ArticlePDF Available

Acoustic Structure and Contextual Use of Calls by Captive Male and Female Cheetahs (Acinonyx jubatus)


Abstract and Figures

The vocal repertoire of captive cheetahs (Acinonyx jubatus) and the specific role of meow vocalizations in communication of this species attract research interest about two dozen years. Here, we expand this research focus for the contextual use of call types, sex differences and individual differences at short and long terms. During 457 trials of acoustic recordings, we collected calls (n = 8120) and data on their contextual use for 13 adult cheetahs (6 males and 7 females) in four Russian zoos. The cheetah vocal repertoire comprised 7 call types produced in 8 behavioural contexts. Context-specific call types (chirr, growl, howl and hiss) were related to courting behaviour (chirr) or to aggressive behaviour (growl, howl and hiss). Other call types (chirp, purr and meow) were not context-specific. The values of acoustic variables differed between call types. The meow was the most often call type. Discriminant function analysis revealed a high potential of meows to encode individual identity and sex at short terms, however, the vocal individuality was unstable over years. We discuss the contextual use and acoustic variables of call types, the ratios of individual and sex differences in calls and the pathways of vocal ontogeny in the cheetah with relevant data on vocalization of other animals.
Content may be subject to copyright.
Acoustic Structure and Contextual Use of
Calls by Captive Male and Female Cheetahs
(Acinonyx jubatus)
Darya S. Smirnova
, Ilya A. Volodin
*, Tatyana S. Demina
, Elena V. Volodina
1Department of Animal Science, Russian State Agrarian UniversityMoscow Timiryazev Agricultural
Academy, Moscow, Russia, 2Department of Vertebrate Zoology, Faculty of Biology, Lomonosov Moscow
State University, Moscow, Russia, 3Scientific Research Department, Moscow Zoo, Moscow, Russia
The vocal repertoire of captive cheetahs (Acinonyx jubatus) and the specific role of meow
vocalizations in communication of this species attract research interest about two dozen
years. Here, we expand this research focus for the contextual use of call types, sex differ-
ences and individual differences at short and long terms. During 457 trials of acoustic
recordings, we collected calls (n = 8120) and data on their contextual use for 13 adult chee-
tahs (6 males and 7 females) in four Russian zoos. The cheetah vocal repertoire comprised
7 call types produced in 8 behavioural contexts. Context-specific call types (chirr, growl,
howl and hiss) were related to courting behaviour (chirr) or to aggressive behaviour (growl,
howl and hiss). Other call types (chirp, purr and meow) were not context-specific. The val-
ues of acoustic variables differed between call types. The meow was the most often call
type. Discriminant function analysis revealed a high potential of meows to encode individual
identity and sex at short terms, however, the vocal individuality was unstable over years.
We discuss the contextual use and acoustic variables of call types, the ratios of individual
and sex differences in calls and the pathways of vocal ontogeny in the cheetah with relevant
data on vocalization of other animals.
Cheetahs (Acinonyx jubatus) are among animals that are most attractive for people due to their
nice appearance and interesting communicative behaviour with conspecifics and humans [1,2].
The cheetahs were intensely studied in relation to their wildlife ecology [37], conservation [8
10], diseases [11], morphology [1214] and genetics [1517]. At the same time, the acoustic
communication is rather poorly investigated for the cheetah.
Previously, based on the acoustic structure, call types of the vocal repertoire have been
described for cheetah cubs [18] and for cheetah adults [19]. The vocal repertoire of adult chee-
tahs comprises eight call types: purr, hiss, growl, chirr, meow, chirp, howl and gurgle [19](Fig
1and S1 Audio). In cheetah cubs younger three months, the vocal repertoire comprised of the
same seven call types as in adults, for the exclusion of the gurgle [18]. For assessing the
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 1/20
Citation: Smirnova DS, Volodin IA, Demina TS,
Volodina EV (2016) Acoustic Structure and
Contextual Use of Calls by Captive Male and Female
Cheetahs (Acinonyx jubatus). PLoS ONE 11(6):
e0158546. doi:10.1371/journal.pone.0158546
Editor: Gianni Pavan, University of Pavia, ITALY
Received: April 1, 2016
Accepted: June 19, 2016
Published: June 30, 2016
Copyright: © 2016 Smirnova et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any
medium, provided the original author and source are
Data Availability Statement: All data and
measurements are available without restrictions from
the paper, figures and Supporting Information. S1
Table contains acoustic measurements (10 acoustic
variables) of cheetah seven call types for describing
the acoustics of call types, in total 194 calls. S2 Table
contains acoustic measurements (9 acoustic
variables) of cheetah meows for estimating the
effects of sex and individuality on the acoustic
variables of meows and for estimating the stability of
vocal individuality in meows, in total 221 calls. Data
from S1 and S2 Tables are sufficient for calculation all
results and statistical comparisons, which are
provided in this paper.
functional role of different call types in the cheetah, a hypothetical scheme relating the acoustic
structure with emotional states of confidence/diffidence and aggressiveness/non-aggres-
siveness, has been proposed [19]. However, this scheme has not yet been confirmed with quan-
titative material on the contextual use of calls.
Whereas the study by Volodina [19] was related to the entire vocal repertoire of adult chee-
tahs, all other studies of cheetah vocalizations were related to particular call types within the
cheetah vocal repertoire. On example of cheetah purr and chirr vocalizations, the mechanism of
vocal vibration, representing an uninterrupted emission of pulsed vocalization for the duration
of both inspiration and expiration phases of breathing, has been investigated [20]. Purring is
caused by rapid twitching of the vocalis muscle, whereas the chirr vocalization is caused with
interaction between the purr and the tonal vocalization produced by the normal vibration of the
vocal folds [20,21]. The acoustics of cheetah purr vocalizations, produced continuously during
the inspiration and expiration phases, were also examined in a few studies [19,2224]. The acous-
tics of agonistic vocalizations of the cheetah have been considered by Eklund with coauthors
[25]. Frustrative meows of captive adult male and female cheetahs and of cheetah cubs, were pre-
liminary described in [26]. The studies focused on searching the vocal indicators of reproductive
state in the cheetah [27,28] revealed that males produce a specific courtingseries of chirrs inter-
spersed with chirps when exposed to urine samples taken from receptive females.
While the bark represents the most characteristic vocalization of canids [2936], the meow
represents the most characteristic vocalization of felids, either wild [3739] or domesticated
[3843]. Consistently, the meow represents a prominent vocalization in their vocal repertoire
of the cheetah [18,19,26]. Cheetah meows occur in different contexts but primarily in the con-
texts related to frustration and discomfort (solicitation, anticipation and separation)
[26,44,45]. Previously, vocal individuality of meow calls was shown in the isolation context for
domestic kittens (Felis catus)[42] and for four adult male cheetahs [44]. At the same time, sex-
specific variation of vocalizations was not yet studied for any felid species, although some vocal
sex dimorphism is expected given the sexual dimorphism in body size. For example, approxi-
mately 20% difference in body mass exists between male and female domestic cats [46], 15%
between male and female captive cheetahs [8] and 22% between male and female wild cheetahs
[14]. Therefore, the vocal apparatus and the sound-producing structures are also expected to
be larger for the larger sex [47,48] and their acoustics (the fundamental and formant frequen-
cies) are expected to be lower in the larger sex [49,50]. However, very close or indistinguishable
acoustics between vocalizations of the larger and smaller sex were reported for some subspecies
of red deer (Cervus elaphus hispanicus [51] and C.e.sibiricus [52]).
The contextual use of different call types represents the powerful tool for examining the
functions of these calls [53]. For the cheetah, the occurrence of different call types across beha-
vioural contexts in captivity was examined only preliminary [45]. The focus of this study is to
integrate the structural and contextual analysis of cheetah vocalizations for more precise assess-
ment of the functional role of different call types in this species. The purposes of this study
were: 1) to estimate quantitatively the contextual use of different call types by adult captive
cheetahs across behavioural contexts, 2) to estimate quantitatively the identity and sex-related
differences in meows at short terms and 3) to estimate the stability of vocal individual traits in
cheetah meows over years.
Material and Methods
Ethics statement
Three authors (IAV, EVV and TSD) are zoo staff members, so no special permission was
required for them and for their Masters student (DSS) to work with animals in Moscow Zoo
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 2/20
Funding: This work was supported by Russian
Science Foundation ( funding to
IAV, DSS and EVV, grant number 14-14-00237. This
research received no additional funding from any
public, commercial or not-for-profit sectors. The
funder had no role in study design, data collection
and analysis, decision to publish, or preparation of
the manuscript.
Competing Interests: The authors have declared
that no competing interests exist.
Fig 1. Spectrogram illustrating seven call types produced by adult (male and female) cheetahs (Acinonyx jubatus)in
captivity. Apurr, Bhiss, Cgrowl, Dchirr, Emeow, Fchirp, Ghowl (S1 Audio). The spectrogram was created at 11025
Hz sampling frequency, Fast Fourier Transform (FFT) 1024, Hamming window, frame 50%, overlap 93.75%.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 3/20
and Volokolamsk Zoo Breeding Station. For access to the cheetahs kept in Yaroslavl and Novo-
sibirsk zoos, the specific permissions were obtained from administration of these zoos by the
request from the President Director of Moscow Zoo V.V. Spitsin for the period of data collec-
tion. Vocalizations were recorded from inside and outside the animal enclosures during zoo
working hours under supervision of zoo staff. Call collector (DSS) did not manipulate the ani-
mals for the purpose of this study. Disturbance of animals was kept to a minimum. No animal
has suffered somehow due to the data collection. The research protocol # 201136 has been
approved by the Committee of Bio-ethics of Lomonosov Moscow State University. We adhered
to the Guidelines for the treatment of animals in behavioural research and teaching(Anim.
Behav., 2006, 71, 245253) and to the laws on animal welfare for scientific research of the Rus-
sian Federation, where the study was conducted.
Study subjects
Spontaneously produced calls of 13 (6 male and 7 female) adult (older 2 years) captive-born
cheetahs of the African subspecies (A.j.jubatus) were recorded in 20122014 in zoos of Russia.
In particular, six animals (Male 1 Seva, Male 2 Adam, Male 3 Kay, Female 7 Eva,
Female 8 Sindi, Female 9 Kimi) were recorded at Volokolamsk Zoo Breeding Station of
Moscow Zoo (Moscow region, Volokolamsk district) in May-August 2012. Five of these ani-
mals (Males 1, 2 and Females 7, 8, 9) were then repeatedly recorded in June-July 2014 and their
calls served for analysis of the stability of the individual acoustic characteristics with time.
Two animals (Male 4 Adday, Female 10 Nayla) were recorded in Yaroslavl Zoo (Jaro-
slavl city) in June 2013. Two animals (Male 5 Kidjan, Female 11 Annay) were recorded in
Novosibirsk Zoo (Novosibirsk city) in July 2013. Three animals (Male 6 Frank, Female 12
Zygota, Female 13 "Winda") were recorded in Moscow Zoo (Moscow city) in October-
November 2014.
Animal housing
Animals were kept in outdoor enclosures (sizes varied from 240 to 500 m
depending on the
zoo) containing the indoor enclosures (warm houses subdivided inside into four departments
4x4 m, individual for each animal). During the day, animals were released into the outdoor
enclosure (together or singly, depending on the zoo), whereas during the night they stayed
inside their individual indoor enclosures. The walls of the indoor enclosures were made of
wire-mesh, so the animals could see the conspecifics in the indoor enclosures. The feeding
occurred twice a day in the individual indoor enclosures, the first one in the morning before
the releasing to the outdoor enclosures and the second one in the evening, before the locking
the exit for the night.
