Courtship and agonistic sounds by the cichlid fish Pseudotropheus zebra

Abstract

Courtship and agonistic interactions in an African cichlid species present a richer diversity of acoustic stimuli than previously reported. Male cichlids, including those from the genus Pseudotropheus (P.), produce low frequency short pulsed sounds during courtship. Sounds emitted by P. zebra males in the early stages of courtship (during quiver) were found to be significantly longer and with a higher number of pulses than sounds produced in later stages. During agonistic intrasexual quiver displays, males produced significantly longer sounds with more pulses than females. Also, male sounds had a shorter duration and pulse period in courtship than in male-male interactions. Taken together, these results show that the acoustic repertoire of this species is larger than what was previously known and emphasize the importance of further research exploiting the role of acoustic stimuli in intra- and interspecific communication in African cichlids.

Full text

Courtship and agonistic sounds by the cichlid fish
Pseudotropheus zebra
J. Miguel Simões,a兲Inês G. Duarte, and Paulo J. Fonseca
Departamento de Biologia Animal e Centro de Biologia Ambiental, Faculdade de Ciências da Universidade
de Lisboa, Bloco C2 Campo Grande, 1749-016 Lisboa, Portugal
George F. Turner
Department of Biological Sciences, University of Hull, HU6 7RX, United Kingdom
M. Clara Amorim
Unidade de Investigação em Eco-Etologia, ISPA, Rua Jardim do Tabaco 34, 1149-041 Lisboa, Portugal
共Received 22 March 2007; revised 6 May 2008; accepted 28 May 2008兲
Courtship and agonistic interactions in an African cichlid species present a richer diversity of
acoustic stimuli than previously reported. Male cichlids, including those from the genus
Pseudotropheus (P.), produce low frequency short pulsed sounds during courtship. Sounds emitted
by P. zebra males in the early stages of courtship 共during quiver兲were found to be significantly
longer and with a higher number of pulses than sounds produced in later stages. During agonistic
intrasexual quiver displays, males produced significantly longer sounds with more pulses than
females. Also, male sounds had a shorter duration and pulse period in courtship than in male–male
interactions. Taken together, these results show that the acoustic repertoire of this species is larger
than what was previously known and emphasize the importance of further research exploiting the
role of acoustic stimuli in intra- and interspecific communication in African cichlids.
©2008 Acoustical Society of America. 关DOI: 10.1121/1.2945712兴
PACS number共s兲: 43.80.Ka 关MCH兴Pages: 1332–1338
I. INTRODUCTION
In recent years much attention has focused on the role of
interspecific mate choice on the impressive rate of speciation
of cichlid fishes from the Great African Lakes that have un-
dergone some of the fastest and most extensive adaptive ra-
diations among vertebrates 共e.g., Turner, 1999;Albertson et
al., 2003兲. Many authors have proposed that sexual selection
driven by female choice acting on male courtship colors may
have been a significant factor on the rapid speciation of these
fishes 共e.g., Couldridge and Alexander, 2001;Genner and
Turner, 2005兲. Males of several African cichlid fishes are
known to produce sounds during courtship 关reviewed in Lo-
bel 共1998兲and Amorim 共2006兲兴and recently acoustic signal-
ing has also been pointed out as a possible mechanism in-
volved in reproductive isolation 共Lobel, 1998;Amorim et al.,
2004兲among African Great Lake cichlids.
Less attention has been given to the role of acoustic
communication in intraspecific mate choice in these fishes.
In the early stages of courtship, male Pseudotropheus (P.)
quiver to females, producing low-frequency short-pulsed
sounds 共Amorim et al., 2004兲, but there are no published
records of sound production associated with behavioral ele-
ments characteristic of the later stages of courtship 共Baerends
and Baerends van Roon, 1950兲, or during agonistic displays.
If there is sufficient intraspecific variability in sound produc-
tion then acoustic communication may play a role in intra-
and intersexual selection and influence the outcome of fights
and mating decisions in Pseudotropheus, as observed in
other animals 共e.g., Ladich et al., 1992兲.
The present study was aimed at investigating the full
acoustic repertoire of P. zebra males and females associated
with both courtship and agonistic contexts.
II. METHODS
A. Fish stocks and maintenance
Twenty adult male and twelve adult female first-
generation offspring bred from a stock of wild-caught adult
Pseudotropheus zebra from Nkhata Bay, Malawi 共11°36⬘N;
34°17⬘E兲were used in this study. After each trial, the fish
were returned to stock tanks. Each tank was fitted with an
external power filter and maintained at 25–27 °C by an in-
ternal 250-W heater, on a 12: 12 h light:dark cycle provided
by room lights. A third of the tank’s water 共pH 7.5–8.5兲was
changed weekly. Fish were fed twice daily with a mixture of
commercial cichlid sticks and koi pellets.
