Aggression in invertebrates
Edward A Kravitz
and Robert Hubery
Invertebrates are outstanding model systems for the study of
aggression. Recent advances and promising new research
approaches are bringing investigators closer to the goal of
integrating behavioral findings with those from other disciplines
of the neurosciences. The presence of highly structured, easily
evoked behavioral systems offer unique opportunities to quantify
the aggressive state of individuals, to explore the mechanisms
underlying the formation and maintenance of dominance
relationships, to investigate the dynamic properties of
hierarchy formation, and to explore the significance of
neural, neurochemical and genetic mechanisms in these
behavioral phenomena.
Addresses

Department of Neurobiology, Harvard Medical School, 220 Longwood
Avenue, Boston, MA 02115, USA
e-mail: edward_kravitz@hms.harvard.edu
y
Bowling Green State University, Biological Sciences, Life Sciences
Building, Bowling Green, OH 43403, USA
e-mail: lobsterman@caspar.bgsu.edu
Current Opinion in Neurobiology 2003, 13:736­743
This review comes from a themed issue on
Neurobiology of behaviour
Edited by Mark Konishi and Randolf Menzel
0959-4388/$ ­ see front matter
ß 2003 Elsevier Ltd. All rights reserved.
DOI 10.1016/j.conb.2003.10.003
Introduction
This review describes several invertebrate models in
which intraspecific aggression is readily evoked in dyadic
interactions between animals; these models enable stu-
dies to be performed at levels ranging from the behavioral
through the physiological and ultimately to the molecular
and genetic levels. Elegant models of interspecific aggres-
sion in invertebrates also exist [1], but these will not be
dealt with here. In the species described here, agonistic
behavior patterns appear to be pre-wired in the nervous
system, as animals with no previous social experience can
engage conspecifics in normal agonistic encounters. Dur-
ing such fights, paired animals exchange highly stereo-
typical behaviors that escalate through different intensity
levels and that, ultimately, result in a decision with
behavioral consequences for both winners and losers. A
common theme in these studies is that amines, peptides
and steroid hormones, substances that function as neu-
romodulators and/or as neurohormones, serve as impor-
tant modulators of aggression.
Aggression in social insects: bees, ants,
wasps and termites
It may be surprising that aggression is seen in social
insects, considering that selfish behavior is rare in groups
with shared reproductive interests, such as honey bee
colonies; however, stereotyped agonistic behavior within
a hive [2] is common during worker policing [3,4]. More-
over, aggression in the context of nestmate recognition
has been explored in ants [5,6], bees [7], wasps [8] and
termites [9] where the determination of self versus non-
self is frequently based on the expression of cuticular
hydrocarbon profiles [10,11]. With pheromonal com-
mands reflecting the collective needs of the colony
[12], many aspects of social behavior are under endocrine
and genetic control, including the reproductive division
of labor [13­15
], investments in reproductive individuals
[16], drone assassinations, queen execution by workers
[17,18] or queen duels [19]. As in many other systems,
agonistic success is fostered by physical superiority [20],
promotes reproductive opportunities [21] and correlates
with amine function [22].
Aggression in other invertebrates: spiders
and dragon flies
Ritualized displays and cues that are predictive of ago-
nistic success enable the assessment of a rival's relative
fighting ability, in particular, in species with dangerous
weapons, such as spiders [23]; the strategies that underlie
aggression and intraspecific, intersexual cannibalism in
this group [24­26] are shaped by the structure of the
population [27]. Dominance enhances feeding opportu-
nities in dragon flies [28] but few physiological studies
that relate specifically to aggression have been carried out
using these models.
Aggression in non-social insects: crickets
Detailed electrophysiological studies have been carried
out in crickets, particularly looking into acoustic signal-
ing. (Singing is used in mating behavior and in aggression
in crickets and other insects [29].) Amine neuron systems
(serotonin, octopamine, dopamine and histamine) have
been fully mapped in cricket nervous systems, including
those systems present in the brain and the ventral nerve
cord (reviewed in [30]). Depletion of nervous system
amines, either globally using reserpine or selectively with
blockers of synthesis specific for serotonin or for octopa-
mine/dopamine, produces alterations in aggression, but
these effects are subtle [31]. Although reserpine depletes
the nervous system of all amines and produces lethargic
behavior, crickets are able to fight at different levels of
intensity. Selective depletion of serotonin had no effects
on fighting behavior, whereas octopamine/dopamine
736
Current Opinion in Neurobiology 2003, 13:736­743 www.current-opinion.com
depletion lowered the intensity of fights. The authors
conclude that octopamine and/or dopamine are not
required for the function of motor circuits involved in
aggression, but that they do seem to play a role in
releasing the behavior in response to appropriate stimuli.
Other studies report that brain levels of serotonin are
lowered in losing males during fights, but only if the wings
that are important in singing behavior are intact [32].
