ArticlePDF Available

Male colour variation in a eurytopic African cichlid: The role of diet and hypoxia

Authors:

Abstract and Figures

Species that cross strong environmental gradients are expected to face divergent selective pressures that can act on sexually-selected traits. In the present study, we examine the role of hypoxia and carotenoid availability in driving divergence in two sexually-selected traits, male colour and reproductive behaviour, in the African cichlid Pseudocrenilabrus multicolor victoriae. Low-dissolved oxygen (DO) (hypoxic) environments are expected to be energetically challenging; given that male nuptial colour expression and courtship displays can be costly, we expected fish in low-DO versus high-DO environments to differ in these traits. First, a field survey was used to describe natural variation in male nuptial colour patterns and diet across habitats divergent in DO. Next, using wild-caught fish from a low-DO and high-DO habitat, we tested for differences in reproductive behaviour. Finally, a laboratory rearing experiment was used to quantify the interaction of DO and diet (low- versus high-carotenoid availability) on the expression of male colour during development. In energetically challenging low-DO environments, fish were more red and, in high-DO environments, fish were typically brighter and more yellow. The frequency of reproductive displays in fish of low-DO origin was 75% lower, although this had no consequence for brooding frequency (i.e. both populations produced the same number of broods on average). Our laboratory rearing study showed carotenoid availability to be important in colour production with no direct influence of DO on colour. Additionally, weak patterns of diet variation across wild populations suggest that other factors in combination with diet are contributing to colour divergence.
Content may be subject to copyright.
Male colour variation in a eurytopic African cichlid:
the role of diet and hypoxia
GEORGIA V. MCNEIL
1
, CAITLIN N. FRIESEN
1
, SUZANNE M. GRAY
1,
*,
AMALIA ALDREDGE
1§
and LAUREN J. CHAPMAN
1,2
1
Department of Biology, McGill University, 1205 Docteur Penfield Avenue, Montreal, QC, H3A 1B1,
Canada
2
Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY, 10460, USA
Received 7 October 2015; revised 21 November 2015; accepted for publication 22 November 2015
Species that cross strong environmental gradients are expected to face divergent selective pressures that can act
on sexually-selected traits. In the present study, we examine the role of hypoxia and carotenoid availability in
driving divergence in two sexually-selected traits, male colour and reproductive behaviour, in the African cichlid
Pseudocrenilabrus multicolor victoriae. Low-dissolved oxygen (DO) (hypoxic) environments are expected to be
energetically challenging; given that male nuptial colour expression and courtship displays can be costly, we
expected fish in low-DO versus high-DO environments to differ in these traits. First, a field survey was used to
describe natural variation in male nuptial colour patterns and diet across habitats divergent in DO. Next, using
wild-caught fish from a low-DO and high-DO habitat, we tested for differences in reproductive behaviour. Finally,
a laboratory rearing experiment was used to quantify the interaction of DO and diet (low- versus high-carotenoid
availability) on the expression of male colour during development. In energetically challenging low-DO
environments, fish were more red and, in high-DO environments, fish were typically brighter and more yellow.
The frequency of reproductive displays in fish of low-DO origin was 75% lower, although this had no consequence
for brooding frequency (i.e. both populations produced the same number of broods on average). Our laboratory
rearing study showed carotenoid availability to be important in colour production with no direct influence of DO
on colour. Additionally, weak patterns of diet variation across wild populations suggest that other factors in
combination with diet are contributing to colour divergence. ©2016 The Linnean Society of London, Biological
Journal of the Linnean Society, 2016, 00, 000000.
ADDITIONAL KEYWORDS: Cichlidae – dietary carotenoids – nuptial coloration – sexual selection –
Uganda.
INTRODUCTION
Divergent selection is an important, often ecologi-
cally-dependent mechanism driving phenotypic vari-
ation among populations from different habitats
(Schluter, 2000; Coyne & Orr, 2004; Rundle & Nosil,
2005; Sobel et al., 2010). Fundamental differences in
physico-chemical properties between habitats (e.g.
temperature, water quality, and physical structure,
including shelter availability), as well as variation
among individuals or populations in biotic interac-
tions (e.g. competition, predation, and/or parasitism),
can drive divergent selection (Schluter, 2000; DeWitt
& Langerhans, 2003; McKinnon et al., 2004; Rundle
& Nosil, 2005). Divergence in traits across gradients
of environmental stressors, especially stressors that
are rapidly increasing as a result of human activi-
ties, are of particular interest as we try to under-
stand consequences of anthropogenic perturbations
on biodiversity (Barrett & Hendry, 2012). In aquatic
systems, hypoxia, or low dissolved oxygen (DO), is a
globally pervasive environmental stressor (Diaz,
*Corresponding author. E-mail: gray.1030@osu.edu
Current address: Department of Integrative Biology, Univer-
sity of Texas, 2401 Speedway, Austin, TX 78712, USA
Current address: School of Environment and Natural
Resources, The Ohio State University, 210 Kottman Hall,
2021 Coffey Road, Columbus, OH 43210, USA
§
Current address: School of Medicine, University of Washing-
ton, 1959 NE Pacific Street, Seattle, WA 98195, USA
1©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
Biological Journal of the Linnean Society, 2016, ,. With 4 figures.
2001) that requires water-breathing aquatic organ-
isms to move to zones of higher DO, improve mecha-
nisms to maximize oxygen uptake from the water,
reduce their metabolic rate and/or energetically
expensive activities to reduce oxygen requirements,
and/or use anaerobic metabolism to bridge the differ-
ence between aerobic capacity and metabolic
demands (Chapman, 2015).
Traits under sexual selection, such as visual com-
munication signals, are expected to be shaped by
divergent environmental conditions and often involve
trade-offs with energetically expensive traits (Bough-
man, 2001; Maan & Seehausen, 2011). The efficacy
of communication signals (e.g. male coloration) is
dependent on background environmental conditions,
signal transmission, and the sensitivity of the signal
receiver (Endler, 1990). For example, if background
lighting varies across populations, the signal that
maximizes male conspicuousness to females is
expected to vary accordingly (Gray & McKinnon,
2007; Gray et al., 2008). Signal transmission may
also be enhanced by signaller behaviour, including
more vigorous or frequent courtship displays (Tuo-
mainen & Candolin, 2011). Because courtship and
other reproductive behaviours in sexually dimorphic
fish species can be energetically expensive (Grantner
& Taborsky, 1998; Ros, Becker & Oliveira, 2006;
Cummings & Gelineau-Kattner, 2009), we expect
trade-offs between these activities and other traits of
attraction (e.g. ornamentation) in challenging envi-
ronments.
Variation in male colour expression across diver-
gent aquatic environments has been extensively
studied with respect to the underwater visual envi-
ronment (e.g. habitat complexity, water clarity
including eutrophication and turbidity, photic compo-
sition of underwater light, and visual sensitivity of
signal receivers (Seehausen, van Alphen & Witte,
1997; Fuller, 2002; Seehausen et al., 2008; Heuschele
et al., 2009; Maan, Seehausen & van Alphen, 2010;
van der Sluijs et al., 2011), geographical variation in
predation (Endler, 1980; Endler & Houde, 1995), and
variation in diet (Kodric-Brown, 1989; Grether,
Hudon & Endler, 2001; Grether, Cummings &
Hudon, 2005a). For example, threespine stickleback
(Gasterosteus aculeatus Linnaeus) males found in
tannin-stained (red-shifted) water have black rather
than red throat coloration compared to males found
in clear waters, maintaining a high contrast between
the male and background light (Boughman, 2001).
Red, orange, and yellow colour displays in verte-
brates are largely derived from carotenoids (Price
et al., 2008); however, vertebrates cannot synthesize
carotenoids de novo and so depend on dietary
uptake, with plants and algae comprising two major
natural sources (Olson, 2006; Leclercq, Taylor &
Migaud, 2010). Carotenoid colour displays are thus
considered costly to produce, and these costs are not
only limited to acquisition of carotenoid-containing
food, but also may be imposed from ecological inter-
actions such as increased intraspecific aggression
(Seehausen & Schluter, 2004; Dijkstra, Seehausen &
Groothuis, 2005; Dijkstra et al., 2011) or vulnerabil-
ity to predators (Godin & McDonough, 2003), or from
metabolic conversion processes, absorption efficien-
cies, and reduced immunocompetance (Olson &
Owens, 1998; Svensson & Wong, 2011; Garratt &
Brooks, 2012; Sefc, Brown & Clotfelter, 2014). Caro-
tenoids are also important antioxidants; thus, fish
experiencing oxidative stress likely trade-off use of
carotenoids for colour and antioxidants (Garratt &
Brooks, 2012). Some evidence suggests that the
expression of yellowred colours is therefore an hon-
est indicator of fitness because only healthy individu-
als (i.e. immunocompetent) are able to display
carotenoid-based colours (Grether, 2000; Pike et al.,
2007; Svensson & Wong, 2011). Other work supports
a potential allocation trade-off between pigmentation
and immunocompetance. For example, redder fight-
ing fish (Betta splendens Regan) appear to invest less
in the immune response and more in colour than
blue fighting fish (Clotfelter, Ardia & McGraw,
2007). In addition to the cost of carotenoid-based col-
our production, access to different forms of carote-
noids may vary in nature. Goodwin (1980) reported
that yellow carotenoids are more widely available in
nature than red carotenoids. Consequently, for most
vertebrates, acquiring red carotenoids may be more
taxing (e.g. requires increased foraging and the con-
version of carotenoids), making red nuptial colours
potentially more costly to produce and maintain than
yellow (Hill, 1996).
Carotenoid colour signals are also affected by
shifts in reproductive costs, brought on by changing
environments, which can lead to the use of other,
more efficient mate attraction strategies (Candolin,
2000; Sparkes, Rush & Foster, 2008). For example,
in threespine stickleback, the strength of selection
on male nuptial coloration was observed to decrease
in eutrophic water (high algal content, green-shifted
light) compared to clear water, whereas there was an
increase in the amount of time and energy males
spent on courtship (Candolin, Salesto & Evers,
2007). This increased courting activity in eutrophic
waters did not result in increased mating success.
Divergent environments, therefore, may lead to
adjustments in reproductive strategies by altering
the costs and benefits of different mate attraction
tactics (Maan & Seehausen, 2011; Tuomainen &
Candolin, 2011).
In the present study, we explore variation in male
coloration, diet, and reproductive behaviour across
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
2G. V. MCNEIL ET AL.
divergent environments in a small, mouth-brooding
cichlid, Pseudocrenilabrus multicolor victoriae (See-
gers), found in the Lake Victoria basin of East Africa
(see Supporting information, Fig. S1). In this sexu-
ally dimorphic species, males have blue lips, and
tend to be bright yellowred ventrally, with a red
patch on the anal fin (see Supporting information,
Fig. S2), whereas females are smaller and far less
colourful. Populations are found in a range of habi-
tats that vary in DO content, turbidity, vegetation
type, and temperature; from hypoxic (low DO) clear,
papyrus-dominated swamps to highly oxygenated
(high DO) turbid rivers (Crispo & Chapman, 2008;
Reardon & Chapman, 2009). An initial observation
and motivation for the present study was that male
colour varies across these environmental gradients
(I. van der Sluijs and L. J. Chapman, unpubl. data).
Previous studies on P. multicolor have focused on a
suite of physiological, biochemical, and morphological
traits that have allowed it to persist across divergent
habitats, with an emphasis on extremes of DO
(Chapman, Galis & Shinn, 2000; Chapman et al.,
2002; Mart
ınez, Chapman & Rees, 2009; Reardon &
Chapman, 2009, 2010). Divergence in these ecologi-
cally important traits (e.g. gill and brain size, meta-
bolic rate, egg size) persists despite high gene flow
between habitats divergent in DO (Crispo & Chap-
man, 2008, 2010). Pseudocrenilabrus multicolor from
low-DO habitats also have higher than normal
testosterone : oestradiol ratios, which is considered
to be induced by impaired enzyme function under
low oxygen (Friesen, Aubin-Horth & Chapman,
2012). Such disruptions could lead to reproductive
impairments by altering behaviour and gonad matu-
ration. Similarly, a recent study in our laboratory
showed that males originating from a low-DO site
performed fewer overall displays (aggression and
mating) when acclimated in the laboratory to low
DO compared to males from the same population
held under high DO conditions (Gotanda, Reardon &
Chapman, 2011). Less courtship may reflect the
energetically demanding nature of oxygen-limited
environments that restrict the use of energetically
expensive behaviours.
