Content uploaded by Marcelo Araya-Salas
Author content
All content in this area was uploaded by Marcelo Araya-Salas on Jan 21, 2018
Content may be subject to copyright.
Is birdsong music? Evaluating harmonic intervals in songs of a Neotropical
songbird
Marcelo Araya-Salas
*
Escuela de Biología, Universidad de Costa Rica, Ciudad Universitaria, Costa Rica
article info
Article history:
Received 4 December 2011
Initial acceptance 15 February 2012
Final acceptance 25 April 2012
Available online 6 June 2012
MS. number: A11-00965R
Keywords:
biomusicology
birdsong
harmonic interval
Microcerculus philomela
music
musical scale
nightingale wren
Signals in animal communication are commonly judged as aesthetically appealing by human standards.
This is particularly common for birdsong, often equated to musical compositions. No formal test,
however, has analysed the harmonic properties of bird vocalizations. Musical intervals are based on the
same physical characteristics of sound that underlie animal vocal signals. Thus, animals may use these
intervals as rules to organize their vocalizations in a similar way as music. I tested a prediction derived
from this hypothesis, that frequency ratios of adjacent notes in birdsong are closer to harmonic intervals
than expected by chance. I determined to what degree the intervals created byadjacent notes of the song
of nightingale wrens, Microcerculus philomela, conform to harmonic intervals. I analysed songs from 81
birds across the entire distribution range of the species, comparing the intervals formed by adjacent
notes to three musical scales: chromatic, major diatonic and major pentatonic. Comparisons were made
based on null model distributions. From 243 comparisons, only six (w2%) were significantly close to
harmonic intervals, suggesting no consistent use of harmonic intervals. The frequency of the notes is the
most varying song parameter in this species. If the frequencies are not determined by harmonic intervals
in this species, it seems less likely that it happens in other birds with more complex song elements.
Documented musical properties in birds might be caused by cultural biases of the listener or misun-
derstanding of the physics of musical compositions.
Ó2012 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Since ancient times, humans have judged animal signals by
human standards of beauty (Darwin 1871;Grammer et al. 2003).
This is particularly recurrent in the appreciation of birdsongs,
because of their resemblance to musical composition (Darwin
1871;Scholes 1938;Hartshorne 1958;Saunders 1959;Hall-
Craggs 1969;Dobson & Lemon 1977;West & King 1990;West
et al. 2004;Baptista & Keister 2005;Tierney et al. 2011). The
resemblance is due to the use of specific tonal qualities, pitch
relationships and phrase duration and rhythm in birdsongs (Marler
1969;Baptista 2004). The parallel between birdsong and music has
inspired many researchers and naturalists to equate the variation in
frequency found in birdsongs to their counterparts in musical
scales. For instance, Baptista & Keister (2005, page 432) in their
paper ‘why birdsong is sometimes like music’wrote: ‘Some bird-
song is pitched to the same scale as Western music, which is one
possible reason for human attraction to these sounds’. Many other
examples can be found in ornithological literature: Saunders (1959)
suggested that white-throated sparrows, Zonotrichia albicollis,have
a perfect fourth (musical interval) between the first and second
notes, Borror & Reese (1956, page 182) suggested that the songs of
the wood thrush, Hylocichla mustelina,‘are so pitched that they
follow our musical scale very accurately’,Hartshorne (1973), that
the canyon wren, Catherpes mexicanus, sings in the chromatic scale
(musical scale with 12 pitches per octave), and Wing (1951), that
the hermit thrush, Catharus guttatus, sings in the pentatonic scale
(musical scale with five pitches per octave).
Although the aforementioned accounts may be useful for
describing the pitch variation among song elements in birdsongs,
they imply that birds use discrete sound intervals, instead of
a continuous range of sounds, to organize their songs. If true, this
would have important implications for understanding the patterns
of song ontogeny and evolution in animal signals. However, that
hypothesis has not been properly tested for any bird species. To my
knowledge, the only study on this matter was made by Dobson &
Lemon (1977), who qualitatively compared the songs of white-
throated sparrows, Z.albicollis,and African boubou shrikes
(Laniarius sp.) with the most common Western musical scales.
