Current Zoology 58 (1): 146−157, 2012

As clear as mud: Turbidity induces behavioral changes in

the African cichlid Pseudocrenilabrus multicolor
Suzanne M. GRAY1*, Laura H. McDONNELL1, Fabio G. CINQUEMANI1**,
Lauren J. CHAPMAN1,2
1
Department of Biology, McGill University, 1205 Dr. Penfield Ave, Montreal, QC, Canada, H3A 1B1
2
Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY 10460, USA

Abstract Aquatic biodiversity is being lost at an unprecedented rate. One factor driving this loss is increased turbidity, an en-
vironmental stressor that can impose behavioral, morphological, and/or physiological costs on fishes. Here we describe the be-
havioral response of a widespread African cichlid, Pseudocrenilabrus multicolor victoriae, to turbidity. We used a split-brood
rearing design to test if F1 offspring reared in turbid water, originating from river (turbid) and swamp (clear) populations, behave
differently than full-sibs reared in clear water. We examined two facets of behavior: (1) behaviors of fish in full sib groups, in-
cluding activity level and social dynamics collected during the rearing period; and (2) male aggressive behavior directed at poten-
tial male competitors after fish had reached maturity; this was done in an experimental set-up independent of the rearing aquaria.
Regardless of population of origin, fish reared in turbid water were marginally less active and performed fewer social behaviors
than those reared in clear water. On the other hand, when tested against a competitor in turbid water, males performed more ag-
gressive behaviors, regardless of population of origin or rearing environment. Our results suggest a plastic behavioral response to
turbidity that may allow P. multicolor to persist over a range of turbidity levels in nature by decreasing activity and general social
behaviors and intensifying reproductive behaviors to ensure reproductive success [Current Zoology 58 (1): 146–157, 2012].
Keywords Cichlidae, Behavioral plasticity, Environmental stressor, Male-male competition

Increased turbidity is an environmental stressor con-
tributing to the loss of diversity in aquatic ecosystems
globally (Ricciardi and Rassmusen, 1999; Richter et al.,
1997). Turbidity, or suspended particulate matter, comes
in two forms: eutrophication or organic turbidity from
algal growth (Smith, 2003), and sedimentary turbidity
resulting from inorganic inputs such as eroded soil and
re-suspended silt (reviewed in Donohue and Garcia
Molinos (2009)). The effects of increased turbidity on
fish can be both direct and indirect. Direct effects in-
clude mortality resulting from deoxygenation of water
associated with massive algal blooms (Bruton, 1985),
loss of condition from decreased foraging (Herbert and
Merkens, 1961; Sigler et al., 1984; Barrett et al., 1992),
and clogged or damaged gills leading to infection
(Herbert and Merkens, 1961; Bruton, 1985; Goldes et
al., 1988; Sutherland and Meyer, 2007). Indirect effects
of increased turbidity are largely due to changes in the
visual environment, through a reduction in the amount
of light available and shifts in the color of underwater
light (Lythgoe, 1979; Utne-Palm, 2002; Seehausen et al.,
2003). This can disrupt communication signals (re-
viewed in van der Sluijs et al. (2011), which can alter
predator-prey interactions (Abrahams and Kattenfeld,
1997), breakdown species recognition signals
(Seehausen et al., 1997), and alter mate choice and re-
productive behavior (e.g., Luyten and Liley, 1985; See-
hausen et al., 1997; Järvenpää and Lindström, 2004;
Candolin et al., 2007; Wong et al., 2007; Maan et al.,
2010; Sundin et al., 2010). In the endemic hap-
lochromine cichlid fishes of Lake Victoria, East Africa,
decreased light transmission and narrowing of the light
spectrum due to eutrophication is thought to constrain
the functional diversity of color signals, leading to a loss
of fish diversity (Seehausen et al., 1997). Furthermore,
changes to the visual environment have lead to a reduc-
tion in the strength of sexual selection in several fish
species, such as threespine sticklebacks Gasterosteus
aculeatus (Candolin et al., 2007; Wong et al., 2007), sand
gobies Pomatoschistus minutus (Järvenpää and Lind-

Received Aug. 1, 2011; accepted Nov. 1, 2011.