Call collection
Calls were recorded indoor and outdoor. The call collector (DSS) was in the same indoor or
outdoor enclosure as animals, but was separated from them with wire-mesh and not entered
into contact with animals. Calls were collected by the focal animal sampling method [54]. For
sound recordings (sampling rate 48 kHz, 16 bit resolution) we used a Zoom-H4n professional
digital recorder (Zoom Corporation, Tokyo, Japan) with built-in microphones. All acoustic
recordings were conducted during daytime in periods of maximal activity of the animals dur-
ing routine procedures (feeding, releasing from indoor to outdoor enclosures, communication
of animals with their keepers and with other cheetahs through the wire-mesh).
We established eight mutually exclusive behavioural contexts in which calls were produced:
1) "Offensive" context (attack or threat aggression toward a human, researcher or keeper, or
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 4/20
toward a conspecific); 2) "Defensive" context (defensive aggression toward a human or conspe-
cific); 3) "Conspecific-Contact" context (friendly close-range communication between conspe-
cifics); 4) "Call-Over" context (distant communication with a conspecific, commonly without
visual contact); 5) "Human-Contact" context (friendly close-range communication with a
human, researcher or keeper); 6) "Release-Soliciting" context (keeper-directed soliciting to
release the animal outdoor or back; 7) "Food-Anticipating" context (feeding anticipation and
arousal when seeing the food and feeding the neighbor animals; 8) "Courting" context (male
courting female as a component of sexual behaviour). Calls were not used as markers for iden-
tifying the contexts. During the recording, the researcher labeled the context by voice. Individ-
uals could be distinguished by their coloration pattern. The distance from microphones to the
animals was 0.510 m. Each recording trial (457 trials in total) lasted 110 minutes, contained
calls from a single individual cheetah and was stored as a wav-file. The number of trials per
cheetah individual was from 6 to 86, on average 35.2 ± 23.7 trials per individual.
Call samples
We tried to collect material maximally balanced by individuals and contexts. However, as vocal
activity of different individuals was unequal, call samples were unevenly distributed by differ-
ent animals. Most indoor recordings contained echo; this did not interfere determining call
types but limited the number of calls potentially applicable for analysis of the acoustic struc-
ture. We prepared four different call samples: 1) for analysis of the occurrence of call types in
different behavioural contexts, 2) for describing the acoustics of call types, 3) for estimating the
effects of sex and individual identity on the acoustics of meows and 4) for estimating the stabil-
ity of the acoustic individuality in meows with time.
For analysis of the occurrence of call types in different behavioural contexts, we analysed a
total of 8120 cheetah calls for their call type using the Avisoft SASLab Pro software (Avisoft
Bioacoustics, Berlin, Germany). Based on frequency and temporal acoustic structure, we subdi-
vided calls into seven types: purr, hiss, growl, chirr, meow, chirp and howl (Fig 1), following
the early classification [19]. In cases when calls had a transitional acoustic structure [19], e.g.
from meow to purr or from howl to growl, each call part was treated as a separate call. For each
call, we determined the behavioural context (among the eight contexts) in which it was pro-
duced. Then we examined the contexts in which the 8120 calls were used, following the criteria
previously adopted by Salmi et al. [53]. We classified a call type as context-specific if it was
given in the same behavioral context more than 65% of cases. Because multiple call types can
be used in the same context, we then examined whether this was the primary call type for this
context, classifying it as signal-specific if it accounted for more than 65% of all calls given dur-
ing that context [53].
For describing the acoustics of call types, occurring in cheetahs in captivity, we selected
from the total massive of recordings from 6 to 60 calls per each of the 7 call types, 183 calls in
total. For each call type, calls were taken from 3 to 12 animals. We took calls of best quality, not
superimposed with other calls and background noise.
For estimating the effects of sex and individuality on the acoustic variables of meows, we
selected 1015 calls per individual from 12 individuals, for the exclusion of Female 13, which
did not produce meows, 151 meows in total. To decrease the effect of pseudoreplication, we
selected calls from different recording trials and within recording trials calls from different
parts of the trial, avoiding taking calls following one after other.
For estimating the stability of vocal individuality in meows with time (over 2 years), we
selected 1015 meows per individual from 5 individuals (Males 1, 3 and Females 7, 8, 9)
recorded in 2012 (69 meows) and in 2014 (70 meows), 139 meows in total.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 5/20
Call analysis
Only calls with high call-to-noise ratio, non-overlapped with background noise or calls of
other individuals, non-disrupted by wind, and clearly identified as belonging to focal individu-
als were included in the analysis. For call analysis, we used Avisoft, with 48 kHz sampling fre-
quency, the Hamming window, FFT length 1024 points, frame 50% and overlap 93.75%. These
settings allowed frequency resolution 46 Hz and time resolution 1.3 ms. All measurements
were made manually and have been exported to Microsoft Excel (Microsoft Corp., Redmond,
For calls of all types, we measured the duration from the screen with the standard marker
cursor in the spectrogram window (Fig 2), for the exclusion of purr, whose duration could last
many minutes. In addition, for all calls of all types, we measured the maximum amplitude fre-
quency (f peak) and three quartiles (q25, q50 and q75), covering respectively 25, 50 and 75% of
call energy (hereafter the lower, medium and upper quartiles) from the mean power spectrum
of each call. For purr, produced continuously at both respiratory phases, the acoustic variables
were measured separately for the inspiration and the expiration call phases. In calls with rhyth-
mic pulsation (chirr, growl and purr call types) we also measured with the standard marker
cursor the pulse rate. For each purr vocalization, three subsequent phases of expiration-inspira-
tion were included in analyses for calculating power variables and pulse rate. In growl, meow,
chirp and howl calls we additionally measured, with the reticule cursor, the initial (f0 beg),
end (f0 end), maximum (f0 max) and minimum (f0 min) fundamental frequencies of each call
(Fig 2).
The acoustic measurements for describing the acoustics of call types are presented in S1
Table. The acoustic measurements for estimating the effects of sex and individuality on the
Fig 2. Measured variables for cheetah meows. Spectrogram (right) and mean power spectrum of the entire call (left). Designations: durationcall duration;
f0 begthe fundamental frequency at the onset of a call; f0 endthe fundamental frequency at the endof a call; f0 maxthe maximum fundamental frequency;
f0 minthe minimum fundamental frequency; f peakthe frequency of maximum amplitude within a call; q25, q50 q75 the lower, the medium and the upper
quartiles, covering respectively 25%, 50% and 75% energy of a call spectrum. The spectrogram was created at 11025 Hz sampling frequency, Fast Fourier
Transform (FFT) 512, Hamming window, frame 50%, overlap 96.87%.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 6/20
acoustic variables of meows and for estimating the stability of vocal individuality in meows
with time are presented in S2 Table.
Statistical analyses
Statistical analyses were made with STATISTICA, v. 6.0 (StatSoft, Tulsa, OK, USA) and R
v.3.0.1 [55]; all means are given as mean ± SD. Significance levels were set at 0.05, and two-
tailed probability values are reported. Only 20 of 224 distributions of measured parameter val-
ues did depart from normality (Kolmogorov-Smirnov test, p>0.05) what allowed us to apply
parametric tests.
We used a two-way ANOVA with Tukey HSD test to compare the acoustics among call
types, with call type as fixed factor and individual as random factor. We used a two-way
ANOVA with Tukey HSD test to compare the acoustics between purr inspiration and expira-
tion phases, with phase as fixed factor and individual as random factor. We used a nested
design of ANOVA with an individual nested within sex to estimate effects of factors individu-
alityand sex, on the acoustic variables of meows, with sex as fixed factor and individual as
random factor.
We used standard procedure of discriminant function analysis (DFA) to calculate the prob-
ability of the assignment of meows to the correct individual. Nine variables, used for the DFA,
showed very low Pearson correlation values to each other. Among the total of 36 pairwise cor-
relations, the R
values were lower 0.2 for 23 comparisons; between 0.2 and 0.4 for 3 compari-
sons; between 0.4 and 0.6 for 4 comparisons; between 0.6 and 0.8 for 5 comparisons, and only
for 1 comparison (f0 end with f0 min) the R
value was 0.91. Then we investigated the stability
of acoustic individuality of meows between years for cheetahs that provided calls in two years.
We classified meows from 2014 with DFA functions derived from 2012, considering the value
of the correct cross-validation as a measure of the retention of individuality over time [5658].
With a 2x2 Yates' chi-squared test, we compared the values of correct assignment of meows to
the correct caller between years.
We used WilksLambda values to estimate how strongly acoustic variables of calls contrib-
ute to discrimination of individuals. To validate our DFA results, we calculated the random val-
ues of correct assignment of calls to individual by applying randomization procedure with
macros, created in R. The random values were averaged from DFAs performed on 1000 ran-
domized permutations on the data sets as described by [59]. Using a distribution obtained by
the permutations, we noted whether the observed value exceeded 95%, 99% or 99.9% of the val-
ues within the distribution [59]. If the observed value exceeded 95%, 99% or 99.9% of values
within this distribution, we established that the observed value did differ significantly from the
random one with a probability p<0.05, p<0.01 or p<0.001 respectively [5760].
Acoustic structure of cheetah calls
From the total sample of cheetah calls, 7228 calls could be classified to distinctive call types,
whereas 446 calls (5.81%, N = 7674 calls) were transitional from one call type to another (thus
including two different call types, one after another without time space between them). In
these cases each call part of the transitional call was treated as a separate call, what resulted in
the total sample of 8120 calls analysed for call type. Most often transitional calls occurred
between purr and meow (267 calls), growl and howl (86 calls) and growl and meow (60 calls).
Based on ANOVA results, we found that all acoustic variables were significantly related to
call type (Table 1). The duration did not differ between the growl and howl, being significantly
higher than in all other call types for the exclusion of purr, the continuous vocalization whose
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 7/20
duration could not be measured. The duration did not differ significantly between the chirr,
meow, chirp and hiss call types (Table 1 and Fig 1).
The pulse rate was minimal in the chirr, intermediate in the purr and maximal in the growl
call type; all differences were found significant. All fundamental frequency variables did not
differ between the growl and howl call types, being significantly lower than for the meow and
chirp call types and for the meow significantly lower than for the chirp (Table 1,Fig 1).
The values of the peak frequency and of the three power quartiles were the highest for the
chirp, lower for the meow and more prominently lower for the chirr. For the purr, growl, howl
and hiss, the values of the peak frequency and of the three power quartiles were the lowest ones
and differed significantly from the three other call types (for the exclusion of q50 and q75 for
the hiss) (Table 1).
For the purr, we also compared the values of acoustic variables between the inspiration and
expiration phases (Table 2). The values of the pulse rate, the peak frequency, and the lower and
medium quartiles were significantly higher for the inspiration phase than for the expiration
phase, whereas the values of the upper quartile did not differ between the inspiration and expi-
ration phases of the purr call type (Table 2).
Context and signal-specific call types
Four of 7 call types were context-specific (i.e., given mostly in one specific context), including
growl, howl and hiss (during Offensive context) and chirr (during Courting context) (Table 3).
All context-specific calls were given in a single context more than in 80% cases, except for
Table 1. Values (mean±SD) of acoustic variables for the cheetah call types.
Acoustic Call type ANOVA
variable Chirr (N = 11) Purr (N = 11) Growl (N = 33) Meow (N = 60) Chirp (N = 6) Howl (N = 19) Hiss (N = 43)
Duration (s) 0.74±0.73
= 38.01; p<0.001
Pulse rate (Hz) 16.59±1.03
= 98.46; p<0.001
f0 beg (kHz) - - 0.16±0.09
= 69.85; p<0.001
f0 end (kHz) - - 0.15±0.03
= 67.76; p<0.001
f0 max (kHz) - - 0.19±0.08
= 74.29; p<0.001
f0 min (kHz) - - 0.14±0.04
= 77.18; p<0.001
f peak (kHz) 0.59±0.33
= 77.58; p<0.001
q25 (kHz) 0.68±0.26
= 80.26; p<0.001
q50 (kHz) 1.19±0.36
= 40.18; p<0.001
q75 (kHz) 2.03±0.44
= 24.88; p<0.001
Results for comparison of acoustics between call types (two-way ANOVA with Tukey HSD test with call type as xed factor and individual identity as random
factor) are given with letters; means sharing the same letter are not signicantly different. N = total number of calls of each type.
