ArticlePDF Available

Iberian Wolf Howls: Acoustic Structure, Individual Variation, and a Comparison with North American Populations


Abstract and Figures

We present a detailed description of the acoustic structure of howls emitted by Iberian wolves and a comparison with published descriptions of North American wolf howls. We recorded and analyzed 176 howls emitted by I I wolves held in captivity in social groups of 1-5 individuals. Our sample included solo howls as well as howls included in choruses. Iberian wolf howls are long (1.1- to 12.8-s) harmonic sounds, with a mean fundamental frequency between 270 and 720 Hz. Our results revealed striking similarities between Iberian and North American wolf howls in all variables analyzed except for the number of discontinuities in the frequency of the howl, which was lower for Iberian wolves. Using discriminant function analysis we could assign 84.7% of howls to the correct individual. Variables related to fundamental frequency (mean and maximum) and the coefficient of fundamental modulation best discriminate individuals. We suggest that Iberian wolves could use howls for individual recognition.
Content may be subject to copyright.
Instituto Cavanilles de Biodiversidad y Biologı´a Evolutiva. Universidad de Valencia, Apdo 22085,
46071—Valencia, Spain (VP, EF)
Fonoteca Zoolo´gica, Departamento de Biodiversidad y Biologı´a Evolutiva, Museo Nacional de Ciencias Naturales,
C/Jose´ Gutie´rrez Abascal, no. 2, 28006—Madrid, Spain (RM)
We present a detailed description of the acoustic structure of howls emitted by Iberian wolves and a comparison
with published descriptions of North American wolf howls. We recorded and analyzed 176 howls emitted by 11
wolves held in captivity in social groups of 1–5 individuals. Our sample included solo howls as well as howls
included in choruses. Iberian wolf howls are long (1.1- to 12.8-s) harmonic sounds, with a mean fundamental
frequency between 270 and 720 Hz. Our results revealed striking similarities between Iberian and North
American wolf howls in all variables analyzed except for the number of discontinuities in the frequency of the
howl, which was lower for Iberian wolves. Using discriminant function analysis we could assign 84.7% of howls
to the correct individual. Variables related to fundamental frequency (mean and maximum) and the coefficient of
fundamental modulation best discriminate individuals. We suggest that Iberian wolves could use howls for
individual recognition.
Key words: Canis lupus, geographic variation, howl, Iberian wolf, individual recognition, vocalizations, wolves
Behavior, like other phenotypic traits, often exhibits geo-
graphic variation within a species (Foster and Endler 1999). In
fact, population comparisons provide some of the best insights
into the causes and mechanisms of adaptive differentiation.
Vocalizations are not an exception. Recent research has revealed
that, far from being invariant, vocalizations often show geo-
graphic variation at macrogeographic or microgeographic scales
(Mundinger 1982). Vocal geographic variation has been
documented for American pikas (Ochotona princeps—Conner
1982), Gunnison’s prairie dogs (Cynomys gunnisoni—Perla and
Slobodchikoff 2002), leopard seals, (Hydrurga leptonyx
Thomas and Golladay 1995), bottlenose dolphins (Tursiops
truncatus—Wang et al. 1995), sperm whales (Physeter
catodon—Weilgart and Whitehead 1997), and Barbary macaques
(Macaca sylvanus—Fischer et al. 1998). Geographic variation in
vocalizations can be based on genetic differences, environmental
differences, or vocal learning (Janik and Slater 1997).
The wolf (Canis lupus) is a wide-ranging social carnivore
with a complex spatial organization for which acoustic
communication plays an important role (Harrington and Asa
2003; Mech 1970). Wolves are found throughout the northern
hemisphere, inhabiting a great variety of habitats. Eurasian and
North American wolves have been isolated for 10,000 years,
since the closing of the Bering land bridge, and wolf
populations show evidence of genetic differentiation on
regional and continental scales (Wayne and Vila´ 2003). Thus,
it is conceivable that the acoustic structure of wolf vocal-
izations shows geographic variation. However, to our knowl-
edge, no attempt to look for variation in wolf vocalizations
among different populations has been made. Most studies of
wolf vocalizations have been carried out with North American
wolves (Coscia 1995; Harrington 1989; Harrington and Mech
1983; Holt 1998; Theberge and Falls 1967; Tooze et al. 1990).
Schassburger (1993) described the vocal repertoire of Eurasian
and North American timber wolves kept in captivity, but the
bulk of the data in his study belonged to North American
wolves and he did not look for geographic differences.
The largest population of wolves in western Europe is found
in the Iberian Peninsula (Boitani 2003). This population has
been isolated from the rest of European wolves for more than
a century (Boitani 2003; Vila´ 1993). Based on morphological
characteristics Iberian wolves may represent a subspecies (i.e.,
Canis lupus signatus) distinct from other European wolves
(Vila´ 1993). There have been only 2 surveys dealing with
* Correspondent:
Ó2007 American Society of Mammalogists
Journal of Mammalogy, 88(3):606–613, 2007
European wolf vocalizations. Kappe (1997) studied the threat
vocalizations emitted by European wolves when competing
over a food item and Feddersen-Petersen (2000) compared the
ontogeny of acoustic communication in European wolves and
in various dog breeds. However, the acoustic structure of
Iberian wolf vocalizations is completely unknown.
Howls allow wolves to communicate over distances up to
several kilometers (Harrington and Asa 2003). Howls have
been described as long harmonic sounds with a fundamental
frequency from 150 Hz to more than 1,000 Hz for adults
(Harrington and Asa 2003). Two types of vocalizations
involving howls have been reported: solo (lone) and chorus
howls. Solo howls are emitted by a single individual (alone or
with other pack members that do not howl). Chorus howls have
been described as a vocalization in which one wolf begins
howling, with other members joining in until several or all
members of a pack are howling together (Joslin 1967). Usually,
chorus howls include not only howls but also other vocal-
izations such as growls, barks, squeaks, and howl variations
such as ‘‘woa-woa howls’’ (Holt 1998).
Recognizing individuals could be advantageous for social
mammals and some long-distance vocalizations do contain
information about individual identity (e.g., African lions
[Panthera leo—McComb et al. 1993], spotted hyenas
[Crocuta crocuta—Holekamp et al. 1999], African bush
elephants [Loxodonta africana—McComb et al. 2000], yellow
baboons [Papio cynocephalus—Fischer et al. 2002], and arctic
foxes [Vulpes lagopus—Frommolt et al. 2003]). The role of
howling in individual recognition in wolves has been explored
in some detail (Theberge and Falls 1967; Tooze et al. 1990).
Tooze et al. (1990) identified vocal signatures in the solo howls
of 7 North American wolves. With respect to chorus howling, it
has been suggested that the initial howls of choruses may
provide signature information about individual or pack identity
(Harrington 1989).
In this study, we analyzed 176 howls from 11 Iberian wolves
held in captivity. Howls were emitted by a single wolf (solo
howls) or by 2 or more wolves howling successively or
simultaneously (chorus howls). We present a detailed de-
scription of the acoustic structure of howls emitted by wolves
belonging to this population and investigate whether howls
provide information regarding the individual identity of the
emitter. Furthermore, we compare our results with those
obtained by Tooze et al. (1990) for 7 timber wolves from North
America to assess whether acoustic structure of howls shows
differences between these 2 populations.
Howls were recorded from 2001 to 2003 from wolves held
in captivity at 3 locations in the Iberian Peninsula: Centro de
Recuperac¸a˜o do Lobo Ibe´rico (CRLI, Malveira, Portugal), La
Dehesa (Riopar, Albacete, Spain), and Fauna Ibe´ rica (El
Rebollar, Valencia, Spain). Ninety-one percent of the howls
included in the analysis were evoked by human imitations of
wolf howling, whereas the remaining 9% were howls that were
emitted spontaneously. Recordings were made during 2
seasons: from January to March (corresponding to the mating
season of wolves in the Iberian Peninsula—M. Barrientos, pers.
comm.) and from September to November. There were
typically 2 recording sessions per day during times when the
wolves howl regularly and are visually identifiable: 0600–1000
h and 1800–2100 h. The wolves were habituated to humans.