B. Experimental protocol
Experiments were conducted between January and Sep-
tember 2005. Trials were conducted in two aquaria 共120
⫻60⫻45 cm high兲placed on top of a concrete plate sup-
ported by two rockwool blocks 共100⫻50⫻30 cm兲. This
setup proved to be effective to minimize external noise trans-
mitted through the building improving considerably noise to
signal ratio at the low frequencies considered in this study
共Fig. 1兲. Each experimental tank was divided transversally
a兲Electronic mail: jsimoes@ispa.pt
1332 J. Acoust. Soc. Am. 124 共2兲, August 2008 © 2008 Acoustical Society of America0001-4966/2008/124共2兲/1332/7/$23.00

by two opaque removable partitions into three compart-
ments: one of 50 cm in the middle and two of 35 cm.
During courtship experiments, a single male was intro-
duced into each of the smaller lateral compartments. These
compartments were provided with terracotta pots that served
as refuges and prospective spawning sites. In the central
compartment, six or seven females were kept permanently.
Males were left visually isolated to acclimatize for a mini-
mum of 36 h prior to the beginning of the recordings. This
period was required for males to become territorial, as shown
by “digging” behavior around the refuge. Before the record-
ing period, all electrical devices were switched off, apart
from the room lights. Then, one of the opaque partitions was
removed, and one male had free access to the females in the
central compartment. During courtship behavior, male P. ze-
bra perform a number of distinct types of displays to the
females, which are not always shown in a fixed order. These
include the behavioral patterns dart, quiver, lead swim, and
circling with the female 共Baerends and Baerends van Roon,
1950;Amorim et al., 2004兲. During the recordings, we noted
which visual displays were accompanied by sound produc-
tion. However, sometimes males would produce sounds
without performing any behavioral display, such as during
swimming or when standing still in the water column. Once
recording was complete, the tested subject was weighed 共wet
mass, M兲, measured 共standard length, SL兲and returned to a
stock tank. Only 12 males and 5 females emitted sounds
suitable for analysis. Male size averaged 107.1 mm SL
关⫾SD 共range兲=⫾11.8 共88.0–122.0兲mm, where SD is stan-
dard deviation兴and 40.5 g M 关⫾9.7 共22.0–57.7兲g兴,
whereas females averaged 103.6 mm SL 关⫾0.02 共100.0–
106.5兲mm兴and 30.4 g M 关⫾2.5 共27.4 – 33.5兲g兴.
Sounds from females were recorded from female–
female interactions that naturally occurred when they were in
the middle compartment isolated from the males. Sounds
from male–male interactions were recorded by placing an-
other male in the middle compartment 共instead of the fe-
males兲, and following a similar procedure to the courtship
sound recordings. Agonistic encounters consisted of frontal
and lateral displays and chasing, occasionally escalating to
physical contact, including biting 共Baerends and Baerends
van Roon, 1950兲. During lateral displays, animals often
quiver, a behavior that is similar to the courtship quiver. To
avoid physical injuries, fish were separated before or at the
first sign of escalation to physical contact.
Recordings lasted 10 min for female–female, 15 min for
male–male, and 20 min for male–female interactions. The
duration of the recording sessions was derived from prelimi-
nary observations. All individuals were identified by natural
features, such as number and position of eggspots, fin length,
and marks on the body and fins.
C. Sound recording and analysis
Sounds were recorded using two High Tech 94 SSQ hy-
drophones 共sensitivity of −165 dB re 1 V
␮
Pa−1, flat fre-
quency response ⫾1dBupto6kHz兲and a Pioneer DVD
Recorder DVR-3100 共⫾1.5 dB from 40 Hz to 2 kHz,
sampled at 48 kHz, 16 bit兲. One hydrophone was placed
above the terracotta pot, where the territorial individuals
would most likely exhibit courtship or agonistic behaviors. A
second hydrophone was placed in the middle of the main
compartment or in the location where individuals would
more actively display at each other. The use of two hydro-
phones improved the probability of recording sounds close to
the sound producer and also provided information on the
degradation of the acoustic signals with distance. Recorded
sounds could be attributed to the subject males because their
intensity varied with distance from the hydrophones and
were consistently associated with particular courtship dis-
plays.
Sounds were analyzed with Adobe Audition 2.0. 共Adobe
Systems Inc., 2005兲and Raven 1.2.1 for Windows 共Cornell
Lab of Ornithology, 2003兲. Only sounds that showed a clear
structure and a high signal-to-noise ratio were considered.
These were typically recorded at a distance of 1-2 body
lengths of the focal fish. The acoustic parameters analyzed
共Fig. 2兲were sound duration; number of pulses in a sound;
mean pulse period of the entire sound 共Mean PP兲; and sound-
peak frequency 共for a description of the acoustic parameters
see Amorim et al., 2004兲. In addition, other parameters also
considered included the mean pulse period of the first five
pulses in a sound 共Initial PP兲, and a second previously unde-
tected sound-peak frequency 共PF1兲typically around 150 Hz,
which is of higher energy than the sound peak in the 450 Hz
region 共PF2兲described by Amorim et al. 共2004兲. PF1 is eas-
ily confounded with background noise if the recording
aquarium is insufficiently acoustically insulated. When com-
paring an uninsulated stock tank with the experimental tanks,
background noises differed by approximately 50 dB at
100 Hz, i.e., around the frequency region of PF1, and by
30 dB at 450 Hz, i.e., around the frequency of PF2 共Fig. 1兲.