In the present study, we examine the hypothesis
that an energetically challenging environment (e.g.
low DO) contributes to a trade-off between reproduc-
tive behaviour and colour displays within a single,
widespread species. Although a diverse and complex
array of environmental and social variables (e.g.
water clarity, predation, etc.) is expected to influence
selection on secondary sexual traits in P. multicolor,
we focus here on dissolved oxygen as a first step in
understanding differences in colour and reproductive
behaviour because we know that this stressor is
associated with plastic and genetic divergence in
other energetically expensive traits in P. multicolor
(Chapman, Albert & Galis, 2008; Reardon & Chap-
man, 2009; Crispo & Chapman, 2010, 2011; Crocker,
Chapman & Martinez, 2013). We predicted that fish
from low-DO habitats would be less active to con-
serve energy and would differ in male nuptial col-
oration from fish found in high-DO environments. To
test these predictions, we first quantified previously
observed differences in male colour among natural
populations and assessed diet in the same popula-
tions. Second, we tested for differences in P. multi-
color reproductive behaviour in wild-caught fish from
two populations (representing divergent oxygen
regimes) held under natural DO conditions in the
laboratory. Third, we used a split-brood rearing
experiment with one population of P. multicolor to
test for effects of DO, carotenoid availability, and the
combined effects of DO and diet on the development
of male nuptial coloration in a focal population.
MATERIAL AND METHODS
STUDY SITES
Pseudocrenilabrus multicolor were collected from
paired sites from two drainages in Uganda, East
Africa (see Supporting information, Fig. S1,
Table S1): the Lake Nabugabo region that lies in the
Lake Victoria basin and the Mpanga River drainage
of western Uganda. These well-characterized paired
sites (Crispo & Chapman, 2008; Friesen et al., 2012;
Crocker et al., 2013) represent extremes of DO con-
tent. Swamps have a naturally low DO as a result of
excess organic matter, whereas rivers tend to be more
oxygenated; thus, the swamps represent naturally
hypoxic sites where fish have likely persisted for
many generations. In the Lake Nabugabo region, fish
were collected from small, hypoxic pools in the Lwa-
munda Swamp and from the well-oxygenated waters
of nearby Lake Kayanja (straight distance of approxi-
mately 9.8 km). DO levels remain low year round in
the Lwamunda Swamp, with a monthly mean of
1.5 mg L
1
, whereas the monthly mean DO is
6.07 mg L
1
in the ecotonal waters of Lake Kayanja
(Chapman et al., 2000; Crispo & Chapman, 2008). In
the Mpanga River, fish were collected from a site
within the extremely hypoxic waters of the Kiaragura
Swamp (Bwera) and from an open-water river site
(Bunoga) where DO levels are fluctuating but gener-
ally high. The Bwera swamp is an extensive,
papyrus-dominated wetland that feeds into the
Mpanga River below Bunoga and has DO levels that
remain extremely low year-round, with monthly
mean of only 0.28 mg L
1
(Crispo & Chapman, 2010).
In comparison, the Bunoga site is characterized by a
much higher mean DO (4.7 mg L
1
) than Bwera,
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 3
although monthly mean values fluctuate (range =
1.79.0 mg L
1
; McNeil, 2012), with the lower oxygen
likely a result of overflow of upstream swamps and
run-off from heavily converted agricultural land
during the rainy season (Crispo & Chapman, 2008;
L. J. Chapman, pers. observ.). We sampled during
dry season conditions, when DO tends to be high in
open-water sites in the Mpanga River system (Crispo
& Chapman, 2008). Other environmental variables
are also expected to vary across these sites, and were
measured during sampling for qualitative compar-
isons (see below).
FIELD SURVEY OF MALE COLOUR AND DIET
Each of the four sites (Nabugabo: Lwamunda and
Lake Kayanja; Mpanga: Bwera and Bunoga) was
sampled over a 1 to 4 day-period between mid-Janu-
ary to mid-February 2011 (first dry season of the
year). Additional males were also collected in June
2011 (second dry season) to supplement sample
sizes for diet analyses. Unbaited standard G min-
now traps, set between 08.00 h and 10.00 h for
approximately 2 h, were used for live capture of
fish. Photographs of a subset of individual male P.
multicolor were taken upon capture with a Canon
PowerShot G11 (macro setting with no flash) in a
Plexi-glass cuvette (21 915 910.5 cm) containing
clear water with a white standard attached to the
cuvette. Fish were gently positioned against the
front window of the cuvette using a grey polyvinyl
chloride sheet as a standard background. Care was
taken to take the pictures under similar conditions
at all sites (e.g. between 10.00 h and 14.00 h, in the
shade). Fish were then euthanized with an overdose
of clove oil (2.0 mL of clove oil solution (1 : 10 clove
oil : ethanol) in 250 ml of water) after which they
were measured (standard length; cm) and weighed
(g) (see Supporting information, Table S1). Eutha-
nized fish were individually tagged, and the intesti-
nal cavity was injected with 10% formalin (to
quickly preserve stomach contents) prior to preserv-
ing the fish in the same medium. The additional
live males from each site were live-transferred to
the field camp for spectral reflectance measure-
ments.
At each of the four sites, point-in-time environ-
mental parameters were measured near fish capture
locations (generally two or three per site) between
10.00 h and 14.00 h. Water temperature (°C) and
DO concentration (mg L
1
) were measured using a
Polaris (Oxyguard) probe. Measures were taken in
the upper 20 cm of the water column; additional
measures were taken near the bottom in water dee-
per than 50 cm. Turbidity (nephelometric turbidity
units; NTU) (LaMotte 2020e) and pH (Vital Sine pH
Meter) were measured for three water samples col-
lected from the same locations at each site and time.
Colour analysis
A large area of the male P. multicolor body displays
yellow, orange or red (see Supporting information,
Fig. S2). Although other colours are present (e.g.
ultraviolet, blue), we were interested in potentially
costly carotenoid based pigments; therefore, descrip-
tions of ‘colour’ are referred to as yellowred colours
(i.e. long wavelength chroma). We used two methods
to quantify colour variation. First, we followed the
colour photo analysis procedure of Maan et al. (2004)
to measure the proportional area of the body covered
by red or yellow pixels and thus quantify the overall
pattern of coloration. Using PHOTOSHOP (Adobe
Systems Inc.), white balance was adjusted to the
white standard in each photograph, then pho-
tographs were cropped to include only the fish’s body
(removing fins and eyes), and colour was assessed
using SIGMASCAN PRO, version 4.0 (Systat Soft-
ware Inc.). The area of fish covered by either red or
yellow pixels was quantified by defining colours
based on hue and saturation thresholds for the col-
our of interest (red: hue =026 and 232255, satura-
tion 4097%; yellow: hue =2745 saturation 4097%,
sensu Maan et al., 2004). The total number of red
and yellow pixels was then divided by the total num-
ber of pixels of the specimen to yield the total pro-
portion of red (% Red) and yellow (% Yellow) pixels
covering the body, respectively. We then calculated
the ratio of % Red : % Yellow and used this metric to
test for colour variation across sites.
Second, to quantify the spectral content of male
colour pattern elements, spectral reflectance mea-
surements were taken on additional males per site
sensu Gray et al. (2011). Briefly, fish were eutha-
nized using an overdose of clove oil solution (as
above), and spectral reflectance measurements were
made immediately using a spectrometer (OEM
S2000; Ocean Optics), tungsten-halogen light source
(LS-1; Ocean Optics), and a bifurcated fibre optic
cable (diameter 600 lm) that connected the probe to
the spectrometer and light source. Measurements
were made relative to a diffuse white Teflon stan-
dard and a dark standard. Six predetermined
patches on the left side of the fish (see Supporting
information, Fig. S2) were measured by placing the
ferule of the bifurcated fibre optic cable at a 45°
angle 1.0 cm over each patch (a black sheath cover-
ing the ferule maintained constant angle and dis-
tance from patch and excluded ambient light).
Reflectance was recorded at 0.33-nm intervals across
380700 nm. Integration time was set at 200 ms,
and the mean reflectance spectrum was calculated
from five scans with boxcar smoothing set at 3 and
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
4G. V. MCNEIL ET AL.
further processing used to smooth curves to 1-nm
intervals (see Gray et al., 2011). The resulting spectra
for patches 4, 5, and 6 (see Supporting information,
Fig. S2) were typical of those indicating carotenoid-
based pigments (Bleiweiss, 2004; Hoffman et al.,
2006), although the presence of pterin-based pigments
cannot be ruled out (Grether et al., 2005a; Johnson &
Fuller, 2015). In accordance with Hoffman et al.
(2006), we determined the reflectance midpoint, R
50
(i.e. percentage reflectance found at the midpoint
between maximum and minimum reflectance) indica-
tive of the intensity of the signal, and the spectral
position of the midpoint, kR
50
, indicative of the ‘colour’
of the signal. Although patches 1, 2, and 3 (found on
the darker dorsal portion of the body) were typically of
similar spectral shape, maximum percentage reflec-
tance was on average <10% and so was not evaluated
further (Hoffman et al., 2006).
Analysis of covariance (ANCOVA) was initially
used to test for an effect of site (low DO versus high
DO) nested within drainage and drainage on the %
Red : % Yellow colour ratio, with standard length as
a covariate. We refer to both Lake Kayanja and the
Bunoga site on the Mpanga River as high-DO sites
because the yearly mean values are much higher
than the nearby swamp sites (Lwamunda, and
Bwera, respectively) (see Supporting information,
Table S2), although it should be noted that Bunoga
does experience occasional low-oxygen conditions
during wet seasons. Ratios and standard lengths
were log-transformed to improve normality and sta-
bilize variance where appropriate. Nonsignificant
interactions were removed from the model. Because
of a large site nested in drainage effect (P<0.001),
we also analyzed each drainage separately using sep-
arate ANCOVA. The spectral reflectance variables
R
50
and kR
50
were analyzed in the same way as the
colour ratio data to evaluate differences in intensity
and colour, respectively, across populations. Environ-
mental data were analyzed qualitatively and used
for broad comparisons across sites. All statistical
analyses were performed using SPSS, version 17.0
(SPSS Inc.) and evaluated at a=0.05.
Diet analysis
For each preserved P. multicolor male, the stomach
(oesophagus and intestine not included) was removed
and cleaned of extra tissues. Stomach mass was
recorded, and fullness was visually assessed in accor-
dance with methods first described by Ball (1961).
This method groups stomachs into five fullness cate-
gories: (1) empty, (2) one-quarter full, (3) half full,
(4) three-quarters full, and (5) completely full. Sto-
machs were then opened and, in accordance with the
points method reviewed by Hyslop (1980), each prey
item was identified to order-level resolution. Based
upon a recent assessment of the carotenoid content
of various taxa (Olson, 2006), prey items were identi-
fied and placed into categories that rank from high
carotenoid content to low: algae, plants, fish, zoo-
plankton, and insects. An additional category called
‘varia’ included sand, detritus, and unidentified
material. Prey groups were then visually assigned a
relative percentage based on the volume that they
occupied, which was then converted to a points score
to account for the importance of prey items in stom-
achs of varying fullness. To do this, the assigned
percentage was multiplied by the stomach fullness.
Diet data were examined by calculating both the
mean percentage volume for each prey type per site
and the total points of each prey type expressed as
percentage of the total points per site. The nonpara-
metric MannWhitney U-test was used to detect dif-
ferences in the percentage volume fish, zooplankton,
insects, algae, plant matter, and varia between low-
and high-DO sites within the two drainages. Diet
similarity among different sites was compared using
the percent similarity index (PSI) (Schoener, 1970),
using the equation:
PSI ¼10:5Xjpxi pyij

ð1Þ
Where pis the proportion of points of prey item ‘i’in
fish from site xand site y.
REPRODUCTIVE BEHAVIOUR OBSERVATIONS
Live fish for this experiment were captured as above
in June 2009 and transported to McGill University,
Canada, and housed in 10 separate 30-L aquaria
under conditions similar to their natural environ-
ment (12 : 12 h light/dark cycle at 25 0.2 °C). Fish
from two distinct populations were held in the labo-
ratory at conditions similar to the yearly mean DO
level assessed for each site using long-term environ-
mental datasets (Lwamunda Swamp =1.5 mg L
1
monthly mean; Lake Kayanja =6.07 mg L
1
monthly mean).We selected these two populations
because the DO conditions in the field are very diver-
gent. Therefore, P. multicolor from Lwamunda
Swamp were held under low DO conditions in the
laboratory from June through December (weekly
mean DO SE =1.2 0.02 mg L
1
) to approximate
natural conditions, maintained by a digital oxygen
controller (Point Four Systems, Inc.) that monitored
DO concentrations and periodically pulsed nitrogen
gas into the aquarium through a diffuser to lower
DO levels. Fish from Lake Kayanja were held under
high DO conditions (weekly mean DO SE =7.3
0.08 mg L
1
) in the laboratory by standard aeration.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 5
All fish were housed in aquaria with three females
and one male, with this sex ratio helping to buffer
aggression against females and to control for male
male competition across aquaria. Fish were fed flake
food once daily ad libitum in the morning. The
brooding status of each female was visually inspected
daily (i.e. a female was brooding if embryos could be
seen in her mouth), and brooding females were iso-
lated in clear plastic baskets at the top of the aquar-
ium to prevent injury from aggression.
