ANIMAL BEHAVIOUR, 1999, 57, 1145–1149
Article No. anbe.1998.1075, available online at http://www.idealibrary.com on
Laterality and cooperation: mosquitofish move closer to a
predator when the companion is on their left side
ANGELO BISAZZA*, ANDREA DE SANTI* & GIORGIO VALLORTIGARA†
*Dipartimento di Psicologia Generale, Universita` di Padova
†Dipartimento di Psicologia, Universita` di Trieste
(Received 22 September 1998; initial acceptance 10 November 1998;
final acceptance 27 December 1998; MS. number: 5998)
Mirror images simulating social partners that cooperated or defected have been used as an experimental
method to test the hypothesis that, while inspecting a predator, pairs of fish play a conditional strategy,
Tit for Tat, in an iterated version of the Prisoner’s Dilemma game. Using this method, we found that
predator inspection was more likely to occur when the mirror image was visible on the left rather than on
the right side of mosquitofish, Gambusia holbrooki. The same occurred even when a videorecorded
stimulus presentation was used, in which sequences of the predator were mixed with their mirror-image
equivalents, thus showing that the asymmetry was not due to behavioural or morphological asymmetries
of the predator itself. Moreover, irrespective of whether they were tested with a cooperative (parallel
mirror) or a defecting (angled mirror) partner, mosquitofish drew closer to the predator when the mirror
was on their left side. These findings suggest that the images seen on the right and left sides by a fish may
evoke different types of social behaviour, probably because of differing modes of analysis of perceptual
information carried out by the left and right sides of the brain; accurate control and balancing of the side
of presentation of visual stimuli during behavioural experiments thus appears to be crucial.

when the appearance of an otherwise familiar object is
changed (Vallortigara & Andrew 1991; Vallortigara et al.
1999). In detour tests, male mosquitofish, Gambusia
holbrooki, showed a right eye preference during lateral
fixation of a stimulus of interest (i.e. females, predator);
the preference disappeared or reversed when more
neutral stimuli were used (i.e. an empty cage or an
opaque barrier; see Bisazza et al. 1997a, 1998b). Are these
asymmetries capable of affecting in any significant way
the normal behaviour of animals? Although this has been
hypothesized to be the case for birds (Andrew 1988;
Workman & Andrew 1989; Vallortigara et al. 1997, 1999),
direct evidence is scant. If behavioural asymmetries prove
to be important in the everyday behaviour of animals,
then the evolutionary pressures and genetic mechanisms
maintaining these asymmetries should be of concern to
evolutionary biologists. In particular, it will be crucial to
establish when and why the advantages of having asymmetric brains could overcome the obvious disadvantages
of displaying evident (and predictable) asymmetric
behavioural patterns.
To investigate this issue we took advantage of an
experimental paradigm that has been widely used in
recent years in the field of behavioural ecology. It is
common among fish that pairs of individuals leave
their shoal in order to approach and inspect a potential
Hemispheric specialization of function has long been
considered unique to the human species. In recent years,
however, a large body of evidence has shown that in a
variety of vertebrate species stimuli are processed and
stored differently in the right and left sides of the brain
(review in Bradshaw & Rogers 1993). Although the field
of hemispheric specialization has been a domain of neurologists and neuropsychologists, there are recent signs of
interest by evolutionary biologists as well (Hori 1993;
Raymond et al. 1996). Apart from handedness, lateral
asymmetries in humans typically manifest in pathological conditions or in highly unnatural settings (see
Bradshaw & Rogers 1993). However, in species with laterally placed eyes and small binocular overlap, preferential
use of the right or left eye has been shown to be associated with inspecting different stimuli or carrying out
different cognitive tasks (Workman & Andrew 1986;
Dharmaretnam & Andrew 1994; Cantalupo et al. 1995;
Vallortigara et al. 1996; Bisazza et al. 1997a, b). For
instance, chicks tend to fixate imprinting objects with the
lateral field of the right eye and to shift to the left eye
Correspondence: A. Bisazza, Dipartimento di Psicologia Generale,
Universita` di Padova, Via Venezia 8, 35135 Padova, Italy (email:
[email protected]). G. Vallortigara is at the Dipartimento di
Psicologia, Universita` di Trieste, Via dell’Universita` 7, 34123 Trieste,
Italy.