Unfortunately, no statistical test was made and no clear insight was
provided on the harmonic properties of birdsongs.
Furthermore, no cogent rationale for the use of similar intervals
in music and birdsong has been proposed. Why should a bird use
*Correspondence and present address: M. Araya-Salas, Department of Biology,
New Mexico State University, 1615 Espina, Las Cruces, NM 88001, U.S.A.
E-mail address: marceloa27@gmail.com.
Contents lists available at SciVerse ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
0003-3472/$38.00 Ó2012 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.anbehav.2012.04.038
Animal Behaviour 84 (2012) 309e313
the same intervals as musicians? Musicians use specific intervals
because of the consonance produced by the combination of their
sounds. This consonance comes from the physical properties of
instruments and human vocal tracts (Benson 2008). Stringed and
wind instruments and human vocal tracts produce a spectrum of
overtones along with a fundamental frequency for each note. In
both instruments and vocal tracts these overtones are harmonics,
which are defined as exact integer multiples of the fundamental
frequency (Benson 2008;Gill & Purves 2009). Different pitch notes
for which the frequencies are proportionally related share most of
their harmonics, which generates the consonance of their interval
(Benson 2008). In other words, musical intervals are formed by
sounds with a higher similarity in the frequencies of their harmonic
spectrum compared to nonmusical intervals. Thus, musical inter-
vals are based not in human conventional arrays of frequencies but
in the physical characteristics of the sounds. The preference in
humans may have evolved through a sensory bias towards intervals
reflecting the spectral characteristics of conspecific vocalizations
(Gill & Purves 2009) or as a honest signal of cognitive abilities
favoured by sexual selection (Miller 2001). Similar hypotheses may
explain the evolution of analogous preferences in other animals
with comparable spectral complexity (see below). Several other
hypotheses have been proposed for the evolution of music in
humans and the present list is not exhaustive (Patel 2010).
Many species of birds, as well as mammals and amphibians,
generate harmonics proportionally related to the fundamental
frequency in their vocalizations in the same way as the musical
instruments we use (Bradbury & Vehrencamp 1998). In addition,
birds seem to have the necessary auditory and neural capacity to
relate harmonically linked sounds. Birds can perceive the harmonic
structure of single elements in their songs (Cynx et al. 1990;Lohr &
Dooling 1998;Doolinget al. 2002) and in some cases use it to encode
biologically relevant information (Morgan & Howse 1973;Aubin &
Bremond 1992). Some bird species even show relative pitch
discrimination as humans do and can discriminate pairs of notes by
their frequency ratios (Weary & Weisman 1991;Weisman et al.
1994;Bottoni et al. 2003). Overall, the available evidence suggests
that birds have the necessary abilities requiredto relate and organize
different song elements by the consonance of their frequencies, or in
other words, to conform songs to harmonic intervals.