∗ Corresponding author. E-mail: suzanne.gray@mail.mcgill.ca, gray.suzanne@gmail.com
** Current address: Department of Biology, University of Toronto at Mississauga, 3359 Mississauga Road North, Mississauga, ON, Canada,
L5L 1C6
© 2012 Current Zoology
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid
ström, 2004), and Lake Victoria cichlids Pundamilia
nyererei (Maan et al., 2010). This occurs when changes
in the signaling environment induce behavioral changes
associated with reproduction, which can ultimately af-
fect population viability.
Behavior is often a plastic phenotypic trait and can be
the first mechanism used to cope with environmental
change and/or stress (West-Eberhard, 2003; Ghalambor
et al., 2010; Sih et al., 2011), i.e. upon contact with an
environmental stressor or novel environment, animals
can adjust their behaviors immediately to minimize
negative effects of an altered environment. Behavioral
changes in activity level, foraging, and antipredator be-
haviors might mitigate impacts of direct physiological
effects of turbidity. On the other hand, increased energy
expenditure on activities related to reproduction may be
necessary to offset difficulties in communication caused
by the indirect or direct effects of increased turbidity.
For example, threespine stickleback males in the Baltic
Sea experiencing increased eutrophication turbidity
spend more time courting than do males in clear water,
with no net gain in reproductive success (Candolin et al.,
2007). It is thought that the decreased ability to see
prompts increased courtship behaviors. While experi-
mental studies like these have investigated the response
of adult fish to increased turbidity, common garden
rearing studies have yet to be used to explore develop-
mental response to divergent turbidity environments.
Social behaviors associated with maintaining hierar-
chical social systems and aggressive reproductive be-
haviors associated with intrasexual competition (usually
between males) in fish can be energetically expensive
(Grantner and Taborsky, 1998; Ros et al., 2006; Cum-
mings and Gelineau-Kattner, 2009). Agonistic interac-
tions associated with competition for females are ex-
pected to be especially costly because of the risk of in-
jury and use of energy stores (Ros et al., 2006; Briffa
and Sneddon, 2007). In stressful environments, such as
high turbidity, we might expect a decrease in overall
activity level. For example, in salmonids, indirect ef-
fects of turbidity include deterioration of social domi-
nance hierarchies, reduced territoriality, and reduced
gill-flaring (Berg and Northcote, 1985). However, we
might also expect to see a trade-off between social
group behaviors and reproductive behaviors in stressful
environments (e.g., polluted water) (Jones and Reynolds,
1997): a decrease in activity and general social beha-
viors to mediate the physiological effects of the stressor
(leading to a weakened hierarchical structure within the
community; Ang and Manica, 2010) and an intensification
147
of reproductive behaviors to ensure reproductive success.
Pseudocrenilabrus multicolor victoriae (Seegers) is a
wide-ranging cichlid found throughout the Lake Victo-
ria basin and Nile River drainage in Eastern Africa. It is
found across a broad range of habitats from the clear,
dark waters of hypoxic (low dissolved oxygen) swamps
to the highly oxygenated, turbid waters of deforested
rivers (Chapman et al., 1996; Crispo and Chapman,
2008; Crispo and Chapman, 2010; Crispo and Chapman,
2011). Extensive work in this system, focusing on dis-
solved oxygen as an environmental stressor, has shown
that P. multicolor responds both plastically and geneti-
cally to low dissolved oxygen (DO) across its range.
There are strong patterns of morphological and physio-
logical variation across populations associated with
dissolved oxygen gradients, such as larger gills and
smaller brains, and lower standard metabolic rates in
females in fish from low oxygen habitats, compared to
those in high oxygen environments (Chapman et al.,
2000; Chapman et al., 2008; Crispo and Chapman, 2010;
Reardon and Chapman 2010; Crispo and Chapman,
2011). In addition to low dissolved oxygen environ-
ments, P. multicolor also experiences spatiotemporal
gradients of turbidity. In particular, river populations are
subject to variably turbid waters across wet and dry
seasons, especially where surrounding land has been
deforested. For example, in a heavily deforested section
of the Mpanga River of western Uganda, turbidity
ranged from 3.8 to 25.9 NTU over a one year period
(Chapman and Gray, unpublished data). Turbidity, like
dissolved oxygen, is expected to act as a strong selective
agent on P. multicolor phenotype.
Like many cichlids, P. multicolor is a highly social
species: high aggressiveness increases social rank
within the local population (Baerends and Baerends-van
Roon, 1950) and intense male-male aggression is used
to defend territories and attract mates (Fernö, 1986). In
low dissolved oxygen environments, P. multicolor
shows decreased activity level compared to populations
from highly oxygenated rivers, and shows behavioral
flexibility to compensate for the physiological costs of
low oxygen availability (Chapman and McKenzie,
2009). For example, male P. multicolor acclimated to
low and high DO for five months showed a decreased
number of reproductive displays under low DO, which
may reflect the high energy cost of these displays in an
environment where oxygen acquisition is challenging
(Gotanda et al., 2011).
We used a common garden rearing design to detect
effects of population of origin and developmental envi-
 148
ronment on the behavioral responses of P. multicolor to
turbidity. We reared F1 offspring of parental stock
originating from clear (swamp) vs. variably turbid (river)
habitats under clear vs. turbid conditions. We looked at
behavioral response to turbidity in two different con-
texts: (a) in a social group of full sibs in the rearing en-
vironment, and, (b) in a male-male competition experi-
ment under both the rearing environment and the recip-
rocal environment.
1 Materials and Methods
1.1 Split-brood rearing experiment
Wild-caught adult fish from natural field populations
were captured at two sites (swamp and river) within the
Mpanga River drainage, western Uganda in June 2008
and live-transferred to McGill University. The swamp
population, which represents a clear water population,
was sampled at Kanyantale, while the river population,
which represents a turbid water population, was sam-
pled from Kamwenge [see Fig. 1 in Crispo and Chapman
(2010) for map]. These two populations are geographi-
cally isolated (Crispo and Chapman, 2008) and divergent
with respect to gill size, brain mass, and body shape in
association with different oxygen regimes (Crispo and
Chapman, 2010; Crispo and Chapman, 2011).
To produce broods with known parentage, a single
male was placed in a 30-L aquarium with three to five
females. This ratio of males to females buffered male
aggression toward females. As soon as one female was
found to be holding eggs in her mouth, the male was
Fig. 1 Experimental apparatus for male-male aggression trials (see text for details)
Note that fish photo size is not to scale with experimental set-up.
Current Zoology Vol. 58 No. 1
removed from the aquarium and isolated so that we
could ensure he did not fertilize other broods. The
brooding female was also isolated in a separate aqua-
rium. This was repeated to produce five sets of known
parents for each population (10 families in total).
Once the eggs hatched and the fry had absorbed most
of their yolk sacks (i.e. just prior to exogenous feeding)
they were removed from the mother’s mouth. The brood
was counted and split equally between a turbid-water
aquarium and a clear-water aquarium. This gave us a
total of 20 aquaria (2 populations × 5 families ×2 treat-
ments). All 20 aquaria (15-L each) were located on the
same shelving unit in our Animal Facility, with broad
spectrum lighting (CoralLife Aqualight 96 W with 50/50
Power Compact Lamp) over every 5 aquaria. For each
family, the two aquaria (turbid, clear) were randomly
assigned from the aquaria assembly.
One week post-separation from the mother, turbidity
was incrementally increased daily in the turbid aquaria.
Turbidity was created using a concentrated Bentonite
clay-water solution (5 g clay in 250 ml water) that was
suspended using water pumps and air stones. Each day
for 4 days we added 10 ml of the turbidity solution, for
a total turbidity solution concentration of 2.7 ml/L. Tur-
bidity was measured in all aquaria weekly using a
LaMotte 2020e turbidimeter and adjusted as necessary
to maintain clear aquaria at < 1.0 NTU (mean ± SE =
0.81 NTU ± 0.01 over 18 months) and turbid aquaria
between 6.5 and 9.8 NTU (mean ± SE = 8.7 NTU ± 0.10
over 18 months). To prevent the clay from being trapped
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid
in filters, these were removed from all pumps (clear and
turbid treatments), which meant that small water
changes had to be made twice weekly to prevent a
build-up of waste. Basic water quality parameters (dis-
solved oxygen, conductivity, temperature, pH, Ammo-
nia and Nitrite) were measured weekly to ensure good
water quality.
In addition to turbidity, the water was colored to
mimic the tea-stained nature of the water in Uganda
where the parental stocks were caught. We used the ave-
rage maximum wavelength from measurements in the
field with a peak in the medium-long wavelength area
of the visible spectrum (van der Sluijs and Chapman,
unpublished) as a guide in developing our tea-stain so-
lution (12 drops green, 12 drops red, 36 drops yellow
Club House Food Colour in 500 ml water). We added
115 ml of this solution to each aquarium (7.7 ml
tea-stain solution/L), and light measurements were