Recordings were made 5–40 m from the wolves, with the
researcher often in full sight of the animals. We analyzed howls
of 11 wolves held in captivity in 8 different enclosures, each
with 1–5 individuals (Table 1). We analyzed solo howls and
howls included in choruses (Table 1). Two types of choruses
were recorded: choruses that included only howls (n¼20); and
choruses that included, in addition to howls, other vocalizations
such as growls, barks, squeaks, and woa-woa howls (n¼22).
Audio recordings were made on TDK SA-60 cassette tapes
(TDK Electronics Corp., New York, New York) using
a Sennheiser MK 66 unidirectional microphone with K-6
power unit (Sennheiser Electric GmbH & Co. kG, Wedemark,
Germany) connected to a Marantz PMD 222 cassette recorder
(Marantz America, Inc., Mahwah, New Jersey). Recordings
were digitized with 44.1-kHz sampling frequency and 16 bits
TABLE 1.—Characteristics of Iberian wolves and number of howls (n) analyzed at 3 wolf recovery centers.
area (m
) Social group Wolf Sex
nSeason (A/B)
CRLI 310 Adult male and adult female C-1 $28 7 15 3/12 0/15 1/14
C-2 #42 7 19 1/18 0/19 0/19
1,830 Adult male and adult female C-3 $30 11 15 1/14 1/14 0/15
C-4 #39 11 16 9/7 0/16 0/16
1,753 Adult male and adult female C-5 $28 11 26 6/20 16/10 14/12
8,387 Four adult males and 1 adult female C-6 $27 7 14 0/14 0/14 0/14
C-7 #43 7 15 0/15 0/15 0/15
La Dehesa 500 Adult male, adult female, and subadult male D-1 #42 4 25 9/16 19/6 1/24
1,500 Two adult males D-2 #36 2 14 4/10 14/0 0/14
1,000 Adult male and 2 adult females D-3 $? 3 11 11/0 0/11 0/11
Fauna Ibe´rica 600 Adult female F-1 $23 2 6 0/6 6/0 0/6
Sample size (n) 11 176 44/132 56/120 16/160
A/B: Autumn/breeding.
S/Ch: Solo/chorus.
S/I: Spontaneous/induced.
in the Fonoteca Zoolo´gica, Museo Nacional de Ciencias
Naturales (CSIC, Madrid, Spain), using Delta 66 (Irwindale,
California) or Digi 001 (Bucks, United Kingdom) digitizer
cards connected to Apple Macintosh G4 computers (Cupertino,
California). Recordings were saved in ‘‘.wav’’ format in CD-
ROM. Recordings were subsequently deposited in the animal
sounds collection of the Fonoteca Zoolo´ gica. Tape recordings,
once digitized, were analyzed using commercially available
software (Spectrogram 7.2, 2002; Visualization Software LLC,
Stafford, Virginia). We generated audiospectrograms conduct-
ing a fast Fourier transform (2,048-point fast Fourier transform;
Hanning window; time step: 10 ms; frequency range: 9,000 Hz;
frequency resolution: 21.5 Hz). We used the cursor to measure
the fundamental frequency and amplitude at intervals of 0.05 s
along the entire length of the howl. For each howl we measured
16 variables (Appendix I), 13 of which have been used in
previous works (Coscia 1995; Tooze et al. 1990). All
procedures complied with guidelines of the American Society
of Mammalogists (Animal Care and Use Committee 1998).
Statistical analyses.— For statistical analyses we used SPSS
(12.0) for Windows (SPSS Inc., Chicago, Illinois) and the R
statistical package (Dalgaard 2002). We used discriminant
function analysis to classify 176 howls from 11 known
individuals. Discriminant function analysis identifies a linear
combination of independent variables that best discriminates
groups from each other. The assumptions of multivariate
normality and equal covariance matrices were not met even
with transformed variables, but discriminant analysis is robust
to violations of these 2 assumptions (Klecka 1980; Selvin
1995). When the assumptions of multivariate normality and
equal covariance are not met, it is advisable to use the leave-
one-out cross-validation results (Huberty 1994). In this method,
each observation is systematically dropped, the discriminant
function is reestimated, and then the excluded observation is
classified (Huberty 1994). Our data set included cases of
temporally close howls, thus violating the independence
assumption. We grouped howls recorded in the same session
and conducted a 1-way analysis of variance for each individual
using session as the independent variable and the acoustic
variables as dependent variables. We found differences in only
1 variable (frequency modulation) of howls from 1 individual
recorded in different sessions. Therefore, we assume that such
a small amount of temporal autocorrelation should not affect
the overall results.
We compared our results with those reported by Tooze and
colleagues (Tooze 1987; Tooze et al. 1990) for a sample of
308 howls recorded from 7 wolves. Because there are errors in
the figures for duration reported in Tooze et al. (1990: table 2;
F. Harrington, pers. comm.), we used the original values
reported in Tooze (1987: table 1.8). Because of nonnormality,
the presence of outliers, and the limited sample size, we used
the Yuen–Welch test for equality of trimmed (a¼0.2) means
to compare 11 variables recorded in both studies (Yuen 1974).
We used sequential correction to account for the number of
pairwise comparisons made (Rice 1989).
Iberian wolf howls were long-duration (1.1- to 12.8-s),
harmonic sounds (1–18 harmonics), with a mean fundamental
frequency between 270 and 720 Hz (Appendix II). Fundamen-
tal frequencies in howls ranged from 92 to 1,116 Hz. The
coefficient of frequency modulation ranged from 0.21 to 6.72,
and the range of the coefficient of frequency variation was
between 2.03 and 44.63. Iberian wolf howls usually had
inflexion points (1–15) and discontinuities (1–8) in the
fundamental frequency. The maximum fundamental frequency
occurred in most cases (79% of the howls analyzed) during the
1st quarter of the howl, and the minimum during the last
quarter (78%). The fundamental peak amplitude occurred in the
1st half of the howl (83%).
The 2 acoustic characteristics that best distinguish each howl
are the presence of frequency discontinuities and frequency
modulations. Thus, howls were arbitrarily assigned to 1 of 4
groups based on these 2 attributes (Fig. 1). The 1st group
consisted of flat howls, which were relatively constant-
frequency howls, without frequency discontinuities and with
low values of both frequency modulation and variation (Table
2). The shape of the audiospectrogram was flat, not wavy. The
2nd group consisted of continuous wavy howls, which were
FIG.1.—Types of howls recorded from wolves held in captivity at 3 locations in the Iberian Peninsula from 2001 to 2003: A) flat, B) breaking,
C) continuous wavy, and D) breaking wavy.
howls without frequency discontinuities and with frequency
modulations (i.e., wavy audiospectrograms). The 3rd group
consisted of breaking howls, which were howls with large (21-
to 250-Hz) frequency discontinuities, and low values of the
coefficient of frequency modulation (Table 2). The audiospec-
trogram was not wavy. The 4th group consisted of breaking
wavy howls, which were howls with large (21- to 250-Hz)
frequency discontinuities and numerous frequency modulations
(Table 2). The audiospectrogram was wavy.
Breaking wavy howls had the greatest coefficients of
frequency modulation and variation, and flat howls had the
lowest (Table 2). Howls with frequency discontinuities had the
longest duration. In most cases, both solo howls and howls
included in choruses had discontinuities, with breaking howls
being the most frequent type of howl (Table 3). Solo howls of
Iberian wolves were significantly shorter than howls included
in a chorus (t¼5.734, d.f. ¼174, P,0.001).
The discriminant function analysis identified mean funda-
mental frequency, maximum frequency of the fundamental,
number of harmonics, and frequency modulation as the most
important discriminating variables. Using discriminant function
analysis with independent variables entered simultaneously, we
could assign 84.7% of howls to the correct individual. The
cross-validation procedure resulted in 72.7% of howls correctly
assigned, a percentage much higher than expected by chance
(10.15%). Using only the howls included in chorus howling,
81.7% of howls were assigned to the correct individual, and the
leave-one-out cross-validation resulted in 72.5% of howls
correctly assigned.