Even though birdsongs might conform to harmonic scales, the
often complex and frequency-modulated structure of the song
elements has made the testing of this hypothesis difficult, basically
because it complicates the selection of a single frequency value as
representative of a particular element. Onlya few bird species seem
to have the adequate element structure and pitch variation
between elements to allow detailed analysis (Dobson & Lemon
1977). An ideal birdsong for analysis in terms of harmonic inter-
vals should have sounds of sustained and almost pure pitch
elements, a variety of frequencies among adjacent elements, and
clear harmonic structure to its elements. The nightingale wren,
Microcerculus philomela, represents an excellent opportunity to test
these ideas. Its song consists of a single stereotyped series of nearly
pure-tone notes at different pitches. The notes show clear harmonic
structure with roughly no pitch repetition among consecutive
notes (Fig. 1). This species also shows microgeographical song
variation with several dialects in an area of just a few kilometres
(M. Araya-Salas, T. F. Wright & G. Barrantes, unpublished data). The
geographical variation is due to different arrays of notes rather than
to variation in the structure of the composing elements (M. Araya-
Salas, T. F. Wright & G. Barrantes, unpublished data). This type of
variation allows us to test the existence of harmonic intervals in
several different arrays of notes. For this study, I tested the
prediction derived from the harmonic birdsong hypothesis, that the
frequency ratio of adjacent notes would be closer to harmonic
intervals than expected by chance. For this I use ‘just intonation’
4.04
FP8 P8P5 m7 M2 M3
4.43
F
P8
P5
P8
m7
M2
M3
Time (s)
18
12
6
140
70
1234567891011
Fre
q
uenc
y
(kHz)
Frequency (kHz)Amplitude (dB)
12 13 14 15 16 17 18 19
4.46
Figure 1. Spectrogram of a segment of the song of a nightingale wren from La Selva Biological Station, Costa Rica. The spectrogram shows the harmonic structure of the different
notes. Below is a spectrogram slice corresponding to the dotted line in the spectrogram view indicating the fundamental pitch and the associated harmonics according to the
nomenclature of Western music (F: fundamental; P8: perfect octave; P5: perfect fifth; M2: major second; 3M: major third). Points over the dotted line indicate the fundamental
frequency (lowest point) and the associated harmonics in the spectrogram slice. Frequencies of the harmonics are exact integer multiples of the fundamental frequency.
M. Araya-Salas / Animal Behaviour 84 (2012) 309e313310
intervals, a particular type of musical interval formed by pairs of
notes related by small integer ratios. Similar integer ratios are
produced between a fundamental note and its harmonics and in
animal vocalizations. Hence, these scales represent the most intu-
itive pattern in which birds might base their songs.
METHODS
Study Species and Sites
The nightingale wren is a resident of the understory of mature
forests from southern Mexico to northern Costa Rica (AOU 1998).
I recorded eight individuals from La Tirimbina Biological
Station (10
24
0
40.87
00
N; 84
6
0
48.40
00
W) with a Sony TCM-2000DV
recorder and a Sennheiser ME66 directional microphone. In addi-
tion, I obtained 73 recordings from bioacoustics archives from
eight sites (Supplementary Table S1). These recordingsbelong to nine
populations from the Caribbean Slope of Costa Rica and eight pop-
ulations from other locations in Central America and Mexico
(Supplementary Fig.S1). Thus, therecordings fairly represent the song
variation across the entire geographical distribution of the species.
I digitalized recordings from La Tirimbina at a sampling rate of
48 kHz. All other recordings were digitalized at a sampling rate of
44.1 kHz. I analysed the songs using Raven 1.2 software (Charif et al.
2004). For spectrogram analysis, I used the Hann window function,
a frame length of 202 samples, a filter bandwidth of 314 Hz, and the
‘Gray scale’colour scheme option. I measured the fundamental
frequency (the lowest frequency in a harmonic series) of each note.
Based on this variable I calculated the interval formed by adjacent
notes from the same song, dividing the fundamental frequency of
the second note of the interval by the fundamental frequency of the
first note (Fig. 2). Thus, the intervals were expressed as ratios of
values between one and two. Ratios higher than two were divided
by integer multiples of two. This calculation allowed me to simplify
the analysis, keeping all the ratios in the same range without
changing the harmonic relation between notes (e.g. a ratio of 3 and
a ratio of 1.5 are harmonically the same: a perfect fifth). Intervals
formed by repeated notes (notes with equal frequency; e.g. intro-
ductory notes) were omitted from the analysis. These notes create
harmonic intervals as a by-product of note repetition.