Removal of the wings results in lowering of brain sero-
tonin in both winners and losers of fights. Injections of the
opioid antagonist naloxone enhanced aggression in losing
male crickets and in females [33], while injection of a
vertebrate m-opioid agonist reduced aggression in male
winners. These results suggest that status-specific effects
must be considered when injecting drugs into winning
and losing animals for this kind of study. Interestingly, it
has been demonstrated that forcing losing male crickets
to fly after agonistic encounters rapidly restores their
willingness to fight [34]. This fact, well known to gam-
blers involved in cricket-fighting, requires an intact ven-
tral nerve cord between the brain and the thoracic
segments of the nerve cord.
Crustacean models of aggression
The first anatomical and physiological studies with crus-
tacean species were performed more than 100 years ago.
Indeed, the eminent figures TH Huxley, S Freud and G
Retzius conducted extensive early anatomical studies on
the nervous systems of these organisms. In the mid-
twentieth century, fundamental questions of synaptic
physiology were answered by B Katz, SW Kuffler, P Fatt,
CAG Wiersma and others, due to the anatomical simpli-
city of their peripheral nervous systems of the crustacean
models. These same systems now provide exciting infor-
mation on neuronal function at a `systems' level; thus,
important studies with the crustacean stomatogastric
ganglion describe, at the level of identified neurons,
how modulation affects the output of a neural network
[35]. A more recent frontier in which, once again, crus-
tacean models offer opportunities that are not readily
available with other species, is in the study of social
behavior. Crustaceans such as crayfish and lobsters appear
to be ideal for exploration of the neural basis of aggression
because: (1) their structurally elegant, modular neural
systems feature relatively few, large aminergic neurons,
whose distribution has been mapped and whose physio-
logical properties have been defined [36
,37
]; (2) the
behaviorally relevant neural circuits have also been
mapped [38
­40] and socially modulated changes in these
circuits that relate to amine neuron function can be
observed [39,41
]; (3) amine levels can be both mon-
itored [42] and experimentally altered [43­46]; (4) stereo-
typed behavioral acts can be represented by quantitative
measures in many contexts [47] and finally; (5) crustacean
individuals maintain a fundamentally solitary existence,
with dominance resulting largely from physical super-
iority. By contrast, fighting success in other systems is
often determined by an ability to form coalitions or by
differential treatment of kin.
Agonistic meetings between crayfish or lobsters in con-
trolled laboratory situations feature a series of highly
structured behavioral acts, with escalation being governed
by strict rules. Fights progress through ritualized visual
displays, antennae whipping, claw lock, wrestling and, if
physical asymmetries are only minor, brief periods of
unbridled claw use [48­50]. The expression of particular
fighting strategies varies with hunger states [51], body
size [52] and previous agonistic success [53]. Although
fighting frequently serves to obtain or defend resources,
such as shelter [54] or mates [55], its occurrence, parti-
cularly in the absence of a resource, suggests an inherent
predisposition towards agonism [56,57].
As in other groups, amine neuron systems (serotonin and
octopamine) are implicated as key physiological regulators
of agonistic behavior and social dominance in crustaceans
[37
,42], but controversy surrounds the experimental
results in this area and their interpretation by different
authors. Acute, experimental injections of serotonin and
octopamine in lobsters (Homarus americanus) produced
postures resembling those seen in dominant (serotonin-
like posture) and subordinate (octopamine-like posture)
animals during and after agonistic encounters [58]. It was
these observations that inspired the detailed examination
of the roles of amines in aggression in crustaceans by
many authors, using a variety of species. Postural changes
and enhancement of aggression were recently reported in
a second lobster species (Munida quadrispina) [59]. Acute
and constant infusion of serotonin in crayfish (Astacus
astacus) produces aggression with a unique specificity:
after a delay of 10­30 min, treated individuals engage
larger opponents in prolonged bouts of fighting, even in
instances that carry substantial risk of injury [46,50,60
­
62]. Conversely, in a different species of crayfish (Pro-
cambarus clarkii), serotonin injections produced postural
changes that did not resemble those seen during agonistic
encounters; serotonin injections also reduced levels of
aggression in agonistic encounters, whereas a serotonin
analogue enhanced aggression [43]. In interpreting these
observations, other authors do not focus on a direct role for
serotonin in decapod aggression; instead, they suggest
that serotonin treatment in H. americanus might indirectly
affect social interactions through an inhibition of retreat
in the losing animals [63]. Crayfish with lowered serotonin
levels (Orconectes rusticus) are indistinguishable from con-
trols in terms of fighting behavior [45], but serotonin-
depleted lobsters showed enhanced levels of aggression
[44], similar to those initially reported when serotonin
levels were raised in lobsters and crayfish [60
]. One
interpretation of such apparently contradictory results
focuses on the possibility that levels of serotonin within
a narrow window of concentration might have to be
released at the correct time and place in the nervous
Aggression in invertebrates Kravitz and Huber 737
www.current-opinion.com Current Opinion in Neurobiology 2003, 13:736­743