0003–3472/99/051145+05 $30.00/0
1999 The Association for the Study of Animal Behaviour
1145

1999 The Association for the Study of Animal Behaviour
1146 ANIMAL BEHAVIOUR, 57, 5
Glass
plate
Parallel
mirror
10
9
8
7
6
5
4
3
2
1
1
2
3
4
5
6
7
Angled
mirror
8
9
10
Plant
Figure 1. Experimental apparatus showing the position of the test fish and the mirrors simulating a cooperative (parallel mirror) and defecting
(angled mirror) companion. The predator is shown in the central tank. The experimental tanks were divided into 10 sectors of equal length.
predator (Magurran & Pitcher 1987; Magurran & Seghers
1990). The risk of being preyed upon is shared if both fish
simultaneously inspect the predator, but not if one of the
fish remains at a distance. Fish are thus believed to face a
classic Prisoner’s Dilemma in this situation, and predator
inspection behaviour has been used as a model to analyse
the evolution of mutual cooperation among unrelated
individuals. In an influential paper, Milinski (1987)
found that sticklebacks, Gasterosteus aculeatus, are more
likely to approach a predator when a mirror is placed
parallel to the tank so that the image appears to swim
along with the fish (simulating a cooperative partner)
than when the mirror is angled so that the image appears
to swim away from the fish (simulating a defecting
partner). Similar results were obtained using guppies,
Poecilia reticulata (Dugatkin 1988) and mosquitofish
(Stephens et al. 1997).
In this study we duplicated Milinski’s original procedure. However, we checked for the effects of positioning a mirror on either the left or the right side of the fish.
In the first experiment we measured the tendency of a
female mosquitofish to inspect a pumpkinseed sunfish,
Lepomis gibbosus, when a parallel mirror simulated a
cooperating companion. To check for possible effects
induced by a lateralized behaviour of the predator itself,
in the second experiment we repeated the test using a
manipulated video playback of the predator. The third
experiment was aimed at estimating to what extent laterality could have produced artefacts in previous studies
using cooperative predator inspection.
METHODS
Subjects
In all experiments we used wild-caught female mosquitofish. They were kept in 150-litre tanks and fed dry
fish food daily. Water temperature was maintained at
252C; artificial lighting was provided 16 h a day.
Gambusia are live-bearing fish native to the southeastern
United States and were introduced in Europe at the
beginning of the 20th century for the control of mosquito
larvae. Pumpkinseed sunfish, the species used as a predator stimulus, is also native to North America and was
introduced in Europe in the 19th century. Both species
are widely distributed in Italy and often found in the
same habitats. Although L. gibbosus feed mainly on invertebrates, they occasionally capture small fish and in the
laboratory readily eat adult mosquitofish. Mosquitofish
and sunfish were caught at the same location (Idrovia
Padova-Venezia, near Camin, northeast Italy) and
released at the site of capture after the experiments.
Apparatus and Procedure
The apparatus used for experiments 1 and 3 (Fig. 1)
consisted of three identical glass tanks (604035 cm),
a central tank containing the predator and two lateral
experimental tanks. Each experimental tank was divided
into 10 sectors of equal length; section 1 was the closest
to the predator. The two experimental tanks were divided
longitudinally with a glass plate and a mirror was leant
against one or other side of this plate. Fish could thus be
tested with the mirror placed on their right or on their left
side. A simulated plant composed of a series of elongated
bars, 1 cm in size and spaced 1 cm apart, was located
between sectors 8 and 9 making available to the fish a
‘safe’ starting area. The fish was dip-netted from a nearby
holding tank and introduced into sectors 10 and 9. We
recorded by direct observation the position of the fish
every 10 s beginning when it first exited from the safe
area. Each fish was tested only once. We computed the
average position occupied by the fish during a 10-min
testing period.