To quantify differences in reproductive behaviours
between the two DO regimes, behavioural observa-
tions were carried out weekly on males and females
in 10 aquaria (low DO, N=5; high DO, N=5) for
three consecutive months (i.e. nine observation peri-
ods per tank from October through December 2009).
All behavioural observations were carried out in the
afternoon, with aquaria selected in random order by
a single observer to avoid observational bias. The
observation of an individual aquarium was preceded
by a 10-min acclimation period during which the
observer sat behind a blind (moveable, black plastic
screen with a viewing strip) to acclimate fish to their
presence.
During each observation period, two sequential
protocols were used. First, scan observations were
used to estimate the amount of time (expressed as a
percentage) that each individual fish engaged in var-
ious behaviours over a set time period (see Support-
ing information, Table S3). For scans, the behaviour
of each fish in a given aquarium was scored simulta-
neously every 10 s for 5 min. Second, a set of focal
observations was performed immediately after the
scans to detect rare behaviours (see Supporting
information, Table S3). Every fish was observed indi-
vidually for 5 min, and the number of times a beha-
viour occurred was quantified to provide an estimate
of the frequency of each type of behaviour. For both
scan and focal data, observations were classified into
six broad behavioural categories: active, stationary,
aggressive, submissive, reproductive, and feeding;
loosely modelled on a previous behavioural study of
cichlids (Sopinka et al., 2009). Each of these broad
categories (except feeding) incorporated a number of
specific behaviours: (1) active: darts, swims, yawns,
creeps, follows; (2) stationary: still, pauses; (3)
aggressive: bites, chases, head shakes, mouth fights,
puffed throat, rams; (4) submissive: flees, submissive
posture; (5) reproductive: mating quiver (Sopinka
et al., 2009). Because only male P. multicolor per-
form mating quivers, female P. multicolor were docu-
mented as performing a reproductive behaviour
when engaged in a courtship display with a male.
The large majority of behaviours quantified during
scan observations were active or stationary
behaviours, whereas aggressive, submissive, and
reproductive behaviours were extremely rare. As
such, scan data were used to quantify the frequency
of active and stationary behaviours, whereas focal
data were used to quantify the frequency of repro-
ductive, aggressive, and submissive behaviours. The
reproductive status of each female was also recorded
(e.g. brooding or nonbrooding) during observation
periods as a proxy for reproductive success, although
the behaviour of brooding females was excluded from
observation during periods of isolation.
The scan data were arcsin square root trans-
formed, and the focal data were square root trans-
formed to improve normality. A repeated measures
analysis of variance (ANOVA) was used to test for a
time effect on each behaviour in the scan and focal
data sets (i.e. response variables =active, stationary,
aggressive, submissive, reproductive behaviours).
Data from females were used exclusively for this
analysis because of missing values in the male data
set. Because time effects were not significant, data
were pooled within aquaria for each sex, and a two-
way ANOVA was then used to detect differences in
the frequency of behaviours and the overall levels of
activity with population and sex as fixed factors
(N=5 hypoxia aquaria and 5 normoxic aquaria 92
sexes). For the frequency of reproductive behaviours,
one outlier (very high frequency of reproductive
behaviour in one high-DO aquarium) was removed to
stabilize variance; however, inclusion of the value
did not alter the results. The effects of body size
(standard length) were not detected, and so body size
was not included as a covariate in the ANOVA. A
chi-squared contingency test was used to compare
the proportion of brooding females across populations
to assess reproductive success over the course of the
experiment.
DIET AND OXYGEN LABORATORY REARING
EXPERIMENT
Parental stock were collected from the Bunoga site
(Mpanga River drainage) and live-transferred to
McGill University in 2009 to conduct a rearing exper-
iment designed to detect effects of hypoxia, carotenoid
availability, and their interaction on environmen-
tally-induced colour variation in male P. multicolor.
We selected this site because previous studies (Crispo
& Chapman, 2010, 2011) have demonstrated a high
degree of developmental plasticity in morphological
traits in response to variation in the oxygen content
of the rearing environment. This site experiences a
mean DO that is relatively high (4.7 mg L-1) but fluc-
tuates (1.79mgL
1
), therefore representing a good
candidate site for where plasticity in hypoxia-induced
colour variation may be likely. Stock fish were housed
in 30-L aquaria in a temperature-controlled room
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
6G. V. MCNEIL ET AL.
held under a 12 : 12 h light/dark cycle at 25°C (using
40 W light bulbs), and fed once per day. To attain
broods, a single male was housed with several
females (two to four) in 30-L aquaria (held under nor-
moxia) until a brooder was observed. Brooding
females were isolated in separate aquaria until fry
were released from their mouths and the male
returned to a stock tank (80 L) so that families were
not related.
We used a full-sib, split-brood factorial experi-
mental design to test the influence of DO and diet
on male colour expression: offspring from each of
seven families (each family represented by a single
brood) were reared in a factorial design with low-
and high-DO crossed with low- and high-carotenoid
diets. This experiment aimed to explicitly test for
environmentally-induced effects of diet and DO on
male colour; it is possible that variation among field
populations may reflect genetic differentiation or
variation induced by other environmental factors,
although this was not tested. Each brood was
divided into four groups and each group randomly
allocated to one of 28 aquaria (30 L), each of which
was randomly assigned to one of the four oxy-
gen 9diet treatments. Fish were culled to a maxi-
mum of five fish per aquarium to prevent
overcrowding and to maintain similar levels of
social interaction across aquaria. Seven days post
release, DO in low-oxygen aquaria was gradually
lowered to 1.3 mg L
1
over a 1-week period. Low-
DO conditions (1.3 mg L
1
) were created by bub-
bling nitrogen gas into the aquarium and covering
75% of the water surface with bubble wrap. DO
levels were maintained using the Point Four System
(as above). High-DO conditions (approximately 7.5
8.0 mg L
1
) were maintained via constant bubbling of
air through the water column.
Newly released fry were fed Hikari First Bites for
2 weeks and then carefully weaned onto experimen-
tal diets during the third week (fed ad libitum). Fish
were fed once per day and were allowed to feed until
satiated, after which time extra food was removed.
Diets were made in our laboratoty based on the clas-
sic H440 fish diet; a diet composed primarily of
casein, gelatine, and dextrin, which has been used in
previous studies investigating the effects of dietary
carotenoids on fishes (Kodric-Brown, 1989; Frisch-
knecht, 1993; Grether et al., 2005b; Lin et al., 2010)
(see Supplemental information for diet preparation
details).
Treatments were randomly distributed across the
available aquaria (i.e. to account for variation in
temperature and light as a result of position in the
aquaria racks). Each aquarium was equipped with a
Hagen Fluval underwater filter. Water quality (DO,
pH, ammonia, nitrite, conductivity) was assessed
weekly, and water changes were used as necessary
to minimize variation among aquaria in all variables
except dissolved oxygen and food.
The largest male P. multicolor in each aquarium
was photographed at four different stages of develop-
ment to document the ontogenetic expression of nup-
tial colours. Initial photographs were taken just
prior to sexual maturity (approximately 4 months)
and continued at 2-month intervals until the end of
the experiment (10 months). Digital photographs
were analyzed as above. ANCOVA was used to detect
variation in male coloration (% Red, % Yellow, and %
Red : % Yellow) with family as a random factor.
Although we always chose the largest, dominant
male to be photographed at each time period, differ-
ent individuals may have been photographed at the
various time points because of mortality or social
changes; therefore, each time period was tested sepa-
rately. Interaction terms between the fixed factors
and the covariate (standard length) were removed
when not significant.
RESULTS
STUDY SITES:ENVIRONMENTAL DATA
During field sampling periods, the two low-DO
swamp sites were characterized by very hypoxic con-
ditions: DO averaged 0.5 mg L
1
in the papyrus-
dominated Bwera swamp site and 1.2 mg L
1
in the
pools of the Lwamunda Swamp (see Supporting
information, Table S2). In comparison, DO averaged
8.5 mg L
1
at the Bunoga site of the Mpanga River
and 7.8 mg L
1
at the Lake Kayanja site (see Sup-
porting information, Table S2). Mean water tempera-
ture ranged from 18.7 to 22.3 °C in January and
February and from 20.5 to 27 °C in June (see Sup-
porting information, Table S2). Turbidity ranged
from 1.48 to 29.8 NTU at low-DO swamp sites and
from 12 to 13 NTU at high-DO sites (see Supporting
information, Table S2). These values represent point-
in-time measurements.
FIELD SURVEY OF MALE COLOUR AND DIET
Male colour
The nested ANCOVA results yielded large drainage
effects (Fig. 1, Table 1) for all colour variables tested
and so drainages were also analyzed separately.
Analysis of photographs revealed that fish from the
low-DO swamp sites had consistently higher ratios of
% Red : % Yellow compared to lake/river sites across
drainages (Fig. 1A). The shape of the spectral reflec-
tance curves suggest that P. multicolor males display
carotenoid-based pigments on the lateral and ventral
portions of their bodies (see Supporting information,
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 7
Fig. S2) (Hoffman et al., 2006). The properties of
these spectra differed between high-DO and low-DO
sites: fish from low-DO sites were significantly dar-
ker (lower R
50
) and red-shifted (on average 21 nm
longer wavelength kR50) relative to lake/river fish,
matching the red-shift observed in the photographic
analysis (Fig. 1B, C, Table 1) and previous unpub-
lished observations.
Diet analysis
Of 150 male P. multicolor captured at the four
sites, 83 had non-empty stomachs (see Supporting
information, Table S4). The dominant items in the
diet of P. multicolor were insects, plant matter,
and fish remains; this general pattern was similar
when expressed as the mean percentage volume or
the percentage of total points (Fig. 2A, B). Within
the Mpanga drainage, fish from the high-DO river
site were characterized by a higher proportion of
fish (MannWhitney U-test, P=0.001) and a lower
Figure 1. Colour variation in male Pseudocrenilabrus
multicolor from two drainages (Mpanga and Nabugabo)
and two sites per drainage [solid circles represent high-
dissolved oxygen (DO) sites, solid triangles represent low-
DO sites]. A, estimated marginal mean SE ratios of %
Red : % Yellow of the body area measured using photo
analysis. High proportions signify more red, whereas
lower scores are more yellow. Ratios were log trans-
formed. Estimated marginal mean SE of (B) reflectance
midpoint (R
50
) and (C) spectral position of midpoint
(kR
50
) measured using spectral reflectance. All pairwise
comparisons within drainage were significant at a=0.05,
except for kR
50
in the Mpanga drainage (Table 1).
Table 1. Results of nested analysis of covariance to detect
colour differences between high- and low-DO sites within
each of two drainages (Nabugabo, Mpanga) sampled for
Pseudocrenilabrus multicolor
Variable d.f. FP
% Red : % Yellow*
Nested Drainage 1,2 0.111 0.771
Site (Drainage) 2,126 36.31 < 0.001
SL 1,126 2.817 0.096
Drainages separate
Nabugabo Site 1,72 32.06 < 0.001
SL 1,72 6.238 0.015
Mpanga Site 1,53 25.15 < 0.001
SL 1,53 0.002 0.969
R
50
Nested Drainage 1,2.4 0.001 0.980
Site (Drainage) 2,21 4.145 0.030
SL 1,21 22.35 < 0.001
Site 9SL
(Drainage)
3,21 5.904 0.004
Drainages separate
Nabugabo Site 1,7 6.050 0.043
SL 1,7 17.69 0.004
Site 9SL 1,7 14.90 0.006
Mpanga Site 1,15 8.299 0.011
SL 1,15 7.076 0.018
kR
50
Nested
Drainage 1,2.02 0.390 0.596
Site (Drainage) 1,26 10.83 < 0.001
Drainages separate
Nabugabo Site 1,10 6.05 0.002
Mpanga Site 1,16 2.354 0.145
*Colour was measured as the % Red : % Yellow ratio
from photos, and the reflectance midpoint (R
50
) and spec-
tral position of the midpoint (kR
50
) from reflectance mea-
surements; standard length (SL) was used as a covariate;
ratios and SL were log-transformed to improve normality
and stabilize variance.
Values shown in bold are significant at a=0.05.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
8G. V. MCNEIL ET AL.
proportion of zooplankton (P<0.001) than fish from
the low-DO Bwera site (Fig. 2B). No other prey cat-
egories differed between the two sites (insect:
P=0.101; plants: P=0.060; algae: P=0.271).
Within the Nabugabo drainage, P. multicolor from
the low-DO Lwamunda Swamp were characterized
by a higher proportion of zooplankton in their diet
(P=0.004) than the high-DO lake site, although
other prey types did not differ significantly between
the two sites (P0.150). Dietary similarity (esti-
mated by the PSI; see Supporting information,
Table S4) averaged 68% and was lowest between
the low-oxygen Bwera site and the two high-DO
sites (Bwera versus Bunoga =57%; Bwera versus
Kayanja 59%). Comparatively, dietary overlap
among the other three sites was relatively high
(6977%).