Temporal features were measured from oscillograms and
sound peak frequencies from power spectra based on 2048
point FFT with a Hamming window applied. Data are pre-
sented in relative units as it was not possible to measure
absolute sound levels.
Statistical analyses were performed using Statistica 7.0
共StatSoft Inc., 2005兲. Nonparametric statistics were used
whenever the assumptions for parametric tests were not met
Fre
q
uenc
y(
Hz
)
0
-20
-40
-60
-80
0300 600 900
(b)
(a)
Re
l
at
i
ve Amp
li
tu
d
e
(d
B
)
FIG. 1. Comparison between spectra of laboratory background noise re-
corded in 共a兲a stock tank, placed on top of a thin layer 共2cm兲of expanded
polystyrene and in 共b兲an experimental tank, placed on top of a thick layer
共50 cm兲of rockwool. Amplitude levels 共dB兲are relative to the maximum
value of the spectra. Sampling frequency 48 kHz, 2048 point FFT, filter
bandwidth 15 Hz, Hamming window, and 50% overlap.
J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra 1333

after applying data transformations. One-way analysis of
variance 共ANOVA兲was used to compare differences among
means of the acoustic parameters of male courtship quiver
sounds. The square root transformation was applied to the
number of pulses to meet the ANOVA assumptions. Spear-
man rank correlation was used to estimate whether courtship
quiver sound parameters were related to male morphological
features 共standard length, weight, and number of eggspots兲.
Twelve males with an average of 17 sounds per male 共⫾12.1
SD兲were considered for these analyses.
The Kruskal–Wallis nonparametric test was used to
compare the acoustic characteristics of sounds produced dur-
ing different stages of courtship 共lead swim, quiver, no visual
display, and circle兲. Because few interactions proceeded to
late stages of courtship, relatively few sounds were recorded
during activities characteristic of such phases. The following
sample sizes were considered: 36 quiver sounds from 9
males; 12 lead-swim sounds from 5 males; 10 circle sounds
from 1 male; and 8 no display sounds from 4 males. Note
that in this analysis, data concerning quiver sounds were re-
stricted to 36 randomly selected sounds from 9 males from
the whole data set 共i.e., 4 sounds per each male兲, to avoid
large imbalances between factor levels sample sizes. Circling
sounds were extremely hard to record, not only due to the
fact that this species rarely got to the ending stages of court-
ship during trials, but also because circling did not always
occur near the hydrophone. Thus, even though a few other
circling interactions were observed, it was possible in only
one case to analyze their uttered sounds. Nevertheless, the
comparison between circling and other sounds seemed nec-
essary to ascertain the variability in the acoustic repertoire of
this species and was included for analyses. Post-hoc pairwise
comparisons were made with Dunn tests to determine differ-
ences between groups of courtship behaviors 共Zar, 1984兲.
The hypothesis that sounds produced by males when
courting females could be different from those produced by
both sexes during agonistic interactions was also tested with
one-way ANOVA. For these analyses, 198 courtship quiver
sounds from the 12 analyzed males 共i.e., all quiver sounds
recorded during courtship interactions兲, as well as 124 ago-
nistic sounds emitted by 9 males and 27 sounds produced in
agonistic contexts by 5 females were considered. An average
of 14 sounds 共⫾4.4兲and 5 sounds 共⫾2.1兲were considered
per male and per female, respectively, in agonistic contexts.
The square root transformation was applied to the number of
pulses, whereas logarithmic transformations were carried out
for sound duration and the mean pulse period of the first five
pulses to meet the requirements of normality and homosce-
dasticity. Because PF2 is correlated with male SL 共see Sec.
III兲, an analysis of covariance 共ANCOVA兲was used to com-
pare this frequency parameter among sexes and social con-
text, having fish SL as a covariate to control for the effect of
male size. Post-hoc pairwise comparisons were made with
Tukey tests for unequal sample sizes.
III. RESULTS
A. Male courtship sounds
During intersexual courtship trials, female sounds were
not detected. Focal male Pseudotropheus zebra varied in
their tendency to court females. Sounds were more fre-
quently produced when individuals from both sexes showed
a greater courtship activity. Eight of the twenty males tested
neither attempted courtship nor produced any sound suitable
for analysis. Only four males displayed late-stage courtship
behavior. Most recorded sounds 共86.4%兲were produced by
males during quivering, the main early stage courtship be-
havior. Because few encounters proceeded to the late stages
of courtship, such as lead swim and circle, there was a rela-
tively scarce sample 共13.2%兲of sounds produced during final
courtship.