In experiment 1, we placed a mirror (5830 cm) parallel to the tank either on the left or the right side. To
control for possible tank effects, use of the left and right
lateral tanks was counterbalanced. Four pumpkinseed
BISAZZA ET AL.: LATERALITY IN PREDATOR INSPECTION 1147
Table 1. Sector position (X±SEM) of mosquitofish in the experimental apparatus in the presence of a live or
recorded predator
Sector position (±SEM)
Experiment
1
2
3
Mirror
Mirror
Mirror
Mirror
parallel (live predator)
parallel (video playback of live predator)
parallel (live predator)
angled away (live predator)
Left
Right
6.35±0.37
4.42±0.49
6.83±0.66
7.92±0.44
7.94±0.30
6.11±0.49
7.91±0.25
8.91±0.22
The experimental tank was divided into 10 sectors, with sector 1 being closest to the predator.
sunfish (average length 104 mm) were alternated in the
role of predator. We tested 40 mosquitofish (mean
lengthSEM=29.20.51 mm), 20 for each treatment (in
each treatment, 10 animals with the right and 10 with the
left lateral tank).
In experiment 2, we presented a videorecorded image
of the predator on a 14-inch monitor in front of a single
experimental tank, with the mirror either on the right
or the left side. The behaviour of a sunfish facing a
mosquitofish was videorecorded. Then, using digital
techniques (Avid Videoshop for Power Macintosh), we
reassembled alternating pieces of original sequences (12–
20 s in length) and their mirror-image versions, thus
generating a 10-min video with no behavioural or morphological asymmetry of the predator. The monitor was
placed 12 cm away from the tank; the size of the video
image of the predator was increased by a factor of 2.5 in
order to compensate partly for the changed retinal size
consequent to the different viewing distance. We tested
24 mosquitofish (mean lengthSEM=28.40.53 mm),
12 for each left or right treatment.
In experiment 3, we used the same apparatus and
procedure as in experiment 1. This time, however, a
mirror parallel to the tank (5830 cm) and a mirror
angled 30 away and much shorter (2830 cm) were
used as independent testing conditions (three sunfish
alternated in the role of predator). We tested 36 mosquitofish (mean lengthSEM=29.20.64 mm), nine for
each left or right treatment.
RESULTS
In experiment 1, predator inspection was more likely to
occur when the cooperative partner was visible on the
left rather than on the right side of the mosquitofish
(Table 1). An ANOVA revealed a significant main effect of
the side of the mirror (F1,36 =10.678, P=0.002). Tank
effect and interaction were not significant (F1,36 =0.0001,
P=0.99; F1,36 =0.214, P=0.65, respectively). There were no
significant effects associated with the four different
pumpkinseed sunfish used as the predator (F3,36 =0.59,
NS).
Since the trials lasted quite a long time (10 min) we also
analysed the data for 2-min intervals, to check for any
effect of habituation to the predator. The ANOVA
revealed significant effects of duration (F4,152 =10.7,
P<0.001) and side of the mirror (F1,38 =11.2, P=0.002) but
no significant interaction (F4,152 =1.20, NS).
In experiment 2, mosquitofish were more likely to
approach the video playback of the predator when their
own mirror image swam on their left rather than on their
right side (F1,22 =5.79, P=0.025; Table 1).