Table 2. Values (mean±SD) of acoustic variables for the cheetah purr inspiration and expiration phases.
Acoustic variable Inspiration (N = 11) Expiration (N = 11) ANOVA
Pulse rate (Hz) 24.57±2.05 20.80±1.74 F
= 64.65; p<0.001
f peak (kHz) 0.25±0.15 0.08±0.03 F
= 18.129; p<0.001
q25 (kHz) 0.21±0.11 0.08±0.03 F
= 22.58; p<0.001
q50 (kHz) 0.43±0.05 0.21±0.11 F
= 40.68; p<0.001
q75 (kHz) 0.91±0.22 0.84±0.69 F
= 0.09; p=0.77
Note: Three subsequent phases of expiration-inspiration measured for each poor vocalization (N = 11) served for calculating the acoustic variables.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 8/20
howl, which were given in a context-specific way in 72.2% cases. The three remaining call types
(meow, chirp and purr) were not found context-specific. The meow was the single call type
occurring in all behavioural contexts (Table 3).
For each of the 8 behavioural contexts we found the signal-specific call types (i.e., the most
common call type used in a specific context) (Table 4). The chirr was given primarily during
the Courting context; the purr was given primarily during the Human-Contact context. The
growl was given primarily during two aggressive contexts: the Offensive and the Defensive.
The meow usage was specific during four contexts: the Conspecific-Contact, the Call-Over, the
Release-Soliciting and the Food-Anticipation. The remaining three call types (chirp, howl and
hiss) were not signal-specific (Table 4).
Effects of identity and sex on meow acoustics
Two-way ANOVA revealed effects of individual identity on all variables of meows, whereas
effects of sex were found only on variables of fundamental frequency (Table 5). Values of all fun-
damental frequency variables were substantially and significantly lower in males than in females
(for instance, the f0 max was 0.85±0.40 kHz in males and 1.07±0.25 kHz in females, Table 5).
We conducted two DFAs (for sex and for individual identity), each DFA based on all the 9
measured variables of meows. The DFA showed the average values of correct assignment to sex
Table 3. Call context-specificity: the percent of calls given in each of context, context-specific calls (i.e., for which >65%) are indicated in bold.
Call Context
type N
Offensive Defensive Conspecic-Contact Call-Over Human-Contact Release-Soliciting Food-Anticipating Courting
Chirr 978 0 0.2 0 0 2.2 0 0.1 97.4
Purr 1045 0 0 1.1 0.5 48.0 2.8 47.7 0
Growl 1485 81.1 9.6 0 0 0 1.9 7.3 0.2
Meow 3867 0.6 0.1 1.3 5.7 1.9 37.1 45.6 7.8
Chirp 73 1.4 2.7 0 26.0 1.4 52.1 8.2 8.2
Howl 212 72.2 6.6 0 0 0 8.0 13.2 0
Hiss 460 87.8 8.9 0 0 0 0 3.3 0
8120 1786 204 60 243 598 1548 2418 1263
= total number of calls of each type recorded in all contexts; N
= total number of calls given in each behavioral context.
Table 4. Call signal-specificity: the percent of 7 call types given in each context, signal-specific calls (i.e., those for which >65% are given in a sin-
gle context) are indicated in bold.
Context Call type
Chirr Purr Growl Meow Chirp Howl Hiss
Offensive 1786 0 0 67.4 1.3 0.1 8.6 22.6
Defensive 204 1.0 0 69.6 1.5 1.0 6.9 20.1
Conspecic-Contact 60 0 18.3 0 81.7 000
Call-Over 243 0 2.1 0 90.1 7.8 0 0
Human-Contact 598 3.7 83.9 0 12.2 0.2 0 0
Release-Soliciting 1548 0 1.9 1.8 92.8 2.5 1.1 0
Food-Anticipating 2418 0.1 20.6 4.5 72.9 0.2 1.2 0.6
Courting 1263 75.5 0 0.2 23.8 0.5 0 0
8120 978 1045 1485 3867 73 212 460
= total number of calls of each type recorded in all contexts; N
= total number of calls given in each behavioral context.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 9/20
of 78.1%, what was significantly higher than the random value 59.4 ± 3.3% (permutation test,
1000 permutations, p<0.001) (Fig 3). In order of decreasing importance, the f0 beg, duration
and q50 were mainly responsible for discrimination of sex for the meows.
At the same time, DFA showed the average values of correct assignment to individual of
59.6%, what was significantly higher than the random value 25.5 ± 3.0% (permutation test,
1000 permutations, p<0.001) (Fig 3). In order of decreasing importance, the f0 max, q75 and
q25 were mainly responsible for discrimination of individuals for the meows. However, the
value of correct assignment varied among individuals from 20% to 73.3%, and for one of 12 indi-
viduals did not differ from the random value. Thus, meows had reliable individual-specific traits
Table 5. Values (meanSD) of the cheetah meow variables and results of nested ANOVA for individual and sex differences.
Acoustic ANOVA Variable values for each sex
variable Individual differences Sex differences Male calls (N = 71) Female calls (N = 80)
Duration (s) F
= 5.43; p<0.001 F
= 0.54; p= 0.46 0.31±0.17 0.34±0.14
f0 beg (kHz) F
= 4.79; p<0.001 F
= 40.19; p<0.001 0.68±0.29 0.92±0.22
f0 end (kHz) F
= 12.01; p<0.001 F
= 33.54; p<0.001 0.62±0.25 0.77±0.16
f0 max (kHz) F
= 13.18; p<0.001 F
= 29.30; p<0.001 0.85±0.40 1.07±0.25
f0 min (kHz) F
= 9.90; p<0.001 F
= 40.83; p<0.001 0.59±0.22 0.76±0.16
f peak (kHz) F
= 11.52; p<0.001 F
= 0.28; p= 0.60 1.22±0.74 1.11±0.35
q25 (kHz) F
= 20.27; p<0.001 F
= 0.01; p= 0.91 1.04±0.51 1.00±0.20
q50 (kHz) F
= 9.32; p<0.001 F
= 1.62; p= 0.21 1.58±0.63 1.43±0.38
q75 (kHz) F
= 4.35; p<0.001 F
= 0.0; p= 0.98 2.38±0.79 2.31±0.64
Note: Individual nested within sex (with sex as xed factor, and individual as random factor); N = 12 cheetahs (6 males and 6 females).
Fig 3. Sex and individual discrimination of the cheetah meows. Green bars indicate values of
discriminant function analysis and yellow bars indicate random values, calculated with the randomization
procedure. Comparisons between observed and random values with permutation tests are shown above the
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 10 / 20
not in all individuals. Therefore, cheetah meows bear reliable cues to sex (higher fundamental fre-
quency in females compared to males) and have a potential to encode individual identity.
Between-year stability of meows
For 5 animals (2 males and 3 females) that provided sufficient number of meows in both 2012
and 2014, we compared the stability of vocal individuality in meows between years (Fig 4).
Within years, DFA showed high values of correct classification of meows to individual (75.4%
in 2012 and 78.6% in 2014) significantly exceeding the random value (44.4 ± 5.1% in 2012 and
44.8 ± 5.3% in 2014, permutation test, p<0.001 in both cases), and did not differ between
years (χ
= 0.06, p= 0.80).
However, cross-validation of meows recorded in 2014 using discriminant functions created
for meows recorded in 2012, revealed a strong decrease in the correct classification of individu-
als (Fig 4). The average value of correct classification dropped to the level expected by chance
alone (42.9%), and became significantly lower compared to call samples from either 2012
= 13.86, p= 0.002) or from 2014 (χ
= 17.25, p<0.001). Thus, in the cheetah, individual
identity of meows was unstable between years.
Vocal repertoire of adult cheetahs
We found that values of acoustic variables were substantially different between call types in the
cheetah. In this sense, the cheetah vocal repertoire can be considered as discretevocal
Fig 4. Discrimination of individual cheetahs by meows in two years (2012 and 2014). Green bars indicate values
of discriminant function analysis and yellow bars indicate random values, calculated with a randomization procedure.
Comparisons between observed and random values with permutation tests and comparisons between 2012 and 2014
meows with χ
tests are shown by brackets above the bars. The red bar indicates the classification value of 2014
meows with discriminant functions created for meows recorded in 2012.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 11 / 20
repertoire, similar to the more or less discretevocal repertoires of other felids [37] and some
ungulates, e.g. red deer [52]. This is distinctive to the more continualvocal repertoires, e.g.,
in red fox (Vulpes vulpes)[36] and in the wild boar (Sus scrofa)[61], in which a noticeable
number of intermediate vocalizations occurred along to distinctive call types. Transitional
forms from one call type to another occurred only in 5.8% cases, similar to [19]. No complicat-
ing features such as nonlinear phenomena or articulation effects [36,62] have been detected in
the cheetah vocalizations by this or previous studies.
The values of acoustic variables were close to those found in the previous study of the cheetah
vocal repertoire [19]. For instance, the average pulse rate of the chirr was 16.6 Hz in this study and
18.4 Hz in [19]; the pulse rate of the purr was 22.7 Hz in this study and 23.5 Hz in [19]; and the
pulserateofthegrowlwas37.4Hzinthisstudyand36.4Hzin[19], although the samples of ani-
mals, the animal housing and the staff were entirely different compared to the former study [19].
For the two phases (expiration and inspiration) of the purr vocalization, we found the differ-
ences in the pulse rate, the peak frequency, and in the lower and medium quartiles (all were sig-
nificantly higher for the inspiration phase than for the expiration phase). The differences in the
pulse rate (24.6 Hz during the expiration phase and 20.8 Hz during the inspiration phase) were
similar with those reported by [20] (26 Hz and 21 Hz respectively), by [24] (20.9 Hz and 18.3
Hz respectively) and by [63] (21.923.4 Hz and 19.320.9 Hz respectively). The average value
of pulse rate for the cheetah purr of 17.5 Hz [23] probable represents the value of pulse rate for
the cheetah chirr, as in another study of this author [22] the average value of pulse rate for the
cheetah chirr (termed the gurgle by [22]) was 16 Hz (ranging from 11 to 20.8 Hz).
Contextual use of call types in the cheetah
Context-specific call types were either related to aggressive behaviour in the Offensive context
(growl, howl and hiss) or to sexual behaviour in the Courting context (chirr). Consistently, ear-
lier studies report that male cheetahs produce the chirrs when courting receptive females,
whereas females use the chirrs for communication with cubs [19,64]. Towards urine samples of
non-receptive (not ready to mate) females, male cheetahs remain silent [27,28]. This helps to
select appropriate time for joining pairs for mating [27,28], as in zoos, males and female chee-
tahs are kept separately [65], otherwise they do not breed. Chirr vocalizations given towards
female urine sample in the heat indicates male competence as a breeder, whereas male silence
in this situation indicates its incapability to mate [27,28].
Among call types that were not context-specific (chirp, purr and meow) the meow takes an
especial place in the vocal repertoire of captive cheetahs. In this study, the meow was the most
often produced call type (47.6% of all calls, Table 3) presented in both sexes and in 12 of 13
individuals. In four of the 8 behavioural contexts, the meow was the most often call type, being
therefore in terminology of [53] the signal-specific call type for these four contexts. Two of
these contexts (Conspecific-Contact and Call-Over) are related to communication between
conspecifics, therefore they might correspond to the natural usage of meows in the wild. Two
remaining contexts (Release-Soliciting and Food-Anticipation) are specific for the captive con-
ditions, as meows given in these contexts were frustration calls directed toward a keeper and
appealing to human help. The regular use of meows for cheetah-human interactions may indi-
cate manipulating the keeper behaviour by the animals. Similarly, domestic cats use meows for
manipulating their owners [38,40,41]. Consistently, domestic dogs (Canis familiaris) exposed
to insoluble tasks, appeal for help to humans, staring on them and producing specific move-
ments [66] and frustrative whining vocalizations [67].