Male quiver sounds had two main sound-peak frequen-
cies at approximately 150 Hz 共PF1兲and 450 Hz 共PF2兲. The
mean quiver sound duration was around 700 ms with ap-
proximately 9 pulses per sound. The mean pulse period was
approximately 90 ms and the initial pulse period circa 80 ms.
There were significant differences between males in all sonic
characteristics measured, except for sound duration 共Table I兲.
Intraindividual variation was generally high, especially for
sound duration and number of pulses and lowest for PF1 and
PF2, as shown by their coefficients of variation 共Table I兲.
Larger males produced quiver courtship sounds with
lower frequencies at PF2 共mass: rS=−0.62, N= 12, P
Frequency (kHz)
C
B
A
(a)
1.8
1.3
0.8
0.3
Time (s)
0.2 0.6 1.0 1.4 1.8
(b)
Relative Units
Re
l
at
i
ve Amp
li
tu
d
e
(d
B
)
PF1 PF2
(c)
Fre
q
uenc
y(
Hz
)
0
-20
-40
-60
1.0
0.5
0.0
-0.5
-1.0
150 350 550 750 950
FIG. 2. 共Color online兲共a兲Oscillogram, 共b兲sonogram, and 共c兲power spec-
trum of a Pseudotropheus zebra male courtship sound, representing some of
the acoustic parameters measured: mean pulse period of the 共A兲first five
pulses, 共Bis an example of a pulse兲number of pulses, and 共C兲sound
duration in 共a兲and peak frequency 1 共PF1兲and 2 共PF2兲in 共c兲. Sampling
frequency 48 kHz, 2048 point FFT, filter bandwidth 15 Hz, Hamming win-
dow, and 50% overlap.
1334 J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra

=0.028; length: rs= −0.81, N= 12, P= 0.001兲. Eggspot num-
ber was not significantly related to male size 共M and SL:
rs=0.50, P⬎0.05兲. Males with larger number of eggspots
tended to make calls with lower PF1 frequencies 共rs=−0.82,
N=9, P= 0.001兲and higher pulse repetition rates, i.e., shorter
pulse periods 共mean pulse period: rs=−0.68, N=9, P=0.04兲.
The duration of sounds and their number of pulses dif-
fered according to the courtship behavior performed by the
males with longer sounds containing more pulses emitted
during quivering bouts 共Table II, Fig. 3兲. The mean pulse
period of the first five pulses was shorter in sounds associ-
ated with quivering than in sounds registered when males
were not displaying, with lead-swim and circle sounds being
intermediate 共Table II, Fig. 3兲. The PF1 also differed signifi-
cantly according to which behavior the sound was associated
with 共Table II兲, but Dunn tests were unable to distinguish any
pair of behavioral categories.
B. Agonistic sounds
Sound production by males during agonistic interactions
frequently occurred after a brief fight, where males would
silently display frontally or laterally. Following such a con-
test, the dominant male 共normally the resident or the larger
fish兲displayed laterally and quivered to the submissive male.
Submissive males rapidly lost their bright colors, becoming
pale. They sometimes bit at the dominant male’s anal fin
eggspots, in a similar manner to a female during courtship.
TABLE I. Characteristics of sounds produced by P.zebra males and females during quiver in inter- and intrasexual interactions 共male–female—courtship
interactions; male–male and female–female—agonistic interactions兲. Means, SD, and range are based on fish means. Coefficients of variation 共COV
=SD/mean⫻100兲represent intraindividual variability of the acoustic parameters. Results for one-way ANOVA testing differences between males for
courtship quiver acoustic parameters, and testing differences between sounds made during different contexts and gender are presented.
Male–female Male–Male Female–female
Differences
between males
共courtship quiver兲
Differences
between
contexts/gender
Sound
parameters
Mean⫾SD
共range兲
COV
共%兲
Mean⫾SD
共range兲
COV
共%兲
Mean⫾SD
共range兲
COV
共%兲F11,186 PF
2,23 P
Duration
共ms兲
671.7⫾135.59
共421.4–856.8兲
60.37 960.5⫾295.29
共549.1–1429.5兲
69.13 524.2⫾152.95
共358.3–732.6兲
72.00 0.87 ns 8.56c0.002
Number of
pulses
8.6⫾1.67
共6.6–12.4兲
51.23 8.7⫾3.48
共4.7–13.8兲
52.59 4.9⫾0.99
共3.8–6.3兲
52.46 1.86a0.047 6.16a0.007
Mean pulse
period 共ms兲
86.8⫾14.37
共67.5–113.3兲
22.32 125.7⫾23.91
共90.1–160.9兲
36.39 123.8⫾27.06
共92.9–165.3兲
39.83 3.69 ⬍0.001 11.10 ⬍0.001
Initial pulse
period 共ms兲
76.7⫾15.31
共52.1–103.6兲
26.43 110.8⫾27.97
共79.2–149.8兲
34.64 116.7⫾34.52
共91.7–176.4兲
41.15 3.77 ⬍0.001 8.45c0.002
PF1 共Hz兲155.6⫾26.20
共129.4–220.7兲
15.17 138.0⫾14.97
共117.2–164.1兲
5.99 143.1⫾6.72
共133.9–152.3兲
8.33 10.80 ⬍0.001 2.04 ns
PF2共Hz兲488.8 ⫾40.84
共423.9–557.8兲
8.77 462.9⫾35.40
共433.6–550.8兲
6.06 480.2⫾29.14
共445.3–525.0兲
8.05 8.13b⬍0.001 1.80bns
Squared root transformation is applied.
