system for normal behavioral modulation. If this is so,
then pharmacologically elevated or lowered levels of
serotonin (or other modulators of behavior) could produce
similar or different behavioral phenotypes, depending on
the species examined and the multiplicity of effects
mediated by the pharmacologically altered substance.
A novel model system: aggression in male
fruit flies
Despite the wealth of information gathered from inverte-
brate models of aggression and that of the roles of amines
such as serotonin and octopamine in this behavior, little, if
any, information exists on exactly how amine neurons
function during fighting. Do their firing rates change? If
so, are the changes seen before, during or after fights?
What are the consequences of such changes? The major
difficulty that is shared by invertebrate and vertebrate
models, is that the physiological activity in these neurons
cannot easily be monitored while animals are fighting;
thus, although hypotheses are abundant, we are left with
the unsatisfactory conclusion that amines clearly are
instrumental in aggression, but how they serve this role
remains unknown.
One experimental approach that is not readily available in
most invertebrate models, is the use of genetics and the
accompanying powerful genetic experimental methods;
thus, there are no inbred lines of lobsters, crickets or
crayfish, and their genomes have not been sequenced.
Molecular approaches, including RNAi, probably can be
used with these species but it would be on an individual
animal basis and would not involve the generation of
experimental lines of, for example, highly aggressive
animals. In all respects, other than the size of their
neurons, fruit flies would be an ideal species to use,
providing that robust patterns of aggression can be trig-
gered in these animals under ethologically acceptable
experimental conditions.
Aggression and territoriality are well-known in certain
Hawaiian species of fruit flies [64,65]; what is less well-
known, however, is that aggression also exists in common
highly inbred species of fruit flies (Drosophila melano-
gaster). Originally described by Sturtevant in 1915 [66],
other investigators have also reported an aggression in
fruit flies during the period 1960­1990 [67­70]. This early
work described many of the comprising components of
the behavior, but in experimentally difficult scenarios
involving many individuals. This was simplified recently
to only a pair of male flies in a chamber with a small food
cup and a headless mated female, illuminated from
above. The food, potential mate and light serve to attract
males to the surface, and within a few minutes both flies
move onto the food cup, where they commence a series of
encounters [71
]. The flies used in these fights are
isolated as soon as they emerge as adults, and kept for
3­4 days in small test tubes containing food. Using these
experimental conditions, an ethogram of the behavior was
assembled, transition matrices were constructed and a
Markov Chain analysis produced a quantitative analysis of
the behavior. Flies meet and engage in encounters about
once a minute. These encounters show varying intensity
levels (Figure 1) and an average duration of 11 s; this
varies proportionately with the intensity level. A hier-
archical relationship is established relatively early in
fights, but losers can continue to re-engage winners for
several hours. With the behavior well characterized in a
relatively simple experimental paradigm, one can now
begin to take advantage of the powerful genetic methods
that are available.
Mutant studies of aggression in flies
A few publications have explored the effects of selected
classical mutations and chemical modifications of neuro-
transmitter levels on fighting behavior of flies. Ebony flies
fail to incorporate b-alanine into their cuticles and black
mutants show decreased synthesis of the same amino
acid. These defects, in normal cuticular tanning, result
in an early onset of courtship and enhanced territoriality
[68]. A more recent study [72] has corroborated the results
of earlier studies, which used the b-alanine mutants, and
has explored the roles of dopamine and serotonin (using
precursors of the amine or blockers of amine synthesis),
octopamine (using a null mutant) and mushroom bodies
(examined using transgenic animals expressing tetanus
Figure 1
(a) (b)
(c) (d)
The components of fighting behavior in adult male fruit flies: (a) wing
threat, in which both wings are elevated for a sustained period of time.
This is often seen early in fights; (b) fencing, in which animals push
off with one of their legs in a sideward or forward direction; (c) boxing,
in which both flies rise on their hind legs and thrash at each other
with their forelegs; (d) chase, seen when a hierarchical relationship
has been established, with the winner pursuing the loser who will either
fly from the food surface or retreat to the edge of the food dish.
Reprinted from [69], in a slightly modified form, with permission.
738 Neurobiology of behaviour
Current Opinion in Neurobiology 2003, 13:736­743 www.current-opinion.com
toxin in output neurons of the mushroom bodies). The
experimental protocol that was used here was compli-
cated, with six males and three mated females in a
chamber at the same time. The scoring system used
was also complicated; however, the results successfully
indicated that octopamine null mutants and flies with
reduced synaptic output from the mushroom bodies
showed reduced aggression, whereas, flies with elevated
dopamine showed somewhat reduced aggression. Altera-
tion of serotonin levels had no effect on aggression. It
should be noted that many important controls were not
included in these studies, which, according to the authors,
were preliminary. Repetition of these studies, using a
more simplified experimental protocol, with conditional
mutant lines of flies would greatly help to interpret the
observed results.