In contrast to the meow, the chirp was the rarest call type comprising only 0.9% of all calls,
and half of them (52%, Table 3) was given in the Release-Soliciting context. In other studies,
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 12 / 20
the cheetahs used chirps for calling towards cubs, mothers, potential mates or group mates
[44,64]. At experimental separations of coalitions of adult males in captivity, the chirps com-
prised 90% of the total of 196 calls [44]. In addition, consistently to this study, both chirps and
meows were used by cheetahs in contexts of food or stroll soliciting [26]. The found in this
study signal-specificity of the purr for the context of friendly close-range communication with
humans (Human-Contact) is consistent to reported data for the cheetah [2,19,24] and for the
domestic cat [23,24].
In this study, cheetahs more often vocalized in discomfort-related contexts: the Offensive,
Defensive, Release-Soliciting and Food-Anticipating contexts comprised 73.3% of calls,
whereas contexts of friendly interactions between animals (Conspecific-Contact, Call-Over
and Courting) or animals and humans (Human-Contact) comprised only 26.7% of all calls
(Table 4). This agrees well with findings that mammals primarily use calls in contexts related
to the negative emotional arousal and vocalize much more rarely when experience positive or
neutral emotions [6870].
For analysis of contextual use of call types, we used a pooled sample of calls from all chee-
tahs. This is the single possible approach for analysis of the contextual use of different vocaliza-
tions in either captivity or in the wild [53]. In captivity, each cheetah is situated in unique
conditions that limit the potential number of possible behavioural contexts. For instance, if an
animal is kept with mates, it can display towards them aggressive, friendly or sexual attitudes.
Otherwise, if an animal is kept singly, any interactive contexts are impossible. Furthermore,
high-ranking individuals may initiate aggressive interactions more often compared to the low-
ranking individuals; some cheetahs intended to interact with people whereas others intended
to avoid them. As the result, in our study different individuals participated in different sets of
situations. In nature, equal time of observations for focal animals also did not help to balance
call sets from different individuals by behavioural contexts [53].
Sex and individuality in meows
We found a strong influence of sex on cheetah meow vocalizations. Sexual differences were
well-expressed and mainly were determined by the values of fundamental frequency (lower in
males than in females), whereas the values of all other vocal variables were indistinguishable
between sexes. The found differences in fundamental frequency of about 20% (Table 5) are
comparable with 15% body mass differences between males and females in captive cheetahs [8]
and with 22% body mass differences between males and females in wild cheetahs [14].
However, the values of call fundamental frequency depend primarily on the length of vocal
folds in the larynx [71,72], which are related in most mammals with linear body dimensions.
In the cheetah, differences in linear body dimensions between males and females in the skull
length, the foreleg length and the hind leg length range between 4.2 to 7.3% depending on the
measure [14]. Therefore, we may expect that differences in size of the larynx between male and
female cheetahs exceed the overall differences of linear body size between sexes. Earlier, sex-
specific dimorphism in size of the larynx exceeding body size between males and females was
reported for humans [73], Mongolian gazelles (Procapra gutturosa)[47], and goitred gazelles
(Gazella subgutturosa)[48]. In humans, this dimorphism results from sexual selection for the
lower-pitched male voices as a component of apparent body size exaggeration for attracting
females and competing with males [74,75]. Nevertheless, in the cheetah, voice pitch differences
reliably reflect the differences in body size between sexes, as in most mammalian and bird spe-
cies with sex dimorphism of body size [76,77], but see [51,52].
In cheetah meows, sex differences were higher whereas individual differences were compa-
rable to other mammals, in which DFAs to identity and sex were applied to the same samples
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 13 / 20
of calls and animals. Sex discrimination higher than random was reported for alarm calls of
adult yellow-bellied marmots (Marmota flaviventris)[78], for contact calls of young goitred
gazelles [79] and for alarm calls of giant otters (Pteronura brasiliensis)[80]. At the same time,
in alarm calls of speckled ground squirrels (Spermophilus suslicus)[78], yellow ground squirrels
(S.fulvous)[78] and chinchillas (Chinchilla lanigera)[81] and in barks of two borzoi breeds of
the domestic dog [35], discrimination to sex was found on the level expected by chance alone,
whereas individual differences were comparable to those in cheetah meows in our study.
Although vocal individuality was well-expressed in cheetah meows, these individualistic fea-
tures were unstable over years. Indeed, all studied species of mammals display poor stability of
individual vocal traits with time: speckled and yellow ground squirrels [56,60,82], domestic
dogs [35], red deer [58,83], fallow deer (Dama dama)[57] and common marmosets (Callithrix
jacchus)[84]. In contrast, stable individual and pair duet vocal signatures were found for the
periods up to five years in some birds: red-breasted geese (Branta ruficollis)[85], red-crowned
cranes (Grus japonensis)[86] and crested auklets (Aethia cristatella)[87]. It seems that bird
calls retain better the individualistic traits compared to calls of mammals, however further
study with more species and call types is necessary to confirm this.
Vocal repertoires of adult and cub cheetahs
The vocal repertoire of the cheetah is primarily stated at birth, as all seven call types (chirr,
growl, meow, chirp, howl, purr and hiss) described in adults in this study and in [19] were pre-
viously described in 15 cheetah cubs aged from 2 days to 3 months [18]. However, whereas the
entire set of call types is already presented in cubs, the acoustic variables changed strongly with
age [18,19]. In vocal repertoire of 1.53 months-old cubs [18], the average duration of the chirr
was 0.42 s (n = 19 chirrs), that is much shorter than 0.74 s in adults in this study (Table 1); the
average duration of the growl was 0.93 s (n = 24 growls), that is much shorter than 2.34 s in
adults in this study (Table 1), and the average duration of the howl was 0.70 s (n = 6), that is
much shorter than 2.11 s in adults in this study (Table 1). At the same time, the average dura-
tion of cub meow was 0.56 s (n = 38 meows), that is substantially longer than 0.32 s in adults in
this study (Table 1). Consistently, the average duration of cub chirp was 0.32 s (n = 139 calls),
that is much longer than 0.11 s in adults in this study (Table 1).
The fundamental frequency variables also differed between the 1.53 months-old cheetah
cubs [18] and adults in this study. The average maximum fundamental frequency of cub
meows was 3.89 kHz (n = 38 meows) that is much higher than 0.94 kHz in adults in this study
(Table 1). The average maximum fundamental frequency of cub chirps was 5.85 kHz (n = 142
chirps) that is much higher than 1.81 kHz in adults in this study (Table 1). The average maxi-
mum fundamental frequency of cub howls was 2.58 kHz (n = 6) that is much higher than 0.31
kHz in adults in this study (Table 1).
For the cheetah, the substantially higher values of maximum fundamental frequency in cubs
than in adults indicate the descending ontogeny of fundamental frequency with age that is
usual for mammals [76,88], some birds [8991] and reptiles [92]. The distinctive ontogenetic
pathways with same-frequency or even lower-frequency calls in the young than in adults were
reported for the Siberian red deer (Cervus elaphus sibiricus)[52], four species of ground squir-
rels [88,9395] and two species of shrews [9698].
All studies of cheetah vocalizations including the present study have been conducted in captiv-
ity. Captive conditions hardly affect the acoustic variables given that cheetah vocal repertoire is
stated at birth. However, the captive conditions might affect somehow the contextual use of
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 14 / 20
vocalization by these animals. Captive conditions include some behavioural contexts that do
not occur in nature (e.g, the contexts involving animal-human communication). At the same
time, some natural contexts may lack in captivity. Further research of vocal behaviour of free-
ranging cheetahs should reveal the natural use of each call type in different ages and sexes of
Supporting Information
S1 Audio. Calls of adult cheetahs. Purr, hiss, growl, chirr, meow, chirp, howl.
S1 Table. Acoustic measurements of cheetah seven call types for describing the acoustics of
call types.
S2 Table. Acoustic measurements of cheetah meows for estimating the effects of sex and
individuality on the acoustic variables of meows and for estimating the stability of vocal
individuality in meows.
We thank the staff of Moscow Zoo, Jaroslavl Zoo, Novosibirsk Zoo and Volokolamsk Zoo
Breeding Station and personally T. Nemtsova and I. Egorov, for their help and support. We
thank the President of Eurasian Regional Association of Zoos and Aquariums (EARAZA), V.
Spitsin, for his administrative help in arrangement of this research in zoos of Russia. We thank
Livio Favaro and the anonymous reviewer for their valuable comments to the manuscript.
Author Contributions
Conceived and designed the experiments: IAV EVV. Performed the experiments: DSS TSD.
Analyzed the data: DSS IAV EVV. Contributed reagents/materials/analysis tools: DSS TSD
IAV EVV. Wrote the paper: DSS IAV EVV.
1. Adamson J. The spotted sphinx. London: Collins and Havill Press; 1969. 313 p.
2. Adamson J. Pippa's challenge. London: Collins and Havill Press; 1972. 175 p.
3. Durant SM, Kelly M, Caro TM. Factors affecting life and death in Serengeti cheetahs: environment,
age, and sociality. Behav Ecol. 2004; 15: 1122. doi: 10.1093/beheco/arg098
4. Mills MGL, Mills MEJ. Cheetah cub survival revisited: a re-evaluation of the role of predation, especially
by lions, and implications for conservation. J Zool. 2013; 292: 136141. doi: 10.1111/jzo.12087
5. Scantlebury DM, Mills MGL, Wilson RP, Wilson JW, Mills MEJ, Durant SM, et al. Flexible energetics of
cheetah hunting strategies provide resistance against kleptoparasitism. Science. 2014; 346: 7981.
doi: 10.1126/science.1256424 PMID: 25278609
6. Swanson A, Caro T, Davies-Mostert H, Mills MG, Macdonald DW, Borner M, et al. Cheetahs and wild
dogs show contrasting patterns of suppression by lions. J Anim Ecol. 2014; 83: 14181427. doi: 10.
1111/1365-2656.12231 PMID: 24724917
7. Welch RJ, Bissett C, Perry TW, Parker DM. Somewhere to hide: Home range and habitat selection of
cheetahs in an arid, enclosed system. J Arid Environm. 2015; 114: 9199. doi: 10.1016/j.jaridenv.
8. Wildt DE, Brown JL, Bush M, Barone MA, Cooper KA, Grisham J, et al. Reproductive status of cheetahs
(Acinonyx jubatus) in North American zoos: The benefits of physiological surveys for strategic planning.
Zoo Biol. 1993; 12: 4580. doi: 10.1002/zoo.1430120107
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 15 / 20
9. Johnson S, Marker L, Mengersen K, Gordon CH, Melzheimer J. Schmidt-Küntzel A, et al. Modeling the
viability of the free-ranging cheetah population in Namibia: an object-oriented Bayesian network
approach. Ecosphere. 2013; 4(7): 90. doi: 10.1890/ES12-00357.1
10. Voigt CC, Thalwitzer S, Melzheimer J, Blanc A- S, Jago M, Wachter B. The conflict between cheetahs
and humans on Namibian farmland elucidated by stable isotope diet analysis. PLoS ONE. 2014; 9(8):
e101917. doi: 10.1371/journal.pone.010 PMID: 25162403
11. Krengel A, Cattori V, Meli ML, Wachter B, Böni J, Bisset LR, et al. Gammaretrovirus-specific antibodies
in free-ranging and captive Namibian cheetahs. Clin Vaccine Immunol. 2015; 22: 611617. doi: 10.
1128/CVI.00705-14 PMID: 25809630
12. Hudson PE, Corr SA, Payne-Davis RC, Clancy SN, Lane E, Wilson AM. Functional anatomy of the
cheetah (Acinonyx jubatus) forelimb. J Anat. 2011; 218: 375385. doi: 10.1111/j.1469-7580.2011.