REPRODUCTIVE BEHAVIOUR OBSERVATIONAL STUDY
Results of repeated measures ANOVA indicated no
significant effect of time on any of the behaviours in
the scan or focal data sets (results not shown). As a
result, data for all behaviours were averaged across
time periods prior to testing for effects of population,
sex, and their interaction. ANOVA on behaviours
quantified using scan data indicated that males
exhibited active behaviours 47% more frequently
(Table 2) and, consequently, stationary behaviours
less frequently than females (Fig. 3A, Table 2).
ANOVA on behaviours quantified using focal data
indicated a 53% higher frequency of aggressive beha-
viour in males than in females (Fig. 3B, Table 2),
and an 88% higher frequency of submissive beha-
viours in females than in males (Table 2). There was
no observed population effect for active, stationary,
aggressive or submissive behaviours; however, the
frequency of reproductive behaviours was 75% lower
in fish from the low-DO population compared to fish
from the high-DO population (Fig. 3C, Table 2).
There was no observed difference in the proportion
of brooders between the populations (v
2
=0.624,
P>0.25).
DIET AND OXYGEN LABORATORY REARING
EXPERIMENT
Within each time period of the rearing experiment,
fish from high-carotenoid treatments were character-
ized by a greater % Yellow than fish from low-carote-
noid treatments (Fig. 4A, Table 3; see also
Supporting information, Fig. S3). Yellow coloration
also increased with time in fish from all treatments.
In comparison, red coloration tended to decrease
with time, although there were no significant treat-
ment effects, except at month 6, when % Red was
greater in the high-carotenoid treatment (Fig. 4B,
Table 3). The % Red : % Yellow ratio support these
trends (Table 3). We found no evidence for a signifi-
cant effect of low DO on coloration, despite success-
fully maintaining consistent low-DO conditions for
10 months (see Supporting information, Table S5).
DISCUSSION
Our field survey of P. multicolor in two drainages of
Uganda indicated repeated differences in male col-
oration between habitats, with P. multicolor males
from low-DO swamp sites having more red and males
from high-DO river and lake sites characterized by a
higher percentage of yellow. Although the number of
sites that we have looked at to date is limited to two
paired comparisons, these repeated results in two
Figure 2. Percentage contribution of prey in the diet of
Pseudocrenilabrus multicolor from two drainages
(Mpanga and Nabugabo) and two sites per drainage cal-
culated as (A) proportion of total points for each prey type
and (B) mean percentage volume of each prey type. Sam-
ple size of non-empty stomachs was 22, 16, 16, and 29 for
Bunoga, Bwera, Kayanja, and Lwamunda, respectively
(for additional information, see Supporting information,
Table S1).
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 9
drainages are suggestive of a larger pattern that will
be tested in the future. Additionally, our behavioural
observations of wild-caught fish held under native
DO conditions show no difference in brooding fre-
quency, despite fish from a low-DO site performing
reproductive behaviours at a significantly lower fre-
quency than fish from a high-DO population. This
suggests that low DO could be inducing a physiologi-
cal trade-off in male P. multicolor, where energeti-
cally expensive courtship behaviour is reduced in
hypoxic habitats but mate attraction is maintained
by increasing red coloration. Interestingly, we did not
find a relationship between DO and redyellow colour
expression in our laboratory rearing experiment,
indicating that development under hypoxia does not
by itself induce greater red coloration, at least in the
one population of fish tested. As expected, there was
a strong link between dietary carotenoid availability
and the production of yellow coloration. We discuss
these findings in the context of ecological divergence
in sexually-selected traits, and we introduce potential
scenarios that may reconcile differences between the
field and laboratory studies.
POTENTIAL LOW OXYGEN-INDUCED TRADE-OFF
BETWEEN COURTSHIP AND COLOUR
Short-term energetic costs associated with male
courtship displays have been shown to be high rela-
tive to resting activity in many taxa (e.g. courting
salamanders: Bennett & Houck, 1983; calling frogs:
Bucher, Ryan & Bartholomew, 1982; Ryan, 1985;
Taigen & Wells, 1985; displaying birds: Oberweger &
Goller, 2001; fireflies: Woods et al., 2007; spiders:
Kotiaho et al., 1998; Hoefler, 2008). In cichlid fishes,
reproductive behaviours are suggested to be particu-
larly costly because breeding rituals are character-
ized by a high level of parental care (females) and
elaborate courtship displays (males) (Baerends &
Baerends van Roon, 1950; Desjardins et al., 2008;
Nakazawa & Yamamura, 2009). Given the poten-
tially high energetic cost of courtship in male P. mul-
ticolor, we expected a decrease in the intensity of
displays under taxing conditions such as low DO,
potentially diverting energetic expenditures to other
traits or behaviours that promote reproductive suc-
cess. The results of our behavioural study support
this hypothesis: the frequency of reproductive dis-
plays was 75% lower in wild-caught fish from a low-
DO site than fish from a high-DO site, although
there was no difference between source populations
in brooding frequency. This suggests that, despite
less active courtship, low-DO males were still attrac-
tive to females. Such a reduction in energy expendi-
ture could also help alleviate oxidative stress,
reducing the use of carotenoids as antioxidants and
freeing them for deposition in the skin as pigments.
SOURCE OF VARIATION IN MALE COLOUR
Although a higher proportion of red colour was corre-
lated with low-DO and yellow with high-DO habitats
in the field, DO availability appeared to have no
direct effect on the amount of yellow and/or red col-
our in fishes of high-DO origin reared under low-
versus high-DO conditions in the laboratory. If DO
was the main driver in colour variation, we expected
fish to be more red in low-oxygen rearing treatments
based on our field survey findings, especially when
carotenoid availability was high. Our findings do not
necessarily exclude this possibility. For example, if
colour expression is largely under genetic control
(low level of plasticity), we would expect the fish
used in the experiment (which originate from the
high-DO Bunoga site) to be more yellow, as we found
to be the case in the wild Bunoga population. Fish
from high-carotenoid treatments were characterized
by a greater % Yellow than fish from low-carotenoid
treatments across all time periods, and fish from all
treatments exhibited more yellow than red. If fish
Table 2. Results of two-way analysis of variance to detect effects of population of origin, sex, and population 9sex
interactions on the behaviours of Pseudocrenilabrus multicolor
Population Sex Population 9Sex
FP*FP FP
Active 0.531 0.477 12.981 < 0.002 0.020 0.890
Stationary 0.987 0.335 12.087 < 0.002 0.053 0.821
Aggressive 0.363 0.555 10.262 0.006 0.500 0.490
Submissive 1.329 0.266 48.995 < 0.001 0.436 0.519
Reproductive 11.898 0.004 0.032 0.478 1.182 0.294
*Values shown in bold are significant at a=0.05; N=20 (two sexes, five replicates per DO treatment) for all beha-
viours, except reproduction where N=19.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
10 G. V. MCNEIL ET AL.
from the low-DO Bwera site were reared in a similar
experiment, we would expect them to exhibit a
higher proportion of red compared to yellow, regard-
less of oxygen treatment, if colour expression has a
large genetic component. In a closely-related species,
P. philander, several sympatric ecotypes have
evolved that can be distinguished by differences in
male nuptial coloration that are similar to popula-
tion-level colour differences observed in the present
study; however, the mechanism driving colour diver-
gence in that system is also unknown (Stelkens &
Seehausen, 2009). Population-level genetic variation
in the differential use of carotenoids has been shown
in a number of taxa, including red and blue fighting
fish morphs (Clotfelter et al., 2007) and red and yel-
low African weaver birds Euplectes spp.(Prager,
Johansson & Andersson, 2009). Thus, it is conceiv-
able that P. multicolor populations also differ in how
they incorporate dietary carotenoids. Additionally,
another class of pigments, pterins, reflect in the long
wavelength end of the spectrum and can be synthe-
sized by vertebrates. In some cases, such as the Tri-
nidad Guppy (Poecilia reticulata), pterins are used in
combination with carotenoids to produce orange
spots (Grether et al., 2005a), whereas, in the bluefin
killifish [Lucania goodei (Jordan)], carotenoid and
pterin pigments are used separately to produce
orange in the caudal fin and yellowred in the anal
fin (Johnson & Fuller, 2015). In the guppy, pterin
synthesis and deposition is heritable (Grether et al.,
2005a). We cannot rule out the presence of pterin-
based pigments in P. multicolor, although the signifi-
cant difference in colour of males reared on low
versus high carotenoid diets means that long wave-
length colour signals are largely carotenoid-depen-
dent in this species. Our rearing experiment was
designed to detect plastic effects of hypoxia and diet
on colour variation. We anticipated plastic effects of
hypoxia on colour variation, given the high levels of
plasticity exhibited by this species in response to
divergent DO environments. We did detect plastic
effects, although these were driven by diet rather
than oxygen availability. Insights into population
effects on colour will be an informative next step,
which could be achieved through common-garden
rearing experiments on multiple populations from
low-DO and high-DO sites and biochemical analysis
to assess pigment composition.
DIET,CAROTENOID AVAILABILITY,AND COLOUR
The presence of dietary carotenoids was clearly impor-
tant for the production of colour in laboratory-reared
P. multicolor. Throughout development, fish fed on
high carotenoid diets produced more red and yellow
colour overall relative to those on low carotenoid diets.
This finding supports a number of studies that have
demonstrated intraspecific differences in colour asso-
ciated with variation in dietary carotenoids (Grether,
Figure 3. Frequency of observed behaviours in male and
female Pseudocrenilabrus multicolor from high-dissolved
oxygen (DO) and low-DO populations (held in the labora-
tory under high-DO and low-DO conditions, respectively).
A, active behaviours calculated from scan data. B, aggres-
sive behaviours calculated from focal data. C, reproductive
behaviours calculated from focal data. Values shown repre-
sent the mean SE (N=5 aquaria per group). Data were
arcsin square root transformed for activity and square root
transformed for aggressive and reproductive behaviours.
Values represent the mean frequency per fish per 5-min
interval averaged across sampling periods. Values were
significantly different at a=0.05 between sexes in (A) and
(B) and between populations in (C) (Table 2).
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 11
Hudon & Millie, 1999; McGraw et al., 2001; Costan-
tini et al., 2005). Interestingly, we demonstrated clear
colour differences between high- and low-DO field
sites but no conclusive trends between our field diet
analysis and habitat. The major difference between
the diet of P. multicolor between low-DO swamp sites
and the high-DO sites was a higher proportion of zoo-
plankton and, in the case of Bwera, a lower proportion
of fish than in high-DO Bunoga site. Although we did
not measure carotenoid concentration in the foods con-
sumed by P. multicolor, there was no evidence that
foods known to contain high carotenoid concentration
(e.g. algae, plants; Olson, 2006) were more abundant
in the diet of swamp-dwelling P. multicolor. This sug-
gests that one or more factors in addition to diet may
be driving the observed colour divergence. Limitations
of our diet study include (1) the point-in-time collec-
tion: it is possible that a seasonal sampling scheme
may have yielded more inter-site variation in diet; (2)
capture methods: minnow traps may have selected for
specific behavioural types, thus not representing the
full dietary breadth of each population; and (3) the
resolution of our diet data: it is possible that the taxa
of plants and algae consumed varied in carotenoid
content across sites.
WATER PROPERTIES AND LIGHT TRANSMISSION
The swamp sites in both drainages are characterized
by clear but tannin-stained water that is reddish in
appearance. Clear waters are characterized by
broad-spectrum, high intensity light that gives back-
ground water a low-intensity bluegreen background.
By contrast, ‘tea-stained’, tannin-rich waters create a
red-shifted, lower intensity background (Lythgoe,
1979). This means that males with dark, red-shifted
colour patterns are found in red-shifted habitats,
whereas males with a higher proportion of yellow
and brighter signals are found in sites with poten-
tially shorter wavelength, broader-spectrum ambient
light. Differences in water colour between sites may
contribute to differences in the perception of fish col-
our by conspecifics because colour perception is
affected by attenuation, absorption and scattering of
light, and the visual sensitivity of the signal receiver
(Endler, 1990). Fish may optimize the conspicuous-
ness of their colour signal by shifting their colour
(i.e. increasing chromatic contrast) or by becoming
brighter (i.e. increasing signal intensity) [e.g. in
sticklebacks: Boughman, 2001; bluefin killifish:
Fuller, 2002; pygmy perch Nannoperca australis
(G
unther): Morrongiello et al., 2010]. In P. multi-
color, it is not clear whether redder fish living in
red-shifted swamp sites are more conspicuous than
yellower fish living in less red-shifted river/lake
sites. It could be that neither is more conspicuous in
its respective environment but that a threshold con-
trast is being maintained across all habitats. As
noted above, male stickleback in red-shifted environ-
ments have black rather than red throat patches,
which would be perceived as a contrasting patch in
that environment (Boughman, 2001). This could be
driven by female choice that has evolved to favour a
certain contrast level, regardless of changes in habi-
tat and light conditions.