Results from ANCOVA using fish SL as a covariate.
Logarithmic transformation is applied.
TABLE II. Characteristics of courtship sounds made by P. zebra during lead swim, quiver, with no associated display and circle. Data are pooled for all
recorded individuals due to the small sample size 共for quiver sounds only a subsample of 4 sounds per male was considered in the analyses—see Sec. II兲.
Coefficients of variation are also given: COV= SD /mean ⫻100. Results for Kruskal–Wallis statistics testing differences between sounds associated with
different courtship behaviors are presented.
Lead swim Quiver No display Circle Kruskal–Wallis
Sound
parameters
Mean⫾SD
共range兲
COV
共%兲
Mean⫾SD
共range兲
COV
共%兲
Mean⫾SD
共range兲
COV
共%兲
Mean⫾SD
共range兲
COV
共%兲HP
Duration
共ms兲
567.3⫾247.14
共214.0–1210.0兲
43.6 1198.6⫾647.32
共298.0–2622.0兲
54.0 481.4⫾381.16
共201.0–1276.0兲
79.2 561.7⫾157.44
共343.0–853.0兲
28.0 25.06 ⬍0.001
Number of
pulses
7.0⫾2.26
共4–12兲
32.2 14.9⫾7.78
共5–33兲
52.3 5.4⫾3.11
共3–12兲
57.9 7.4⫾1.07
共6–9兲
14.5 29.07 ⬍0.001
Mean pulse
period 共ms兲
91.5⫾20.35
共57.0–119.6兲
22.2 86.8⫾18.46
共60.3–132.6兲
21.3 100.4⫾13.33
共78.7–116.1兲
13.3 82.6⫾20.72
共52.7–134.7兲
25.1 6.26 ns
Initial pulse
period 共ms兲
82.7⫾20.08
共51.0–110.3兲
24.3 69.2⫾21.67
共37.8–124.8兲
31.3 90.8⫾13.35
共70.8–114.3兲
14.7 76.9⫾24.26
共52.6–141.4兲
31.6 10.73 0.01
PF1共Hz兲132.8 ⫾15.26
共117.2–164.1兲
11.5 149.5⫾30.66
共109.4–257.8兲
20.5 128.9⫾17.72
共117.2–164.1兲
13.7 140.6⫾0.00
共140.6–140.6兲
0.0 8.52 0.04
PF2共Hz兲459.0 ⫾52.37
共375.0–539.1兲
11.4 488.7⫾62.07
共398.4–585.9兲
12.7 454.1⫾48.43
共375.0–539.1兲
10.7 471.1⫾30.15
共421.9–492.2兲
6.4 1.60 ns
J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra 1335

Commonly, dominant males produced sounds during such
agonistic quivering. In female-female encounters, sounds
were generally produced during agonistic quivers, often by
females that showed sexual readiness or during mouthbrood-
ing, which also seemed to be more aggressive 共three out of
five females producing recorded sounds were mouthbrood-
ing兲.
Sounds produced in male–female, male–male, and
female–female encounters differed significantly in all tempo-
ral parameters but not in the frequency domain 共Table I, Figs.
4and 5兲. Male sounds were longer and included more pulses
than those emitted by females; moreover, male sounds also
differed in duration according to social context 共Fig. 4兲.
Courting male sounds also showed significantly shorter ini-
tial and mean pulse periods than agonistic sounds by either
sex 共Fig. 4兲.
IV. DISCUSSION
A. Male courtship sounds
The present study has shown that Pseudotropheus zebra
males produce sounds not only in the early stages of court-
ship, during quiver, but also during courtship displays that
occur closer to spawning. Moreover, the sound production in
the presence of females but without any other noticeable be-
havioral display, consistent with observations on another
Malawian haplochromine cichlid Tramitichromis interme-
dius, suggests that sound can be a purposely generated uni-
modal courtship display 共Ripley and Lobel, 2004兲.
Courtship sounds varied in their characteristics accord-
ing to the associated courtship behavior, being longer and
with a higher pulse rate during quivering 共Fig. 3兲. Although
only a small sample size of late stage courtship sounds was
FIG. 3. Variation of courtship sound parameters in Pseudotropheus zebra
males during lead swim 共LS兲, quiver 共Q兲, sounds produced with no apparent
body movement 共X兲, and circle 共C兲. Groups that are significantly different
共
␣
=0.05兲are indicated by different letters 共results from Tukey tests兲. Both
“Mean PP” and “Initial PP” refers to mean values of pulse periods; whereas
the first is the mean of the pulse periods throughout the entire sound, the
second indicates the mean of the first five pulses. Note that comparisons
considered data pooled for all males due to the small sample sizes obtained
for LS, X, and C. Only a subsample of quiver sounds was considered for the
analyses 共see methods兲.