01344.x PMID: 21332715
13. Hudson PE, Corr SA, Wilson AM. High speed galloping in the cheetah (Acinonyx jubatus) and the rac-
ing greyhound (Canis familiaris): spatio-temporal and kinetic characteristics. J Exp Biol. 2012; 215:
24252434. doi: 10.1242/jeb.066720 PMID: 22723482
14. Marker LL, Dickman AJ. Morphology, physical conditions, and growth of the cheetah (Acinonyx jubatus
jubatus). J Mammal. 2003; 84: 840850. doi: 10.1644/BRB-036
15. Merola M. A reassessment of homozygosity and the case for inbreeding depression in the cheetah,
Acinonyx jubatus: implications for conservation. Conserv Biol. 1994; 8: 961971. doi: 10.1046/j.1523-
16. Castro-Prieto A, Wachter B, Melzheimer J, Thalwitzer S, Hofer H, Sommer S. Immunogenetic variation
and differential pathogen exposure in free-ranging cheetahs across Namibian farmlands. PLoS ONE
2012; 7(11): e49129. doi: 10.1371/journal.pone.0 PMID: 23145096
17. Charruau P, Fernandes C, Orozco-Terwengel P, Peters J, Hunter L, Ziaie H, et al. Phylogeography,
genetic structure and population divergence time of cheetahs in Africa and Asia: evidence for long-term
geographic isolates. Mol Ecol. 2011; 20: 706724. doi: 10.1111/j.1365-294X.2010.04986.x PMID:
18. Volodina EV. Infantile features revealed from sound structure in the cheetah in captivity: keys to ani-
malsself-esteem as being strong or weak. Scientific researches in Zoos. 1998; 10: 143159. (in Rus-
sian). Available:
19. Volodina E.V. Vocal repertoire of cheetah Acinonyx jubatus (Carnivora, Felidae) in captivity: sound
structure and search for means of assessing the state of adult animals. Entomol Rev. 2000; 80(2):
s368s378. Available:
20. Frazer Sissom DE, Rica DA, Peters G. How cats purr? J. Zool. 1991; 223: 6778. doi: 10.1111/j.1469-
21. Weissengruber GE, Forstenpointner G, Peters G, Kübber-Heiss A, Fitch WT. Hyoid apparatus and
pharynx in the lion (Panthera leo), jaguar (Panthera onca), tiger (Panthera tigris), cheetah (Acinonyx
jubatus) and domestic cat (Felis silvestris f. catus). J Anat. 2002; 201: 195209. doi: 10.1046/j.1469-
7580.2002.00088.x PMID: 12363272
22. Peters G, Tonkin-Leuhausen BA. Evolution of acoustic communication signals of mammals: friendly
close-range vocalizations in Felidae (Carnivora). J Mammal Evol. 1999; 6: 129157. doi: 10.1023/
23. Peters G. Purring and similar vocalizations in mammals. Mammal Review. 2002; 32: 245271. doi: 10.
24. Eklund R, Peters G, Duthie ED. An acoustic analysis of purring in the cheetah (Acinonyx jubatus) and
in the domestic cat (Felis catus). In: Schötz S, Ambrazaitisg, editors. Proceedings from Fonetik 2010.
Lund: Lund University; 2010. pp. 1722. Available:
25. Eklund R, Peters G, Weise F, Munro S. An acoustic analysis of agonistic sounds in wild cheetahs. In:
Eriksson A, Abelin A, editors. Proceedings Fonetik 2012. Gothenburg: University of Gothenburg,
2012. pp. 3740. Available:
26. Volodina EV. Cheetahs vocalizations in a frustration-provoking situation of movementsrestraint in cap-
tivity. Scientific researches in Zoos. 1994; 4: 7077. (in Russian). Available: http://www.bioacoustica.
27. Volodina EV. The apply of bioacoustical methods in management and breeding of the cheetah in cap-
tivity. Scientific researches in Zoos. 1994; 4: 92100. (in Russian). Available: http://www.bioacoustica.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 16 / 20
28. Volodina EV, Volodin IA. Vocalizations associated with reproductive behaviour in rare Felidae species.
Scientific researches in Zoos. 1996; 6: 142184. (in Russian). Available:
29. Schassburger RM. Wolf vocalization: an integrated model of structure, motivation and ontogeny. In:
Frank H, editor. Man and wolf. Dordrecht: Dr W. Junk Publishers; 1987. pp. 313346.
30. Volodin IA, Volodina EV, Isaeva IV. Vocal repertoire in the dhole Cuon alpinus (Carnivora, Canidae) in
captivity. Entomol Rev. 2001; 81(2): S346S361.
31. Yin S. A new perspective on barking in dogs (Canis familiaris). J Comp Psychol. 2002; 116: 189193.
doi: 10.1037/07.35.7036.116.2.189 PMID: 12083615
32. Darden SK, Dabelsteen T, Pedersen SB. A potential tool for swift fox Vulpes velox conservation: individ-
uality of long-range barking sequences. J Mammal. 2003; 84: 14171427. doi: 10.1644/BEM-031
33. Frommolt K- H, Goltsman ME, MacDonald DW. Barking foxes, Alopex lagopus: field experiments in
individual recognition in a territorial mammal. Anim Behav. 2003; 65: 509518. doi: 10.1006/anbe.
34. Yin S, McCowan B. Barking in domestic dogs: context specificity and individual identification. Anim
Behav. 2004; 68: 343355. doi: 10.1016/j.anbehav.2003.07.016
35. Chulkina MM, Volodin IA, Volodina EV. Individual, intersexual and interbreed variability of barks in dog
Canis familiaris (Carnivora, Canidae). Zool Zh. 2006; 85: 544555. (In Russian). Available: http://www.
36. Gogoleva SS, Volodin IA, Volodina EV, Trut LN. To bark or not to bark: Vocalization in red foxes
selected for tameness or aggressiveness toward humans. Bioacoustics. 2008; 18: 99132. doi: 10.
37. Peters G. Vergleichende Untersuchung zur Lautgebund einiger Feliden (Mammalia, Felidae). Zeit.
Zool. Spixiana. 1978; 1: 1283.
38. Nicastro N. Perceptual and acoustic evidence for species-level differences in meow vocalizations by
domestic cats (Felis catus) and African wild cats (Felis silvestris lybica). J Comp Psychol. 2004; 118:
287296. doi: 10.1037/0735-7036.118.3.287 PMID: 15482056
39. Yeon SC, Kim YK, Lee SS, Suh EH, Houpt KA, Lee HC, et al. Differences between vocalization evoked
by behavioural stimulation in feral cats and house cats. Behav Process. 2011; 87: 183189. doi: 10.
40. Nicastro N, Owren MJ. Classification of domestic cat (Felis catus) vocalizations by naïve and experi-
enced human listeners. J Comp Psychol. 2003; 117: 4452. doi: 10.1037/0735-7036.117.1.44 PMID:
41. McComb K, Taylor AM, Wilson C, Charlton BD. The cry embedded within the purr. Current Biol. 2009;
19: 507508. doi: 10.1016/j.cub.2009.05.033
42. Scheumann M, Roser A-E, Konerding W, Bleich E, Hedrich H-J, Zimmermann E. Vocal correlates of
sender-identity and arousal in the isolation calls of domestic kitten (Felis silvestris catus). Frontiers in
Zool. 2012; 9: 36. doi: 10.1186/1742-9994-9-36
43. Ellis SLH, Swindell V, Burman OHP. Human classification of context-related vocalizations emitted by
familiar and unfamiliar domestic cats: an exploratory study. Anthrozoös. 2015; 28: 625634. doi: 10.
44. Ruiz-Miranda CR, Wells SA, Golden R, Seidensticker J. Vocalizations and other behavioral responses
of male cheetahs (Acinonyx jubatus) during experimental separation and reunion trials. Zoo Biol. 1998;
17: 116. doi: 10.1002/(SICI)1098-2361(1998)17:1<1::AID-ZOO1>3.0.CO;2-D
45. Smirnova DS, Demina TS, Volodina EV, Volodin IA. The occurrence of different call types in different
behavioural contexts in captive cheetahs (Acinonyx jubatus). Scientific researches in Zoos. 2015; 31:
104112. (in Russian). Available:
46. Pontier D, Rioux N, Heizmann A. Evidence of selection on the orange allele in the domestic cat Felis
catus: the role of social structure. Oikos. 1995; 73: 299308. doi: 10.2307/3545954
47. Frey R, Riede T. The anatomy of vocal divergence in North American elk and European red deer. J
Morphol. 2013; 274: 307319. doi: 10.1002/jmor.20092 PMID: 23225193
48. Efremova KO, Frey R, Volodin IA, Fritsch G, Soldatova NV, Volodina EV. The postnatal ontogeny of
the sexually dimorphic vocal apparatus in goitred gazelles (Gazella subgutturosa). J. Morphol. 2016;
doi: 10.1002/jmor.20538
49. Fitch WT, Hauser MD. Unpacking honesty: Vertebrate vocal production and the evolution of acoustic
signals. In: Simmons A, Fay RR, Popper AN, editors. Acoustic Communication, Springer Handbook of
Auditory Research. New York: Springer; 2002. pp. 65137.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 17 / 20
50. Efremova KO, Volodin IA, Volodina EV, Frey R, Lapshina EN, Soldatova NV. Developmental changes
of nasal and oral calls in the goitred gazelle Gazella subgutturosa, a nonhuman mammal with a sexually
dimorphic and descended larynx. Naturwissenschaften. 2011; 98: 919931. doi: 10.1007/s00114-011-
0843-7 PMID: 21976026
51. Volodin I, Matrosova V, Volodina E, Garcia AJ, Gallego L, Márquez R, et al. Sex and age-class differ-
ences in calls of Iberian red deer during rut: reversed sex dimorphism of pitch and contrasting roars
from farmed and wild stags. Acta Ethol. 2015; 18: 1929. doi: 10.1007/s10211-013-0179-8
52. Volodin IA, Sibiryakova OV, Volodina EV. Sex and age-class differences in calls of Siberian wapiti Cer-
vus elaphus sibiricus. Mammal Biol. 2016; 81: 1020. doi: 10.1007/s10211-013-0179-8
53. Salmi R, Hammerschmidt K, Doran-Sheehy DM. Western gorilla vocal repertoire and contextual use of
vocalizations. Ethology. 2013; 119: 831847. doi: 10.1111/eth.12122
54. Altmann J. Observational study of behavior: sampling methods. Behaviour. 1974; 49: 227265. doi:
10.1163/156853974X00534 PMID: 4597405
55. R foundation for statistical computing. Available:
56. Matrosova VA, Volodin IA, Volodina EV, Vasilieva NA, Kochetkova AA. Between-year stability of indi-
vidual alarm calls in the yellow ground squirrel Spermophilus fulvus. J Mammal. 2010; 91: 620627.