Comparing our results with those reported by Tooze and
colleagues (Tooze 1987; Tooze et al. 1990) we found
statistically significant differences only for the variable Abrupt
(Table 4). The Iberian wolf howls analyzed have fewer
frequency discontinuities than the howls recorded from 7
North American timber wolves. Nevertheless, the 2 data sets
are not homogeneous. All the wolves in our study were adults,
whereas in the study of Tooze and colleagues (Tooze 1987;
Tooze et al. 1990) 2 individuals were yearlings. However, an
age effect seems unlikely because the results are similar when
only data from adult wolves are compared (Table 4).
Iberian wolf howls can be classified into 4 types (flat,
continuous wavy, breaking, and breaking wavy howls) based
on the 2 criteria that best define howl shape in the
audiospectrogram: the presence of discontinuities in the
fundamental frequency and the pattern of frequency modula-
tion. Before our study, 2 types of howls had been reported for
North American wolves: flat howls and breaking howls.
Although these 2 types were singled out as representing the
extremes seen in frequency modulation, there is a fair degree of
variation within each type (revised in Harrington and Asa
2003). It is unclear to what extent the 4 howl types identified in
our study for descriptive purposes represent, to the wolves,
natural or functionally distinct vocalizations. Using discrimi-
nant function analysis with howl type as the grouping variable
(results not shown) we could assign 89% of howls to the
correct type, showing that the 4 howl types have a different
acoustic structure. It has been proposed that variation in howls
may be related to general arousal or to the sequence of the howl
TABLE 2.—Shape variables (mean and range [maximum
minimum value]) for each type of howl. Variables are described in
Appendix I.
Variable Statistic
Howl type
Flat Breaking
X6SD 5.7 62.5 18.3 66.9 19.5 67.1 23.4 68.1
Range 8.0 32.7 32.5 26.5
X6SD 125 657 279 688 318 6137 437 6167
Range 187 425 550 572
X6SD 0.7 60.5 1.3 60.6 1.4 61.5 1.9 60.8
Range 1.8 3.7 6.2 3.3
X6SD 0.0 60.0 2.1 61.1 0.0 60.0 3.3 61.9
Range 0.0 7.0 0.0 7.0
X6SD 1.1 61.8 2.2 62.0 1.6 61.7 5.9 63.0
Range 6.0 9.0 5.0 14.0
X6SD 5.5 62.3 6.9 62.3 5.3 61.8 6.4 62.6
Range 8.4 9.7 6.5 10.0
TABLE 3.—Frequency of the different types of howls identified in this study. See Table 1 for characteristics of individual wolves.
Solo Chorus
Flat Breaking Continuous wavy Breaking wavy Total Flat Breaking Continuous wavy Breaking wavy Total
C-1 0 0 0 0 0 1 13 0 1 15
C-2 0 0 0 0 0 1 13 2 3 19
C-3 0 1 0 0 1 0 4 10 0 14
C-4 0 0 0 0 0 2 7 2 5 16
C-5113 1 11609 1 010
C-6 0 0 0 0 0 1 10 1 2 14
C-7 0 0 0 0 0 0 13 0 2 15
D-1 1 17 0 1 19 0 4 1 1 6
D-2 0 1 3 10 14 0 0 0 0 0
D-3 0 0 0 0 0 0 1 1 9 11
F-1 4 2 0 0 6 0 0 0 0 0
No. howls 6 34 4 12 56 5 74 18 23 120
% 10.71 60.71 7.14 21.43 100 4.17 61.67 15.00 19.17 100
in a chorus, among other factors (Harrington 1989; Harrington
and Asa 2003). The fact that we have identified solo howls of
the 4 types suggests that, under certain circumstances, wolves
can emit highly modulated howls not necessarily integrated in
a chorus. It would be interesting to investigate whether the
different howl types reported in this study are functionally
distinct and if so, how they are produced, under what
circumstances they arise, and what information might they
Tooze et al. (1990) found individual differences among the
solo howls of 7 North American timber wolves. Our results
show that solo and chorus howls of 11 Iberian wolves are
individually distinct. The acoustic structure of wolf howls
allowed us to identify individuals, and wolves could use this
information for individual recognition. Our results agree with
those obtained in Tooze et al. (1990) in emphasizing variables
related to fundamental frequency (mean and maximum) and
coefficient of fundamental modulation as the variables that best
discriminate individuals. Frequency characteristics usually
encode individuality because they are mostly determined by
the characteristics of an animal’s vocal apparatus (Fitch 1997).
The fundamental frequency is one of the acoustic features that
best discriminate among individuals in other mammal vocal-
izations, including isolation calls of Amazonian manatees
(Trichechus inunguis—Sousa-Lima et al. 2002), calls of
African bush elephants (L. africana—McComb et al. 2003),
domestic dog barks (Canis familiaris—Yin and McCowan
2004), and the whistle call of dholes (Cuon alpinus—Durbin
1998). Frequency modulation also plays an important role in
individual recognition as reported in chirps of Belding’s
ground squirrels (Spermophilus beldingi—McCowan and
Hooper 2002), whistles of bottlenose dolphins (T. truncatus
Caldwell and Caldwell 1965), and calls of subantarctic fur
seal pups (Arctocephalus tropicalis—Charrier et al. 2002).
Transmission characteristics of the atmosphere impose con-
straints on acoustic communication, and frequency modulation
represents one of the best ways to encode information in long-
range vocal signals (Wiley and Richards 1978). Although
fundamental frequency is highly determined by morphological
characteristics, acoustic features related to the shape of the
spectrum (e.g., frequency modulation) are determined by
details of the opening and closing movement of the vocal
folds (Rubin and Vatikiotis-Bateson 1998). Morphological
characteristics of the vocal apparatus and the development of
an individually specific howling technique could be the basis
for individual recognition by means of howling in wolves, as
it has been suggested for coyotes (Canis latrans—Mitchell
2004). Nevertheless, the fact that acoustic structure of howls is
individually specific does not imply that wolves use this
information for individual recognition. To confirm this hypoth-
esis would require playback experiments (e.g., Frommolt et al.
2003; McComb et al. 2001; Mitchell 2004).
Animal vocalizations commonly vary over the geographic
range of the species. However, our results reveal many
similarities between the acoustic structure of howls of Iberian
and North American wolves (Harrington 1989; Harrington and
Mech 1978; Theberge and Falls 1967; Tooze et al. 1990).
Furthermore, when comparing our results with those obtained
by Tooze and colleagues (Tooze 1987; Tooze et al. 1990), we
only found significant differences in the number of frequency
discontinuities. This difference could be due to Iberian wolves
emitting relatively fewer breaking howls than North American
wolves. However, this explanation seems unlikely considering
that most howls produced by Iberian wolves have frequency
discontinuities (Table 3). Alternatively, Iberian and North
American wolves could be producing a similar proportion of
breaking howls but those of Iberian wolves would have fewer
discontinuities per howl. Further data, including a larger
sample of howls and individuals, will be required to assess the
importance of these interpopulation differences.
Thus, in spite of possible genetic, morphological, or
environmental differences, it seems that Iberian wolf howls
and North American timber wolf howls show few detectable
differences, at least with the variables used in this study. Lack
of geographic differences in vocalizations has been reported for
other large mammals, such as West Indian manatees
(Trichechus manatus—Nowacek et al. 2003), and in the songs
of gibbons (Hylobates—Marshall and Marshall 1976). Genetic
differences are not always correlated with variation in vocal
signals. For example, on a microgeographic scale, Wright et al.