FIG. 4. Variation of the acoustic parameters of quiver sounds emitted in
courtship 共male–female兲and agonistic interactions 共male–male and female–
female兲by Pseudotropheus zebra. Groups that are significantly different
共
␣
=0.05兲are indicated by different letters 共results from Dunn tests兲.
FIG. 5. Oscillograms of sounds produced associated with different contexts
and gender: 共A兲male courtship quiver, 共B兲circle, 共C兲male agonistic quiver,
and 共D兲female agonistic quiver. Sampling frequency 48 kHz.
1336 J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra

recorded, the present results indicate that acoustic communi-
cation is more diversified during the courting activities than
previously reported. In other cichlids, sound production
seems mostly restricted to male quivering during the early
stages of courtship 共Ripley and Lobel, 2004, reviewed in
Amorim et al., 2004兲, except in the Mozambique tilapia
Oreochromis mossambicus that produces sounds throughout
courtship 共Amorim et al., 2003兲.InO. mossambicus, sounds
are longer and with a faster pulse rate during the late-
courtship behavior of tail wagging 共Baerends and Baerends
van Roon, 1950兲than during other courtship activities
共Amorim et al., 2003兲. Although performed in different
phases of courtship, the quivering of P. zebra and the tail
wagging of O. mossambicus are probably equivalent in func-
tion. Both consist of displays in which males quiver their
bodies vigorously, simultaneously emitting sounds in close
proximity of the female, and may convey information of
their quality and motivation. In addition, both P. zebra and
O. mossambicus males may quiver and tail wag in all stages
of courtship 共Baerends and Baerends van Roon, 1950兲, par-
ticularly when females begin to wander out of a male’s core
spawning area.
Male motivation and quality may be advertised by
higher calling rates, longer calls, and higher pulse repetition
rates that are likely to be more energetically expensive. At
least in some species these parameters may be assessed by
females during mate choice. For example, in the gray tree
frog Hyla versicolor, females prefer longer male calls with a
higher pulse number to shorter calls 共Gerhardt et al., 2000兲,
and this parameter is an indicator of male genetic quality
共Welch et al., 1998兲. In fishes, Thorson and Fine 共2002兲dem-
onstrated that males Opsanus beta call faster at twilight,
shortening and simplifying their multiboop calls, suggesting
a tradeoff between call repetition rate and complexity in fe-
male choice. In invertebrates, pulse number and rate, to-
gether with sound frequency, are the most important acoustic
features involved in female choice 共e.g., Simmons, 1988兲.
Other sound parameters may transmit additional infor-
mation relevant to female mate choice. Quiver sounds dif-
fered considerably among individual males, for example
with larger males producing lower frequencies at PF2. This
parameter was also the one that showed the least intraindi-
vidual variation 共Table I兲, probably because it may be depen-
dent on male size 共Lobel, 2001;Amorim et al., 2003兲rather
than motivation. Male size is often regarded as an indication
of higher fitness and in cichlids may be related to social
status and breeding success 共e.g., Oliveira et al., 1996兲.
The association of courtship quiver sound parameters
共PF1 and pulse period兲with the number of eggspots in the
anal fin is less obviously explicable. Perhaps these param-
eters are independent indicators of some common cause,
such as overall male fitness. Eggspot number was correlated
with fish size in other cichlids 共Goldschmidt, 1991兲, although
it does not seem the case in P. zebra perhaps because of
species differences or the restricted size of fish used. Females
of several haplochromine cichlids are known to choose mat-
ing partners on the basis of their eggspot number 共Could-
ridge and Alexander, 2001兲and in some Pseudotropheus spe-
cies, females prefer a larger number of eggspots 共Couldridge
and Alexander, 2001兲.InP. zebra, lower sound-peak fre-
quency at PF2 and especially higher pulse rate may indicate
better male condition and could be used with additional vi-
sual cues from the eggspots in mate sexual selection.
B. Agonistic sounds
Sound produced by both sexes during agonistic contexts
is described in this study for P. zebra for the first time and
has been documented for a number of cichlid species 共re-
viewed by Lobel, 1998;Amorim, 2006兲. We found several
significant differences in the sounds produced in agonistic
context by males and females 共Fig. 4兲. Aggressive males
produced significantly longer and more pulsed sounds than
females. In addition, male sounds also differed according to
the social context. Courtship sounds were shorter and also
had a faster pulse repetition rate than male agonistic sounds
共Fig. 4兲. In line with our observations, in the croaking goura-
mis 共Trichopsis vittata兲, where only females produce sounds
during mating 共Brittinger, 1991;Ladich, 2007兲, female court-
ship croaks are also produced at a faster rate than the aggres-
sive croaks produced by both sexes 共Brittinger, 1991兲. Simi-
larly, the intervals between the double pulses that make up a
croak also differ between sexes and social context 共Brit-
tinger, 1991兲. Although there are relatively few published
quantitative comparisons of the influence of sex and social
context on fish sounds, taken together, the results with T.