Data Analysis
For each bird, I tested the harmony of the entire number of
intervals (the intervals from all the songs). To carry out the analysis
Ifirst calculated for each frequency interval the distance to the
closest harmonic interval. The distance was considered as the
absolute difference between the observed interval ratio and the
closest harmonic interval ratio. For instance, two intervals with
values of 1.53 and 1.47 would have the same distance of 0.03 to the
closest harmonic interval (1.5; a perfect fifth). In the harmonic
scales (described below), the range between harmonic intervals is
related to the magnitude of the intervals: the higher the magnitude
of the harmonic interval, the higher the range among subsequent
intervals. To account for this, I standardized the distance to the
closest harmonic interval by converting distances to percentages,
with 100% representing the closest distance (or an exact harmonic
interval) and 0% the furthest distance (exactly in the middle of two
harmonic intervals).
I used three musical scales derived from ‘just intonation’inter-
vals to test the conformity of the songs to harmonic intervals
(Fig. 3aec): chromatic scale (12 notes), major diatonic scale (seven
notes) and major pentatonic scale (five notes). The chromatic scale
has the 12 intervals used in Western music, including consonant
and dissonant intervals (Fig. 3a). The other two scales (diatonic and
pentatonic) are subsets of the chromatic scale containing only the
intervals with the smallest integer ratios (called consonant inter-
vals). These intervals are formed by notes that share most of their
harmonics. The pentatonic scale contains the fundamental (or
octave), major second, major third, perfect fifth and major sixth
(Fig. 3c), which are the intervals derived from the first five
harmonics. These are also the harmonics that are emphasized in the
nightingale wren whistles (Fig. 1). The diatonic scale contains,
additionally, the perfect fourth and major seventh (Fig. 3b), which
are derived from the sixth and seventh harmonics, respectively. The
chromatic scale also contains another type of interval created by
the secondary relationship between consonant intervals; these
intervals are called dissonant. For instance, if a tonic note is played,
followed by a major third, and then a major sixth is played from
that major third, the interval created by the last note and the tonic
note is a minor second, a dissonant interval. Hence the chromatic
scale would occur only if the notes have a harmonic relation to
notes two or more positions away in the song. Diatonic and
pentatonic scales would occur if the birds build harmonic intervals
only between immediately consecutive notes. Therefore, the scales
were selected under the criteria of assessing two levels of conso-
nance (pentatonic and diatonic scales) and evaluating the possible
harmonic relation among nonconsecutive notes (chromatic scale).
Although other subsets of intervals from the chromatic scale might
be formed, they would occlude some consonant intervals or mix
consonant and dissonant intervals with no clear criteria.
I used a null model to compare the observed intervals (repre-
sented by the percentage of the distance to the closest harmonic
interval) to expected nonharmonic values. Intervals from nonhar-
monic songs are expected to be uniformly distributed around
harmonic interval values. Hence, for each individual, I randomly
created a uniformly distributed variable of 10 000 values between
0 and 100. I then took 1000 subsets from the generated variable
Figure 2. Spectrogram of three notes of the nightingale wren showing the calculation
of interval ratios between adjacent notes. Dotted lines indicate the fundamental
frequency of the notes (frequency with the highest energy). Interval ratios were
defined as the ratio of the fundamental frequency of the second note to the first note of
the interval.
Figure 3. Musical notation of three scales used to assess the conformity of the songs to
harmonic intervals: (a) chromatic scale (12 notes), (b) major diatonic scale (seven
notes) and (c) major pentatonic scale (five notes). F: fundamental; m: minor; M: major;
P: perfect; A: augmented. Numbers represent the ordinal position of the note
(e.g. M2 ¼major second).
M. Araya-Salas / Animal Behaviour 84 (2012) 309e313 311
(with replacement), each one having a sample size equal to the
number of intervals of the individual to be compared. Each subset
was compared to the observed set of values with a Student’sttest.
The proportion of the 1000 tests in which the observed values were
not significantly different from the uniform distribution was used
as the associated statistical probability. These analyses were carried
out in R version 2.9.0 (R Development Core Team 2009).