In experiment 3, the ANOVA revealed a significant
main effect of the angle of the mirror (F1,32 =5.82,
P=0.022) and of its right–left position (F1,32 =5.68,
P=0.023). There was no significant interaction between
the angle of the mirror and the right–left position
(F1,32 =0.10, P=0.921). There were no significant effects
associated with the three different pumpkinseed sunfish
used as the predator (F2,33 =1.82, NS). The fish were more
likely to approach the predator with the parallel mirror
than with the angled mirror, thus duplicating Milinski’s
classic findings. However, irrespective of whether they
were tested with a cooperative (parallel mirror) or a
defecting (angled mirror) partner, fish drew closer to the
predator when the mirror was on their left side.
DISCUSSION
Our results show that in mosquitofish the propensity to
inspect a predator depends on the side on which a social
partner happens to be located, and that this effect cannot
be accounted for in terms of morphological or behavioural asymmetries of the predator itself. These findings
support recent claims suggesting homology of cerebral
lateralization in teleosts and tetrapods: in spite of substantial differences between species in the general structure of the nervous system, there is growing evidence that
the functional specializations of the two halves of the
brain may be conserved throughout a wide evolutionary
spectrum (Bisazza et al. 1998a; Miklosi et al. 1998; Robins
et al. 1998; Vallortigara et al. 1998).
The most likely explanation for the reported asymmetry is that mosquitofish tend to use the lateral parts of
the right and left eyes for different purposes. When faced
with a vertical-bar barrier through which a target was
visible, female mosquitofish turned to the left so that
they were using their right eye to look at the target when
the target was a predator and vice versa when the target
was conspecifics of the same sex (Bisazza et al. 1998b).
Results suggest that these lateral biases arise as a consequence of preferential right- or left-eye use during
1148 ANIMAL BEHAVIOUR, 57, 5
sustained viewing of biologically relevant visual stimuli
(Bisazza et al. 1997a). Considering that during the present
tests mosquitofish typically swam very close to the mirror, it seems likely that positioning the mirror on the left
side produced the best arrangement of monocular lateral
stimulation, with the right eye fixating the predator and
the left eye monitoring the companion. At present, we
have no data to disentangle the roles played by right-eye
use in fixation of the predator and left-eye use in fixation
of the companion; in principle, either would suffice for
producing the reported asymmetry, but it is probable that
both contribute.
The use of conditional strategies such as Tit for Tat rests
on a monitoring of the cooperative or defecting behaviour of the other member of the pair and on the recognition of its identity. There is widespread evidence
that among birds and mammals the neural structures in
the right side of the brain have a dominant role
in the recognition of individual conspecifics (review in
Vallortigara & Andrew 1994a). It is unknown whether
similar specializations hold for lower vertebrates such as
fish but, if so, that may explain why it is the left eye
(mainly feeding neural structures on the right side of the
brain) that is used to monitor the behaviour of the image
on the mirror. Alternative explanations for the hypothesis that fish in Milinski’s procedure are playing Tit for
Tat have, however, been proposed (e.g. that they are
simply schooling; Masters & Waite 1990). These different
hypotheses could perhaps be tested with reference to the
differential modes of analysis of perceptual information
between the two halves of the brain.
Whatever the final outcome of the theoretical discussion associated with the use of predator inspection
responses as a way to investigate strategies of cooperation
(Masters & Waite 1990; Reboreda & Kacelnik 1990;
Milinski et al. 1997), it is apparent from our data that
without proper control and counterbalancing of the
side of presentation of visual stimuli, lateralization of
response may affect the results. Our attempt to duplicate
Milinski’s overall procedure showed that when the mirror
was on the left, fish drew closer to the predator with
either a cooperative (parallel mirror) or a defecting
(angled mirror) partner (Table 1). Variation arising from
left–right positioning of the mirror was equivalent to that
arising from the differential cooperativeness of mirror
images. This suggests that in previous studies any partner’s influence was properly assessed only when the
mirror was on the same side of the fish being treated; if
this was not the case, a zeroing or an artefactual increase
of the phenomenon might have occurred (assuming, of
course, that lateralization is there in other species of fish
as well).