Alternatively, differences in turbidity between
sites could contribute to differences in the colour and
availability of light, thus influencing the strength of
selection on male colours (e.g. Maan et al., 2010).
For example, Maan et al. (2010) show that, in the
cichlid Pundamilia nyererei, female preference for
male nuptial coloration is diminished in turbid com-
pared to clear waters, and this is associated with
more drab coloured males in turbid habitats. In the
P. multicolor system, low-DO sites tend to be clear
(i.e. low suspended sediments/turbidity) but tea-stained
Figure 4. Development of (A) % Yellow and (B) % Red coloration over 10 months in Pseudocrenilabrus multicolor males
reared under low- or high-DO crossed with low-carotenoid or high-carotenoid food. Beginning at sexual maturity (ap-
proximately 4 months), colour was measured every 2 months and time periods were analyzed separately for treatment
effects and their interaction. % Red scores were square root transformed for analysis; and untransformed values are pre-
sented.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
12 G. V. MCNEIL ET AL.
Table 3. Results of analysis of covariance used to test for effect of diet, dissolved oxygen, and their interaction on %
Red coloration, % Yellow coloration, and the ratio of % Red : % Yellow coloration in Pseudocrenilabrus multicolor with
standard length (SL) as a covariate
Color Stage (months)* Effect d.f. FP
% Red
4 Oxygen 1,12 2.269 0.158
Carotenoids 1,12 0.145 0.710
Oxygen 9Carotenoids 1,12 3.768 0.076
SL 1,12 0.314 0.586
6 Oxygen 1,15 0.340 0.568
Carotenoids 1,15 6.711 0.020
Oxygen 9Carotenoids 1,15 1.186 2.930
SL 1,15 0.463 0.507
SL 9Carotenoids 1,15 8.186 0.016
8 Oxygen 1,15 1.234 0.284
Carotenoids 1,15 0.011 0.918
Oxygen 9Carotenoids 1,15 0.546 0.471
SL 1,15 0.996 0.334
10 Oxygen 1,16 0.419 0.527
Carotenoids 1,16 0.299 0.592
Oxygen 9Carotenoids 1,16 1.250 0.280
SL 1,16 0.256 0.620
% Yellow 4 Oxygen 1,12 2.520 0.138
Carotenoids 1,12 17.48 0.001
Oxygen 9Carotenoids 1,12 1.064 0.323
SL 1,12 1.414 0.257
6 Oxygen 1,16 1.27 0.277
Carotenoids 1,16 26.660 < 0.001
Oxygen 9Carotenoids 1,16 0.552 0.468
SL 1,16 0.268 0.612
8 Oxygen 1,15 0.142 0.712
Carotenoids 1,15 24.000 < 0.001
Oxygen 9Carotenoids 1,15 0.214 0.650
SL 1,15 0.021 0.886
10 Oxygen 1,16 2.672 0.122
Carotenoids 1,16 17.115 0.001
Oxygen 9Carotenoids 1,16 0.013 0.909
SL 1,16 1.647 0.218
% Red : % Yellow
§
4 Oxygen 1,12 1.373 0.264
Carotenoids 1,12 0.696 0.420
Oxygen 9Carotenoids 1,12 0.696 0.666
SL 1,12 0.020 0.889
6 Oxygen 1,16 0.265 0.614
Carotenoids 1,16 0.206 0.656
Oxygen 9Carotenoids 1,16 0.311 0.585
SL 1,16 6.460 0.022
8 Oxygen 1,15 0.993 0.335
Carotenoids 1,15 2.740 0.119
Oxygen 9Carotenoids 1,15 0.088 0.771
SL 1,15 0.032 0.861
10 Oxygen 1,16 1.428 0.249
Carotenoids 1,16 4.584 0.048
Oxygen 9Carotenoids 1,16 0.311 0.585
SL 1,16 0.0030 0.954
*Measurements were taken at different developmental stages starting at 4, 6, 8, and 10 months post-hatch. Interaction
terms between fixed factors and standard length were tested for and removed from the model when not significant.
Values shown in bold are significant at a=0.05.
A square root transformation was applied to the % Red scores.
§
A log10 transformation was applied to the % Red : % Yellow scores.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 13
(i.e. high dissolved organic carbon), whereas high-DO
sites tend to be variably turbid (Crispo & Chapman,
2008; L. J. Chapman, pers. observ.). We might then
expect the red-shifted colour pattern that we observe
in low-DO swamp habitats to be a reflection of
female preference for high contrast males, and the
more yellow male coloration observed in high-DO riv-
ers and lakes to reflect decreased visibility and
female preference. We know that increased turbidity
results in labile behavioural responses where male
P. multicolor exhibit more aggression toward com-
petitors (Gray et al., 2012). Quantification of the
background light environments and the interaction
between turbidity and DO regimes are currently
being explored.
CONCLUSIONS
In this initial examination of secondary sexual trait
divergence in wild-caught P. multicolor, we report
the first evidence of male colour divergence across
environmental gradients, contributing to our general
understanding of divergence across extreme environ-
ments. The finding that P. multicolor males from
low-DO habitats have a lower reproductive display
frequency, in combination with observed male colour
variation, provides support for a possible trade-off,
where male P. multicolor from low-DO populations
maintain levels of mate attraction by increasing
levels of (potentially) more costly nuptial colours (i.e.
red), and reducing energetically expensive courtship
displays. In addition, our laboratory study supports
the role of carotenoid availability in the diet as a dri-
ver of phenotypically plastic colour variation as
expected, and suggests that colour production may
have a fixed genetic component, and/or that other
environmental factors (e.g. light environment) cou-
pled with low DO are important in driving colour
divergence among populations in the field. Our low-
DO sites are chronically and naturally hypoxic and
our body of work on P. multicolor suggests that the
fish have adapted over generations to this extreme
environment. Hypoxia is likely to be an increasingly
important ecological variable in aquatic ecosystems
as the intensity and frequency of hypoxic events con-
tinue to rise globally (Diaz, 2001). Thus, continued
research examining the role of hypoxia as a driver of
adaptive divergence may provide important insights
into potential effects of the global hypoxia crisis on
the maintenance of fish diversity.
ACKNOWLEDGEMENTS
We thank D. Green, A. Hendry, R. Krahe, members
of the Chapman Laboratory, and several anonymous
reviewers for their comments on earlier versions of
the manuscript; C. Verdone-Smith for help with
graphics; P. A. Omeja, D. Twinomugisha, and Kibale
Fish and Monkey Project field assistants T. Molton,
C. Walsh, and T. Tran for help with fieldwork; and
fish laboratory personnel L. Grigoryeva, L.H.
MacDonnell, J. Hunter, and K. Gong. This work was
approved by the McGill University Animal Care
Committee (Protocol #5659) and received permission
from the Ugandan Government. Funding was pro-
vided by Quebec Centre for Biodiversity Science and
National Science and Engineering Council of Canada
(NSERC) graduate fellowships to GVM and CNF;
NSERC PDF to SMG; NSERC Discovery Grant to
LJC; and Canada Research Chair funds (LJC).
REFERENCES
Baerends GP, Baerends van Roon JM. 1950. An introduc-
tion to the study of the ethology of the cichlid fishes. Beha-
viour, Supplement (1): 1243.
Ball JN. 1961. On food of the brown trout of Llyn Tegid. Pro-
ceedings of the Zoological Society of London 137: 599622.
Barrett RDH, Hendry AP. 2012. Evolutionary rescue. In:
Wong B, Candolin U, eds. Behavioral responses to a chang-
ing world. Oxford: Oxford University Press, 216233.
Bennett AF, Houck LD. 1983. The energetic cost of court-
ship and aggression in a plethodontid salamander. Ecology
64: 979983.
Bleiweiss R. 2004. Novel chromatic and structural biomark-
ers of diet in carotenoid-bearing plumage. Proceedings of
the Royal Society of London Series B, Biological Sciences
271: 23272335.
Boughman JW. 2001. Divergent sexual selection enhances
reproductive isolation in sticklebacks. Nature 411: 944948.
Bucher TL, Ryan MJ, Bartholomew GA. 1982. Oxygen
consumption during resting, calling, and nest building in
the frog Physalaemus pustulosus.Physiological Zoology 55:
1022.
Candolin U. 2000. Changes in expression and honesty of
sexual signalling over the reproductive lifetime of stickle-
backs. Proceedings of the Royal Society of London Series B,
Biological Sciences 267: 24252430.
Candolin U, Salesto T, Evers M. 2007. Changed environ-
mental conditions weaken sexual selection in sticklebacks.
Journal of Evolutionary Biology 20: 233239.
Chapman LJ 2015. Low-oxygen lifestyles. In: Riesch R,
Tobler M, Plath M, eds. Extremophile fishes: ecology,
evolution, and physiology of teleosts in extreme environ-
ments. Basel: Springer, 934.
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: 387393.
Chapman LJ, Chapman CA, Nordlie FG, Rosenberger
AE. 2002. Physiological refugia: swamps, hypoxia tolerance
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
14 G. V. MCNEIL ET AL.
and maintenance of fish diversity in the Lake Victoria
region. Comparative Biochemistry and Physiology A: Molec-
ular and Integrative Physiology 133: 421437.
Chapman LJ, Albert J, Galis F. 2008. Developmental plas-
ticity, genetic differentiation, and hypoxia-induced trade-
offs in an African cichlid fish. The Open Evolution Journal
2: 7588.
Clotfelter ED, Ardia DR, McGraw KJ. 2007. Red fish,
blue fish: trade-offs between pigmentation and immunity in
Betta splendens.Behavioral Ecology 18: 11391145.
Costantini D, Dell’Omo G, Casagrande S, Fabiani A,
Carosi M, Bertacche V, Marquez C, Snell H, Tapia W,
Gentile G 2005. Inter-population variation of carotenoids
in Galapagos land iguanas (Conolophus subcristatus). Com-
parative Biochemistry and Physiology B: Biochemistry and
Molecular Biology 142: 239244.
Coyne JA, Orr HA 2004. Speciation. Sunderland, MA: Sin-
auer.
Crispo E, Chapman LJ. 2008. Population genetic structure
across dissolved oxygen regimes in an African cichlid fish.
Molecular Ecology 17: 21342148.
Crispo E, Chapman LJ 2010. Temporal variation in popu-
lation genetic structure of a riverine African cichlid fish.
Journal of Heredity 101: 97106.
Crispo E, Chapman LJ 2011. Hypoxia drives plastic diver-
gence in cichlid body shape. Evolutionary Ecology 25: 949
964.
Crocker C, Chapman LJ, Martinez M. 2013. Natural vari-
ation in enzyme activity of the African cichlid Pseudocreni-
labrus multicolor victoriae.Comparative Biochemistry and
Physiology Part B: Biochemistry and Molecular Biology
164: 5360.
Cummings ME, Gelineau-Kattner R. 2009. The energetic
costs of alternative male reproductive strategies in
Xiphophorus nigrensis.Journal of Comparative Physiology
A195: 935946.
Desjardins JK, Stiver KA, Fitzpatrick JL, Milligan N,
van der Kraak GJ, Balshine S. 2008. Sex and status in
a cooperative breeding fish: behavior and androgens. Behav-
ioral Ecology and Sociobiology 62: 785794.
DeWitt TJ, Langerhans RB. 2003. Multiple prey traits,
multiple predators: keys to understanding complex
community dynamics. Journal of Sea Research 49: 143155.
Diaz RJ. 2001. Overview of hypoxia around the world. Jour-
nal of Environmental Quality 30: 275281.
Dijkstra PD, Seehausen O, Groothuis TGG. 2005. Direct
malemale competition can facilitate invasion of new colour
types in Lake Victoria cichlids. Behavioural Ecology and
Sociobiology 58: 136143.
Dijkstra PD, Wiegertjes GF, Forlenza M, van der Sluijs I,
Hofmann HA, Metcalfe NB, Groothuis TGG. 2011. The
role of physiology in the divergence of two incipient cichlid
species. Journal of Evolutionary Biology 24: 26392652.
Endler JA. 1980. Natural selection on colour patterns in
Poecilia reticulata.Evolution 34: 7691.
Endler JA. 1990. On the measurement and classification of
colours in studies of animal colour patterns. Biological
Journal of the Linnean Society 41: 315352.
Endler JA, Houde AE. 1995. Geographic variation in
female preferences for male traits in Poecilia reticulata.
Evolution 49: 456468.
Friesen CN, Aubin-Horth N, Chapman LJ. 2012. The
effect of hypoxia on sex hormones in an African cichlid
Pseudocrenilabrus multicolor victoriae.Comparative Bio-
chemistry and Physiology A: Molecular and Integrative
Physiology 162: 2230.
Frischknecht M. 1993. The breeding coloration of male stick-
lebacks (Gasterosteus aculeatus) as an indicator of energy
investment in vigour. Evolutionary Ecology 7: 439450.