Much more powerful genetic methods are available for
behavioral studies on aggression with flies, however, than
have been reported thus far. These include conditional
expression of mutations in the fly brain, whenever and
wherever desired. One example is the GAL4/UAS sys-
tem, originally described by Brand and Perrimon [73]. In
this method, a cross between two transgenic lines of flies
(one expressing the transcription factor GAL4 in subtypes
of neurons, the other expressing the binding site for
GAL4, driving expression of any desired gene) yields
progeny in which desired genes are expressed in subtypes
of neurons in the fly brain. In preliminary studies, we have
expressed GAL4 in dopamine and serotonin neurons or
selectively in dopamine neurons to drive expression of a
temperature-sensitive mutant form of the protein dyna-
min in those cells [74
]. This protein behaves normally at
258, becoming mutated at 308 and leading to a rapid block
of vesicle recycling and hence of synaptic transmission at
the elevated temperature [75
]. In preliminary studies
(unpublished observations), we have observed decreased
numbers of encounters between flies at the elevated
temperature, using both GAL4 lines. Further experi-
ments with this and related technology should yield
valuable information about the role of subtypes of neu-
rons in aggression in flies.
Aggression in female flies
As with male flies, Sturtevant was the first to mention
aggression in female flies, in this case, directed towards
males: `occasionally a female seems to frighten off a male
by spreading her wings and moving quickly towards him'
[66]. A second mention of aggression in females came in a
study comparing mating success in ebony and light males
of freshly isolated D. melanogaster [65]. A more complete
study of aggression in females was published last year,
however, by Ueda and Kidokoro, using the common
Canton-S strain of D. melanogaster [76
]. They observed
and scored three patterns of female aggression, including
`approach', `lunge' and `wings erect'. The authors report
that these patterns `are similar to those of male aggression
in this species' but that they differ from the patterns
shown in female rejection behavior during courtship.
Females show enhanced fighting if live yeast are growing
on the surface of the food dish; conditioning female flies
to the yeast food source $12 hours before their fights,
reduced the extent of fighting. This was interpreted as a
possible adaptation to the enriched food source. Higher
levels of aggression were seen when newly emerged
females were held in isolation before the fights, in com-
parison with females held in groups of ten in a vial. Thus,
both housing conditions and the quality of the food source
influence fighting behavior in female flies, and, as with
males, newly emerged females began to show aggression
only a day after their emergence as adults.
Recently, we confirmed those behavioral patterns
reported by Ueda and Kidokoro [76
], but also noted
that specific differences are seen between the behavioral
patterns in male and female D. melanogaster (unpublished
observations). Certain patterns are seen predominantly in
male flies, not in females (extended-duration wing-threat,
boxing, tussling, holding and chasing), whereas, other
patterns are seen mainly in females and not males (lung-
ing with head butting, front limb fencing in an elevated
posture). The remaining offensive and defensive patterns
beyond these (for ethograms of male fighting behavior see
[69]) are shared and seen in both males and females.
Of particular relevance here, is a report by Lee and Hall
of a new behavioral phenotype seen in fruitless ( fru)
mutant male flies [77]: enhanced head interactions.
Fruitless male flies court other males, forming long court-
ship chains, and show abnormal patterns of wing vibra-
tion during courtship, but otherwise appear capable of
mating with females, depending on the fru mutant sub-
type [78]. In addition to enhanced head interactions,
fru mutant males reportedly do not display boxing, which
is a common high intensity component of male fruit fly
fighting behavior. The head­head interactions are not
seen during the first day after eclosion; their occurrence
increases to a maximum 4­5 days later, in parallel with
the chaining behavior.
These results raise an interesting question: does the fru
mutation, with its well-studied effects on mating beha-
vior, also direct the expression of female patterns of
fighting behavior in the brains of male fru mutants?