(2005) did not find concordance between vocal dialects and
population genetic structure in the yellow-naped parrot
TABLE 4.—Yuen–Welch test for equality of trimmed (a¼0.2) means to compare acoustic features of North American timber wolves (Tooze
1987; Tooze et al. 1990) and Iberian wolf howls. Statistically significant differences are set in boldface. See Appendix I for descriptions of
Meanf Maxf Minf Range Cofv Cofm Dur Changf Abrupt Posmax Narm
All wolves
Yuen’s test statistic 0.84 0.25 1.68 0.93 2.90 0.49 1.25 0.07 7.93 1.91 2.40
d.f. 7.73 7.96 7.92 9.99 7.67 9.49 6.96 6.72 10.00 9.51 9.70
P0.425 0.809 0.132 0.372 0.021 0.638 0.252 0.944 ,0.001 0.087 0.038
Critical Pvalue (Rice 1989) 0.013 0.025 0.007 0.01 0.005 0.017 0.008 0.05 0.005 0.006 0.006
Adult wolves only
Yuen’s test statistic 0.64 0.15 2.15 1.02 3.77 0.53 1.55 0.67 6.76 1.86 2.64
d.f. 2.65 2.73 3.49 5.29 4.36 7.96 6.69 2.89 6.21 3.74 5.24
P0.571 0.894 0.107 0.351 0.017 0.609 0.167 0.554 ,0.001 0.141 0.044
Critical Pvalue (Rice 1989) 0.017 0.05 0.006 0.01 0.005 0.025 0.008 0.013 0.005 0.007 0.006
(Amazona auropalliata). Geographic variation in vocalizations
can also arise because of environmental differences. Selec-
tion could shape the structure of long-distance acoustic signals
to maximize transmission through different environments
(Blumstein and Turner 2005; Morton 1975). Both North
American timber wolves and Iberian wolves live in mountain-
ous and forested areas. It is possible that the similarities found
in their howls are due to selection for acoustic characteristics
that maximize their transmission in similar habitats. Further
research including samples of vocalizations from wolves living
in different environments is needed to clarify this issue.
We thank F. Petrucci-Fonseca and S. Pinho (CRLI ), J. Escudero
(La Dehesa), and R. Peris (Fauna Ibe´rica) for allowing us to obtain
recordings from the captive wolves under their care. The Fonoteca
Zoolo´ gica of the Museo Nacional de Ciencias Naturales (CSIC) in
Madrid provided support for the sound analyses (projects CGL2005-
0092/BOS, CGL2005-25130-E/BOS, and CGL2004-21489-E, Minis-
terio de Educacio´ n y Ciencia). We thank M. Kramer for help with the
statistical analyses.
ANIMAL CARE AND USE COMMITTEE. 1998. Guidelines for the capture,
handling, and care of mammals as approved by the American
Society of Mammalogists. Journal of Mammalogy 79:1416–1431.
BLUMSTEIN, D. T., AND A. C. TURNER. 2005. Can the acoustic
adaptation hypothesis predict the structure of Australian birdsong?
Acta Ethologica 15:35–44.
BOITANI, L. 2003. Wolf conservation and recovery. Pp. 317–340 in
Wolves: behavior, ecology, and conservation (L. D. Mech and L.
Boitani, eds.). University of Chicago Press, Chicago, Illinois.
CALDWELL, M. C., AND D. K. CALDWELL. 1965. Individualized whistle
contours in bottlenose dolphins (Tursiops truncatus). Nature
CHARRIER, I., N. MATHEVON,AND P. JOUVENTIN. 2002. How does a fur
seal mother recognize the voice of her pup? An experimental study
of Arctocephalus tropicalis. Journal of Experimental Biology
CONNER, D. A. 1982. Geographic variation in short calls of pikas
(Ochotona princeps). Journal of Mammalogy 63:48–52.
COSCIA, E. M. 1995. Ontogeny of timber wolf vocalizations: acoustic
properties and behavioural contexts. Ph.D. dissertation, Dalhousie
University, Halifax, Nova Scotia, Canada.
DALGAARD, P. 2002. Introductory statistics with R. Springer-Verlag,
New York.
DURBIN, L. S. 1998. Individuality in the whistle call of the Asiatic wild
dog Cuon alpinus. Bioacoustics 9:197–206.
FEDDERSEN-PETERSEN, D. U. 2000. Vocalization of European wolves
(Canis lupus lupus L.) and various dog breeds (Canis lupus f. fam.).
Archiv fu¨ r Tierzucht. Archives of Animal Breeding 43:387–397.
2002. Acoustic features of male baboon loud calls: influences of
context, age, and individuality. Journal of the Acoustical Society of
America 111:1465–1474.
FISCHER, J., K. HAMMERSCHMIDT,AND D. TODT. 1998. Local variation
in Barbary macaque shrill barks. Animal Behaviour 56:623–629.
FITCH, W. T. 1997. Vocal tract length and formant frequency
dispersion correlate with body size in rhesus macaques. Journal of
the Acoustical Society of America 102:1213–1222.
FOSTER, S. A., AND J. A. ENDLER. 1999. Geographic variation in
behavior. Perspectives on evolutionary mechanisms. Oxford
University Press, Oxford, United Kingdom.
Barking foxes, Alopex lagopus: field experiments in indi-
vidual recognition in a territorial mammal. Animal Behaviour
HARRINGTON, F. H. 1989. Chorus howling by wolves: acoustic
structure, pack size and the Beau Geste effect. Bioacoustics
HARRINGTON, F. H., AND C. S. ASA. 2003. Wolf communication. Pp.
66–103 in Wolves: behavior, ecology, and conservation (L. D.
Mech and L. Boitani, eds.). University of Chicago Press, Chicago,
HARRINGTON, F. H., AND L. D. MECH. 1978. Wolf vocalization. Pp.
109–132 in Wolf and man: evolution in parallel (R. L. Hall and
H. S. Sharp, eds.). Academic Press, New York.
HARRINGTON, F. H., AND L. D. MECH. 1983. Wolf pack spacing:
howling as a territory-independent spacing mechanism in a territorial
population. Behavioral Ecology and Sociobiology 12:161–168.
HOLEKAMP, K. E., ET AL. 1999. Vocal recognition in the spotted hyaena
and its possible implications regarding the evolution of intelligence.
Animal Behaviour 58:383–395.
HOLT, T. D. 1998. A structural description and reclassification of the
wolf, Canis lupus, chorus howl. M.S. thesis, Dalhousie University,
Halifax, Nova Scotia, Canada.
HUBERTY, C. J. 1994. Applied discriminant analysis. John Wiley &
Sons, Inc., New York.
JANIK, V. M., AND P. J. B. SLATER. 1997. Vocal learning in mammals.
Advances in the Study of Behavior 26:59–99.
JOSLIN, P. W. B. 1967. Movements and home sites of timber wolves in
Algonquin Park. American Zoologist 7:279–288.
KAPPE, T. 1997. Subjective resource value and the intensity of threat
vocalizations in European wolves (Canis lupus lupus). Mammalian
Biology 62(Suppl. 2):97–100.
KLECKA, W. R. 1980. Discriminant analysis. Sage University Papers
Series on Quantitative Applications in the Social Sciences. Sage,
Newbury Park, California.
MARSHALL, J. T., JR., AND E. R. MARSHALL. 1976. Gibbons and their
territorial songs. Science 193:235–237.
2001. Matriarchs act as repositories of social knowledge in African
elephants. Science 292:491–494.
extensive networks of vocal recognition in African elephants.
Animal Behaviour 59:1103–1109.
lions can identify potentially infanticidal males from their roars.
Proceedings of the Royal Society of London, B. Biological Sciences
Long-distance communication of acoustic cues to social identity in
African elephants. Animal Behaviour 65:317–329.
MCCOWAN, B., AND S. L. HOOPER. 2002. Individual acoustic variation
in Belding’s ground squirrel alarm chirps in the High Sierra
Nevada. Journal of the Acoustical Society of America 111:1157–
MECH, L. D. 1970. The wolf: the ecology and behaviour of an
endangered species. Natural History Press, Doubleday, New York.
MITCHELL, B. R. 2004. Coyote vocal communication and its
application to the selective management of problem individuals.
Ph.D. dissertation, University of California, Berkeley.
MORTON, E. S. 1975. Ecological sources of selection on avian sounds.
American Naturalist 109:17–34.
MUNDINGER, P. C. 1982. Microgeographic and macrogeographic
variation in the acquired vocalizations of birds. Pp. 147–208 in
Acoustic communication in birds (D. E. Kroodsma and E. H.
Miller, eds.). Vol. 2. Academic Press, New York.
D. A. MANN. 2003. Intraspecific and geographic variation of West
Indian manatee (Trichechus manatus spp.) vocalizations. Journal of
the Acoustical Society of America 114:66–69.
PERLA, B. S., AND C. N. SLOBODCHIKOFF. 2002. Habitat structure and
alarm call dialects in Gunnison’s prairie dog (Cynomys gunnisoni).
Behavioral Ecology 13:844–850.