Fuller RC. 2002. Lighting environment predicts the relative
abundance of male colour morphs in bluefin killifish (Luca-
nia goodei) populations. Proceedings of the Royal Society of
London Series B, Biological Sciences 269: 14571565.
Garratt M, Brooks RC. 2012. Oxidative stress and condi-
tion-dependent sexual signals: more than just seeing red.
Proceedings of the Royal Society of London Series B, Biolog-
ical Sciences 279: 31213130.
Godin JG, McDonough HE. 2003. Predator preference for
brightly colored males in the guppy: a viability cost for a
sexually selected trait. Behavioral Ecology 14: 194200.
Goodwin TW. 1980. Nature and distribution of carotenoids.
Food Chemistry 5: 313.
Gotanda KM, Reardon EE, Chapman LJ. 2011. Hypoxia
and male behaviour in an African cichlid Pseudocrenilabrus
multicolor victoriae.Journal of Fish Biology 78: 20852092.
Grantner A, Taborsky M. 1998. The metabolic rates associ-
ated with resting, and with the performance of agonistic,
submissive and digging behaviours in the cichlid fish
Neolamprologus pulcher (Pisces: Cichlidae). Journal of
Comparative Physiology B 168: 427433.
Gray SM, McKinnon JS. 2007. Linking color polymorphism
maintenance and speciation. Trends in Ecology and Evolu-
tion 22: 7179.
Gray SM, Dill LM, Tantu FY, Loew ER, Herder F,
McKinnon JS. 2008. Environment-contingent sexual selec-
tion in a color polymorphic fish. Proceedings of the Royal
Society of London Series B, Biological Sciences 275: 1785
1791.
Gray SM, Lisney TJ, Hart F, Tremblay M, Hawryshyn
CW. 2011. The effects of handling time, ambient light and
anaesthetic method on the standardized measurement of
fish colouration. Canadian Journal of Fisheries and Aquatic
Sciences 68: 330342.
Gray SM, McDonnell LH, Cinquemani FG, Chapman
LJ. 2012. As clear as mud: turbidity induces behavioral
changes in the African cichlid Pseudocrenilabrus multicolor.
Current Zoology 58: 143154.
Grether GF. 2000. Carotenoid limitation and mate prefer-
ence evolution: a test of the indicator hypothesis in guppies
(Poecilia reticulata). Evolution 54: 17121724.
Grether GF, Hudon J, Millie DF. 1999. Carotenoid limita-
tion of sexual coloration along an environmental gradient
in guppies. Proceedings of the Royal Society of London Ser-
ies B, Biological Sciences 266: 13171322.
Grether GF, Hudon J, Endler JA. 2001. Carotenoid
scarcity, synthetic pteridine pigments and the evolution of
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 15
sexual coloration in guppies (Poecilia reticulata). Proceed-
ings of the Royal Society of London Series B, Biological
Sciences 268: 12451253.
Grether GF, Cummings ME, Hudon J. 2005a. Counter-
gradient variation in the sexual coloration of guppies (Poe-
cilia reticulata): drosopterin synthesis balances carotenoid
availability. Evolution 59: 175188.
Grether GF, Kolluru GR, Rodd FH, de la Cerda J, Shi-
mazaki K. 2005b. Carotenoid availability affects the devel-
opment of a colour-based mate preference and the sensory
bias to which it is genetically linked. Proceedings of the
Royal Society of London Series B, Biological Sciences 272:
21812188.
Heuschele J, Mannerla M, Gienapp P, Candolin U.
2009. Environment-dependent use of mate choice cues in
sticklebacks. Behavioral Ecology 20: 12231227.
Hill GE. 1996. Redness as a measure of the production cost
of ornamental coloration. Ethology, Ecology and Evolution
8: 157175.
Hoefler CD. 2008. The costs of male courtship and potential
benefits of male choice for large mates in Phidippus clarus
(Araneae: Salticidae). Journal of Arachnology 36: 210212.
Hoffman EA, Schueler FW, Jones AG, Blouin MS. 2006.
An analysis of selection on a colour polymorphism in the
northern leopard frog. Molecular Ecology 15: 26272641.
Hyslop EJ. 1980. Stomach contents analysis a review of
methods and their application. Journal of Fish Biology 17:
411429.
Johnson AM, Fuller RC. 2015. The meaning of melanin,
carotenoid, and pterin pigments in the bluefin killifish,
Lucania goodei.Behavioural Ecology 26: 158167.
Kodric-Brown A. 1989. Dietary carotenoids and male mating
success in guppy: an environmental component to female
choice. Behaviour Ecology and Sociobiology 25: 393401.
Kotiaho JS, Alatalo RV, Mappes J, Nielsen MG, Parri S,
Rivero A. 1998. Energetic costs of size and sexual signal-
ing in a wolf spider. Proceedings of the Royal Society of
London Series B, Biological Sciences 265: 22032209.
Leclercq E, Taylor JF, Migaud H. 2010. Morphological
skin color changes in teleosts. Fish and Fisheries 11: 159
193.
Lin SM, Nieves-Puigdoller K, Brown AC, McGraw KJ,
Clotfelter ED. 2010. Testing the carotenoid trade-off
hypothesis in the polychromatic Midas Cichlid, Amphilo-
phus citrinellus.Physiological and Biochemical Zoology 83:
333342.
Lythgoe JN. 1979. The ecology of vision. Oxford: Clarendon
Press.
Maan ME, Seehausen O. 2011. Ecology, sexual selection
and speciation. Ecology Letters 14: 591602.
Maan ME, Seehausen O, Soderberg L, Johnson L, Rip-
meester EAP, Mrosso HDJ, Taylor MI, van Dooren
TJM, van Alphen JJM. 2004. Intraspecific sexual selection
on a speciation trait, male coloration, in the Lake Victoria
cichlid Pundamilia nyererei.Proceedings of the Royal Society
of London Series B, Biological Sciences 271: 24452452.
Maan ME, Seehausen O, van Alphen JJM. 2010. Female
mating preferences and male coloration covary with water
transparency in a Lake Victoria cichlid fish. Biological
Journal of the Linnean Society 99: 398406.
Mart
ınez ML, Chapman LJ, Rees BB. 2009. Population
variation in hypoxic responses of the cichlid Pseudocreni-
labrus multicolor victoriae.Canadian Journal of Zoology
87: 188194.
McGraw KJ, Hill GE, Stradi R, Parker RS. 2001. The
influence of carotenoid acquisition and utilization on the
maintenance of species-typical plumage pigmentation in
male American goldfinches (Carduelis tristis) and northern
cardinals (Cardinalis cardinalis). Physiological and Bio-
chemical Zoology 74: 843852.
McKinnon JS, Mori S, Blackman BK, David L, Kingsley
DM, Jamieson L, Chou J, Schluter D. 2004. Evidence
for ecology’s role in speciation. Nature 429: 294298.
McNeil GV. 2012. Hypoxia, carotenoids, and colour expres-
sion in the widespread African cichlid fish Pseudocrenilab-
rus multicolor victoriae. Thesis, Department of Biology,
McGill University.
Morrongiello JR, Bond NR, Crook DA, Wong BBM.
2010. Nuptial coloration varies with ambient light environ-
ment in a freshwater fish. Journal of Evolutionary Biology
23: 27182725.
Nakazawa T, Yamamura N. 2009. Theoretical considera-
tions for the maintenance of interspecific brood care by a
Nicaraguan cichlid fish: behavioral plasticity and spatial
structure. Journal of Ethology 27: 6773.
Oberweger K, Goller F. 2001. The metabolic cost of bird-
song production. Journal of Experimental Biology 204:
33793388.
Olson VA. 2006. Estimating nutrient intake in comparative
studies of animals: an example using dietary carotenoid
content in birds. Oikos 112: 620628.
Olson VA, Owens IPF. 1998. Costly sexual signals: are car-
otenoids rare, risky or required? Trends in Ecology and
Evolution 13: 510514.
Pike TW, Blount JD, Bjerkeng B, Lindstrom J, Metcalfe
NB. 2007. Carotenoids, oxidative stress and female mating
preference for longer lived males. Proceedings of the Royal
Society of London Series B, Biological Sciences 274: 1591
1596.
Prager M, Johansson EI, Andersson S. 2009. Differential
ability of carotenoid C4-oxygenation in yellow and red
bishop species (Euplectes spp.). Comparative Biochemistry
and Physiology Part B: Biochemistry and Molecular Biology
154: 373380.
Price AC, Weadick CJ, Shim J, Rodd FH. 2008. Pig-
ments, patterns and fish behaviour. Zebrafish 5: 297307.
Reardon EE, Chapman LJ. 2009. Hypoxia and life-history
traits in a eurytopic African cichlid. Journal of Fish Biology
75: 17951815.
Reardon EE, Chapman LJ 2010. Energetics of hypoxia in
a mouth-brooding cichlid: evidence for interdemic and
developmental effects. Physiological and Biochemical Zool-
ogy 83: 414423.
Ros AFH, Becker K, Oliveira RF. 2006. Aggressive beha-
viour and energy metabolism in a cichlid fish, Oreochromis
mossambicus.Physiology and Behavior 89: 164170.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
16 G. V. MCNEIL ET AL.
Rundle HD, Nosil P. 2005. Ecological speciation. Ecology
Letters 8: 336352.
Ryan MJ. 1985. Energetic efficiency of vocalization by the
frog Physalaemus pustulosus.Journal of Experimental Biol-
ogy 116: 4752.
Schluter D. 2000. The ecology of adaptive radiation. Oxford:
Oxford University Press.
Schoener TW. 1970. Nonsynchronous spatial overlap of
lizards in patchy habitats. Ecology 51: 408418.
Seehausen O, Schluter D. 2004. Malemale competition
and nuptial-color displacement as a diversifying force in
Lake Victoria cichlid fishes. Proceedings of the Royal Soci-
ety of London Series B, Biological Sciences 271: 13451353.
Seehausen O, van Alphen JJM, Witte F. 1997. Cichlid
fish diversity threatened by eutrophication that curbs sex-
ual selection. Science 277: 18081811.
Seehausen O, Terai Y, Magalhaes IS, Carleton KL,
Mrosso HDJ, Miyagi R, van der Sluijs I, Schneider
MV, Maan ME, Tachida H, Imai H, Okada N. 2008.
Speciation through sensory drive in cichlid fish. Nature
455: 620626.
Sefc KM, Brown AC, Clotfelter ED. 2014. Carotenoid-
based coloration in cichlid fishes. Comparative Biochemistry
and Physiology A: Molecular and Integrative Physiology
173: 4251.
van der Sluijs I, Gray SM, Amorim MCP, Barber I, Can-
dolin U, Hendry AP, Krahe R, Maan ME, Utne-Palm
AC, Wagner HJ, Wong BBM. 2011. Communication in
troubled waters: the evolutionary implications of changing
environments on fish communication systems. Evolutionary
Ecology 25: 623640.
Sobel JM, Chen GF, Watt LR, Schemske DW. 2010. The
biology of speciation. Evolution 64: 295315.
Sopinka NM, Fitzpatrick JL, Desjardins JK, Stiver KA,
Marsh-Rollo SE, Balshine S. 2009. Liver size reveals
social status in the African cichlid Neolamprologus pulcher.
Journal of Fish Biology 75: 116.
Sparkes T, Rush V, Foster SA. 2008. Reproductive costs,
condition and carotenoid-based color in natural populations
of threespine stickleback (Gasterosteus aculeatus). Ecology
of Freshwater Fish 17: 292302.
Stelkens RB, Seehausen O. 2009. Phenotypic divergence
but not genetic distance predicts assortative mating among
species of a cichlid fish radiation. Journal of Evolutionary
Biology 22: 16791694.
Svensson PA, Wong BBM. 2011. Carotenoid-based signals in
behavioural ecology: a review. Behaviour 148: 131189.
Taigen TL, Wells KD. 1985. Energetics of vocalization by
an anuran amphibian (Hyla versicolor). Journal of Compar-
ative Physiology B 155: 163170.
Tuomainen U, Candolin U. 2011. Behavioural responses to
human-induced environmental change. Biological Reviews
86: 640657.
Woods WA Jr, Hendrickson H, Mason J, Lewis SM.
2007. Energy and predation costs of firefly courtship sig-
nals. American Naturalist 170: 702708.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publisher’s web-
site:
Data S1. Supplemental information.
Table S1. Sample size and mean SE standard length (SL) (cm) of Pseudocrenilabrus multicolor collected
from four field sites in Uganda for stomach content analysis and color photo analysis (note some fish were
used for both analyses).
Table S2. Point-in-time measurements of various environmental characteristics for four field sites (SE)
where Pseudocrenilabrus multicolor were sampled during dry season months in 2011 in Uganda.
Table S3. Description and classification of behaviours for Pseudocrenilabrus multicolor of swamp (low DO)
and lake (high DO) origin collected and scored during observational periods.
Table S4. Percent Similarity Index among the diets of Pseudocrenilabrus multicolor collected from four sam-
pling sites across two drainages in Uganda.