Fish use more than their eyes when interacting with
schooling partners (e.g. lateral line, olfaction). In natural
situations fish will thus have a variety of sensory information that could compensate for any laterality arising
solely from vision. None the less, the possibility that
behavioural lateralization produces asymmetries in the
organization of social structures such as fish schools
deserves consideration (parallel lateralization in other
sensory modalities may also be possible, see Vallortigara
& Andrew 1994b for birds). Predators could exploit
marked asymmetry in social behaviour during the detection of danger and schooling of prey and this therefore
does not appear prima facie to be adaptive. However, the
phenomenon might have arisen by some combination of
the peculiar advantages of having individually asymmetric brains and the need to maintain synchronized social
behaviours within the group (Rogers 1989). Obviously,
direct extrapolation of our findings to real life situations
is difficult. When two fish are engaged in cooperative
predator inspection, the fish that has the partner to its
left would be expected to approach the predator more
closely; but it is unclear what the other fish (which has a
partner on its right side) would do. In contrast to a mirror
image, a real partner can modify its speed or its direction
of movement. Fish within a small school may move
continuously in order to occupy the most ‘favourable’
positions in terms of visual processing during sustained
viewing, for example conspecifics on the left side, with
the ‘free’ and more risky right flank monitored by the
right eye. Actually, it could be that any tendency to have
conspecifics on the left side is simply a by-product of
preferential right use in monitoring the most potentially
dangerous parts of the environment. At the level of each
individual within a school, the best strategy would be
having partners on both sides, but given that some
individual must, eventually, occupy the more risky
extreme positions, it would be advantageous to do so on
the ‘right’ side. These are at present speculations, but we
believe that placing the study of brain lateralization into
an ecoethological framework promises to be a fruitful
enterprise.
Acknowledgments
We thank R. J. Andrew, A. Pilastro, L. Regolin and L. J.
Rogers for reading and commenting on the manuscript.
The research was supported by Italian MURST 40% to A.B.
References
Andrew, R. J. 1988. The development of visual lateralization in the
domestic chick. Behavioural Brain Research, 29, 201–209.
Bisazza, A., Pignatti, R. & Vallortigara, G. 1997a. Detour tests
reveal task- and stimulus-specific behavioural lateralization in
mosquitofish (Gambusia holbrooki ). Behavioural Brain Research, 89,
237–242.
Bisazza, A., Pignatti, R. & Vallortigara, G. 1997b. Laterality in
detour behaviour: interspecific variation in poeciliid fishes. Animal
Behaviour, 54, 1273–1281.
Bisazza, A., Rogers, L. J. & Vallortigara, G. 1998a. The origins of
cerebral asymmetry: a review of evidence of behavioural and brain
lateralization in fishes, amphibians, and reptiles. Neuroscience and
Biobehavioral Reviews, 22, 411–426.
Bisazza, A., Facchin, L., Pignatti, R. & Vallortigara, G. 1998b.
Lateralization of detour behaviour in Poeciliid fishes: the effect of
species, gender and sexual motivation. Behavioural Brain Research,
91, 157–164.
Bradshaw, J. L. & Rogers, L. J. 1993. The Evolution of Lateral
Asymmetries, Language, Tool Use and Intellect. San Diego:
Academic Press.
BISAZZA ET AL.: LATERALITY IN PREDATOR INSPECTION 1149
Cantalupo, C., Bisazza, A. & Vallortigara, G. 1995. Lateralization of
predator-evasion response in a teleost fish. Neuropsychologia, 33,
1637–1646.
Dharmaretnam, M. & Andrew, R. J. 1994. Age- and stimulusspecific use of right and left eyes by the domestic chick. Animal
Behaviour, 48, 1395–1406.
Dugatkin, L. A. 1988. Do guppies play tit for tat during predator
inspection visits? Behavioral Ecology and Sociobiology, 23, 395–
399.