In addition, I analysed the melodies of musical compositions to
validate my method for detecting harmonic structure in acoustic
signals. I analysed songs from continuous pitch instruments. Unlike
instruments such as piano and flute, continuous pitch instruments
are not constrained to play notes at discrete intervals, allowing the
production of any possible interval in an octave, similar to an avian
syrinx and a human vocal tract. Hence, the intervals produced by
continuous pitch instruments represent an actual selection of
specific intervals by the musicians. Musical compositions were
obtained from commercial recordings or solo performances freely
available online (Supplementary Table S2). Compositions belong to
three musical styles: classical, jazz and popular music. A minimum
of 40 notes per song were included in the analysis. Intervals formed
by notes of equal frequency were discarded. Melodies were ana-
lysed in the same way as bird recordings.
Ethical Note
This research involved recording from nightingale wrens in the
wild. No birds were captured, and no behavioural experiments
were implemented. The study was conducted in compliance with
regulations of the Department of Environment and Energy of the
government of Costa Rica.
RESULTS
A total of 81 birds were used in the analysis. These birds
belonged to 16 song type populations: eight in Northern Costa Rica
and the rest from other sites across the species’distribution range
(see Supplementary Fig. S1). I carried out 243 comparisons (81 for
each of the three scales). A mean of 96 (range 15e970) intervals
were used in each test. Only five birds (w6%) had intervals that
were significantly closer to harmonic intervals than the expected
random distribution (2.5% of the 243 comparisons). In three cases
the intervals conformed to the pentatonic scale (MG8: P¼0.013;
TO5: P¼0.002; OC1: P¼0.037). One of these birds also conformed
to the diatonic scale (OC1 ¼0.044). Two other birds conformed to
0.5
0.4
0.3
0.2
0.1
0
0.5
Density
0.3
0.4
0.2
0.1
0
0.5
0.6
0.7
0.8
0.9
0.4
0.3
0.2
0.1
00 20 40 60 80 100
Distance to the closest harmonic interval (%)
Harmonic interval
(a) (b) (c)
(d) (e) (f)
(g) (h) (i)
0 204060801000 20406080100
Harmonic interval Harmonic interval
‘Non−
harmonic’
birds
‘Harmonic’
birds
Musical
compositions
Figure 4. Distance to harmonic intervals (chromatic scale) of adjacent note ratios from three nightingale wrens whose adjacent note ratios were nonsignificantly close to harmonic
intervals (aec, ‘nonharmonic birds’), three birds whose adjacent note ratios were significantly close to harmonic intervals (def, ‘harmonic birds’) and three musical compositions
(gei). Distances are shown as percentages. In the histograms, 100% represents the closest distance (a perfect harmonic interval; thick lines) and 0% represents the furthest distance
(exactly in the middle of two harmonic intervals).
M. Araya-Salas / Animal Behaviour 84 (2012) 309e313312
the chromatic scale (LS8: P¼0.016; LS10: P¼0.032). Only two
birds belonged to the same population.
In contrast, my method successfully detected the harmonic
structure in musical compositions. From the 24 melodies analysed,
all were significantly close to intervals from the chromatic scale
(P<0.003 in all cases), and 21 were also significantly close to
intervals of both pentatonic and diatonic scales. Differences in the
distribution of frequency ratios in songs of nightingale wrens and
musical compositions are graphically shown in Fig. 4.
DISCUSSION
Nightingale wren songs are not organized by the same rules used
in musical composition. Overall, the frequency relation between
adjacent notes did not fit with harmonic intervals, nor was thereany
consistency in the particular scale used in the few cases (2.5% of all
examined) where harmonic scales were detected. The few songs
that conformed to musical scales are probably an unintended
matching to harmony, which can be expected based on the large
sample size (81 birds, 243 comparisons). Given the uncommon use
of harmonic intervals in this species, my resultsstrongly suggest that
it is not an intentional feature of these birdsongs.
My method was successful in detecting harmonic intervals in
human musical compositions. All the melodies conformed to harmonic
intervals. My results, however, did not show a perfect match toall three
harmonic scales. The mismatch between some melodies and the
pentatonic and diatonic scales is most likely due to the use of intervals
not included in these scales but included in the chromatic scale.