vittata and the present study data with P. zebra suggest that
temporal parameters of fish sounds may contain information
on the motivation and gender of the sound producer. Sounds
may carry information about male quality or motivation,
which may influence the outcome of contests, in a manner
similar to that proposed for female mating decisions. Play-
back of conspecific aggressive sounds may inhibit aggression
in Cichlasoma 共now Archocentrus兲centrarchus, a Central
American cichlid fish 共Schwarz, 1974兲, whereas Trichopsis
males that vocalized during contests had an increased chance
of winning 共Ladich et al., 1992兲. In other taxa, a classical
example is provided by male toads, Bufo bufo, that settle
contests for the possession of females by signaling body size
and hence fighting ability with call frequency 共Davies and
Halliday, 1978兲.
Sound production by female P. zebra was noted in a
previous study, but not analyzed or compared with sounds
made by males 共Amorim et al., 2004兲. Sound was only pro-
duced when females appeared to be sexually receptive 共when
the ovipositor was visible兲or mouthbrooding, both situations
where females typically become more aggressive. Similarly,
sound production by mouthbrooding females has been docu-
mented for O. mossambicus 共Marshall, 1971兲. In another
cichlid fish, A. centrarchus, both sexes made sounds during
the breeding cycle in an aggressive context 共Schwarz, 1980兲.
Females of this substrate spawning species emit sounds
mainly during brood defence but also during nest preparation
before spawning. It has been suggested that sound produc-
tion by female fish may be more frequent than previously
thought, perhaps because the sound producing apparatus is
often less developed than in the male, resulting in weaker
J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra 1337

vocalizations, which are harder to detect 共Hawkins, 1993;
Ladich, 2007兲.
C. Concluding remarks
The variation in sounds we have documented indicates
that P. zebra vocalizations have the potential to carry infor-
mation about sex, size, motivation, and other fitness param-
eters that may play a role in sexual selection. Although ab-
solute sound pressure levels have yet to be measured, it is
clear that the sounds made by P. zebra are of low amplitudes
and attenuate severely within short distances from the sender
共Fig. 6兲, and it is unlikely that they are used to attract mating
partners or to repel rivals at distance 共Krebs et al., 1978兲.
More probably, and consistent with the behavioral contexts
in which the sounds were observed, acoustic signals may be
important during close-range encounters already initiated on
the basis of visual signals. As females may reject males at
this stage of a courtship sequence, and territorial rival males
may decide to flee or continue fighting, close range sounds
may play a major and complex role in the social behavior of
P. zebra and other African cichlid fishes.
This study is a detailed description of sounds produced
during courtship and agonistic interactions in the cichlid
Pseudotropheus zebra and reveals an acoustic repertoire
richer than previously thought. It emphasizes the need of
additional research to clarify the behavioral functions of the
sounds that may have also played a role in the rapid specia-
tion of African cichlids.
ACKNOWLEDGMENTS
The authors are grateful to J. Simões senior for the lo-
gistic support, N. Wreathall for helping with the fish ship-
ment, and all of those who helped building the tanks and
maintaining the fish. This study was supported by the Natural
Environment Research Council 共NERC兲that initially funded
our collections of fish from Malawi, Program No. POCTI-
ISFL-4-329/FCT, 共UI&D 331/94兲/FCT and a Grant No.
POSI SFRH/BPD/14570/2003/FCT 共MCA兲.
Albertson, R. C., Streelman, J. T., and Kocher, T. D. 共2003兲. “Directional
selection has shaped the oral jaws of Lake Malawi cichlid fishes,” Proc.
Natl. Acad. Sci. U.S.A. 100, 5252–5257.
Amorim, M. C. P. 共2006兲. “Diversity of sound production in fish,” Commu-
nication in Fishes, edited by F. Ladich, S. P. Collin, P. Moller, and B. G.
Kapoor 共Science Publishers, Enfield兲, Vol. I, pp. 71–104.
Amorim, M. C. P., Fonseca, P. J., and Almada, V. C. 共2003兲. “Sound pro-
duction during courtship and spawning of the cichlid Oreochromis mossa-
mbicus: male-female and male-male interactions,” J. Fish Biol. 62,658–
672.
Amorim, M. C. P., Knight, M. E., Stratoudakis, Y., and Turner, G. F. 共2004兲.
“Differences in sounds made by courting males of three closely related
Lake Malawi cichlid species,” J. Fish Biol. 65, 1358–1371.
Baerends, G. P., and Baerends van Roon, J. M. 共1950兲. “An introduction to
the study of the ethology of cichlid fishes,” Behaviour 1, 1–235.
Brittinger, W. 共1991兲. “The significance of the sound generating organ in the
behavioural context in Trichopsis vittatus 共Cuvier and Valencienne, 1831兲
共Belontiidae, Teleostei兲,” M.Sc. thesis, Univ. of Vienna, Vienna, Austria
共in German兲.
