The use of playbacks can influence encounters
with birds: an experiment
André Magnani Xavier de Lima1,2 and James Joseph Roper1
1 Universidade Federal do Paraná, Graduate Program in Ecology and Conservation, Caixa Postal 19031, CEP 81531‑980, Curitiba, PR, Brasil.
E‑mail: andremxlima@uol.com.br; jjroper@gmail.com
2 Instituto de Pesquisa e Conservação da Natureza – Idéia Ambiental, Rua Euclides Bandeira, 1635, CEP 80530‑020, Curitiba, PR, Brasil.
Recebido em: 17/07/2008. Aceito em: 15/06/2009.
ResuMo: o uso de Playbacks pode influenciar encontros com aves: um experimento. A reprodução do repertório sonoro
“playback” costuma ser utilizada para atrair aves tanto para visualizá‑las como para capturá‑las. O uso combinado de playback e
captura pode afetar o comportamento das aves de modo que a continuidade do uso pode não gerar os resultados desejados na coleta
de dados. Neste trabalho estudamos efeitos do uso desta técnica na detectabilidade das aves subseqüente ao uso de playback. Aves
expostas a playbacks tendem a se aproximar menos do observador nas amostragens subseqüentes. Estas aves ainda respondem de
modo mais breve aos playbacks em relação a indivíduos que nunca tiveram contato prévio com esta técnica. É possível que a captura
em si tenha menor ou pouca importância nos comportamentos posteriores das aves. Considerando o amplo uso de playbacks,
recomendamos atenção para as análises em estudos populacionais que utilizam este método.
PALAvRAs-ChAve: playback, viés amostral, Conopophaga melanops, Drymophila squamata, Myrmeciza squamosa, experimento
AbsTRACT: Playback of recorded bird songs is often used to attract birds for sighting as well for captures. The combined use
of playbacks and captures may in fact change the behavior of birds such that the response to playbacks does not reflect responses
to natural songs. Here we studied whether the use of playbacks changes detectability of the birds. Birds previously exposed to
playbacks (experienced birds) did not approach the sound source as closely as did naive birds in playback trials. The response time
of experienced birds is also shorter than that of naive birds. Limited data suggest that the biased response to playbacks is unrelated
to whether the bird was captured. Considering the common use of playback, we recommend caution in the analysis in population
studies in which playbacks are used.
Key-WoRds: playback, bias, Black‑cheeked Gnateater, Scaled Antbird, Squamate Antbird, experiment.
The use of playbacks may be a useful tool for a
variety of studies of bird behavior, ecology, and popu‑
lation dynamics. Studies of vocalization and behavior
in birds often use playbacks of recorded bird songs to
gather data of the role of those songs (Ficken and Fick‑
en 1970, Verner and Milligan 1971, Simpson 1984, Ri‑
gelski and Moldenhauer 1996, Kroodsma 2005). Other
studies, especially those of survival, reproductive suc‑
cess and population dynamics, often require that birds
be captured and uniquely color‑banded (Kearns et  al.
1998, Bayne and Hobson 2002, Hansen and Slagsvold
2003, Roper 2005). In these studies, recorded songs
may be used to attract birds to nets to improve capture
probabilities. Later, marked birds must then be found
again to gather the relevant data, where re‑sighting
marked birds is equivalent to recapture. If birds are dif‑
ficult to find or bands difficult to read, researchers may
again use playbacks to attract the birds for sighting or
recapture.
If birds that were previously exposed to playbacks
behave differently than unexperienced birds (denominat‑
ed here as naive birds), then this bias must be taken into
consideration when analyzing data. If this bias exists, then
birds exposed to playbacks are likely to respond to play‑
backs differently than they would to songs of neighboring
birds. Also, if sighting of birds is required and birds that
were previously captured by the use of playbacks behave
differently than naive birds, then using playbacks will re‑
sult in a biased sampling. Here we tested whether birds
that had been exposed to playbacks and capture respond‑
ed differently to playbacks than did naive birds.
MeThods
This study took place in the Salto Morato Nature
Reserve (25°10’S, 48°18’W) in southern Brazil. The re‑
serve comprises coastal tropical Atlantic Forest, relatively
ARTIGo Revista Brasileira de Ornitologia, ##(#):###-4
xxxxxxxxxx de ####

close to the city of Guaraqueçaba, in the state of Paraná.
Experimental use of playbacks included three species of
insectivorous, permanently territorial, understory forest
birds: the Black‑cheeked Gnateater (Conopophaga mela-
nops, Conopophagidae), the Scaled Antbird (Drymophila
squamata) and the Squamate Antbird (Myrmeciza squa-
mosa, both Thamnophilidae).