RICE, W. R. 1989. Analyzing tables of statistical tests. Evolution
RUBIN, P., AND E. VATIKIOTIS-BATESON. 1998. Measuring and modeling
speech production. Pp. 251–290 in Animal Acoustic Communica-
tion (S. L. Hopp, M. J. Owren, and C. S. Evans, eds.). Springer-
Verlag, Berlin Heidelberg, Germany.
SCHASSBURGER, R. M. 1993. Vocal communication in the timber wolf,
Canis lupus, Linnaeus: structure, motivation, and ontogeny.
Advances in Ethology Series 30. Paul Parey Scientific Publishers,
Berlin, Germany.
SELVIN, S. 1995. Practical biostatistical methods. 1st ed. Duxbury
Press, Belmont, California.
Signature information and individual recognition in the isolation
calls of Amazonian manatees, Trichechus inunguis (Mammalia:
Sirenia). Animal Behaviour 63:301–310.
THEBERGE, J. B., AND J. B. FALLS. 1967. Howling as a means of
communication in timber wolves. American Zoologist 7:331–
THOMAS, J. A., AND C. L. GOLLADAY. 1995. Geographic variation in
leopard seal (Hydrurga leptonyx) underwater vocalizations. Pp.
201–221 in Sensory systems of aquatic mammals (R. A. Kastelein,
J. A. Thomas, and P. E. Nachtigall, eds.). De Spil Publishers,
Woerden, The Netherlands.
TOOZE, Z. J. 1987. Some aspects of the structure and func-
tion of long-distance vocalizations of timber wolves (Canis
lupus). M.Sc. thesis, Dalhouise University, Halifax, Nova Scotia,
Individually distinct vocalizations in timber wolves, Canis lupus.
Animal Behaviour 40:723–730.
´, C. 1993. Aspectos morfolo´ gicos y ecolo´ gicos del lobo Ibe´ rico
(Canis lupus L.). Ph.D. dissertation, Universidad de Sevilla, Sevilla,
WANG, D., B. WURSING,AND W. E. EVANS. 1995. Whistles of
bottlenose dolphins: comparisons among populations. Aquatic
Mammals 21:65–77.
´. 2003. Molecular genetic studies of
wolves. Pp. 218–238 in Wolves: behavior, ecology, and conserva-
tion (L. D. Mech and L. Boitani, eds.). University of Chicago Press,
Chicago, Illinois.
WEILGART, L., AND H. WHITEHEAD. 1997. Group-specific dialects and
geographical variation in coda repertoire in South Pacific sperm
whales. Behavioral Ecology and Sociobiology 40:277–285.
WILEY,R.H.,AND D. G. RICHARDS. 1978. Physical constraints on acoustic
communication in the atmosphere: implications for the evolution of
animal vocalizations. Behavioral Ecology and Sociobiology 3:69–94.
dialects, sex-biased dispersal, and microsatellite population struc-
ture in the parrot Amazona auropalliata. Molecular Ecology
YIN, S., AND B. MCCOWAN. 2004. Barking in domestic dogs: context
specificity and individual identification. Animal Behaviour 68:
YUEN, K. K. 1974. The two-sample trimmed tfor unequal population
variances. Biometrika 61:165–170.
Submitted 18 June 2006. Accepted 29 November 2006.
Associate Editor was Floyd W. Weckerly.
Description of structural variables used in analysis of wolf howls and units of measure (in parentheses).
Variable type Abbreviation Structural variables
Frequency Meanf Mean frequency of the fundamental at 0.05 intervals over the duration (Hz)
Maxf Maximum frequency of the fundamental (Hz)
Minf Minimum frequency of the fundamental (Hz)
Range Range of the fundamental: Range ¼Maxf Minf (Hz)
Cofm Coefficient of frequency modulation: Cofm ¼Pn1
Meanf 100
Cofv Coefficient of frequency variation: Cofv ¼ð SD
Changf Number of inflexions of the fundamental
Abrupt Number of discontinuities of the fundamental
Posmax Position in the howl at which the maximum frequency occurs: Posmax ¼time of
Maxf /Dur
Posmin Position in the howl at which the minimum frequency occurs: Posmin ¼time of
Minf /Dur
Endf Frequency at the end of the fundamental (Hz)
Dur Duration of the howl measured at the fundamental (s)
Narm Maximum number of harmonics to 2,000 Hz
Amplitude Frecpaf Fundamental at its amplitude peak (Hz)
Pospaf Position in the howl at which Frecpaf occurs: Pospaf ¼time of Frecpaf/Dur
Coidv Coefficient of amplitude variation at the fundamental frequency (%)
Parameters of Iberian wolf howls analyzed in this study. Range denotes minimum and maximum values for each variable. See Table 1 for characteristics of each wolf and Appendix I for
descriptions of variables.
Wolf Meanf Maxf Minf Range Cofm Cofv Narm Changf Abrupt Dur Posmax Posmin Pospaf
X6SD 373 632 430 641 234 655 196 657 1.1 60.5 11.8 65.3 4.5 60.5 3.5 62.9 2.5 61.9 8.3 62.5 0.48 60.34 0.38 60.48 0.31 60.18
Range 320431 367504 108335 103274 0.52.0 5.220.3 4.05.0 0.010.0 0.06.0 3.811.0 0.000.99 0.001.00 0.060.63
X6SD 474 648 596 660 318 675 278 697 0.8 60.3 15.7 67.1 3.1 60.6 2.3 61.8 2.1 61.9 8.0 61.6 0.04 60.06 0.90 60.17 0.26 60.12
Range 360588 496712 173524 114539 0.41.2 6.238.0 2.05.0 0.06.0 0.08.0 4.610.6 0.000.02 0.421.00 0.100.57
X6SD 466 625 587 623 346 617 241 629 0.8 60.2 18.1 63.1 2.9 60.8 1.3 61.6 0.6 60.9 5.8 61.3 0.05 60.06 0.88 60.15 0.25 60.27
Range 417517 547647 302376 205302 0.51.0 12.922.4 2.05.0 0.05.0 0.02.0 3.07.6 0.000.17 0.511.00 0.070.80
X6SD 357 639 472 675 222 653 250 6100 1.2 60.3 15.4 65.1 4.1 60.9 3.6 62.8 1.6 61.1 8.3 61.9 0.30 60.28 0.64 60.45 0.39 60.22
Range 274405 296570 114319 68456 0.41.7 4.423.0 2.05.0 0.010.0 0.03.0 4.811.7 0.000.92 0.001.00 0.020.80
X6SD 395 647 562 674 240 674 324 670 1.3 60.3 21.4 65.1 4.9 60.4 2.6 62.6 1.6 60.9 7.3 61.8 0.08 60.10 0.96 60.07 0.21 60.24
Range 345533 389672 92349 114456 0.82.1 10.032.8 4.05.0 0.09.0 0.03.0 2.79.5 0.020.52 0.701.00 0.010.86
X6SD 407 634 524 660 294 638 229 678 1.1 60.5 15.1 65.8 2.9 60.4 1.9 62.6 1.9 61.2 6.6 61.7 0.14 60.14 0.75 60.41 0.40 60.28
Range 340454 388615 205324 64410 0.32.6 2.824.4 2.03.0 0.08.0 0.04.0 2.18.8 0.000.41 0.001.00 0.010.85
X6SD 332 647 459 670 201 619 258 665 1.3 60.4 17.3 65.6 5.1 60.9 2.5 62.6 3.2 61.9 8.2 62.4 0.08 60.21 0.95 60.10 0.21 60.15
Range 273424 342583 173227 137367 0.62.0 8.327.5 3.06.0 0.08.0 1.08.0 4.212.8 0.000.77 0.631.00 0.020.51
X6SD 409 639 568 665 246 650 322 6100 2.0 60.9 21.5 68.2 4.6 60.6 2.2 61.8 1.6 60.7 4.4 61.3 0.13 60.09 0.95 60.19 0.22 60.15
Range 359491 456661 159342 114456 0.84.7 6.331.7 3.06.0 0.07.0 0.03.0 1.17.1 0.000.41 0.031.00 0.040.56
X6SD 598 652 961 6111 340 644 621 6138 3.0 61.4 31.6 67.7 3.6 61.0 6.5 63.6 3.1 62.6 4.9 61.6 0.22 60.15 0.62 60.48 0.32 60.20
Range 524711 6911116 281432 259777 1.26.7 14.444.6 2.05.0 1.015.0 0.08.0 1.36.5 0.020.53 0.001.00 0.050.68
X6SD 570 652 744 679 326 641 418 670 1.6 60.4 21.9 65.7 3.0 60.5 3.0 61.8 2.5 61.4 4.3 61.3 0.18 60.14 0.96 60.10 0.25 60.14
Range 463641 615842 251388 319540 1.12.2 16.130.1 2.04.0 1.07.0 0.05.0 2.86.6 0.060.53 0.671.00 0.040.46
X6SD 666 660 731 634 543 695 188 678 0.9 60.5 7.0 65.3 2.5 60.6 1.7 62.4 0.5 60.8 4.3 60.9 0.24 60.24 0.51 60.51 0.41 60.10
Range 565723 684775 387638 91297 0.21.7 2.016.9 2.03.0 0.06.0 0.02.0 3.35.7 0.000.57 0.001.00 0.220.53
... Howling is a principle means of spacing between wolves and their packs. It serves as a territory-independent spacing mechanism that will result in the use of exclusive territories when coupled with strong, year-round site attachment, but with floating, exclusive, buffer-areas about migratory packs (Harrington and Mech 1983).This long-range signal has been found to encode individual, group, subspecies and species difference (Hennelly et al. 2017;Kershenbaum et al. 2016;Palacios et al. 2007Palacios et al. , 2015Passilongo et al. 2010;Zaccaroni et al. 2012). One technique that can be used for monitoring of wolves in general is a howling survey. ...