Are male- and female-specific patterns of fighting beha-
vior specified in the brains of fruit flies after gender has
been defined by the sex-determination hierarchy of
genes? Is there an aggression-determination hierarchy
of genes? Are male- and female-specific patterns of
aggression defined as units in the brains of flies or can
components of the behavior (like head butting in females
and tussling in males) be transferred as individual mod-
ules into male and female brains, using methods of
genetic manipulation? The fruit fly model offers exciting
Aggression in invertebrates Kravitz and Huber 739
www.current-opinion.com Current Opinion in Neurobiology 2003, 13:736­743
Table 1
Aggression studies in various species of invertebrates (a partial listing, including only articles cited in this publication)
Ethology Physiology Amines References
INSECTS (social)
Bees 
Apis mellifera   [2,13,14,15
,19]
Apis florea  [3]
Apis cerana  [4]
Andrena scotica  [7]
Panurgus calcaratus  [7]
Bombus terrestris   [22]
Ants
Formica pratensis  [5]
Formica exsecta  [12]
Linepithema humile  [6,10,11,17]
Soleenopsis invicta  [16,18]
Cataglyphis niger  [20]
Odontomachus brunneus  [21]
Termites
Reticulitermes spp  [9]
Wasps
Vespa crabro  [8]
INSECTS (non-social)
Dragon flies
Pachidiplex longipennis  [28]
Crickets 
Gryllus bimaculatus   [29­34]
Fruit flies 
Drosophila melanogaster  y
 [67-70,71
,72,76
,77,78]
Drosophila simulans  [70]
Drosophila sylvestris  [64,65]
Drosophila heteroneura  [64,65]
Drosophila ampelophila  [66]
CRUSTACEANS
Lobsters 
Homarus americanus    [36
,37
,44,49,58,60
,61,63]
Munida quadrispina   [59]
Crabs
Carcinus maenas  
 [42]
Crayfish 
Orconectes rusticus   [45,46,48,50­52,55]
Astacus astacus  
 [60
,62]
Procambarus clarkii    [39,41
,43,56]
Pacifastacus leniusculus    [40,54]
OTHER ARTHROPODS
Spiders 
Dolomedes triton  [26]
Misumena vatia  [27]
Portia spp.  [23]
Latrodectus hasselti  [24]
The table lists species in which: (i) ethological studies on aggression have been conducted (ethology); (ii) physiological and/or anatomical studies
directly relating to aggression have been conducted (physiology); a role of amines in aggression has been demonstrated (amines).

Anatomical and/or electrophysiological studies using intracellular and/or extracellular recording methods have been conducted in these species,
but these are not directly related to aggression.
y
Genetic studies related to aggression have been conducted.
740 Neurobiology of behaviour
Current Opinion in Neurobiology 2003, 13:736­743 www.current-opinion.com
possibilities for the identification of genes involved in
producing patterns of a complex behavior like aggression
in nervous systems. The shaping of these patterns, by
experience and by hormones, throughout the lives of such
organisms can then be fully elucidated.
Conclusions
Aggression is seen and has been studied behaviorally in
many invertebrate species (Table 1), including social and
non-social insects, other arthropods and crustaceans. In
many of these species, detailed physiological and anato-
mical studies are now possible; they have been performed
in relatively few of the insect species (mainly crickets and
bees), but more widely among the crustaceans. Crusta-
cean species are exciting because they enable the study of
aggression to the level of the individual synapses, neurons
and circuits that are key to this behavior. In all inverte-
brate systems examined thus far, altering the levels or
function of amine neurons causes important changes in
aggression; in one particular model, changes in status are
accompanied by alterations in the function of certain
synaptic regions relating to amines. Despite this progress,
which goes far beyond our knowledge of vertebrate sys-
tems, no clear picture has yet emerged of exactly how
amines modulate or alter the behavior. Recent experi-
ments using fruit flies as a model for aggression have
introduced the possibility of adding powerful genetic
methods to the armamentarium of tools that are available
for the study of aggression in invertebrates. Ultimately,
this will allow investigators more insight into how this
complex pattern of behavior is assembled in the nervous
system and should shed further light on the role of
modulators in the behavior.
Acknowledgements
It is difficult to summarize as large a field as aggression in invertebrates in a
short review of this type, particularly one that tries to concentrate on
recent literature. Moreover, studies in this large field use many species of
animals and range from purely ethological examinations of agonistic behavior
to physiological, molecular and genetic studies viewed from a
neuroethological perspective. With the space constraints facing us, we
recognized in advance that we could not do justice to the studies of all
the excellent investigators who work in this field, and apologize to those
of our colleagues who feel slighted by our somewhat amine-centric review
of the field. EA Kravitz was supported by grants from NSF (IBN 90730),
NIGMS (GM65595) and a starter grant from the Gund Fund of the Mind,
Brain Behavior Program at Harvard University. R Huber was supported by
grants from NSF (IBN 9874608) and NIMH (62557). Mostly, we
acknowledge the outstanding contributions of the dedicated young
undergraduate and graduate students and post-doctoral fellows who have
worked with us over the years on the studies discussed here.
References and recommended reading
Papers of particular interest, published within the annual period of
review, have been highlighted as:
 of special interest
 of outstanding interest
1. Haspel G, Libersat F: Wasp venom blocks central cholinergic
synapses to induce transient paralysis in cockroach prey.
J Neurobiol 2003, 54:628-637.