Birds were captured beginning in July 2006. By Sep‑
tember 2006, males of C. melanops (n = 10), D. squama-
ta (n = 6) and M. squamosa (n = 3) and three females of
C. melanops were captured and uniquely color‑banded, all
prior to the beginning of the experiment proposed in the
current study. Captures were carried out by using play‑
backs to attract birds to the mist‑nets. All of these were
randomly exposed at least two times to playbacks of three
different recorded songs. Songs destined to be used as
playbacks were recorded nearby, but not in the study area,
prior to the beginning of the experiment. Three distinctive
recorded territorial songs per species were used for all tests.
Since these three species are territorial and pairs
travel together, playbacks exposed both members of each
pair to the playback. Thus, we consider each pair as naive
or unexperienced (if there was no attempt at capture with
the use of playback) or experienced (at least one mem‑
ber of the pair captured after the use of playback), ac‑
cording to the area where the playbacks were conducted.
Also, since members of a pair travel together, exposing
one to playbacks also exposes the other. Thus, when only
one member of a pair was color‑banded, we assumed the
other was the mate who was also exposed. Observations
indicate that this assumption is valid since pairs remain
together while on an established territory within one
breeding season (Lima and Roper in review).
experiment
Two areas (each ~20 ha, separated by > 300 m) were
used in the experiment, one with experienced birds (previ‑
ously exposed to playbacks) and the other with naive birds
(never exposed to playback). Playbacks of each of the three
species were used to attract birds during a nine day interval
in late October 2006, in the breeding season of these spe‑
cies. In the experienced area, playbacks were used in 12
locations, and in the naive area playbacks were used in 22
locations. Playback locations were used only once each in
the naive area. All playback locations were a minimum of
100 m from each other. Each day of the experiment the
location of the playback was selected randomly, with the
constraint that adjacent territories were never used on the
same or the following day. Each day 20 locations were used.
The experiment was carried out throughout the day
at all 20 daily playback points and began with two rep‑
etitions of the recorded song. Playbacks lasted for 6 min
after the initial response (defined as singing nearby or
FIGuRe 1: A comparison between Naive and Experienced birds of
the Nearest Approach Distance (A) and Stay Time (B) for Conopoph-
aga melanops and Drymophila squamata (only Distance, C), showing
that experienced birds tend to remain farther from the sound source
(A, C), and they leave the area sooner (B), than do naive birds. Aver‑
ages and 95% confidence intervals are shown (A and C, all P < 0.05).
In the distance comparison, values were log transformed prior to test‑
ing to normalize the residuals, and therefore the Y‑axis is logarithmic.
2 The use of playbacks can influence encounters with birds: an experiment
André Magnani Xavier de Lima and James Joseph Roper
Revista Brasileira de Ornitologia, ##(#), ####

approaching the loudspeaker < 30 m). During the 6 min,
the closest distance that the bird approached the loud‑
speaker was measured (Minimum Approach Distance).
Duration of the response (Stay Time) was timed, begin‑
ning with the first sign of response (song or approach) and
lasted until the bird lost interest and flew away (> 30 m),
or 6 min, whichever came first. Immediately after 6 min
the experiment was moved to the next location and re‑
peated. If there was no response by the target species to
the playback within 5 min, the experiment was moved to
the next location.
The unit of analysis was each individual response. If
both birds of the pair responded, only that of the first in‑
dividual seen was used for the main analysis. In the test of
the dependency of the capture and playback exposure, we
included responses of unmarked males mated to the three
marked females of C. melanops. Minimum Approach Dis‑
tance (meters) was compared between naive and experi‑
enced birds with t‑tests. Since Stay Time was censored
(that is, not continuous because it has a maximum cut‑off
at 360 seconds) we used Proportional Hazards Survival
Analysis to compare the two treatments. All tests used a
decision rule of α = 0.05.
ResuLTs
Distance was log‑normally distributed and so analy‑
ses were based on loge‑transformed data; in the description
the values were back transformed. Naive C. melanops ap‑
proached more closely (2.8 m versus 11.1 m for experi‑
enced birds, t18 = 4.34, P < 0.05, Figure 1A) and stayed
longer (302.6  s versus 157.8  s, log‑likelihood ratio
χ2 = 6.18, d.f. = 1, P < 0.05, Figure 1B) than experienced
birds. naive D.  squamata also approached more closely
(3.3 m versus 6.8 m, t14  =  2.89, P < 0.05, Figure  1C).
Stay Time for D. squamata was similar for naive and ex‑
perienced birds and most birds stayed the entire 6 min (2
experienced birds had shorter stay times). Sample size for
M. squamosa did not allow statistical comparison, yet the
one naive individual approached more closely than did
the three experienced individuals (3.0 m versus 5.5‑8.0 m
respectively).