Nightingale wren song is unusual among birdsongs in the extent
to which frequency alone, rather than duration or patterns of
frequency modulation, is the dominant difference among notes.
Thus, variation in note frequencies is the primary way to generate
song complexity (otherwise the song would be a series of equally
pitched whistles). This variation creates different note arrays,
which constitute the variation among song types (M. Araya-Salas,
T. F. Wright & G. Barrantes, unpublished data). If the harmonic
relationship between notes is not used as a rule to organize notes in
the song of this species, it seems even less likely that this occurs in
other birds with more complex song elements, in which song
variability seems to be emphasized in the production and combi-
nation of different types of elements (e.g. ascending or descending
whistles, broad-bandwidth elements, trills, etc.). Nevertheless, the
harmonic birdsong hypothesis remains to be tested in species with
more complex song elements.
This work represents the first quantitative analysis for testing
whether the frequency shifts in a songbird’s vocalizations conforms
to the harmonic intervals associated with human musicand provides
no support for a significant role of harmonic intervals in the organi-
zation of the songs.This pattern might also occur in otherbird species
where versatility is not based on tonal variation between repeated
elements. The documented musical properties in other species may
be caused bya cultural bias of the human listenertowardsrecognizing
the occasional harmonic interval or by a simple misunderstanding of
the physicsof music composition, rather thanan actual characteristic
of the songs. This studyshould help bioacousticsresearchers to clarify
some notions about the acoustic structure of music and to recognize
the implicit assumptions when equating music and birdsong.
Acknowledgments
I thank Eduardo Chacon, Erick Fuchs, Federico Bolaños and
Gilbert Barrantes for stimulating discussion of ideas on early stages
of this manuscript, and Elizabeth Hobson, Timothy Wright, Johel
Chaves, Sean Ehlman and anonymous referees for comments on the
manuscript. I am also grateful to Borror Laboratory of Bioacoustics,
Bioacoustics Laboratory of Universidad de Costa Rica, Florida
Museum of Natural History, Macaulay Library, Alex Villegas, David
Bradley, Gerardo Obando and Julio Sanchez for access to recordings
and the Organization for Tropical Studies and Tirimbina Biological
Reserve for logistic support. Finally, I thank Evelyn Cubillo for help
and support during the development of this manuscript.
Supplementary Material
Supplementary material for this article is available, in the online
version, at doi:10.1016/j.anbehav.2012.04.038.
References
AOU (American Ornithologists’Union) 1998. Checklist of North American Birds. 7th
edn. Washington, D.C.: American Ornithologists' Union
Aubin, T. & Bremond, J. C. 1992. Perception of distress call harmonic structure by
the starling (Sturnus vulgaris). Behaviour,120,151e163.
Baptista, L. F. 2004. Nature and its nurturing in avian vocal development. In:
Nature’s Music, the Science of Birdsong (Ed. by H. Marler & H. Slabbekoorn), pp.
39e60. Amsterdam: Elsevier.
Baptista, L. F. & Keister, R. A. 2005. Why birdsong is sometimes like music.
Perspectives in Biology and Medicine,48, 426e443.
Benson, D. 2008. Music: a Mathematical Offering. 1st edn. Cambridge: Cambridge
University Press.
Borror, D. J. & Reese, C. R. 1956. Vocal gymnastics in wood thrush songs. Ohio
Journal of Science,56,177e182.
Bottoni, L., Massa, R. & Lenti Boero, D. 2003. The grey parrot (Psittacus erithacus)
as musician: an experiment with the temperate scale. Ethology Ecology and
Evolution,15,133e141.
Bradbury, J. & Vehrencamp, S. L. 1998. Principles of Animal Communication. Sun-
derland, Massachusetts: Sinauer.
Charif, R. A., Clark, C. W. & Fristrup, K. M. 2004. Raven. 1.2 User’s Manual. Ithaca,
New York: Cornell Laboratory of Ornithology.