Table S5. Average water temperature (°C) and dissolved oxygen concentration (mg L
1
)(SE) for the four
treatments used in the laboratory rearing experiment. Mean values were calculated from weekly measure-
ments taken over 10 months.
Figure S1. (A) The Mpanga River system of Western Uganda where samples were collected from the Bwera
swamp site (low DO) and the Bunoga river site (high DO); and, (B) the Nabugabo region where samples were
collected from the Lwamunda Swamp (low DO) and Lake Kayanja (high DO) (map modified from Crispo &
Chapman, 2008).
Figure S2. Example spectral reflectance measurements taken at six predetermined patches on the left side of
four wild-caught Pseudocrenilabrus multicolor (one for each site: A) Bunoga, B) Bwera, C) Lake Kayanja, and
D) Lwamunda). Spectra are colour coded with patch location on fish photo in (A). Fish photos correspond to
the specific fish used for spectral analysis of field populations.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
COLOUR VARIATION IN A EURYTOPIC CICHLID 17
Figure S3. Example spectral reflectance measurements were taken at six predetermined patches on four
lab-reared Pseudocrenilabrus multicolor, one from each treatment group (A) high DO-high carotenoid, (B) low
DO-high carotenoid, (C) high DO-low carotenoid, (D) low DO-low carotenoid), at the end of the lab-rearing
experiment. Spectra are colour coded with patch location on the fish in (A). Fish photos are of the fish that
were measured for spectral analysis.
©2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, ,
18 G. V. MCNEIL ET AL.
... While there is strong evidence for trait change associated with anthropogenically disrupted environments across wild populations at large spatial scales (e.g. Chapman et al., 2000;Lorenzon et al., 2001;McNeil et al., 2016;Winchell et al., 2016), our understanding of responses to rapid human-induced environmental change at finer, microgeographic spatial scales, within wild populations, is limited (Richardson et al., 2014). Here, we use a correlative approach to test for divergence (although we do not explicitly test if differences are genetic and/or plastic) in male nuptial coloration associated with variation in human-induced changes to water clarity within a single population of an African cichlid. ...
... Stream sites tend to be more variable than swamp sites due to intense human alteration of the landscape (e.g. deforestation and agriculture; Chapman et al., 2000;McNeil et al., 2016). Populations from these major habitat types are plastically, and in some cases genetically, divergent in a number of traits associated with low dissolved oxygen and high turbidity (e.g. ...
... Populations from these major habitat types are plastically, and in some cases genetically, divergent in a number of traits associated with low dissolved oxygen and high turbidity (e.g. Chapman et al., 2000;Chapman et al., 2008;Crispo & Chapman, 2010a;Gray et al., 2012;McNeil et al., 2016;Oldham et al., 2018;Reardon & Chapman, 2010). For example, fish from swamps tend to have smaller brains, larger gills (Chapman et al., 2000;Chapman et al., 2008;Crispo & Chapman, 2010a), reduced metabolic rates (Reardon & Chapman, 2010), and display a lower frequency of courtship behaviors (McNeil et al., 2016) relative to stream populations. ...
Article
Full-text available
Human activities are altering natural ecosystems, leading to widespread environmental change that can vary across spatiotemporal scales, thus creating dynamic, novel conditions at both large and small scales. In highly disturbed aquatic systems, elevated turbidity is one common stressor that alters the sensory environment of fishes and can disrupt communication, including mate choice, driving population‐level shifts in visual communication traits such as nuptial coloration. At a smaller, within‐population scale, we can expect similar adaptive divergence to a heterogeneous visual landscape. Using the cichlid fish, Pseudocrenilabrus multicolor, we investigated within‐population variation in diet and nuptial coloration by sampling fish from microhabitats within a relatively small site (~0.14 km2). These visual microhabitats are affected by different types of human disturbance at a very small scale leading to significant differences in water clarity (i.e. turbidity). We used three, non‐mutually exclusive working hypotheses to test if (1) males in low turbidity invest more in carotenoid‐based coloration (economy of pigments hypothesis), (2) fish from low‐turbidity sites eat more carotenoid‐rich foods (diet hypothesis), and (3) fish are habitat matching. Stomach content analyses revealed relatively high overlap in diet across microhabitats; however, fish from stations with the lowest turbidity consumed relatively more plant material (high in carotenoid content) than fish captured at high‐turbidity stations. Males from clearer waters displayed significantly more carotenoid‐based, red and yellow coloration than fish found in microhabitats with higher turbidity, similar to between‐population color variation in this species. Furthermore, larger fish displayed more carotenoid coloration overall, but there was no difference in mean male size among microhabitats suggesting that fish were not sorting into microhabitats. Our results suggest that within‐population variation in nuptial coloration could be associated with microhabitat heterogeneity in the visual landscape driven by turbidity, a diet with more carotenoid‐rich prey items, or a combination of both. In highly disturbed aquatic systems, elevated turbidity is one common stressor that alters the sensory environment of fishes and can disrupt communication. Using a cichlid fish we investigated within‐population variation in diet and nuptial coloration across highly disturbed sites. We found that, even at a microgeographic scale, changes in the visual landscape caused by elevated turbidity are associated with male color.
... In aquatic ecosystems, hypoxia (i.e., low dissolved oxygen, DO) and turbidity (i.e., suspended particulates) are metabolically and visually challenging conditions, respectively, that have been shown individually to generate divergence in behavioral, life history, and physiological traits in aquatic organisms (e.g., van der Sluijs et al. 2011;Chapman 2015). For example, in contrast to normoxic environments, hypoxic environments tend to favor fish that are less active (Abrahams et al. 2005;Gotanda et al. 2011;McNeil et al. 2016), have relatively larger gills and smaller brains (Chapman et al. 2008;Crispo and Chapman 2010), lower metabolic rates (Reardon and Chapman 2010;Crocker et al. 2013), and shorter brooding times due to the challenge of extracting oxygen at very low concentrations. ...
... We also examined whether individuals exhibited consistent individual differences in these behavioral traits (i.e., personality) both within and across populations. This species is broadly distributed across the Nile River basin of East Africa where it is found living under a variety of environmental conditions, the extremes of which include the dense interior of hypoxic (low DO), clear swamps to normoxic (high DO) but turbid lake edges and rivers (e.g., Chapman et al. 2002;Crispo and Chapman 2008;McNeil et al. 2016). Additionally, river sites experience fluctuating environmental conditions due to degradation of the surrounding landscape leading to excessive sediment run-off that is pulsed into the system during intense rain events. ...
... In their home environment, swamp fish face energetic challenges associated with low DO (Chapman 2015) and have adapted (through both plastic and genetic mechanisms) to these environments by having lower routine metabolic rates, larger gills, and smaller brains than river fish from high DO habitats (Reardon and Chapman 2009;Crispo and Chapman 2010). Furthermore, swamp fish have also been shown to be less active and display fewer reproductive behaviors than river fish when in their low DO home environment (e.g., Gotanda et al. 2011;McNeil et al. 2016). Therefore, it might make sense to expect swamp fish to demonstrate a slower pace of life relative to river fish, but only under home conditions. ...
Article
Full-text available
Animals are increasingly faced with human-induced stressors that vary in space and time, thus we can expect population-level divergence in behaviors that help animals to cope with environmental change. However, empirical evidence of behavioral trait divergence across environmental extremes is lacking. We tested for variation in behavioral traits among two populations of an African cichlid fish [Pseudocrenilabrus multicolor victoriae (Seegers 1990)] that experience extremes of dissolved oxygen and turbidity and are known to vary in a number of physiological and life history traits associated with these stressors. Using a common garden rearing experiment, F1 progeny from wild-caught parents originating from a swamp (low dissolved oxygen (DO), clear) and a river (high DO, turbid) were reared in high DO, clear water. Predator simulation assays were conducted to test for (a) variation in boldness, general activity, and foraging activity between populations, (b) differences in correlations between behaviors within and across populations, and (c) repeatability of behaviors. There was strong evidence for divergence between populations, with swamp fish being more bold (i.e., leaving refuge sooner after a simulated predator attack) and active (i.e., spent more time out of refuge) than river fish. Across populations there were positive correlations between foraging activity and both boldness and general activity; however, within populations, there was only a strong positive relationship between foraging activity and boldness in the river population. Here, we have demonstrated that populations that originate from drastically different environments can produce progeny that exhibit measurable differences in behaviors and their correlated relationships even when reared under common conditions.
... A. Degrees of freedom for the F-test are shown as a subscript. Algal quality (ng g −1 ) is not equivalent to dividing total algal carotenoid abundance (ng) by algal mass (g) because carotenoids concentrations were not successfully obtained from all algae samples Environ Biol Fish the link between dietary carotenoid availability and fish coloration.McNeil et al. (2016) revealed that male carotenoid-based coloration in the eurytopic African cichlid Pseudocrenilabrus multicolor victoriae varied with dissolved oxygen content of the water, but not in any measurable way with fish diet.Deutsch (1997) found no evidence that male Malawi cichlids varied in coloration as a function of Differences in tissue caro ...
Article
Full-text available
Carotenoid pigments have myriad functions in fish, including coloration and immunity. The “carotenoid trade-off hypothesis” posits that dietary limitation of carotenoids imposes constraints on animals to allocate to one function at the expense of another. This hypothesis rarely has been tested in fish. We quantified tissue carotenoids in breeding and non-breeding female convict cichlids in Lake Xiloá, Nicaragua. This species is reverse sexually dichromatic such that females possess carotenoid-based coloration that males lack. We also collected algae samples near nest sites to assess carotenoid availability, recorded water depth, and examined cichlids’ behavioral interactions with pair mates, conspecifics, heterospecific competitors, and predators. Each of these, we predicted, would mediate potential carotenoid trade-offs. We found that non-breeding females had significantly higher levels of carotenoids in their integument, liver, and gonads compared to breeding fish. We found that algae and total carotenoids declined with depth across our study transects at 9, 11, 13, and 15 m, but the concentration of carotenoids (ng carotenoid g⁻¹ algae, or algal quality) did not vary with depth. Furthermore, relationships among carotenoid concentrations of the three tissue types did not vary with depth, and female color status (orange or not) was not affected by behavioral interactions with other community members, reproductive status, or water depth. Our results support previous studies showing that carotenoid pigmentation may serve a signal function that facilitates the establishment of non-breeding females within the breeding population. Our study also uncovered no evidence indicating that carotenoids are limiting in the diet of breeding female convict cichlids.
... Thus, when fish reduce swimming activity under hypoxia, it has been inferred to be an energy-saving response (Metcalfe and Butler 1984;Fischer et al. 1992;Chapman and McKenzie 2009). In an earlier study of P. multicolor, McNeil et al. (2016) collected adult fish from a low-oxygen swamp site and a high-oxygen lake site and held these fish under low-and high-oxygen conditions, respectively, in the lab for 6 mo, after which behavioral observations were recorded. McNeil and colleagues found no difference in overall activity levels; however, the frequency of reproductive behaviors was 75% lower in fish from the low-DO population compared to those from the high-DO population, which they inferred as an energy-saving strategy given the potentially high cost of reproductive displays. ...
Article
Hypoxia and climate warming are pervasive stressors in aquatic systems that may have interactive effects on fishes because both affect aerobic metabolism. We explored independent and interactive effects of dissolved oxygen (DO) and temperature on thermal tolerance, behavior, and fitness-related traits of juvenile F1 offspring of the African cichlid Pseudocrenilabrus multicolor. Fish were reared in a split brood design with four treatments (low or high DO; cool or hot temperature); thermal tolerance, growth, and condition were measured after 1 month in the rearing treatments, following which behavioral traits were measured over 3.6 months. Critical thermal maximum was higher in fish reared under hot conditions, but was not affected by hypoxia. There was an interactive effect of DO and temperature on agitation temperature (temperature at which fish show behavioral signs of thermal stress) and the thermal agitation window (oC between the onset of agitation and final loss of equilibrium). Fish reared and tested under hot, normoxic conditions showed a higher agitation temperature; while fish reared and tested under hot, hypoxic conditions showed the largest thermal agitation window. Fish grew more quickly in length under hot than under cool conditions, and more quickly under normoxic than hypoxic conditions. Fish reared under cool, normoxic conditions were characterized by higher condition than other groups. Both cool and hypoxic rearing conditions reduced activity and aggression. These results highlight the importance of integrating physiological tolerance measures with sub-lethal behavioral effects of hypoxia and high temperature to gain a fuller understanding of species responses to multiple stressors.