doi: 10.1644/09-MAMM-A-143.1
57. Briefer E, Vannoni E, McElligott AG. Quality prevails over identity in the sexually selected vocalisations
of an ageing mammal. BMC Biology. 2010; 8: 35. doi: 10.1186/1741-7007-8-35 PMID: 20380690
58. Sibiryakova OV, Volodin IA, Matrosova VA, Volodina EV, Garcia AJ, Gallego L, et al. The power of oral
and nasal calls to discriminate individual mothers and offspring in red deer, Cervus elaphus. Frontiers in
Zool. 2015; 12: 2. doi: 10.1186/s12983-014-0094-5
59. Solow AR. A randomization test for misclassification probability in discriminant analysis. Ecology.
1990; 71: 23792382. doi: 10.2307/1938650
60. Matrosova VA, Volodin IA, Volodina EV, Vasilieva NA. Stability of acoustic individuality in the alarm
calls of wild yellow ground squirrels Spermophilus fulvus and contrasting calls from trapped and free-
ranging callers. Naturwissenschaften. 2010; 97: 707715. doi: 10.1007/s00114-010-0686-7 PMID:
61. Garcia M, Gingras B, Bowling DL, Herbst CT, Boeckle M, Locatelli Y, et al. Structural classification of
wild boar (Sus scrofa) vocalizations. Ethology. 2016; 122: 329342. doi: 10.1111/eth.12472 PMID:
62. Wilden I, Herzel H, Peters G, Tembrock G. Subharmonics, biphonation, and deterministic chaos in
mammal vocalization. Bioacoustics. 1998; 9: 171196. doi: 10.1080/09524622.1998.9753394
63. Eklund R, Peters G, Weise F, Munro S. A comparative acoustic analysis of purring in four cheetahs. In:
Eriksson A, Abelin A, editors. Proceedings Fonetik 2012. Gothenburg: University of Gothenburg,
2012. pp. 4144. Available:
64. Volodina EV. Vocalization as an emotional indicator in the cheetah Acinonyx jubatus in captivity. Scien-
tific researches in Zoos. 1997; 8: 149162. (in Russian). Available:
65. Wielebnowski N, Brown JL. Behavioral correlates of physiological estrus in cheetahs. Zoo Biol. 1998;
17: 193209. doi: 10.1002/(SICI)1098-2361(1998)17:3<193::AID-ZOO4>3.0.CO;24
66. Miklósi A, Kubinyi E, Topal J, Gacsi M, Virányi Z, Csanyi V. A simple reason for a big difference: wolves
do not look back at humans, but dogs do. Current Biol. 2003; 13: 763766. doi: 10.1016/S0960-9822
67. Volodina EV, Volodin IA, Filatova OA. The occurrence of nonlinear vocal phenomena in frustration
whines of the domestic dog (Canis familiaris). Advances in bioacoustics II, Dissertationes SASA, Clas-
sis IV: Historia Naturalis. 2006; 47: 257270. Available:
68. Jürgens U. Vocalization as an emotional indicator: a neuroethological study in the squirrel monkey.
Behaviour. 1979; 69: 88117. doi: 10.1163/156853979X00412 PMID: 112992
69. Fichtel C, Hammerschmidt K, Jürgens U. On the vocal expression of emotion, a multi-parametric analy-
sis of different states of aversion in the squirrel monkey. Behaviour. 2001; 138: 97116. doi: 10.1163/
70. Volodin IA, Volodina EV, Gogoleva SS, Doronina LO. Indicators of emotional arousal in vocal emis-
sions of the humans and nonhuman mammals. J General Biol. 2009; 70: 210224. (in Russian) Avail-
71. Titze IR. Principles of voice production. Englewood Cliffs, NJ: Prentice-Hall; 1994. 354 p.
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 18 / 20
72. Riede T, Brown C. Body size, vocal fold length, and fundamental frequencyimplications for mammal
vocal communication. Nova Acta Leopoldina NF. 2013; 111: 295314.
73. Fitch WT, Giedd J. Morphology and development of the human vocal tract: a study using magnetic res-
onance imaging. J Acoust Soc Am. 1999; 106: 15111522. doi: 10.1121/1.427148 PMID: 10489707
74. Apicella CL, Feinberg DR, Marlowe FW. Voice pitch predicts reproductive success in male hunter-gath-
erers. Biol. Lett. 2007; 3: 682684. doi: 10.1098/rsbl.2007.0410 PMID: 17895219
75. Puts DA, Hodges CR, Cardenas RA, Gaulin SJC. Men's voices as dominance signals: vocal fundamen-
tal and formant frequencies influence dominance attributions among men. Evol. Hum. Behav. 2007;
28: 340344. doi: 10.1016/j.evolhumbehav.2007.05.002
76. Ey E, Pfefferle D, Fischer J. Do age- and sex-related variations reliably reflect body size in non-human
primate vocalizations? A review. Primates. 2007; 48: 253267. doi: 10.1007/s10329-006-0033-y
PMID: 17226064
77. Patel R, Mulder RA, Cardoso GC. What makes vocalisation frequency an unreliable signal of bodysize
in birds? A study on black swans. Ethology. 2010; 116: 554563. doi: 10.1111/j.1439-0310.2010.
78. Matrosova VA, Blumstein DT, Volodin IA, Volodina EV. The potential to encode sex, age and individual
identity in the alarm calls of three species of Marmotinae. Naturwissenschaften. 2011; 98: 181192.
doi: 10.1007/s00114-010-0757-9 PMID: 21221515
79. Volodin IA, Lapshina EN, Volodina EV, Frey R, Soldatova NV. Nasal and oral calls in juvenile goitred
gazelles (Gazella subgutturosa) and their potential to encode sex and identity. Ethology. 2011; 117:
294308. doi: 10.1111/j.1439-0310.2011.01874.x
80. Leuchtenberger C, Sousa-Lima R, Ribas C, Magnusson WE, Mourão G. Giant otter alarm calls as
potential mechanisms for individual discrimination and sexual selection. Bioacoustics. 2016; doi: 10.
81. Moreno-Gomez FN, Leon A, Velasquez NA, Penna M, Delano PH. Individual and sex distinctiveness in
bark calls of domestic chinchillas elicited in a distress context. J Acoust Soc Am. 2015; 138: 1614
1622. doi: 10.1121/1.4929750 PMID: 26428799
82. Matrosova VA, Volodin IA, Volodina EV. Short-term and long-term individuality in speckled ground
squirrel alarm calls. J Mammal. 2009; 90: 158166. doi: 10.1644/08-MAMM-A-032.1
83. Reby D, Andre-Obrecht R, Galinier A, Farinas G, Cargnelutti B. Cepstral coefficients and hidden Mar-
kov models reveal idiosyncratic voice characteristics in red deer (Cervus elaphus) stags. J Acoust Soc
Am. 2006; 120: 40804089. doi: 10.1121/1.2358006 PMID: 17225434
84. Jorgensen DD, French JA. Individuality but not stability in marmoset long calls. Ethology. 1998; 104:
729742. doi: 10.1111/j.1439-0310.1998.tb00107.x
85. Volodin IA, Klenova AV, Volodina EV. Modelling bioacoustical monitoring through years with captive
population of the red-breasted goose. Casarca, Bulletin of the Goose, Swan and Duck Study Group of
Northern Eurasia. 2008; 11: 2246. Available:
86. Klenova AV, Volodin IA, Volodina EV. Examination of pair-duet stability to promote long-term monitor-
ing of the endangered red-crowned crane (Grus japonensis). J Ethol. 2009; 27: 401406. doi: 10.1007/
87. Klenova AV, Zubakin VA, Zubakina EV. Inter- and intra-season stability of vocal individual signatures in
a social seabird, the crested auklet. Acta Ethol. 2012; 15: 4152. doi: 10.1007/s10211-011-0120-y
88. Matrosova VA, Volodin IA, Volodina EV, Babitsky AF. Pups crying bass: vocal adaptation for avoidance
of age-dependent predation risk in ground squirrels? Behav Ecol Sociobiol. 2007; 62: 181191. doi:
89. Klenova AV, Volodin IA, Volodina EV, Postelnykh KA. Voice breaking in adolescent red-crowned
cranes Grus japonensis. Behaviour. 2010; 147: 505524. doi: 10.1163/000579509X12591315521811
90. Klenova AV, Goncharova MV, Bragina EV, Kashentseva TA. Vocal development and voice breaking in
demoiselle cranes (Anthropoides virgo). Bioacoustics. 2014; 23: 247265. doi: 10.1080/09524622.
91. Adret P. Call development in captive-reared Pied Avocets, Recurvirostra avosetta. J Ornithol. 2012;
153: 535546. doi: 10.1007/s10336-011-0774-2
92. Chabert T, Colin A, Aubin T, Shacks V, Bourquin SL, Elsey RM, et al. Size does matter: crocodile moth-
ers react more to the voice of smaller offspring. Sci. Rep. 2015: 5: 15547. doi: 10.1038/srep15547
PMID: 26493940
93. Swan DC, Hare JF. Signaler and receiver ages do not affect responses to Richardsons ground squirrel
alarm calls. J Mammal. 2008; 89: 889894. doi: 10.1644/07-MAMM-A-228.1
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 19 / 20
94. Volodina EV, Matrosova VA, Volodin IA. An unusual effect of maturation on the alarm call fundamental
frequency in two species of ground squirrels. Bioacoustics. 2010; 20: 8798. doi: 10.1080/09524622.
95. Schneiderová I, Schnitzerová P, Uhlíková J, Brandl P, Zouhar J, MatějůJ. Differences in alarm calls of
juvenile and adult European ground squirrels (Spermophilus citellus): Findings on permanently marked
animals from a semi-natural enclosure. Zoo Biol. 2015; 34: 503512. doi: 10.1002/zoo.21233 PMID:
96. Schneiderová I. Vocal repertoire ontogeny of the captive Asian house shrew Suncus murinus suggests
that the male courtship call develops from the caravanning call of the young. Acta Theriol. 2014; 59:
149164. doi: 10.1007/s13364-013-0141-1
97. Volodin IA, Zaytseva AS, Ilchenko OG, Volodina EV. Small mammals ignore common rules: A compari-
son of vocal repertoires and the acoustics between pup and adult piebald shrews Diplomesodon pul-
chellum. Ethology. 2015; 121: 103115. doi: 10.1111/eth.12321
98. Zaytseva AS, Volodin IA, Mason MJ, Frey R, Fritsch G, Ilchenko OG, et al. Vocal development during
postnatal growth and ear morphology in a shrew that generates seismic vibrations, Diplomesodon pul-
chellum. Behav Process. 2015; 118: 130141. doi: 10.1016/j.beproc.2015.06.012
Cheetah Vocal Repertoire
PLOS ONE | DOI:10.1371/journal.pone.0158546 June 30, 2016 20 / 20
... The most studied felid species regarding infant vocalization is the domestic cat (Felis catus) [33,34] and the call parameters of kittens showed developmental changes, and motivational valuation of the rearing situations [35]. Regarding large cats, the vocal repertoire of calls of adult and cub cheetahs (Acinonyx jubatus) has been systematically described [36,37]. Calls of cubs of many other species have also been shown to have developmental dynamics during ontogeny [38][39][40][41][42][43][44]. ...
... According to the preliminary experiment and reference researches [37,51], six behavioral contexts in which calls were produced were preset: (1) Isolation (when a newborn cub without good eyesight was separated from its mother, or when a cub with good eyesight was separated from its mother and littermates; e.g., the cub was crawling or walking around, frequently emitting long distance calls and looking around until it was tired or met its mother, littermates or breeders, or received food from breeders. The cubs opened their eyes at 21 ~ 25 days old.); (2) Offensive-to-Human (when a cub threatened or attacked a human, e.g., by crouching, showing bare teeth, pushing ears back, and rushing at human with raised paw and slapping); (3) Offensive-to-Conspecific (when a cub threatened or attacked a conspecific, e.g., by crouching, showing bare teeth, pushing ears back, rushing at a conspecific with raising paw and slapping); (4) Conspecific-Contact (friendly greeting with conspecific, e.g., by using head or nose to touch conspecific's face); (5) Human-Contact (friendly greeting using head or nose through the net in front of a human, while walking to and fro); and (6) Play (friendly playing with conspecifics, e.g., chasing, pouncing, gently slapping and biting other cubs or being slapped and bit by other cubs). ...
... Following previous research [37,51], for the analysis of the relationships between call type and behavioral context, a call type was classified as context-specific if it was produced in the same behavioral context in more than 65% of the cases. Because multiple calls can cooccur in the same behavioral context, a standard of 65% of occurrence was used to determine whether this was the primary call type for this context (signal-specific call). ...