made periodically throughout the experiment using a
S2000 spectrometer (Ocean Optics) to ensure stability
in water color.
Brood size ranged from 18 to 48 fish (mean ± SD, 28
± 8.5), such that each aquarium initially held between 9
and 24 fish (mean ± SD, 14 ± 4.2). For the first six
weeks fish were fed fry food (Hikari Tropical First bites)
twice daily; at eight weeks they were fed a mixture of
crushed flake food (TetraMinPro Tropical Crisps) and
fry food twice daily; and, at 24 weeks they were fed
only flake food once per day. Due to the small aquarium
size and rapid growth of the fish, broods were reduced
to 8 fish (at six weeks post-release) and then to 5 fish (at
14 weeks post-release) per aquarium. Fish removed
from aquaria were euthanized with clove oil (2.0 ml
1:10 clove oil:ethanol solution in 500 ml water). Some
fish started to show signs of reproductive maturity after
nine weeks post-release (males: red anal fin spot; fe-
males: eggs in mouth).
1.2 Activity and social behaviors in full sib groups
Behavioral observations that focused on general so-
cial activities in full sib groups (activity level,
non-aggressive social interactions, aggressive behaviors)
began at least nine weeks after the experimental condi-
tions started (between October and December 2009)
using scan and focal observations on all aquaria. The
sequence of aquaria to be observed was randomized.
Due to the limited visibility within the turbid aquaria, a
white, perforated plastic board was used to restrict fish
to the front third of the aquarium during the observation
period in both clear and turbid aquaria. Fish were ha-
bituated to placement of the board daily for two weeks
149
preceding observational trials. A curtain was used to
separate experimental aquaria from any disturbances in
the room. After setting up an aquarium for observation
(i.e. placing the white board in the aquarium), the ob-
server would leave the curtained area and return after a
minimum of 15-min. The observer would then take po-
sition directly in front of the aquarium and wait mo-
tionless for another 5 min to allow fish to become ac-
climated to the observer’s presence. Preliminary tests
showed that these acclimation times were sufficient for
fish to resume regular behaviors (e.g., social interactions
resumed).
A scan technique was used to record activity (either
active or not active) of all fish in the aquarium. Activity
was recorded at 10-s intervals for 5 min and averaged.
Due to differences in fish density (across time due to
culling and among aquaria due to the age of the fish),
we calculated the mean proportion of active fish (i.e.
mean number of fish active over 5 min, divided by the
total number of fish, averaged over three observations
per aquarium). Immediately following the 5-min scan,
the observer performed a 3−5 min focal observation
(minimum = 3 min, maximum = 5 min observation time)
during which individual behaviors (see Table 1) of one
haphazardly selected fish (including both sexes and ju-
veniles) were recorded. After the observation period, a
water sample was taken to measure turbidity (see Gray
et al. (2011) for measurement protocol). All observa-
tions were conducted between 08:00 and 16:00 h.
The scan and focal observations were repeated three
times for each aquarium over a three month period. Us-
ing the focal data, we assessed two categories of behav-
ior of full sib groups reared in divergent turbidity envi-
ronments: non-aggressive social (tail wagging, displays
(directed toward any other fish), yawning), and aggres-
sive (charges, bites) behaviors, and then calculated the
proportion of each type of behavior relative to all be-
haviors observed (per min per fish). The individual be-
haviors we observed and recorded were performed by
the focal fish, in the context of the family group, mean-
ing they were not performed in a reproductive context
(compared with male aggression behaviors, described
below). The data for each focal observation were aver-
aged for each aquarium (n = 19). Note that one aquar-
ium could not be observed from a straight angle due to
its placement on the shelves and so was not included in
the analyses.
For scan and focal observations, all data were square
root transformed. For the three variables of interest (ac-
tivity, social, and aggressive behaviors) we used a
 150 Current Zoology
mixed model ANOVA with treatment and population as
fixed factors and family as a random factor nested in
population. In all cases family was non-significant (Pac-
tivity = 0.305, Psocial = 0.519, Paggressive = 0.138) and so
was removed from the models. Because one aquarium
could not be observed, the aquarium holding the sib-
lings from that family in the opposite treatment was also
excluded from the initial analyses that included family,
but was used for subsequent two-way ANOVAs. All
data were analyzed using SPSS 17.0.
1.3 Male-male aggression
Male aggression trials were performed to explore the
response of individual males to competitors under both
clear and turbid conditions, controlling for rearing con-
dition. Each pair was from the same rearing condition
(clear or turbid) but different population (river or swamp).
Males were size matched to eliminate a confounding
effect of body size in aggressive interactions. We tested
each pair under both clear and turbid conditions to assess
if rearing environment influenced the fish’s ability to
perform in clear or turbid waters (e.g., does a fish reared
in turbid water respond to a competitor more when the
water is clear than when it is turbid?). Males were iso-
lated from other males for at least one week prior to the
trials, although they still had contact with females. These
trials were conducted at least one year post-hatching
(September to December 2010) to ensure that all males
were sexually mature and of a similar size.
The experimental aquarium (72 cm × 18 cm × 18 cm)
was divided into four compartments, separated by per-
forated and transparent walls, allowing both visual and
olfactory communication (Fig.1). The two central com-
partments were used to hold experimental males. Each of
these compartments was visually divided into four equal
lateral sections using marks on the front of the tank so
that the observer could record the position (i.e. distance)
of each fish from the central barrier. Each of the end
compartments contained a water pump and air stone that
ensured suspension of the turbidity solution during turbid
trials. A separate stimulus aquarium, holding a mixture of
females and juveniles, was placed adjacent to the ex-
perimental male compartments (see Fig. 1). The re-
maining walls of the stimulus and the experimental
aquaria were covered with black plastic to prevent any
other visual stimuli. Prior to the beginning of all trials,
we randomly determined which treatment would be used
for the first trial for each male pair. If the upcoming trial
was turbid, 110 ml of turbidity solution was added to the
experimental aquarium; if it was clear, it was left as-is;
water was changed between all trials (see below). Solid
Vol. 58 No. 1
white dividers were placed on either side of the central
transparent divider between the experimental compart-
ments the night before a trial, and sized-matched males
were removed from their aquaria and placed in the
compartments (side was randomized). An opaque plastic
cover (i.e. bubble wrap) was placed loosely on top of the
compartments to reduce disturbance during the overnight
acclimation phase. Males remained in the experimental
set-up for 24 hours prior the start of the trial. One hour
before the start of the trial a curtain surrounding the
set-up was closed (preventing disturbance), and the ex-
perimental area was vacated for 30 minutes. After this
period, the observer entered the experimental area quie-
tly and sat in front of the experimental aquarium.
Prior to the start of the trial, the section in which each
male was located within its compartment (i.e. distance
away from the central partition) and the presence of any
territorial pits in the sand were noted. The plastic covers
and dividers were gently removed simultaneously from
each side of the transparent wall, marking the start of the
trial. The behavior of both males was observed for 20
min. Specifically, the occurrence, timing, and location of
four behaviors (lateral display, charging, quivering, and
tail beating; Table 1) were noted when directed at the
male competitor. The observer also took note of the time
elapsed before the males noticed each other as well as the
section each male was in when this occurred. If the males
did not notice one another, or if neither male had moved
by the 10-min mark, the trial was aborted and attempted
at a later date. At the end of the trial, the final behavior
and location of each male were recorded.
The male in the left compartment was immediately
transferred to a separate container filled with clear water.
This process was repeated with the male from the right
compartment. A water sample was obtained from the
central compartment of the experimental aquarium to
provide a measure of turbidity for the trial (for both clear
and turbid experiments). The experimental aquarium was
drained, filled, and rinsed three times in preparation for
the second treatment condition (clear vs. turbid). The
opaque dividers and bubble wrap were replaced, and the
males were put back into the experimental aquarium in
opposite compartments. They were left in the experi-
mental set-up for 24 hours prior to beginning the second
trial for which the same methods but the alternate con-
dition were used.
We used a repeated measures ANOVA to test if males
reared under different conditions and from parents
originating from two different populations (river vs.
swamp) behaved differently in clear and turbid water
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid 151