Hori, M. 1993. Frequency-dependent natural selection in the handedness of scale-eating cichlid fish. Science, 260, 216–219.
Magurran, A. E. & Pitcher, T. J. 1987. Provenance, shoal size and
the sociobiology of predator-evasion behaviour in minnow shoals.
Proceedings of the Royal Society of London, Series B, 229, 439–465.
Magurran, A. E. & Seghers, B. H. 1990. Population differences in
predator recognition and attack cone avoidance in the guppy,
Poecilia reticulata. Animal Behaviour, 40, 443–452.
Masters, W. & Waite, T. 1990. Tit-for-tat during predator inspection, or shoaling? Animal Behaviour, 39, 603–604.
Miklo`si, A., Andrew, R. J. & Savage, H. 1998. Behavioural lateralization of the tetrapod type in the zebrafish (Brachydanio rerio).
Physiology and Behavior, 63, 127–135.
Milinski, M. 1987. TIT FOR TAT in sticklebacks and the evolution of
cooperation. Nature, 325, 433–435.
Milinski, M., Lu
¨ thi, J. H., Eggler, R. & Parker, G. A. 1997.
Cooperation under predation risk: experiments on costs and
benefits. Proceedings of the Royal Society of London, Series B, 264,
831–837.
Raymond, M., Pontier, D., Dufour, A. & Møller, A. P. 1996.
Frequency-dependent maintenance of left-handedness in
humans. Proceedings of the Royal Society of London, Series B, 263,
1627–1633.
Reboreda, J. & Kacelnik, A. 1990. On cooperation, tit-for-tat and
mirrors. Animal Behaviour, 40, 1188–1189.
Robins, A., Lippolis, G., Bisazza, A., Vallortigara, G. & Rogers, L. J.
1998. Lateralized agonistic responses and hindlimb use in toads.
Animal Behaviour, 56, 875–881.
Rogers, L. J. 1989. Laterality in animals. International Journal of
Comparative Psychology, 3, 5–25.
Stephens, D. W., Anderson, J. P. & Benson, K. E. 1997. On the
spurious occurrence of Tit for tat in pairs of predator-approaching
fish. Animal Behaviour, 53, 113–131.
Vallortigara, G. & Andrew, R. J. 1991. Lateralization of response to
change in a model partner by chicks. Animal Behaviour, 41,
187–194.
Vallortigara, G. & Andrew, R. J. 1994a. Differential involvement of
right and left hemisphere in individual recognition in the domestic
chick. Behavioural Processes, 33, 41–58.
Vallortigara, G. & Andrew, R. J. 1994b. Olfactory lateralization in
the chick. Neuropsychologia, 32, 417–423.
Vallortigara, G., Regolin, L., Bortolomiol, G. & Tommasi, L. 1996.
Lateral asymmetries due to preferences in eye use during visual
discrimination learning in chicks. Behavioural Brain Research, 74,
135–143.
Vallortigara, G., Andrew, R. J., Sertori, L. & Regolin, L. 1997.
Sharply timed behavioural changes during the first 5 weeks of life
in the domestic chick (Gallus gallus). Bird Behavior, 12, 29–40.
Vallortigara, G., Rogers, L. J., Bisazza, A., Lippolis, G. & Robins, A.
1998. Complementary right and left hemifield use for predatory
and agonistic behaviour in toads. NeuroReport, 9, 3341–3344.
Vallortigara, G., Regolin, L. & Pagni, P. 1999. Detour behaviour,
imprinting and visual lateralization in chicks. Cognitive Brain
Research, 7, 307–320.
Workman, L. & Andrew, R. J. 1986. Asymmetries of eye use in
birds. Animal Behaviour, 34, 1582–1584.
Workman, L. & Andrew, R. J. 1989. Simultaneous changes in
behaviour and in lateralization during the development of male
and female domestic chicks. Animal Behaviour, 38, 596–605.