Full-text available
Background The Amur tiger (Panthera tigris altaica) is the largest and one of the most endangered cats in the world. In wild and captive cats, communication is mainly dependent on olfaction. However, vocal communication also plays a key role between mother and cubs during the breeding period. How cubs express their physiological and psychological needs to their mother and companions by using acoustic signals is little known and mainly hindered by the difficult process of data collection. Here, we quantitatively summarized the vocal repertoire and behavioral contexts of captive Amur tiger cubs. The aim of the present work was to investigate the behavioral motivations of cub calls by considering influential factors of age, sex, and rearing experiences. Results The 5335 high-quality calls from 65 tiger cubs were classified into nine call types (Ar-1, Ar-2, Er, eee, Chuff, Growl, Hiss, Haer, and Roar) produced in seven behavioral contexts. Except for Er, eight of the nine call types were context-specific, related to Play (Ar-2, eee, and Roar), Isolation (Ar-1), Offensive Context (Haer, Growl, and Hiss), and a friendly context (Chuff). Conclusions The results suggest that cubs are not quiet, but instead they express rich information by emitting various call types, which are probably crucial for survival in the wild. We herein provide the first detailed spectrogram classification to indicate vocal repertoires of calls and their coding with respect to behavioral contexts in Amur tiger cubs, and we pave the steps for revealing their social communication system, which can be applied for conservation of populations. These insights can help tiger managers or keepers to improve the rearing conditions by understanding the feline cubs’ inner status and needs by monitoring their vocal information expressions and exchanges.
... Animals use acoustic communication to transmit information about several specific situations (e.g., alarm, reproductive and social status). Despite the differences in the sound-generating apparatus of different animals [3,7,8], sound patterns can be handled in a common manner. The similarity among the various sound recognition schemes comes from the fact that a sound source has a very distinctive and characteristic way to distribute its energy over time on its composing frequencies, which constitutes its so-called spectral signature. ...
... The maximum number of k-means iterations for initialization was 50 while both the EM and Baum-Welch algorithms [47] had an upper limit of 25 iterations with a threshold of 0.001 between subsequent iterations. In Algorithm 1, the number of states of the HMMs was selected from the set s ∈ {3, 4, 5, 6} and the number of Gaussian functions from the set g ∈ {2, 4,8,16,32,64,128, 256} using the validation set alone. ...
Full-text available
Cats employ vocalizations for communicating information, thus their sounds can carry a widerange of meanings. Concerning vocalization, an aspect of increasing relevance directly connected withthe welfare of such animals is its emotional interpretation and the recognition of the production context.To this end, this work presents a proof of concept facilitating the automatic analysis of cat vocalizationsbased on signal processing and pattern recognition techniques, aimed at demonstrating if the emissioncontext can be identified by meowing vocalizations, even if recorded in sub-optimal conditions. Werely on a dataset including vocalizations of Maine Coon and European Shorthair breeds emitted in threedifferent contexts: waiting for food, isolation in unfamiliar environment, and brushing. Towards capturing theemission context, we extract two sets of acoustic parameters, i.e., mel-frequency cepstral coefficients andtemporal modulation features. Subsequently, these are modeled using a classification scheme based ona directed acyclic graph dividing the problem space. The experiments we conducted demonstrate thesuperiority of such a scheme over a series of generative and discriminative classification solutions. Theseresults open up new perspectives for deepening our knowledge of acoustic communication betweenhumans and cats and, in general, between humans and animals.
... Lion vocalizations consisted of snarls and growls (shorter distance sounds made during mating or aggression, not long distance calls like roars; Makin et al. 2019), wild dog vocalizations were comprised of "twitter" and "hoo" calls (shorter and medium distance contact calls respectively; Webster et al. 2010) and those of cheetahs were shorter distance sounds related to courting or aggression (Smirnova et al. 2016). To comprise an optimal, non-threatening control composed of familiar, benign heterospecific animal vocalizations (Hettena et al. 2014), we used the vocalizations of three locally abundant species of birds, the African Hoopoe (Upupa africana), Pearl-Spotted Owlet (Glaucidium perlatum), and African Wood Owl (Strix woodfordi), broadcast during diel, crepuscular, and nocturnal hours, respectively. ...
Experiments have begun demonstrating that the fear (antipredator behavioral responses) large carnivores inspire in ungulates can shape ecosystem structure and function. Most such experiments have focused on the impacts of either just one large carnivore, or all as a whole, rather than the different impacts different large carnivores may have in intact multi-predator-prey systems. Experimentally testing the relative fearfulness ungulates demonstrate toward different large carnivores is a necessary first step in addressing these likely differing impacts. We tested the fearfulness ungulates demonstrated to playbacks of lion (Panthera leo), African wild dog (Lycaon pictus), cheetah (Acinonyx jubatus) or non-predator control (bird) vocalizations, in Greater Kruger National Park, South Africa. Ungulates ran most to lions, then wild dogs, and then cheetahs, demonstrating a very clear hierarchy of fear. Those that did not run looked toward the sound more on hearing large carnivores than controls, looking most on hearing lions. Notably, prey species-specific population level kill rates by each predator did not predict the patterns observed. Our results demonstrate that different large carnivores inspire different levels of fear in their ungulate prey, pointing to differing community-level impacts, which we discuss in relation to the ongoing worldwide decline and loss of large carnivores.
... Within-year acoustic stability of identity signals has been found in two species of marmosets (Jones 553 et al., 1993;Jorgensen and French, 1998), but it may be modified over the longer time periods of 554 several years, hindering reliable identification on a longer time scale (Jorgensen and French, 1998). 555 Similarly, calls of individual cheetahs were stable within a single year but changed over years 556 (Smirnova et al., 2016). Short-term acoustic signatures (lasting from days to weeks) were reported in 557 two different ground squirrel species (Matrosova et al., 2009(Matrosova et al., , 2010. ...
Full-text available
Many studies have revealed that animal vocalizations, including those from mammals, are individually distinctive. Therefore, acoustic identification of individuals (AIID) has been repeatedly suggested as a non-invasive and labor efficient alternative to mark-recapture identification methods. We present a pipeline of steps for successful AIID in a given species. By conducting such work, we will also improve our understanding of identity signals in general. Strong and stable acoustic signatures are necessary for successful AIID. We reviewed studies of individual variation in mammalian vocalizations as well as pilot studies using acoustic identification to census mammals and birds. We found the greatest potential for AIID (characterized by strong and stable acoustic signatures) was in Cetacea and Primates (including humans). In species with weaker acoustic signatures, AIID could still be a valuable tool once its limitations are fully acknowledged. A major obstacle for widespread utilization of AIID is the absence of tools integrating all AIID subtasks within a single package. Automation of AIID could be achieved with the use of advanced machine learning techniques inspired by those used in human speaker recognition or tailored to specific challenges of animal AIID. Unfortunately, further progress in this area is currently hindered by the lack of appropriate publicly available datasets. However, we believe that after overcoming the issues outlined above, AIID can quickly become a widespread and valuable tool in field research and conservation of mammals and other animals.
... Individualistic calls serve for maintaining parent-offspring relationship (Klenova et al., 2009;Sèbe et al., 2011;Sibiryakova et al., 2015;Volodin et al., 2019b), for maintaining group cohesion Rendall, 1997, 2001) and for facilitating mate recognition (Klenova et al., 2011;Curé et al., 2016). Reliability of individual vocal signature depends on call type and can change within season (Matrosova et al., 2009), between years (Smirnova et al., 2016;Matrosova et al., 2010;Schneiderová et al., 2017) and along development (Klenova et al., 2009;Lapshina et al. 2012;Favaro et al., 2014). The most powerful factor affecting the developmental changes of acoustic individuality might be the ontogenetic changes of vocal morphology (Lungova et al., 2015;Volodin et al., 2017a;. ...
Acoustic individuality is present in diverse taxa of mammals and birds, becoming especially prominent in those age groups for which discriminating conspecifics by voice is critically important. This study compares, for the first time, the ontogenetic changes of acoustic individuality of ultrasonic and audible calls (USVs and AUDs) across 12 age-classes (from neonates to adults) in captive yellow steppe lemmings Eolagurus luteus. We found that, in this rodent species, the isolation-induced USVs and AUDs are not individually distinct at any age. We discuss that this result is unusual, because discriminating individuals by individualistic vocal traits may be important for such a social species as yellow steppe lemming. We also discuss the potential role of acoustic individuality in studies including rodent models.
... This difference between the exhalatory and inhalatory phases is reminiscent of the koala male advertising calls [66] and the striped possum Dactylopsila trivirgata mating calls [48], produced at both phases of respiration, in spite of the distinctive mechanism of sound production in the koala [46,67]. In contrast, clear pulsation is evident at both exhalatory and inhalatory phases of the purring vocalizations of felids [68,69], probably because of the involvement of active muscle contractions for producing the exhalatory and inhalatory phases of these calls [70]. In male impala, there is a gradual shortening of both exhalations and inhalations from continuous roars via interrupted roars towards pant-roars and further to the pant parts of the pant-roars. ...
Full-text available
Background The rutting vocal display of male impala Aepyceros melampus is unique for its complexity among ruminants. This study investigates bouts of rutting calls produced towards potential mates and rival males by free-ranging male impala in Namibia. In particular, a comparison of male rutting and alarm snorts is conducted, inspired by earlier findings of mate guarding by using alarm snorts in male topi Damaliscus lunatus . Results Rutting male impala produced 4–38 (13.5 ± 6.5) rutting calls per bout. We analyzed 201 bouts, containing in total 2709 rutting calls of five types: continuous roars produced within a single exhalation-inhalation cycle; interrupted roars including few exhalation-inhalation cycles; pant-roars distinctive by a pant-phase with rapidly alternating inhalations and exhalations; usual snorts lacking any roar part; and roar-snorts starting with a short roar part. Bouts mostly started and ended with usual snorts. Continuous roars were the shortest roars. The average duration of the exhalatory phase was longest in the continuous roars and shortest in the pant-roars. The average fundamental frequency (49.7–51.4 Hz) did not differ between roar types. Vocal tract length, calculated by using measurements of the first four vocal tract resonances (formants), ranged within 381–382 mm in all roar types. In the studied male impala, rutting snorts within bouts of rutting calls were longer and had higher values of the upper quartile in the call spectra than alarm snorts produced towards potential danger. Conclusions Additional inhalations during the emission of the interrupted and pant-roars prolong their duration compared to the continuous roars but do not affect the fundamental frequency or the degree of larynx retraction while roaring. Alarm snorts are separated from one another by large intervals, whereas the intervals between rutting snorts within bouts are short. Sometimes, rutting snorts alternate with roars, whereas alarm snorts do not. Therefore, it is not the acoustic structure of individual snorts but the temporal sequence and the occasional association with another call type that defines snorts as either rutting or alarm snorts. The rutting snorts of male impala may function to attract the attention of receptive females and delay their departure from a male’s harem or territory.
... Mammalian calls are offprints of individual vocal apparatus of a caller [10] and therefore by default provide information about caller's individual identity at level higher than by chance [11][12][13][14][15][16][17]. Call variables may provide general information about body size [18,19] and particular information about body mass [20] and body condition [21][22][23][24]. In addition, acoustic traits may reflect sexual dimorphism [14,[25][26][27][28]. ...
Full-text available
Objectives: This is the first study of the sonic and ultrasonic vocalization in a Dipodidae rodent. For the small-sized quadrupedal northern birch mouse Sicista betulina, phylogenetically related to the bipedal jerboas (Dipodidae), we report null results for ultrasonic vocalization and investigate the acoustic cues to individual identity, sex and body size in the discomfort-related high-frequency tonal sonic calls. Results: We used a parallel audio recording in the sonic and ultrasonic ranges during weighting adult northern birch mice before the scheduled hibernation in captivity. The sonic (audible) high-frequency tonal calls (ranging from 6.21 to 9.86 kHz) were presented in all individuals (7 males and 4 females). The ultrasonic calls lacked in the recordings. Two-way nested ANOVA revealed the effects of caller individual identity on all 10 measured acoustic variables and the effects of sex on four out of 10 measured acoustic variables. Discriminant function analyses with 10 acoustic variables included in the analysis showed 85.5% correct assignment of calls to individual and 79.7% correct assignment of calls to sex; both values significantly exceeded the random values (23.1% and 54.3%, respectively) calculated with randomization procedure. Body mass did not differ between sexes and did not correlate significantly with the acoustic variables.