Table 1 Description of behaviors* observed in tests of response to divergent water turbidity levels in Pseudocrenilabrus multicolor

Behavior Description Experiment1

A quick, forward movement directed towards another fish with mouth opened or closed and without contact
Charge2 1, 2
(specifically toward the competing male in the aggression trials).
A quick forward movement directed towards another fish (the other males in aggression trials) with contact
Bite 1
being made between the attacker’s mouth and the receiver (only observed in social group experiment).
Body horizontal, fins close to body. Tail is quickly beaten back and forth, while fish remains in a steady posi-
Tail Beating 2
tion (directed toward competitors).
Tail Wagging3 A form of tail beating, performed in a non-aggressive interaction. 1
Fish opens mouth wide (as if it were yawning) while the dorsal fin becomes erect (only observed in social
Yawn 1
group experiment).
May be performed at varying degrees of intensity. At highest degree, the dorsal and anal fins are fully erect
Lateral Display while the caudal and pelvic fins are fully spread. Performed while in a perpendicular position to another fish 1, 2
(towards the other male in aggression trials).
Fish approach each other face-to face and open their mouths. Both adversaries try to grip the jaws of the
Frontal Display other. Once a firm grip has been established, they push and pull, beating heavily with their tails and fins (only 1
observed in social group experiment).
Quiver Rapid shaking of the entire body (i.e. not just the tail), generally perpendicular to a competitor or female. 2
* Modified from Baerends and Baerends-van Roon (1950).
1
Similar behaviors were observed in two experiments, but in different contexts noted within the definition. Experiment 1 = social behavior in full sib
groups; Experiment 2 = aggressive behavior in male-male competition.
2
Baerends and Baerends von-Roon (1950) do not use the term charge, but it is implied in their terminology “butting” and “mouth fighting”.
3
Baerends and Baerends von-Roon (1950) use tail wagging for both reproductive and non-reproductive behaviors; here we use it only in the social
context.

when faced with a potential competitor (and in the
presence of females). Each of the four behaviors ob-
served and the total behaviors combined were analyzed
separately, with trial (clear vs. turbid) as the repeated
measure, and rearing environment and population of
origin as between-subjects fixed factors. All behavioral
data were square root transformed.
difference of 0.24 cm (± 0.059 SE) and 0.44 mg (± 0.09
SE) between males within pairs. The difference in stan-
dard length and in mass between paired males did not
differ between rearing treatments (standard length: F1, 10
= 0.953, P = 0.352; mass: F1, 10 = 3.756, P = 0.081). All
males used in the experiment displayed nuptial colora-
tion, including a red spot on the anal fin, blue lips, and