... The acoustic structure of howls reported in this study is similar to that found in earlier studies (mean fundamental frequency 274-908 Hz in Italian wolves (Passilongo et al. 2010); 270-720 Hz in Iberian wolves (Palacios et al. 2007); 320-670 Hz in North American timber wolves (Tooze et al. 1990). ...
... It is a means of long-distance intraspecific communication used for territory marking, assembling of the pack, enhancing its cohesion, and for individual and pack recognition (cf. Theberge and Falls 1967;Nowak et al. 2007;Palacios et al. 2007Palacios et al. , 2015Passilongo et al. 2015). It can nevertheless be expected that animals (as a potential prey) and humans perceive howling as a signal of danger and get alerted. ...
Acoustic signals serving intraspecific communication by predators are perceived by potential prey as warning signals. We analysed the acoustic characteristics of howling of wolves and found a striking similarity to the warning sounds of technical sirens. We hypothesize that the effectivity of sirens as warning signals has been enhanced by natural sensory predisposition of humans to get alerted by howling of wolves, with which they have a long history of coexistence. Psychoacoustic similarity of both stimuli seems to be supported by the fact that wolves and dogs perceive the sound of technical sirens as a relevant releasing supernormal stimulus and reply to it with howling. Inspiration by naturally occurring acoustic aposematic signals might become an interesting example of biomimetics in designing new warning sound systems.
... Sixteen acoustic variables (Table 2) were extracted from the howls using Praat (Praat, Amsterdam, Netherlands) [48] and a customized script in MATLAB (Mathworks Inc., Nattick, MA, USA) [49] developed previously [40], where 12 variables have been used in other studies [27,32,36,40]. The vocal parameters were measured by extracting the fundamental frequency (f0) contour of the calls using a cross-correlation method (Sound: To Pitch (cc) time step of 0.005 s, pitch floor 75 Hz, pitch ceiling 1200 Hz) [50]. ...
... Arctic wolves showed an overall correct classification of 95% and several wolves from both Arctic and Eurasian subspecies reached 100% correct classification. These results are in agreement with other studies showing that it is possible to identify individual wolves within subspecies based on their howls [27,36,40]. ...
Full-text available
Simple Summary: This study evaluates the use of acoustic devices as a method to monitor wolves by analyzing different variables extracted from wolf howls. By analyzing the wolf howls, we fo-cused on identifying individual wolves, subspecies. We analyzed 170 howls from 16 individuals from the three subspecies: Arctic wolves (Canis lupus arctos), Eurasian wolves (C.l. lupus), and Northwestern wolves (C.l. occidentalis). We assessed the potential for individual recognition and recognition of three subspecies: Arctic, Eurasian, and Northwestern wolves. Abstract: Wolves (Canis lupus) are generally monitored by visual observations, camera traps, and DNA traces. In this study, we evaluated acoustic monitoring of wolf howls as a method for monitoring wolves, which may permit detection of wolves across longer distances than that permitted by camera traps. We analyzed acoustic data of wolves' howls collected from both wild and captive ones. The analysis focused on individual and subspecies recognition. Furthermore, we aimed to determine the usefulness of acoustic monitoring in the field given the limited data for Eurasian wolves. We analyzed 170 howls from 16 individual wolves from 3 subspecies: Arctic (Canis lupus arctos), Eurasian (C. l. lupus), and Northwestern wolves (C. l. occidentalis). Variables from the fundamental frequency (f0) (lowest frequency band of a sound signal) were extracted and used in discri-minant analysis, classification matrix, and pairwise post-hoc Hotelling test. The results indicated that Arctic and Eurasian wolves had subspecies identifiable calls, while Northwestern wolves did not, though this sample size was small. Identification on an individual level was successful for all subspecies. Individuals were correctly classified with 80%-100% accuracy, using discriminant function analysis. Our findings suggest acoustic monitoring could be a valuable and cost-effective tool that complements camera traps, by improving long-distance detection of wolves.
... Since chorus howls are also used during the breeding season to communicate with pups, we expect sex-and agerelated differences in the vocal behavior of pack members during a chorus. Finally, similar to the vocal cues encoded in howls, we predict that the vocal usage of each individual should be unique and distinctive (Palacios et al. 2007;Watson et al. 2018). To test these hypotheses, we studied how different pack members contribute to a chorus howl, taking advantage of a known scenario with wolves held in captivity, with the information about gender, age, social status, and individual given by the wolf centers and wolf curators. ...
... For instance, individual distinction in calls has been reported for meerkats (Suricatta suricatta, Townsend et al. 2010), wild agile gibbon (Hylobates agilis, Oyakawa et al. 2007), horned guan (Oreophasis derbianus, González-García et al. 2017), and coyotes (Canis latrans, Mitchell et al., 2006). In the case of wolves, Tooze et al. (1987), Palacios et al. (2007), and Root-Gutteridge et al. (2014) concluded that howls uttered by wolves have individual differences. ...
Wolf packs perform group vocalizations called chorus howls. These acoustic signals have a complex structure and could be involved in functions such as strengthening of social bonds, territory advertisement, or spacing between packs. We analyzed video recordings of 46 chorus howls emitted by 10 packs of wolves held in captivity, in order to investigate whether sex, age, social status, pack, or individual influence the way wolves participate in a chorus. We found that, during a chorus, wolves vocalized 63% of the time, with the howl being the most common vocalization (36% of the chorus duration), followed by woa (13.5%), other vocalizations (11.8%), and bark (1.7%). The main factor affecting the vocal behavior of wolves was age, since young wolves vocalized less and uttered shorter acoustic signals than adults. The discriminant analysis carried out with the wolves of Cañada Real pack assigned 89.3% of the cases to the correct individual, which is much better than the assignment expected by chance, suggesting that individuals could have a unique vocal usage during a chorus howl, mainly due to the use of howls and woa-woa howls. Based on our results, we propose that in the context of a chorus the woa-woa howl is important, although further research is needed to address this issue properly.
... Analyses of wolf howls have shown that wolves display individual variation in howl characteristics (e.g., fundamental frequency; Tooze et al. 1990); individual wild eastern grey wolves (C. l. lycaon) can be identified (Root-Gutteridge et al. 2014); and wolves may recognize one another by their distinct howls (Palacios et al. 2007(Palacios et al. , 2015. The complexity and distinctiveness of wolf howls and our increasing ability to record and analyze them suggests that wolf howls may be used to identify and precisely count the number of wolves within a wild pack (Root-Gutteridge et al. 2014, Rocha et al. 2015, Palacios et al. 2016. ...