2. Bernasconi G, Ratnieks FLW, Rand E: Effect of `spraying' by
fighting honey bee queens (Apis mellifera L.) on the temporal
structure of fights. Insectes Sociaux 2000, 47:21-26.
3. Halling LA, Oldroyd BP, Wattanachaiyingcharoen W, Barron AB,
Nanork P, Wongsiri S: Worker policing in the bee Apis florea.
Behav Ecol Sociobiol 2001, 49:509-513.
4. Oldroyd BP, Halling LA, Good G, Wattanachaiyingcharoen W,
Barron A, Nanork P, Wongsiri S, Ratnieks FLW: Worker policing
and worker reproduction in Apis cerana. Behav Ecol Sociobiol
2001, 50:371-377.
5. Beye M, Neumann P, Chapuisat M, Pamilo P, Moritz RFA:
Nestmate recognition and the genetic relatedness of nests
in the ant Formica pratensis. Behav Ecol Sociobiol 1998,
43:67-72.
6. Roulston TH, Buczkowski G, Silverman J: Nestmate
discrimination in ants: effect of bioassay on aggressive
behavior. Insectes Sociaux 2003, 50:151-159.
7. Paxton RJ, Kukuk PF, Tengo¨ J: Effects of familiarity and
nestmate number on social interactions in two communal
bees, Andrena scotica and Panurgus calcaratus
(Hymenoptera, Andrenidae). Insectes Sociaux 1999, 46:109-118.
8. Ruther J, Sieben S, Schricker B: Nestmate recognition in social
wasps: manipulation of hydrocarbon profiles induces
aggression in the European hornet. Naturwissenschaften 2002,
89:111-114.
9. Polizzi JM, Forschler BT: Factors that affect aggression among
the worker caste of Reticulitermes spp. subterranean termites
(isoptera: rhinotermitidae). J Insect Behav 1999, 12:133-146.
10. Liang D, Blomquist GJ, Silverman J: Hydrocarbon-released
nestmate aggression in the Argentine ant, Linepithema humile,
following encounters with insect prey. Comp Biochem Physiol B
Biochem Mol Biol 2001, 129:871-882.
11. Suarez AV, Holway DA, Liang D, Tsutsui ND, Case TJ:
Spatiotemporal patterns of intraspecific aggression in the
invasive Argentine ant. Anim Behav 2002, 64:697-708.
12. Brown W, Liautard C, Keller L: Sex-ratio dependent execution of
queens in polygynous colonies of the ant Formica exsecta.
Oecologia 2003, 134:12-17.
13. Giray T, Robinson GE: Common endocrine and genetic
mechanisms of behavioral development in male and worker
honey bees and the evolution of division of labor. Proc Natl Acad
Sci USA 1996, 93:11718-11722.
14. Pearce AN, Huang ZY, Breed MD: Juvenile hormone and
aggression in honey bees. J Insect Physiol 2001, 47:1243-1247.
15.

Schulz DJ, Sullivan JP, Robinson GE: Juvenile hormone and
octopamine in the regulation of division of labor in honey bee
colonies. Horm Behav 2002, 42:222-231.
Clear evidence of endocrine control of caste in honeybees, with emphasis
of specific roles for octopamine and juvenile hormone. Treatment with
either octopamine or juvenile hormone induces foraging behavior. The
results suggest further that juvenile hormone influences foraging behavior
in part by increasing levels of octopamine in the brain, but octopamine
also exerts actions that are independent of juvenile hormone.
16. Klobuchar EA, Deslippe RJ: A queen pheromone induces
workers to kill sexual larvae in colonies of the red imported
fire ant (Solenopsis invicta). Naturwissenschaften 2002,
89:302-304.
17. Reuter M, Balloux F, Lehmann L, Keller L: Kin structure and queen
execution in the Argentine ant Linepithema humile. J Evol Biol
2001, 14:954-958.
18. Vander Meer RK, Alonso LE: Queen primer pheromone affects
conspecific fire ant (Solenopsis invicta) aggression. Behav Ecol
Sociobiol 2002, 51:122-130.
19. Gilley DC: The behavior of honey bees (Apis mellifera ligustica)
during queen duels. Ethology 2001, 107:601-622.
20. Nowbahari E, Fé né ron R, Malherbe MC: Effect of body size on
aggression in the ant, Cataglyphis niger (hymenoptera;
formicidae). Aggressive Behav 1999, 25:369-379.
21. Powell S, Tschinkel WR: Ritualized conflict in Odontomachus
brunneus and the generation of interaction-based task
allocation: a new organizational mechanism in ants.
Anim Behav 1999, 58:965-972.
Aggression in invertebrates Kravitz and Huber 741
www.current-opinion.com Current Opinion in Neurobiology 2003, 13:736­743