2 Results Table 2 The effect of rearing treatment (clear vs. turbid) and
population of origin (river vs. swamp) on the behavior of P. multi-
2.1 Activity and social behaviors in full sib groups
color in full sib groups
Scan data indicated that fish reared in turbid water
F df P-value
and tested in a social environment of full sibs showed
marginally lower activity levels (P = 0.055) than fish a) Activity

reared in clear water (Table 2, Fig. 2A). Population of treatment 4.349 1, 15 0.055
origin (river vs. swamp) did not have an effect on activ- population 1.500 1, 15 0.239
ity levels, and there was no interaction between rearing treatment*population 0.191 1, 15 0.668
treatment and population of origin. Focal data showed b) Social behaviors
that fish reared in turbid water performed significantly
treatment 8.493 1, 15 0.011
fewer non-aggressive social behaviors (tail-wagging,
population 0.001 1, 15 0.981
yawning, and displays) than fish reared in clear water,
regardless of population of origin (Table 2, Fig. 2B). treatment*population 0.538 1, 15 0.474

There was no effect of treatment, population, or their c) Aggressive behaviors

interaction on aggressive behavior (Table 2, Fig. 2C). treatment 1.099 1, 15 0.311
2.2 Male-male aggression population 0.935 1, 15 0.639
Males ranged in size from 4.1 to 6.1 cm standard treatment*population 0.638 1, 15 0.437
length (mean ± SE, 5.15 cm ± 0.10) and 2.55 to 6.99 mg
Data were derived from scan (activity) and focal (social and aggres-
wet mass (mean ± SE, 4.28 mg ± 0.22). Males were sive behaviors) observations, averaged over three independent obser-
size-matched using both measures, with a mean size vations per aquarium.
 152
Fig. 2 Proportion (mean ± SE) of active fish (A), social behaviors (B), and aggressive behaviors (C) from river and swamp
origin-populations reared for 9 to 24 weeks in clear or turbid treatment in small full sib groups
Means were calculated per tank from scanning observations. Data were square root transformed.
yellow ventral coloration, indicating sexual maturity
(Fernö, 1986). Territorial pits in the sand were common
(>75% of trials). The average (± SE) turbidity recorded
for clear trials was 0.47 NTU (0.14) and 5.65 NTU (0.77)
for turbid trials.
We assessed the number of each individual type of
behavior and the total number of aggressive behaviors
performed by a male when tested against a male in clear
water and in turbid water. Repeated measures ANOVA
indicated a significant effect of experimental condition
(clear vs. turbid) on three of the individual behaviors
(lateral display, tail beats, charge, Fig. 3) and for all
behaviors combined (Table 3). The number of quivers
did not differ between clear and turbid trials and was only
observed 25 times (out of 471 total behaviors observed,
or 0.005%). In no case did rearing condition or popula-
tion of origin have a significant effect on the number of
behaviors performed. Fish tested in turbid water per-
formed more aggressive behaviors than when they were
tested in clear water, regardless of rearing condition or
population of origin (Fig. 4). There was no difference
between treatments in the section where males first no-
ticed their competitor (F1, 20 = 0.711, P = 0.409).
3 Discussion
Results of this study indicate plastic behavioral re-
sponse to turbidity in P. multicolor, and no significant
population effects. Fish reared in turbid water in full sib
groups were marginally less active and performed fewer
social behaviors than those reared in clear water, re-
gardless of population of origin. On the other hand,
when tested against a competitor in turbid water, males
performed more aggressive behaviors than when paired
Current Zoology Vol. 58 No. 1
against the same male in clear water, regardless of
population of origin or rearing environment. Since be-
havior can be a quickly activated reaction to an envi-
ronmental stressor (Timmerman and Chapman, 2004), it
is not surprising that behavioral response to turbidity is
flexible. Nonetheless, these findings have important
implications for the ability of fishes to respond to in-
creasing levels of human-induced turbidity. The behav-
ioral plasticity exhibited by P. multicolor may facilitate
its persistence over a range of turbidity levels in nature
by decreasing both overall activity and general social
behaviors, and by intensifying reproductive behaviors to
ensure reproductive success. It is possible that in a
group, more aggressive behaviors are used in a social
context for establishing dominance hierarchies, while
non-aggressive social behaviors are used for maintain-
ing social hierarchies. Since the full sib groups were
reared together, hierarchies were likely already estab-
lished, and behaviors used only in their maintenance.
Thus, turbidity did not influence aggression in that con-
text, whereas in the male-male competition trials the
motivation to interact aggressively with a novel com-
petitor would have been increased.
Our behavioral plasticity results support two,
non-mutually exclusive ideas: (1) Turbidity alters the
visual environment, and males subsequently increase
aggressive behaviors directed at male competitors. This
is consistent with studies on other fish species that have
shown increased courtship behaviors in turbid condi-
tions at a similar level of turbidity as that used in our
study (Candolin et al., 2007; Engström-Öst and Can-
dolin, 2007; Wong et al., 2007). (2) In addition, there
appears to be a trade-off with respect to energy spent on
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid 153

basic activity and social behaviors and energy spent on
aggressive male-male interactions. This idea is consis-
tent with basic cost-benefit theory for sexually selected
traits (Candolin and Heuschele, 2008). We explore both
of these ideas below.
The social behaviors that we observed to decrease
under turbid conditions in small full sib groups of P.
multicolor were those generally associated with estab-
lishment and maintenance of social hierarchies in
groups, including males, females, and juvenile fish.
These types of behavior can be expensive (Grantner and
Taborsky, 1998). Reducing the amount of social interac-
tions might weaken hierarchies but may be an adaptive
mechanism to cope with environmental stress. Activity
level was only marginally lower in turbid- vs.
clear-reared fish; again, this may reflect energy costs
associated with maintenance activity, but the weak re-
sponse may not be surprising given the low level of

Fig. 3 Number (mean ±SE) of individual aggressive behaviors: tail beats(A), lateral displays (B), charges (C), and quivers
(D) performed by males when paired with a size-matched male under clear (black circles) and turbid (grey circles) condi-
tions
Paired males were from the same rearing condition (clear or turbid) but different population of origin (river or swamp). Population of origin was not
a significant factor in any case and so populations were pooled within rearing treatment in the figures for clarity. Behaviors are defined in Table 1.
Data were square root transformed.
 154 Current Zoology Vol. 58 No. 1