Cynx, J., Williams, H. & Nottebohm, F. 1990. Timbre discrimination in zebra finch (Tae-
niopygia guttata)songsyllables.Journal of Comparative Psychology,104 ,303e308.
Darwin, C. 18 71. The Descent of Man and Selection in Relation to Sex. London:
J. Murray.
Dobson, C. W. & Lemon, R. E. 1977. Bird song as music. Journal of the Acoustical
Society of America,61, 888e890.
Dooling, R. J., Leek, M. R., Gleich, O. & Dent, M. L. 2002. Auditory temporal
resolution in birds: discrimination of harmonic complexes. Journal of the
Acoustical Society of America,112 ,748e759.
Gill,K. Z. & Purves,D. 2009. A biological rationale for musicalscales. PLoS One,4, e8144.
Grammer, K., Fink, B., Møller, A. & Thornhill, R. 2003. Darwinian aesthetics:
sexual selection and the biology of beauty. Biological Reviews,78, 385e407.
Hall-Craggs, J. 1969. The aesthetic content of bird song. In: Bird Vocalizations: Their
Relations to Current Problems in Biology and Psychology (Ed. by W. H. Thorpe), pp.
367e381. London: Cambridge University Press.
Hartshorne, C. 1958. The relation of bird song and music. Ibis,10 0,421e444.
Hartshorne, C. 197 3. Born to Sing. Bloomington: Indiana University Press.
Lohr, B. & Dooling, R. J. 1998. Detection of changes in timbre and harmonicity in
complex sounds by zebra finches (Taeniopygia guttata) and budgerigars
(Melopsittacus undulatus). Journal of Comparative Psychology,112,36e47.
Marler, P. 1969. Tonal quality of bird sounds. In: Bird Vocalizations (Ed. by
R. A. Hinde), pp. 5e18. Cambridge: Cambridge University Press.
Miller, G. 2001. Evolution of human music through sexual selection. In: The Origins
of Music (Ed. by N. L. Wallin, B. Merker & S. Brown), pp. 329e360. Cambridge,
Massachusetts: MIT Press.
Morgan, P. A. & Howse, P. E. 1973. Avoidance conditioning of jackdaws (Corvus
monedula) to distress calls. Animal Behaviour,21, 481e491.
Patel, A. D. 2010. Music, biological evolution, and the brain. In: Emerging Disciplines
(Ed. by M. Bailar), pp. 91e144. Houston, Texas: Rice University Press.
R Development Core Team 2009. R: a Language and Environment for Statistical
Computing. Vienna, Austria: R Foundation for Statistical Computing. http://
www.R-project.org.
Saunders,A. A. 1959.Octaves and kilocycles in birdsongs. Wilson Bulletin,71,280e282.
Scholes, P. A. 1938. The Oxford Companion to Music. Oxford: Oxford University Press.
Tierney, A. T.,Russo,F. A. & Patel,A. D. 2011.The motor originsof human andavian song
structure. Proceedingsof the National Academy of Sciences, U.S.A.,108,15510e15 515.
Weary, D. M. & Weisman, R. G. 1991. Operant discrimination of frequency and
frequency ratio in the black-capped chickadee (Parus atricapillus). Journal of
Comparative Psychology,105, 253e259.
Weisman, R., Njegovan, M. & Ito, S. 1994. Frequency ratio discrimination by zebra
finches (Taeniopygia guttata) and humans (Homo sapiens). Journal of Compara-
tive Psychology,108, 363e372.
West, M. J. & King, A. P. 1990. Mozart’s starling. American Scientist,78,106e114.
West, M. J., King, A. P. & Goldstein, M. H. 2004. Singing, socializing, and the music
effect. In: Nature’s Music, the Science of Birdsong (Ed. by H. Marler &
H. Slabbekoorn), pp. 274e387. Amsterdam: Elsevier.
Wing, L. 1951.Notes on the song-seri es of a hermit thrush in the Yukon. Auk,68,189e193.
M. Araya-Salas / Animal Behaviour 84 (2012) 309e313 313