... AudioMoths may be used near homesites to confirm reproduction and document pup persistence (Palacios et al. 2016), even if the exact location of the homesite is not known. Rigorous individual identification of wild wolves by their vocalizations is very complex (Palacios et al. 2007, Root-Gutteridge et al. 2014, and not all wolves are always present at homesites; therefore, it will likely still prove challenging to obtain precise pack counts, at least in the foreseeable future. Nevertheless, AudioMoths are a reasonable alternative to determine an estimated pack size at homesites, when scats for genetics either cannot be collected because of land access restrictions or because the precise location of the homesite is unknown, when the genetic sample size will be insufficient on account of rapid scat degradation in certain environments (Stenglein et al. [2011] reported that 50 noninvasive genetic samples from rendezvous site areas detected 65-100% of pack members, 100 samples detected 90-100%, and 150 samples detected 100%), or when the budget does not allow for analyzing sufficient scats. ...
As part of a broader trial of noninvasive methods to research wild wolves (Canis lupus) in Minnesota, USA, we explored whether wolves could be remotely monitored using a new, inexpensive, remotely deployable, noninvasive, passive acoustic recording device, the AudioMoth. We tested the efficacy of AudioMoths in detecting wolf howls and factors influencing detection by placing them at set distances from a captive wolf pack and compared those recordings with real-time, on-site howling data between 22 May and 17 June 2019. We identified 1,531 vocalizations grouped into 428 vocal events (236 solo howl series and 192 chorus howls). The on-site AudioMoth correctly recorded 100% of chorus and solo howls that were also documented in real-time. The remote array detected 49.5% of chorus and 11.9% of solo howls (≥1 unit detected the event). The closest remote AudioMoth (0.54 km, 0.33 mi) detected 37% of choruses and 8.9% of solo howls. Chorus howls (9.4%) were detected at the farthest unit (3.2 km, 2.0 mi). Favorable wind (carrying source howls to the remote units) and calm (no wind) conditions increased detectability and detection distance of chorus howls. Temperature was inversely related to detection. Given the detection distances we observed, AudioMoths are probably useful in studying specific sites during periods when wolves move less frequently (e.g., during late spring and summer at homesites or potentially during winter at kill sites of very large prey). AudioMoths would also be useful in a passive sampling array (e.g., occupancy studies), especially when used in concert with other methods such as camera-trapping. Additional research should be conducted in areas with different environmental variables (e.g., wind, temperature, habitat, topography) to determine performance under varying conditions and also when fitted with a parabolic dish.
... Individual variability in vocalization is widespread in the Canidae family. For example, it has been shown in the howling of wolves (Canis lupus) (Palacios et al. 2007;Root-Gutteridge et al. 2014a, b;Tooze et al. 1990), and researchers also found individual-specific variation in dog (Canis familiaris) barks (Molnár et al. 2008;Yin and McCowan 2004). Although these acoustic differences might be used to identify others, the presence of individually distinctive cues does not necessarily indicate the functioning of IR (see Schibler and Manser 2007;Yorzinski 2017). ...
Full-text available
Investigation of individual recognition (IR) is difficult due to the lack of proper control of cues and previous experiences of subjects. Utilization of artificial agents (Unidentified Moving Objects: UMOs) may offer a better approach than using conspecifics or humans as partners. In Experiment 1, we investigated whether dogs are able to develop IR of UMOs (that is stable for at least 24 h) or that they only retain a more generalised memory about them. The UMO helped dogs to obtain an unreachable ball and played with them. One day, one week or one month later, we tested whether dogs display specific behaviour toward the familiar UMO over unfamiliar ones (four-way choice test). Dogs were also re-tested in the same helping context and playing interaction. Subjects did not approach the familiar UMO sooner than the others; however, they gazed at the familiar UMO earlier during re-testing of the problem solving task, irrespectively of the delay. In Experiment 2, we repeated the same procedure with human partners, applying a two-way choice test after a week delay, to study whether lack of IR was specific to the UMO. Dogs did not approach the familiar human sooner than the unfamiliar, but they gazed at the familiar partner earlier during re-testing. Thus, dogs do not seem to recognise an individual UMO or human after a short experience, but they remember the interaction with the novel partner in general, even after a long delay. We suggest that dogs need more experience with a specific social partner for the development of long-term memory.
... fundamental frequency; Tooze et al. 1990), and individuals can be identified via unique acoustic parameters (Root-Gutteridge et al. 2013;Sadhukhan et al. 2021). Palacios et al. (2007Palacios et al. ( , 2014 reported that wolves may recognise one another by their distinct howls, suggesting that wolf howling could have been selected for individual recognition, as has been reported for other canid loud-calls (Balieiro and Monticelli 2019). ...
We studied the spontaneous vocal behaviour of captive wolves at the International Wolf Center (IWC) in Minnesota (spring 2019 and winter 2020), and the Centro del Lobo Ibérico Félix Rodríguez de la Fuente (CLIFRF) in Spain (winter 2020). We used AudioMoth recording devices to record wolf howling 24 h/day. We identified 412 solo howl series and 403 chorus howls and found differences between wolves at the two centres. Vocal rates for North American wolves at the IWC (7.8 chorus howls/day in spring and 4.8 chorus howls/day in winter) were higher than rates obtained for Iberian wolves from CLIFRF (3.8 chorus howls/day in winter). Howling rates obtained in our study were similar to those obtained for captive Mexican wolves and greater than those reported for wild wolves. Hourly distribution of howling was also different between centres. The greatest howling activity identified at IWC was at pre-sunrise, while at CLIFRF the peak occurred at sunset. Weather conditions had little influence on the vocal behaviour of the captive wolves we studied. We show the potential of passive recorders to study topics of animal acoustic communication, such as vocal rates and temporal patterns, that have not been deeply addressed due to technological constraints.
... municate over long distances and have specific functions, such as conveying information about pack members, location, or territory advertisement (Harrington and Asa 2003, Palacios et al. 2007, Zaccaroni et al. 2012, Watson et al. 2018. Researchers conduct acoustic surveys using human-simulated howls or broadcasted, previously recorded wolf howls to detect and monitor wolf packs via their vocal responses (Blanco and Cortés 2012, Passilongo et al. 2015, Palacios et al. 2017. ...
During summer 2019, we recorded an apparent vocal interaction, lasting just under 4 min, between a pair of Great Horned Owls (Bubo virginianus) and a gray wolf (Canis lupus) in Yellowstone National Park. To our knowledge, this is the first report of such an acoustic interaction in the scientific literature. The increased use of passive acoustic recorders, which record spontaneous vocalizations emitted by animals over long periods, will allow us to better document and study the importance of such interspecific interactions.
... Identifying individual wolves from their howls close this gap of implementing the CMR technique for the population assessment of this elusive and challenging to track species 7,25,27 . While a few studies have established that howls carry individuality information 38 and known howls can be distinguished from each other 39,45,71 , no study has been successful before in identifying unknown individuals from a set of howls. Furthermore, attempts to count the number of individuals present in a recording have been limited by difficulties in minimising confidence intervals 18,72 . ...
Full-text available
Previous studies have posited the use of acoustics-based surveys to monitor population size and estimate their density. However, decreasing the bias in population estimations, such as by using Capture–Mark–Recapture, requires the identification of individuals using supervised classification methods, especially for sparsely populated species like the wolf which may otherwise be counted repeatedly. The cryptic behaviour of Indian wolf (Canis lupus pallipes) poses serious challenges to survey efforts, and thus, there is no reliable estimate of their population despite a prominent role in the ecosystem. Like other wolves, Indian wolves produce howls that can be detected over distances of more than 6 km, making them ideal candidates for acoustic surveys. Here, we explore the use of a supervised classifier to identify unknown individuals. We trained a supervised Agglomerative Nesting hierarchical clustering (AGNES) model using 49 howls from five Indian wolves and achieved 98% individual identification accuracy. We tested our model’s predictive power using 20 novel howls from a further four individuals (test dataset) and resulted in 75% accuracy in classifying howls to individuals. The model can reduce bias in population estimations using Capture-Mark-Recapture and track individual wolves non-invasively by their howls. This has potential for studies of wolves’ territory use, pack composition, and reproductive behaviour. Our method can potentially be adapted for other species with individually distinctive vocalisations, representing an advanced tool for individual-level monitoring.