Table 3 The effect of trial condition (clear vs. turbid) on indi-

vidual male-male aggressive behaviors in Pseudocrenilabrus mul-
ticolor originating from two populations (swamp vs. river) and
reared under two treatments (clear vs. turbid)
Within-Subjects F df P-value
a) lateral displays
trial 13.14 1, 20 0.002
trial*population 0.316 1, 20 0.580
trial*rearing treatment 0.552 1, 20 0.466
trial*population*rearing treatment 1.949 1, 20 0.178
Between-Subjects
population 1.923 1, 20 0.181
rearing treatment 0.179 1, 20 0.677
population*rearing treatment 1.178 1, 20 0.291
b) tail beats
Within-Subjects
trial 11.14 1, 20 0.003
trial*population 0.227 1, 20 0.639
trial*rearing treatment 1.841 1, 20 0.190
Trial*population*rearing treatment 0.119 1, 20 0.734
Between-Subjects
population 0.747 1, 20 0.398
Fig. 4 Total number (mean ± SE) of aggressive behaviors
rearing treatment 0.878 1, 20 0.360
performed by large adult males under clear (black circles)
population*rearing treatment 0.081 1, 20 0.779
and turbid (grey circles) conditions when paired against a
c) charges
size-matched male from the same rearing condition (clear
Within Subjects
or turbid) but different population of origin (river or
trial 11.23 1, 20 0.003
swamp)
trial*population 0.001 1, 20 0.977 Paired males were allowed to interact through a clear, perforated bar-
trial*rearing treatment 0.827 1, 20 0.374 rier (see Fig. 1). Data were square root transformed.
trial*population*rearing treatment 0.276 1, 20 0.605
Between-Subjects
population 1.101 1, 20 0.306
turbidity used in this study (~ 9 NTU). This level of
rearing treatment 0.191 1, 20 0.667 turbidity is comparable to levels used in other studies
population*rearing treatment 0.120 1, 20 0.733 (e.g., Wong et al., 2009, ~ 3.5 NTU; Engström-Öst and
d) quivers Candolin, 2007, ~ 10-15 NTU; Heuschele and Candolin,
Within-Subjects 2007, ~ 6 NTU); however, it is much lower than values
trial 3.208 1, 20 0.088 reached at field sites where we have captured P. multi-
trial*population 0.019 1, 20 0.891 color (> 26 NTU, Gray and Chapman, unpublished
trial*rearing treatment 0.124 1, 20 0.729 data). In our experimental set-up, such high levels of
trial*population*rearing treatment 6.976 1, 20 0.016 turbidity would have limited our ability to accurately
Between-Subjects
observe mating behaviors. Therefore, our estimates of
population 0.250 1, 20 0.622
behavioral plasticity with respect to activity and social
rearing treatment 2.133 1, 20 0.160
behaviors in small full sib groups may be conservative,
population*rearing treatment 1.483 1, 20 0.237
but clearly support a reduction in these behaviors under
e) total behaviors
Within-Subjects
increased turbidity.
trial 12.02 1, 20 0.002 In contrast to activity and social behaviors in full sib
trial*population 0.030 1, 20 0.865 groups, we found that the level of aggressiveness in
trial*rearing treatment 1.101 1, 20 0.307 males when encountering another male significantly and
trial*population*rearing treatment 0.741 1, 20 0.400 quite dramatically increased under turbid conditions.
Between-Subjects For example, when faced with the same competitor,
population 1.231 1, 20 0.280 males increased their level of aggressive behaviors by
rearing treatment 0.096 1, 20 0.760 77% in turbid water (Fig. 3, 4), independent of popula-
population*rearing treatment 0.178 1, 20 0.678 tion of origin and rearing treatment. Other studies have
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid
shown similar trends of increased courtship displays
with high turbidity. For example, Candolin and col-
leagues (2007) found that male threespine sticklebacks
increased their courtship activity in turbid (highly eu-
trophied) water. However, the observed increase in
courtship activity did not result in the same level of re-
sponse from females as when male sticklebacks courted
in clear water (Candolin et al., 2007). This suggests that
although male sticklebacks are compensating for re-
duced visibility in turbid conditions, the behavioral
change may be maladaptive or inadequate for the level
of the environmental stressor. In this case we might ex-
pect alternate sensory cues to be used instead of vis-
ual-based cues. In fact, this appears to be one way that
threespine sticklebacks in the Baltic Sea are dealing
with increased turbidity: in turbid water females rely
more on male olfactory than visual cues to choose a
mate (Heuschele et al., 2009).
Male nuptial coloration is likely an important signal
for both male-male competition and female mate choice
in P. multicolor. Male P. multicolor generally have in-
tense yellow coloration ventrally, blue lips, and a
prominent red spot on the anal fin, compared with
brown-grey coloration in females (see fish photos in Fig.
1). Since visual signals are likely impaired by even low
levels of turbidity (Utne-Palm, 2002), males may need
to highlight the extent of their color patterns to other
males (and to females) via increased behavioral signa-
ling. If color signals are masked by turbidity, increased
displays may ensure honest signaling of the males’ abili-
ty to deal with the environmental stressor (Wong et al.,
2007).
The energetic cost of aggressive/agonistic behaviors
in the context of male-male competition can be very
high (Grantner and Taborsky, 1998; Ros et al., 2006;
Briffa and Sneddon, 2007). In the cichlid Oreochromis
mossambicus, aggressive behaviors displayed by adult
males to its mirror image increased energy expenditure
up to five times greater than resting levels (Ros et al.,
2006). Thus, a further increase in such behaviors in the
face of an environmental stressor suggests the impor-
tance of the behaviors and that a trade-off in energy
expenditure with other behaviors is likely.
Population of origin was not a significant driver of
altered behavior at any level, suggesting that the diffe-
rences in behavior we observed are plastic rather than
