera). Behav Ecol Sociobiol2001, 49:214-219.

57. Gempe T, Stach S, Bienefeld K, Beye M: Mixing of honeybeeswith different genotypes affects individual worker behaviorand transcription of genes in the neuronal substrate. PLoSONE 2012:7.


























































































































58. Cook CN, Breed MD: Social context influences the initiationand threshold of thermoregulatory behaviour in honeybees.Anim Behav 2013, 86:323-329.

59. Donaldson-Matasci MC, DeGrandi-Hoffman G, Dornhaus A:Bigger is better: honeybee colonies as distributedinformation-gathering systems. Anim Behav 2013, 85:585-592.

60. Dornhaus A, Powell S, Bengston S: Group size and its effects oncollective organization. Annu Rev Entomol 2012, 57:123-141.

61.�

Manfredini F, Riba-Grognuz O, Wurm Y, Keller L, Shoemaker D,Grozinger CM: Sociogenomics of cooperation and conflictduring colony founding in the fire ant Solenopsis invicta. PLOSGenet 2013, 9:e1003633.

This study examined global gene expression patterns in ant queens whichfounded colonies either individually or in pairs. Expression patterns wereregulated primarily by social context (e.g., queens founding singly versuswith a co-foundress, or co-foundresses that switched versus maintaineddominance rank) than an individual’s own social rank/reproductive status.

62. Wang J, Ross KG, Keller L: Genome-wide expression patternsand the genetic architecture of a fundamental social trait.PLoS Genet 2008:4.

63.�

Amsalem E, Malka O, Grozinger C, Hefetz A: Exploring the role ofjuvenile hormone and vitellogenin in reproduction and socialbehavior in bumble bees. BMC Evol Biol 2014, 14:45.

This study demonstrated that expression levels of an egg-yolk protein,vitellogenin, are regulated by social interactions rather than an individual’sovary activation levels, demonstating that social factors can potentiallyoverride individual variation arising from physiological differences, andthe function of vitellogenin may have been co-opted during the evolutionof sociality.


dx.doi.org/10.1016/j.cois.2014.09.010


64. Guttal V, Romanczuk P, Simpson SJ, Sword GA, Couzin ID:Cannibalism can drive the evolution of behaviouralphase polyphenism in locusts. Ecol Lett 2012,15:1158-1166.

65. Peeters C, Ito F: Colony dispersal and the evolution of queenmorphology in social Hymenoptera. Annu Rev Entomol 2001,46:601-630.

66. Grozinger CM, Richards J, Mattila HR:: From moleculesto societies: mechanisms regulating swarmingbehavior in honey bees (Apis spp.). Apidologie 2013,45:327-346.

67. Stroeymeyt N, Franks NR, Giurfa M:: Knowledgeableindividuals lead collective decisions in ants. J Exp Biol 2011,214:3046-3054.

68. Cao TT: High social density increases foraging and scoutingrates and induces polydomy in Temnothorax ants. Behav EcolSociobiol 2013, 67:1799-1807.

69.��

Yang M, Wei Y, Jiang F, Wang Y, Guo X, He J, Kang L:MicroRNA-133 inhibits behavioral aggregation bycontrolling dopamine synthesis in locusts.PLoS Genet 2014, 10:e1004206.

Excellent study elucidating a causal factor in how locusts detect popula-tion density changes and then enact the physiological and morphologicalchanges associated with the shift to gregarious phase.

70. Scharf I, Modlmeier AP, Fries S, Tirard C, Foitzik S: Characterizingthe collective personality of ant societies: aggressive coloniesdo not abandon their home. PLoS ONE 2012, 7:e33314.















































me and we: the interplay between individual and group

Documents