Chapter
For fishes, the availability of dissolved oxygen (DO) can affect performance and fitness traits and influence distribution patterns. Hypoxia occurs naturally in habitats characterized by low mixing and/or light limitation such as dense wetlands and profundal zones of deep lakes. In addition, human activities are increasing the frequency and extent of aquatic hypoxia through eutrophication and pollution. Thus, it has become increasingly important to understand consequences of hypoxia for fishes and mechanisms that facilitate persistence in low-DO habitats. With strong specialization in some cichlid species and high levels of intraspecific variation in others, cichlids have been a key group for exploring strategies for dealing with hypoxia. These include behavioral responses (e.g. aquatic surface respiration), evolution of mechanisms to maximize oxygen uptake and delivery, metabolic depression, use of anaerobic metabolism, and air breathing in a few species. Despite the diversity of strategies that have permitted some cichlids to persist under extreme hypoxia, low DO can incur potential costs such as smaller body size. Such costs may be offset by benefits of hypoxic habitats such as reduced predation risk. This review details the mechanisms used by cichlids for tolerating hypoxia and the costs and benefits of hypoxia tolerance.
Chapter
The visual ecology of cichlids has contributed greatly to our understanding of mechanisms driving spectacular, colorful cichlid radiations. Interactions between the underwater light environment, the transmission of visual signals, and the visual sensitivity of the signal receiver are integral to the processes driving this diversity. Researchers recognized the importance of vision early in the study of African cichlids, citing the diversity of habitats in which cichlids are found and brilliant male nuptial coloration as potential forces shaping visual differentiation. Later work focused more on visual systems, adapted to the local light environment, as drivers of color pattern diversification. Most recently, researchers have focused on the evolution of visual systems under both ecological and sexual selection and the mechanisms of spectral tuning by investigating opsin gene expression and co-expression across the cichlid phylogeny. In this chapter, I describe the historical context of cichlid vision research, the diversity in cichlid visual ecology, and the current state of our understanding of cichlid visual ecology. Additionally, I discuss the possible consequences of human-induced changes to the underwater visual environment for cichlid diversity and suggest avenues for future research.
Article
Full-text available
Animals often need to signal to attract mates and behavioural signalling may impose substantial energetic and fitness costs to signallers. Consequently, individuals often strategically adjust signalling effort to maximize the fitness payoffs of signalling. An important determinant of these payoffs is individual state, which can influence the resources available to signallers, their likelihood of mating and their motivation to mate. However, empirical studies often find contradictory patterns of state-based signalling behaviour. For example, individuals in poor condition may signal less than those in good condition to conserve resources (ability-based signalling) or signal more to maximize short-term reproductive success (needs-based signalling). To clarify this relationship, I systematically searched for published studies examining animal sexual signalling behaviour in relation to six aspects of individual state: age, mated status, attractiveness, body size, condition and parasite load. Across 228 studies and 147 species, individuals (who were predominantly male) invested more into behavioural signalling when in good condition. Overall, this suggests that animal sexual signalling behaviour is generally honest and ability-based. However, the magnitude of state-dependent plasticity was small and there was a large amount of between-study heterogeneity that remains unexplained.
Preprint
Full-text available
Background. Efficient communication requires that signals are well transmitted and perceived in a given environment. Natural selection therefore drives the evolution of different signals in different environments. In addition, environmental heterogeneity at small spatial or temporal scales may favour phenotypic plasticity in signaling traits, as plasticity may allow rapid adjustment of signal expression to optimize transmission. In this study, we explore signal plasticity in the nuptial coloration of Lake Victoria cichlids, Pundamilia pundamilia and Pundamilia nyererei. These two species differ in male coloration, which mediates species-assortative mating. They occur in adjacent depth ranges with different light environments. Given the close proximity of their habitats, overlapping at some locations, plasticity in male coloration could contribute to male reproductive success but interfere with reproductive isolation. Methods. We reared P. pundamilia , P. nyererei, and their hybrids under light conditions mimicking the two depth ranges in Lake Victoria. From photographs, we quantified the nuptial coloration of males, spanning the entire visible spectrum. In experiment 1, we examined developmental colour plasticity by comparing sibling males reared in each light condition. In experiment 2, we assessed colour plasticity in adulthood, by switching adult males between conditions and tracking coloration for 100 days. Results. We found that nuptial colour in Pundamilia did respond plastically to our light manipulations, but only in a limited hue range. Fish that were reared in light conditions mimicking the deeper habitat were significantly greener than those in conditions mimicking shallow waters. The species-specific nuptial colours (blue and red) did not change. When moved to the opposing light condition as adults, males did not change colour. Discussion. Our results show that species-specific nuptial colours, which are subject to strong divergent selection by female choice, are not plastic. We do find plasticity in green coloration, a response that may contribute to visual conspicuousness in darker, red-shifted light environments. These results suggest that light-environment-induced plasticity in male nuptial coloration in P. pundamilia and P. nyererei is limited and does not interfere with reproductive isolation.
Preprint
Full-text available
Background. Efficient communication requires that signals are well transmitted and perceived in a given environment. Natural selection therefore drives the evolution of different signals in different environments. In addition, environmental heterogeneity at small spatial or temporal scales may favour phenotypic plasticity in signaling traits, as plasticity may allow rapid adjustment of signal expression to optimize transmission. In this study, we explore signal plasticity in the nuptial coloration of Lake Victoria cichlids, Pundamilia pundamilia and Pundamilia nyererei. These two species differ in male coloration, which mediates species-assortative mating. They occur in adjacent depth ranges with different light environments. Given the close proximity of their habitats, overlapping at some locations, plasticity in male coloration could contribute to male reproductive success but interfere with reproductive isolation. Methods. We reared P. pundamilia , P. nyererei, and their hybrids under light conditions mimicking the two depth ranges in Lake Victoria. From photographs, we quantified the nuptial coloration of males, spanning the entire visible spectrum. In experiment 1, we examined developmental colour plasticity by comparing sibling males reared in each light condition. In experiment 2, we assessed colour plasticity in adulthood, by switching adult males between conditions and tracking coloration for 100 days. Results. We found that nuptial colour in Pundamilia did respond plastically to our light manipulations, but only in a limited hue range. Fish that were reared in light conditions mimicking the deeper habitat were significantly greener than those in conditions mimicking shallow waters. The species-specific nuptial colours (blue and red) did not change. When moved to the opposing light condition as adults, males did not change colour. Discussion. Our results show that species-specific nuptial colours, which are subject to strong divergent selection by female choice, are not plastic. We do find plasticity in green coloration, a response that may contribute to visual conspicuousness in darker, red-shifted light environments. These results suggest that light-environment-induced plasticity in male nuptial coloration in P. pundamilia and P. nyererei is limited and does not interfere with reproductive isolation.
Article
The effects of common anaesthetics on the hue, saturation and brightness measurements of the poeciliid fish Girardinus metallicus were investigated in two experiments. For both experiments the coloration of four body regions was measured from digital images of the same males obtained under three conditions: (1) control (in a water‐filled chamber); (2) anaesthetised with MS‐222; (3) anaesthetised with eugenol (clove oil). In experiment 1 anaesthetised fish were photographed out of water. In experiment 2 all photographs were taken in a water‐filled chamber. Anaesthetics altered coloration in both experiments. In the more methodologically consistent experiment 2 we found significantly different hue, increased saturation and decreased brightness in anaesthetic v. control conditions, consistent with darkening caused by the anaesthetics. The body regions differed in coloration consistent with countershading but did not differentially change in response to anaesthesia. These findings suggest that photographing fish in a water‐filled chamber without anaesthetic is preferable for obtaining digital images for colour analysis and that multiple body regions of fish should be measured when assessing coloration patterns meaningful in behavioural contexts, to account for the gradients caused by countershading. We are encouraged that some researchers employ such methods already and caution against using anaesthetics except when absolutely necessary for immobilisation. This article is protected by copyright. All rights reserved.
Article
Full-text available
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].
Article
Males (mean mass 1.7 g) called and amplexing pairs built foam nests in respirometer chambers. Mean oxygen consumption () of resting males during the day was 0.26 ml h⁻¹, and at night it was 0.53 ml h⁻¹. Mean of males that could hear other males calling but that were not themselves calling was 0.70 ml h⁻¹. Mean of calling males was 1.13 ml h⁻¹. The energy cost per call (whine) decreases as whine rate increases. Mean per frog during nest building was 2.03 ml h⁻¹. The individual energy cost incurred by male and female during nest building could not be separated. The data on oxygen consumption during sustained calling and nest building offer an opportunity for measuring voluntarily sustained elevated levels of aerobic metabolism in anurans. During calling and nest building mean aerobic metabolic scope was 1.23 and 1.67 ml h⁻¹, respectively. The corresponding factorial scope of about 5.7 is within the range of published values for anurans undergoing forced activity. Because there is a high energy cost associated with reproductive activities in Physalaemus, and presumably in other anurans, any interpretations of aerobic and anaerobic metabolic patterns in frogs and toads should take into account reproductive, as well as predatory and escape, behavior.
Article
Male Physalaemus pustulosus consume, on average, 1.2 μl of oxygen in the production of a single call, which is equivalent to an energy input per call of 0.024 J. The total power of complex cells, which can have a varying number of components, ranged from 0.36 to 0.46 mW. The total acoustic energy contained in these complex calls ranged from 0.12 to 0.30 mJ. The energetic efficiency of the vocalization ranged from 0.5 to 1.2%, which is similar to the range estimated for some other animals. The low energetic efficiency of vocalization by these frogs is due, in part, to the fact that the wavelengths of the call are too long, relative to the size of the frog, to be radiated efficiently. Although shorter wavelengths (higher frequencies) are radiated by the frog at relatively greater intensities, longer wavelengths (lower frequencies) attenuate less rapidly in the environment. It is suggested that selection generated by the acoustics of the environment favours calls with lower frequencies, but the morphology of the animal sets a lower limit to which frequencies can evolve.
Article
Male bluefin killifish (Lucania goodei) exhibit extensive color variation in their fins, but the utility of this variation has not yet been determined. We collected males from multiple populations and spectrophotometrically determined the pigment types responsible for fin coloration. We determined that the orange coloration in the caudal fin is caused by carotenoid pigmentation. In contrast, color in the anal fin is either pterin based (yellow and red) or structural (blue) with a melanic fin border. As these colors have different developmental origins, the potential for complex signaling is high. Therefore, we sought to determine whether behavior, reproductive success, or health correlated with pigmentation. Males with more melanin on the anal fin were more dominant and had higher spawning success. Male-male aggression was greater between males with similar-sized melanic borders, indicating that melanic markings function as badges of status between males. Caudal carotenoid pigmentation did not correlate with dominance, but this highly labile ornament was correlated with body condition, parasite infection, and spawning success, suggesting a role in intersexual selection by signaling health to potential mates. Similar results were found for caudal fin coloration using digital photography. Pterin pigmentation in the anal fin was not related to dominance but was related to overall spawning levels and parasite infection, suggesting that pterin pigmentation may also signal immune status. Thus, the coloration of male bluefin killifish provides multiple messages to multiple receivers through these 3 pigments (melanin, pterin, and carotenoid) that have distinct developmental origins.
Article
We examined the preferences of female guppies (Poecilia reticulata) from 11 localities in Trinidad with respect to male color-pattern elements, body shape and size, and overall color and brightness contrast. Females are on average more attracted to males from their own population than from alien populations, and populations appear to vary in the criteria used in female choice. Multiple-regression analysis suggests that mate-preference criteria vary among localities in intensity, sign, and the number of traits used. Although preference estimators and color-pattern parameters are unique to each population, only orange, black, and color contrast showed a correlation between degree of male trait and degree of preference for that trait. There is a clear effect of water color and a possible effect of predation intensity. The results are discussed in light of various models of sexual selection and the early stages of speciation.
Article
Sympatric native Anolis species with similar structural habitats but contrasting climatic habitats are closer in head and body size on species-rich than on depauperate islands. In two localities, sympatric Anolis species with differential occurrences in sun or shade sought lower, more shaded perches during midday, resulting in partly nonsynchronous utilization of the vegetation by the two species. The second observation may be related to the first in the following way: nonsynchronous spatial overlap could dictate relatively great resource overlap for species coinhabiting patchy or edge areas, requiring great differences between the species in prey size in addition to those in climatic habitat. The extent of such overlap on small depauperate islands could be greater if these contained a greater proportion of patchy or edge habitats (with respect to insolation), or if climatic preferences were broader and more overlapping than on large, species-rich islands. In each locality, the relatively more shade-inhabiting species occurred more often on larger perches and on lower perches than did the other species. In both species of the Bermudan pair, adult males occupied higher and larger perches, and in grahami, shadier perches, than did female-sized individuals. The statistical significance of these and other differences was evaluated using several unweighted @g^2 procedures, Cochran's weighted @g^2 test and a partitioning technique for analyzing interactions among variables in complex contingency tables. The last method is described in detail in the papaer by Fienberg, immediately following this one.
Article
Methods for analysing fish stomach contents are listed and critically assessed with a view to their suitability for determining dietary importance—this term is defined. Difficulties in the application of these methods are discussed and, where appropriate, alternative approaches proposed. Modifications which have practical value are also considered. The necessity of linking measurements of dietary importance to stomach capacity is emphasized and the effects of differential digestion upon interpretation of stomach contents outlined. The best measure of dietary importance is proposed as one where both the amount and bulk of a food category are recorded.