genetic. In the male-male aggression trials, the results
also indicated that the observed behavioral plasticity
was not a developmental phenotype, but rather a
short-term, quickly activated response to increased tur-
155
bidity (i.e. the developmental environment was not as
important as current environment with respect to in-
trasexual competition). However, we did not test activ-
ity and social interactions in the alternate condition to
the rearing environment and so do not know if those
behaviors are as flexible as aggressive male-male inter-
actions. A similar observation to our aggression results
was reported in male guppies Poecilia reticulata that
alter mating tactics under low light conditions. Male
guppies switched from using sigmoidal courtship dis-
plays to using non-consensual pelvic thrusts in low light,
regardless of whether they were reared in low or high
light environments, suggesting that flexible behavioral
plasticity is more important in responding to changes in
the visual environment (Chapman et al., 2009). Such
flexibility in reproductive behaviors may allow these
species to persist, even under increasingly hu-
man-induced environmental change (Lane et al., 2011;
Sih et al., 2011).
The apparent trade-off we observed with respect to
energy spent on social behaviors in small full sib groups
vs. intrasexual reproductive behaviors points to a
physiological cost of turbidity driven by development
under turbid conditions. Though we have not yet di-
rectly explored physiological response to development
under turbidity in P. multicolor, we do know that low
DO is a significant driver in shaping population varia-
tion in phenotype, and that many morpho-physiological
traits show a high level of developmental plasticity (e.g.,
gill size; Crispo and Chapman (2010)). Across the
range of aquatic systems inhabited by P. multicolor, low
DO and low sedimentary turbidity seem to be coupled:
hypoxic swamps tend to have clear, tannin-stained water,
whereas well-oxygenated rivers tend to be more turbid.
The combined influence of both stressors likely shapes
population-level phenotypic differences. Future research
will examine combined effects of two stressors on the
behavior, morphology, and physiology of this species.
Acknowledgements We thank the Kibale Fish and Monkey
Project field assistants and the logistical assistance of P. Omeja
and D. Twinomugisha for collection of adult fish in Uganda. I.
van der Sluijs, E. Reardon, J. Hunter, L. Grigoryeva, M. Bieber,
and E. Crispo were instrumental in setting-up and maintaining
the rearing experiment. Collection and export permits were
provided by the Uganda National Council on Science and
Uganda Wildlife Authority. All procedures were approved by
the McGill University Animal Care Committee (protocol
#5069). Funding was provided by National Science and Engi-
neering Research Council of Canada (NSERC) Postdoctoral
Fellowship and the American Cichlid Association (SMG),
 156 Current Zoology
NSERC and Canada Research Chair funds (LJC).
References
Abrahams M, Kattenfeld M, 1997. The role of turbidity as a con-
straint on predator-prey interactions in aquatic environments.
Behavioral Ecology and Sociobiology 40: 169–174.
Ang TZ, Manica A, 2010. Aggression, segregation and stability in
a dominance hierarchy. Proceedings of the Royal Society of
London B-Biological Sciences 277: 1337–1343.
Baerends GP, Baerends-van Roon JM, 1950. An Introduction to
the Study of the Ethology of Cichlid Fishes. Leiden: E.J. Brill.
Barrett JC, Grossman GD, Rosenfeld J, 1992. Turbidity-induced
changes in reactive distance of rainbow trout. Transactions of
the American Fisheries Society 121: 437–443.
Berg L, Northcote TG, 1985. Changes in territorial, gill-flaring,
and feeding behaviour in juvenile coho salmon Oncorhynchus
kisutch following short-term pulses of suspended sediment.
Canadian Journal of Fisheries and Aquatic Sciences 42:
1410–1417.
Briffa M, Sneddon LU, 2007. Physiological constraints on contest
behaviour. Functional Ecology 21: 627–637.
Bruton MN, 1985. The effects of suspensoids on fish. Hydrobi-
ologia 125: 221–241.
Candolin U, Heuschele J, 2008. Is sexual selection beneficial
during adaptation to environmental change? Trends in Ecology
and Evolution 23: 446–452.
Candolin U, Salesto T, Evers M, 2007. Changed environmental
conditions weaken sexual selection in sticklebacks. Evolution-
ary Biology 20: 233–239.
Chapman BB, Morrell LJ, Krause J, 2009. Plasticity in male
courtship behaviour as a function of light intensity in guppies.
Behavioral Ecology and Sociobiology 63: 1757–1763.
Chapman LJ, Albert J, Galis F, 2008. Developmental plasticity,
genetic differentiation, and hypoxia-induced trade-offs in an
African Cichlid Fish. The Open Evolution Journal 2: 75–88.
Chapman LJ, Chapman CA, Chandler M, 1996. Wetland ecotones
as refugia for endangered fishes. Biological Conservation 78:
263–270.
Chapman LJ, Galis F, Shinn J, 2000. Phenotypic plasticity and the
possible role of genetic assimilation: Hypoxia-induced
trade-offs in the morphological traits of an African cichlid.
Ecology Letters 3: 387–393.
Chapman LJ, McKenzie DJ, 2009. Behavioural responses and
ecological consequences. In: Richards JG, Farrel AP, Brauner
CJ ed. Hypoxia in Fishes. San Diego: Elsevier, 26–77.
Crispo E, Chapman LJ, 2008. Population genetic structure across
dissolved oxygen regimes in an African cichlid fish. Molecular
Ecology 17: 2134–2148.
Crispo E, Chapman LJ, 2010. Geographic variation in phenotypic
plasticity in response to dissolved oxygen in an African cichlid
fish. Journal of Evolutionary Biology 23: 2091–2103.
Crispo E, Chapman LJ, 2011. Hypoxia drives plastic divergence
Vol. 58 No. 1
in cichlid body shape. Evolutionary Ecology 25: 949–964.
Cummings ME, Gelineau-Kattner R, 2009. The energetic costs of
alternative male reproductive strategies in Xiphophorus ni-