... In addition to olfaction-based tests, research highlighting individual recognition of wolves (Canis spp.) by their unique howls (Palacios et al., 2007;Palacios et al., 2015;Root-Gutteridge et al., 2014;Tooze et al., 1990) suggests auditory self-recognition tests may also be a fruitful line of inquiry. However, there are also potential issues with auditory-self recognition tests. ...
Mirror self-recognition (MSR) tests have been conducted in a variety of species to assess whether these animals exhibit self-awareness. To date, the majority of animals that have convincingly passed are highly social mammals whose wild counterparts live in complex societies, though there is much debate concerning what constitutes “passing” and what passing means in terms of self-awareness. Amid recent reports that a fish (cleaner wrasse, Labroides dimidiatus) passed, it is intriguing that a mammal as highly social, tolerant, attentive, and cooperative as the gray wolf (Canis lupus) has reportedly failed the test. Given the many possible reasons for failure, we were interested in reexamining wolves as a case study of MSR in socially complex mammals as part of a broader overview of the MSR test. We aimed to elucidate the wolves’ responses at various stages of the MSR test to pinpoint potential problem areas where species-specific modifications to the test may be needed. We evaluated 6 socialized, captive gray wolves during July 2017. At a minimum, wolves did not respond to their reflection as an unfamiliar conspecific. Unfortunately, the wolves rapidly lost interest in the mirror and were uninterested in the applied marks. We note limitations of the MSR test for this species, recommend changes for future MSR tests of wolves, discuss other emerging self-cognizance methods for socially complex canids, and highlight the need for a suite of ecologically relevant, potentially scalable self-cognizance methods. Our findings and recommendations may aid in understanding self-cognizance in other MSR-untested, highly social, cooperatively-hunting, coursing, terrestrial carnivores such as African wild dogs (Lycaon pictus), spotted hyenas (Crocuta crocuta), and African lions (Panthera leo).
Full-text available
Barking in domestic dogs still remains a topic of controversial discussions. While some authors assess dog-barking an acoustic means of expression becoming more and more sophisticated during domestication, others name this sound type "non-communicative". Vocal repertoires as works on individual sound types are rare, however, and there has been almost no work done on low-intensity, close-range vocalizations, yet such types of vocalization are especially important with the more social canids, hence, with the human-dog-communication and understanding of dogs. Most of the investigations published so far are based on auditive sound impressions and lack objectivity. The principal method used in this study was sonagraphic. This facilitates the identification of sounds and reveales, whether subjective classification can be verified by objectively measured parameters. Finally, meanings, functions and emotions were examined for all the major sounds described and are discussed in terms of relationships between sound structure and signal function, signal emission and social context as behavioural response, and overlapping channels of communication. Ontogeny of acoustic communication in 11 European wolves has been compared to various dog breeds (8 Standard Poodles, 8 Toy Poodles, 15 Kleine Münsterländer, 11 Weimaraner Hunting Dogs, 16 Tervueren, 12 American Staffordshire Terriers, and 13 German Shepherds, 12 Alaskan Malamutes, and 9 Bull Terriers) from birth up to 8 (12) weeks resp. 4 (12) months of age. Noisy and harmonic sound groups were analysed separately as overriding units. Following parameters were used: fmax=maximum of spectrographic pictured sounds (Hz), xfo=mean of the lowest frequency band of harmonic sounds (Hz), xfd=mean of the frequency of strongest amplitude of noisy sounds (Hz), delta f=frequency range of sounds (Hz), duration of sounds (ms). Statistical analysis was run on "Statistica", Release 4,0. Within the sound type barking 2 to 12 subunits were classified in the different breeds, according to their context-specific spectrographic design, and behavioural responses. Categories of function / emotion include f.e. social play, play soliticing, exploration, caregiving, social contact and "greeting", loneliness, and agonistc behaviours. "Interaction" was the most common category of social context for masted barkings (56% of occurences). Especially close-range vocalizations, concerning the major sound type of most domestic dogs, the bark, evolved highly variable. However, the ecological niche of domestic dogs is highly variable, just as the individual differences in the dogs are, which seem to be breed-typical to a great extent. Thus, complexity within the dog's vocal repertoire, and therefore enhancement of its communicative value, is achieved by many subunits of bark, some standing for specific motivations, informations and expressions. Complexity within the dogs' vocal repertoire is extended by the use of mixed sounds in the barking context. Transitions and gradations to a great extend occur via bark sounds: harmonic, intermediate and noisy subunits.
In the past, behavior was assumed to be largely invariant within species, particularly those elements of behavior used as criteria of mate choice or in species recognition (see Magurran this volume, Verrell this volume). As is obvious from this volume, geographic variation could well be the common condition rather than the exception, and this applies to the full spectrum of behavioral phenotypes. Not only must students of behavior avoid typological thinking (Mayr 1963), but those wishing to infer similarity of behavior among populations must demonstrate the similarity just as surely as those interested in exploring population differentiation must demonstrate the differences. Behavior is as much a phenotype as is morphology; it is the expression of the combined effects of genotype and environment. Like other traits, behavior varies geographically because it is subject to geographically varying conditions and, hence, to natural selection, gene flow, and genetic drift. The chapters in this book provide examples of this variation, of the underlying genetic bases for the differences, and in many cases, the causes of the geographic variation. The study of geographic variation in behavior is in very early stages and lags well behind research on geographic variation in other kinds of traits (Endler 1977, 1986, 1995). Consequently, we cannot answer with assurance many of the questions we would like to be able to answer. However, we can take a first step using the insights offered by the research presented in this book. Before doing so, we briefly address some of the methodological issues that emerged over the course of the research because many are specific to the study of behavior or of geographic variation. We hope this will help others avoid problems encountered in these early studies. Many of the methodological issues discussed in the chapters in this book are related to the difficulty of working with behavioral characteristics that are extremely labile and responsive to environmental conditions. The remainder are issues related to the interpretation of data collected to assess patterns and causes of geographic variation. We will examine them in turn.
In human communication, the speech system is specialized for rapid transfer of information (Liberman et al. 1967; Mattingly and Liberman 1988). Significant events in the acoustic signal can occur in an overlapped or parallel fashion due to the coproduction of speech gestures. One result is that aspects of the signal corresponding to different linguistic units, such as consonants and vowels, often cannot be isolated in the acoustic stream. One way to help tease apart the components of the speech signal is to consider the physical system that gives rise to the acoustic information: The acoustic encoding of phonetic information is viewed in light of the flexibility inherent in the production apparatus, particularly the human supralaryngeal vocal tract, in which individual articulators or groups of articulators can function semi-independently. In this chapter we review this approach. First, we show how the analysis of speech acoustics has benefited by treating the sound production system as one in which the contributions of physical acoustic sources and physiologically determined filters are combined. We then discuss how acoustic diversity has resulted in a desire to find articulatory simplicity. In the process, we review some of the methods used to examine articulatory activity, and also describe in detail a particular attempt at modeling the coordination of the speech articulators. Finally, we consider some recent attempts to explore the links between production, perception, and acoustics in a dynamic-systems approach and in connectionist models.
Conference Paper
When competing for resources wolves use optical and acoustical threat signals of graded intensity. The evolution of threat displays and especially the function of graded threat signals has been the subject of intense discussion over the last decades. While most studies focus on the optimisation of a cost-benefit ratio for an individual's resource gain through threat displays, only little is known about the way the subjective resource value influences threat behaviour. To elucidate this problem, an experimental study was performed with 6 hand-reared wolves at the age of 3 to 5 month, where each wolf as an owner of a food resource was confronted with a second wolf as an intruding competitor. The animals were tested in well-fed condition and in hungry condition on two subsequent days. An analysis of their threat vocalisations revealed, that these were longer in duration and lower in frequency when the wolves were hungry.
Short calls of pikas (Ochotona princeps) living in seven localities in California, Utah, and New Mexico were recorded and analyzed for variation in fundamental frequency, note duration, internote interval, and number of notes per call. Significant differences were found only between widely separated groups. It is suggested that variations in pika short calls are examples of geographic variation, and do not represent vocal dialects. Differences in call parameters between widely separated populations resulted from independent evolutionary histories maintained by geographic barriers to interbreeding. The major constraint on this variability was retention of cues for sound localization.