Couldridge, V. C. K., and Alexander, G. J. 共2001兲. “Color patterns and
species recognition in four closely related species of Lake Malawi
cichlid,” Behav. Ecol. 13, 59–64.
Davies, N. B., and Halliday, T. R. 共1978兲. “Deep croaks and fighting assess-
ment in toads, Bufo bufo,” Nature 共London兲74, 683–685.
Genner, M. J., and Turner, G. F. 共2005兲. “The mbuna cichlids of Lake
Malawi: a model for rapid speciation and adaptive radiation,” Fish Fish. 6,
1–34.
Gerhardt, H. C., Tanner, S. D., Corrigan, C. M., and Walton, H. C. 共2000兲.
“Female preference functions based on call duration in the gray tree frog
共Hyla versicolor兲,” Behav. Ecol. 11, 663–669.
Goldschmidt, T. 共1991兲. “Egg mimics in Haplochromine cichlids 共Pisces,
Perciformes兲from Lake Victoria,” Ethology 88, 177–190.
Hawkins, A. 共1993兲. “Underwater sound and fish behaviour,” Behavior of
Teleost Fishes, 2nd ed., edited by T. Pitcher 共Chapman and Hall, London兲,
pp. 129–169.
Krebs, J., Ashcroft, R., and Webber, M. 共1978兲. “Song repertoires and ter-
ritory defence in the great tit,” Nature 共London兲271, 539–542.
Ladich, F., Brittinger, W., and Kratochvil, H. 共1992兲. “Significance of ago-
nistic vocalization in the croaking gourami 共Trichopsis vittatus, Teleo-
stei兲,” Ethology 90, 307–314.
Ladich, F. 共2007兲. “Females whisper briefly during sex: context- and sex-
specific differences in sounds made by croaking gouramis 共Teleosts兲,”
Anim. Behav. 73, 379–387.
Lobel, P. S. 共1998兲. “Possible species specific courtship sounds by two sym-
patric cichlid fishes in Lake Malawi, Africa,” Environ. Biol. Fishes 52,
443–452.
Lobel, P. S. 共2001兲. “Acoustic behavior of cichlid fishes,” J. Aquaricult.
Aquat. Sci. 9, 167–186.
Marshall, J. A. 共1971兲. “Sound production by Tilapia mossambica 共Pisces:
Cichlidae兲,” Am. Zool. 11, 632.
Oliveira, R. F., Almada, V. C., and Canário, A. V. M. 共1996兲. “Social modu-
lation of sex steroid concentrations in the urine of male cichlid fish Oreo-
chromis mossambicus 共Teleostei: Cichlidae兲,” Horm. Behav. 30, 2–12.
Ripley, J. L., and Lobel, P. S. 共2004兲. “Correlation of acoustic and visual
communication in the Lake Malawi cichlid Tramitichromis intermedius,”
Environ. Biol. Fishes 71, 389–394.
Schwarz, A. 共1974兲. “Sound production and associated behavior in a cichlid,
Cichlasoma centrarchus,” Z. Tierpsychol 35, 147–156.
Schwarz, A. L. 共1980兲. “Sound production and associated behavior in a
cichlid fish, Cichlasoma centrarchus. II. Breeding pairs,” Environ. Biol.
Fishes 5, 335–342.
Simmons, L. W. 共1988兲. “The calling song of the field cricket, Gryllus
bimaculatus 共De Geer兲: constraints on transmission and its role in inter-
male competition and females choice,” Anim. Behav. 36, 380–394.
Thorson, R. F., and Fine, M. L. 共2002兲. “Crepuscular changes in emission
rate and parameters of the boatwhistle advertisement call of the gulf toad-
fish, Opsanus beta,” Environ. Biol. Fishes 63, 321–331.
Turner, G. F. 共1999兲. “Explosive speciation of African cichlid fishes,” Evo-
lution of Biological Diversity, edited by A. E. Magurran and R. M. May
共Oxford University Press, Oxford兲, pp. 217–229.
Welch, A. M., Semlitsch, R. D., and Gerhardt, H. C. 共1998兲. “Call duration
as a reliable indicator of genetic quality in the gray tree frog,” Science
280, 1928–1930.
Zar, J. H. 共1984兲.Biostatistical Analysis, 2nd ed. 共Prentice–Hall, Englewood
cliffs, NJ兲.
Relative Units
Time (s)
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(b)
(a)
-
1.0
0.5
0.0
-0.5
-1.0
-
0.5
0.0
-0.5
FIG. 6. Oscillograms of a courtship sound produced by a Pseudotropheus
zebra male recorded at a distance of 共a兲5 cm and 共b兲40 cm from the
hydrophone, in this case, sound attenuation was approximately 20 dB. Sam-
pling frequency 48 kHz.
1338 J. Acoust. Soc. Am., Vol. 124, No. 2, August 2008 Simões et al.: Acoustic repertoire of Pseudotropheus zebra