grensis. Journal of Comparative Physiology A - Neuroethology
Sensory Neural and Behavioral Physiology 195: 935–946.
Donohue I, Garcia Molinos J, 2009. Impacts of increased sedi-
ment loads on the ecology of lakes. Biological Reviews 84:
517–531.
Engström-Öst J, Candolin U, 2007. Human-induced water turbidity
alters selection on sexual displays in sticklebacks. Behavioral
Ecology 18: 393–398.
Fernö A, 1986. The aggressive behavior between territorial pseu-
docrenilabrus multicolor (Pisces, Cichlidae) with unlimited or
only visual contact. Behavioural Processes 13: 353–366.
Ghalambor CK, Angeloni LM, Carroll SP, 2010. Behavior as
phenotypic plasticity. In: Westneat DF, Fox CW ed. Evolution-
ary Behavioral Ecology. Oxford: Oxford University Press,
90–107.
Goldes SA, Ferguson HW, Moccia RD, Daoust PY, 1988. Histo-
logical effects of the inert suspended clay kaolin on the gills of
juvenile rainbow trout Salmo gairneri Richardson. Journal of
Fish Diseases 11: 23–33.
Gotanda KM, Reardon EE, Chapman LJ, 2011. Hypoxia and male
behaviour in an African cichlid Pseudocrenilabrus multicolor
victoriae. Journal of Fish Biology 78: 2085–2092.
Grantner A, Taborsky M, 1998. The metabolic rates associated
with resting, and with the performance of agonistic, submis-
sive and digging behaviours in the cichlid fish Neolam-
prologus pulcher (Pisces: Cichlidae). Journal of Comparative
Physiology B-Biochemical Systemic and Environmental
Physiology 168: 427–433.
Gray SM, Sabbah S, Hawryshyn CW, 2011. Experimentally in-
creased turbidity causes behavioural shifts in Lake Malawi
cichlids. Ecology of Freshwater Fish OnLineFirst
Herbert WM, Merkens JC, 1961. The effect of suspended mineral
solids on the survival of trout. International Journal of Air and
Water Pollution 5: 56–74.
Heuschele J, Candolin U, 2007. An increase in pH boosts olfac-
tory communication in sticklebacks. Biology Letters 3:
411–413.
Heuschele J, Mannerla M, Gienapp P, Candolin U, 2009. Envi-
ronment-dependent use of mate choice cues in sticklebacks.
Behavioral Ecology 20: 1223–1227.
Järvenpää M, Lindström K, 2004. Water turbidity by algal blooms
causes mating system breakdown in a shallow-water fish, the
sand goby Pomatoschistus minutus. Proceedings of the Royal
Society of London B 271: 2361–2365.
Jones JC, Reynolds JD, 1997. Effects of pollution on reproductive
behaviour of fishes. Reviews in Fish Biology and Fisheries 7:
463–491.
Lane JE, Forrest MNK, Willis CKR, 2011. Anthropogenic influ-
ences on natural animal mating systems. Animal Behaviour 81:
 GRAY SM et al.: Turbidity induces behavioral changes in the African cichlid
909–917.
Luyten PH, Liley NR, 1985. Geographic variation in the sexual
behavior of the guppy Poecilia reticulata (Peters). Behaviour
95: 164–179.
Lythgoe JN, 1979. The Ecology of Vision. Oxford: Clarendon
Press.
Maan ME, Seehausen O, van Alphen JJM, 2010. Female mating
preferences and male coloration covary with water transpa-
rency in a Lake Victoria cichlid fish. Biological Journal of the
Linnean Society 99: 398–406.
Reardon EE, Chapman LJ, 2010. Hypoxia and energetics of
mouth brooding: Is parental care a costly affair? Comparative
Biochemistry and Physiology A - Molecular and Integrative
Physiology 156: 400–406.
Ricciardi A, Rassmusen JB, 1999. Extinction rates of North Ameri-
can freshwater fauna. Conservation Biology 13: 1220–1222.
Richter BD, Braun DP, Mendelson MA, Master LL, 1997. Threats
to imperiled freshwater fauna. Conservation Biology 11:
1081–1093.
Ros AFH, Becker K, Oliveira RF, 2006. Aggressive behaviour and
energy metabolism in a cichlid fish Oreochromis mossambicus.
Physiology & Behavior 89: 164–170.
Seehausen O, van Alphen JJM, Witte F, 1997. Cichlid fish diver-
sity threatened by eutrophication that curbs sexual selection.
Science 277: 1808–1811.
Seehausen O, Van Alphen JJM, Witte F, 2003. Implications of
eutrophication for fish vision, behavioural ecology, and species
coexistence. In: Crisman TL, Chapman LJ, Chapman CA,
Kaufman L ed. Conservation, Ecology and Management of
African Freshwaters. Gainesville: University Press of Florida,
268–287.
Sigler JW, Bjornn TC, Everest FH, 1984. Effects of chronic tur-
bidity on density and growth of steelheads and coho salmon.
157
Transactions of the American Fisheries Society 113: 142–150.
Sih A, Ferrari MCO, Harris DJ, 2011. Evolution and behavioural
responses to human-induced rapid environmental change.
Evolutionary Applications 4: 367–387.
Smith VH, 2003. Eutrophication of freshwater and coastal marine
ecosystems: A global problem. Environmental Science and
Pollution Research 10: 126–139
Sundin J, Berglund A, Rosenqvist G, 2010. Turbidity hampers
mate choice in a pipefish. Ethology 116: 713–721.
Sutherland AB, Meyer JL, 2007. Effects of increased suspended
sediment on growth rate and gill condition of two southern
Appalachian minnows. Environmental Biology of Fishes 80:
389–403.
Timmerman CM, Chapman LJ, 2004. Behavioral and physiologi-
cal compensation for chronic hypoxia in the sailfin molly
Poecilia latipinna. Physiological and Biochemical Zoology 77:
601–610.
Utne-Palm AC, 2002. Visual feeding of fish in a turbid environ-
ment: Physical and behavioural aspects. Marine and Freshwa-
ter Behaviour and Physiology 35: 111–128.
van der Sluijs I, Gray SM, Amorim MCP, Barber I, Candolin U et
al., 2011. Communication in troubled waters: The evolutionary
implications of changing environments on fish communication
systems. Evolutionary Ecology 25: 623–640.
West-Eberhard MJ, 2003. Developmental Plasticity and Evolution.
Oxford: Oxford University Press.
Wong BBM, Candolin U, Lindström K, 2007. Environmental
deterioration compromises socially enforced signals of male
quality in three-spined sticklebacks. American Naturalist 170:
184–189.
Wong BBM, Järvenpää M, Lindström K, 2009. Risk-sensitive
mating decisions in a visually compromised environment. Bi-
ology Letters 5: 600–602.