Early deprivation of adult influence is known to have long‐lasting effects on social abilities, notably communication skills, as adults play a key role in guiding and regulating the behavior of youngsters, including acoustic repertoire use in species in which vocal production is not learned. Cheetahs grow up alongside their mother for 18 months, thus maternal influences on the development of social skills are likely to be crucial. Here, we investigated the impact of early maternal deprivation on vocal production and use in 12 wild‐born cheetahs, rescued and subsequently hand‐reared either at an early (less than 2 months) or a later stage of development. We could distinguish 16 sound types, produced mostly singly but sometimes in repeated or multitype sound sequences. The repertoire of these cheetahs did not differ fundamentally from that described in other studies on adult cheetahs, but statistical analyses revealed a concurrent effect of both early experience and sex on repertoire use. More specifically, early‐reared males were characterized by a high proportion of Purr, Meow, and Stutter; early‐reared females Mew, Growl, Hoot, Sneeze, and Hiss; late‐reared males Meow, Mew, Growl, and Howl; and late‐reared females mostly Meow. Our study demonstrates therefore the long‐term effects of maternal deprivation on communication skills in a limited‐vocal learner and its differential effect according to sex, in line with known social differences and potential differential maternal investment. More generally, it emphasizes the critical importance to consider the past history of the subjects (e.g., captive/wild‐born, mother/hand‐reared, early/late‐mother‐deprived, etc.) when studying social behavior, notably acoustic communication.
Full-text available
Non-vocal, or unvoiced, signals surprisingly have received very little attention until recently especially when compared to other acoustic signals. Some sounds made by terrestrial vertebrates are produced not only by the larynx but also by the syrinx. Furthermore, some birds are known to produce several types of non-syrinx sounds. Besides mechanical sounds produced by feathers, bills and/or wings, sounds can be also produced by constriction, anywhere along the pathway from the lungs to the lips or nostrils (in mammals), or to the bill (in birds), resulting in turbulent, aerodynamic sounds. These noises often emulate whispering, snorting or hissing. Even though hissing sounds have been studied in mammals and reptiles, only a few studies have analyzed hissing sounds in birds. Presently, only the hissing of small, nesting passerines as a defense against their respective predators have been studied. We studied hissing in domestic goose. This bird represents a ground nesting non-passerine bird which frequently produces hissing out of the nest in comparison to passerines producing hissing during nesting in holes e.g., parids. Compared to vocally produced alarm calls, almost nothing is known about how non-vocal hissing sounds potentially encode information about a caller's identity. Therefore, we aimed to test whether non-vocal air expirations can encode an individual's identity similar to those sounds generated by the syrinx or the larynx. We analyzed 217 hissing sounds from 22 individual geese. We calculated the Potential for Individual Coding (PIC) comparing the coefficient of variation both within and among individuals. In addition, we conducted a series of 15 a stepwise discriminant function analysis (DFA) models. All 16 acoustic variables showed a higher coefficient of variation among individuals. Twelve DFA models revealed 51.2-54.4% classification result (cross-validated output) and all 15 models showed 60.8-68.2% classification output based on conventional DFA in comparison to a 4.5% success rate when classification by chance. This indicates the stability of the DFA results even when using different combinations of variables. Our findings showed that an individual's identity could be encoded with respect to the energy distribution at the beginning of a signal and the lowest frequencies. Body weight did not influence an individual's sound expression. Recognition of hissing mates in dangerous situations could increase the probability of their surviving via a more efficient anti-predator response.
Full-text available
Percentages of five call categories were compared among social partners in two contexts: courting male-female in a heat and mother-offsprings. Percentages of two of them, chirr and miaowing, that both function as calling calls, did differ significantly among the social partners in both the contexts. The chirr was prevalent in courting males and mothers, whereas the miaowing - in females in a heat and offsprings. It was proposed based on literature data and the calls usage, that the miaowing is related with emotional state unsure if themselves, whereas the chirr is related with self-confident emotional state. In zoo management practice, proportion of miaowing/chirr vocalisations may reveal the prevalence of a certain emotion in a cheetah throughout longitudinal social interactions. In rapidly changing behavior sequences dynamic vocalisation characteristics (variation of duration, intercalls intervals, transition frequencies) may be more important as emotional indicators.
Full-text available
Many mammalian acoustic displays are generated by vibrations of the vocal folds. The cardinal acoustic index of-cal parameters that determine fundamental frequency is critical for an accurate analysis of acoustic communication systems. Though the relationship between F0 and body size is well established in some vertebrate taxa (anura for example), it remains ambiguous in mammals. Here we review the relationship between F0 and four physiological parameters that govern it (subglottal pressure, vocal fold length, tension, and volume), and we address how variations within these parameters produce unusual results regarding the relationship between F0 and body size.
Full-text available
In this study we assessed the extent of psychological attachment between male cheetahs living in same sex pairs in zoos by recording the behavior and vocal-izations of two male coalitions (siblings and nonsiblings) during four experimental separations and reunions of each coalition. Both coalitions showed higher vocalization rates and walking rates during separations than during reunions, and during separations cheetahs spent less time resting and more time vocaliz-ing, walking, or standing than during baseline observations. Compared to the nonsibling coalition, the sibling coalition showed a significantly higher vocal-ization rate during separations and more affiliative behavior during reunions. The most common calls emitted during separations were chirps, followed by eeaows and stutters. The chirps showed the highest level of individual distinc-tiveness. Eeaows comprised a significantly higher percentage of the calls during separations for nonsiblings that for siblings. The only vocalization heard during reunion was the stutter. We hypothesize that chirps emitted during separations communicate desire to reunite, individual identity, and have a structure that facilitates locating the caller. The results of this study suggest that male cheetahs, both sibling and nonsibling, develop strong psychological attachments to each other. The separation of existing coalitions can create stressful conditions for coalition members. We suggest that raising and maintaining coalitions of male cheetahs in coalitions in zoos is a viable husbandry technique.
Full-text available
This study quantitatively documents the progressive development of sexual dimorphism of the vocal organs along the ontogeny of the goitred gazelle (Gazella subgutturosa). The major, male-specific secondary sexual features, of vocal anatomy in goitred gazelle are an enlarged larynx and a marked laryngeal descent. These features appear to have evolved by sexual selection and may serve as a model for similar events in male humans. Sexual dimorphism of larynx size and larynx position in adult goitred gazelles is more pronounced than in humans, whereas the vocal anatomy of neonate goitred gazelles does not differ between sexes. This study examines the vocal anatomy of 19 (11 male, 8 female) goitred gazelle specimens across three age-classes, that is, neonates, subadults and mature adults. The postnatal ontogenetic development of the vocal organs up to their respective end states takes considerably longer in males than in females. Both sexes share the same features of vocal morphology but differences emerge in the course of ontogeny, ultimately resulting in the pronounced sexual dimorphism of the vocal apparatus in adults. The main differences comprise larynx size, vocal fold length, vocal tract length, and mobility of the larynx. The resilience of the thyrohyoid ligament and the pharynx, including the soft palate, and the length changes during contraction and relaxation of the extrinsic laryngeal muscles play a decisive role in the mobility of the larynx in both sexes but to substantially different degrees in adult females and males. Goitred gazelles are born with an undescended larynx and, therefore, larynx descent has to develop in the course of ontogeny. This might result from a trade-off between natural selection and sexual selection requiring a temporal separation of different laryngeal functions at birth and shortly after from those later in life. J. Morphol., 2016. © 2016 Wiley Periodicals, Inc.
Full-text available
Acoustic variation can convey identity information, facilitate social interactions among individuals and may be useful in identifying sex and group affiliation of senders. Giant otters live in highly cohesive groups with exclusive territories along water bodies defended by the entire group by means of acoustic and chemical signals. Snorts are harsh alarm calls, emitted in threat contexts, which commonly elicit the cohesion and the alert behaviour of the members of the group. The aim of this study was to determine whether giant otter snorts have potential to be used for individual discrimination. We tested this hypothesis by verifying if the acoustic characteristics of snorts vary between two study areas, among social groups and individuals, and between males and females. Snort acoustic variables did not differ significantly among study areas, but varied significantly among groups, individuals and between sexes, with higher discrimination between sexes. The frequency of formants (F1–F5) and formant dispersion (DF) potentially allow identity coding among groups, individuals and sexes. The stronger sex discrimination of snorts may be related to information on body size carried by formant frequencies and dispersion, indicating acoustic sexual dimorphism in giant otters. Acoustic differences among groups and individuals are more likely learned, since we did not find evidence for a genetic signal encoded in the snort variables measured. We conclude that the snorts carry information that could be used for individual or group recognition.
Full-text available
Determining whether a species’ vocal communication system is graded or discrete requires definition of its vocal repertoire. In this context, research on domestic pig (Sus scrofa domesticus) vocalizations, for example, has led to significant advances in our understanding of communicative functions. Despite their close relation to domestic pigs, little is known about wild boar (Sus scrofa) vocalizations. The few existing studies, conducted in the 1970s, relied on visual inspections of spectrograms to quantify acoustic parameters and lacked statistical analysis. Here, we use objective signal processing techniques and advanced statistical approaches to classify 616 calls recorded from semi-free ranging animals. Based on four spectral and temporal acoustic parameters — quartile Q25, duration, spectral flux, and spectral flatness — extracted from a multivariate analysis, we refine and extend the conclusions drawn from previous work and present a statistically validated classification of the wild boar vocal repertoire into four call types: grunts, grunt-squeals, squeals, and trumpets. While the majority of calls could be sorted into these categories using objective criteria, we also found evidence supporting a graded interpretation of some wild boar vocalizations as acoustically continuous, with the extremes representing discrete call types. The use of objective criteria based on modern techniques and statistics in respect to acoustic continuity advances our understanding of vocal variation. Integrating our findings with recent studies on domestic pig vocal behavior and emotions, we emphasize the importance of grunt-squeals for acoustic approaches to animal welfare and underline the need of further research investigating the role of domestication on animal vocal communication.
Full-text available
Vocalization is a broad potential source of information about the internal state of animals. This information may provide a correct estimate of animal behavior and of the most favorable conditions for holding and breeding them in captivity. The sound indicators of animal's welfare may be applied without traumatic manipulations, which is especially important for such a rare and endangered species as the cheetah (Acinonyx jubatus). Eight sound types were distinguished and attributed to three classes - pulsed, tonal, and noisy. The classification proposed is discussed with respect to various mechanisms of producing different sounds. A hypothetical diagram of the correlation between the sound structure and states of confidence/diffidence and aggressiveness/nonaggressiveness in the cheetah is considered.
Full-text available
Dog barks are complex in their structures and may vary from purely tonal to noisy even within individual barking bouts. The possibility to recognize individuality, sex, and breed in Borsoi and Hortaj Windhound breeds based on their acoustical cues was studied. Throughout 2002-2004, barks of 18 Hortaj and 9 Borsoi dogs (kept in a kennel) were recorded in the standard situation, when the same known person approached them. Discriminant analysis of the data on 1329 barks from 11 Hortaj and 9 Borsoi dogs showed 63.5% of their correct assignment to the dogs that exceeded the random value (9.3%). The average value of the correct assignment to sex (358 barks from 12 females and 375 barks from males) was as low as 67.9% (with a random of 58.7%). The value of correct assignment to breed (630 barks from 16 Hortaj and 630 from 8 Borsoi dogs) was only 71.6% (random 54.3%). These results suggest that barks provide for the information concerning individuality of a dog, and to a lesser extent, its sex or breed. The greater breed-dependent than sex-dependent differences in barks arise from greater differences in sizes between the breeds than between sexes. Barks of 3 females and 2 males of Hortaj dogs were tested in order to estimate stability of acoustic features responsible for the recognition of individuality. They were recorded twice with an interval of more than 11 months. The cross-validation analysis of barks recorded during 2003-2004 using discriminant functions of 2002 showed only 38.9% of correct assignment to dog. One can propose that a directional shift in bark characteristics took place due to replacement of a cagemate.