While the sample size was small to test for the effects
of actual capture, with C. melanops a simple comparison
was possible for Minimum Approach Distance and Stay
Time. Uncaptured but experienced individuals (n  =  2)
were within the ranges of values of the captured individu‑
als (Figure 2).
dIsCussIon
The use of playbacks may result in changes in the
responses of the birds being studied. These responses
may influence the probabilities of resighting and recap‑
ture, when playbacks are used. These changes have im‑
plications for studies of avian behavior and ecology when
playbacks are used. In studies that specifically examine
bird song (Ficken and Ficken 1970, Martin 1980, Rich‑
ards 1981, Rigelski and Moldenhauer 1996, Kroodsma
2005), the repeated use of playbacks may have caused
biased responses by the birds which would lead to sub‑
sequent biased analyses. We suggest that the response to
the first exposure to a playback might be the only unbi‑
ased response by territorial forest birds. The use of dif‑
ferent recorded songs for each trial may resolve this issue
and only further testing will confirm this possibility (e.g.,
Dabelsteen and McGregor 1996). In nature, singing in‑
terlopers are usually visible to and interactive with the
territory occupant. Once the bird arrives at the location
of the sound‑source, the playback may no longer elicit
correct or typical responses of the territory owner, since
there is no subsequent behavioral interaction.
Behavioral studies that use playbacks to elicit re‑
sponses must also take into consideration that the biased
“experiment effect” may be important. For example, if
censuses include seasonal (e.g., Yahner and Ross 1995)
or annual (Boscolo et al. 2006) estimates of bird abun‑
dance, then a bias may be expected that decreases with
time since the last capture or playback trial. Due to the
short time interval of this study, it is unclear whether this
bias decreases with time, or whether birds become more
accustomed to playbacks and change their response even
FIGuRe 2: Approach distance (left axis, circles) and Stay Time (right
axis, triangles) for C. melanops compared between captured‑and‑band‑
ed birds (white) and attempt‑to‑capture birds (black), both exposed to
playbacks, showing that despite the small sample size, both treatments
fall within the same range, suggesting that the effect shown in Figure 1
is due to playback and not capture per se.
3The use of playbacks can influence encounters with birds: an experiment
André Magnani Xavier de Lima and James Joseph Roper
Revista Brasileira de Ornitologia, ##(#), ####

further. Many ornithologists recognize that birds may
learn to ignore playbacks. Indeed, there are anecdotal re‑
ports of popular birding locations where birds no longer
respond to playbacks because of their intense and contin‑
ued use (personal communications with several ornithol‑
ogists and online discussion groups, Neoorn‑l archives,
www.museum.lsu.edu/~Remsen/NEOORNintro).
For mark‑and‑recapture studies, this bias has several
implications. First capture probabilities will depend on
whether the birds are naive or experienced with respect
to playbacks (Amstrup et al. 2005). Also, if playbacks are
involved, recapture probabilities should be assumed to be
different from initial capture probabilities. If birds may
be captured or found both with and without using play‑
backs, then the use of playbacks must be considered in
the model. In other words, if other birds show the same
patterns as these, then capture and recapture probabilities
both depend whether playbacks were used, and the ap‑
propriate adjustments must be made in the models (Am‑
strup et  al. 2005). On the other hand, if playbacks are
only used to find individuals and not estimate population
parameters (e.g., Roper 2003, 2005) then the associated
bias will be unimportant.
Similarly, distance methods for estimating popula‑
tion size should not include the use of playbacks due to
the assumptions of the methods (Buckland et al. 2004).
Biases would come from two sources that both violate the
assumptions: experienced birds will remain farther from
the observer, and naive birds will stay longer in the area,
increasing the probability of encounter.
Additionally, playbacks, perhaps surprisingly, may
have a greater influence on behavior than does bird cap‑
ture (Figure 2), suggesting that these effect are due to the
playback itself and not due to the capture of the birds.
Perhaps, from the bird perspective, to be held is less
memorable than an apparent and strange interloper in
the territory. Further study with larger sample sizes will
be needed to adequately test this.
In summary, the use of playbacks for some aspects
of study must be avoided, such as estimating population
size. For understanding behavior associated with songs,
the use of playbacks is more complicated, since response
may only occur after the use of playbacks. Nevertheless,
the use of playbacks and the interpretation of the results
require well‑designed experiments that control for the in‑
troduced bias (examples in Kroodsma 2005).
ACKnoWLedGeMenTs
We would like to thank all field assistants (Flora HML,
Arthur GN and Michel VG) and the Field Ecology 2006 course of
the Graduate Program in Ecology and Conservation of the Federal
University of Paraná, Brazil, for conceptual and moral support with
this experiment. IBAMA/CEMAVE provided licenses for capture
and banding. CAPES provided a fellowship to AMXL. We thank the
Fundação O Boticário para a Proteção de Natureza for the use of Salto
Morato Nature Reserve, where the research took place, and financial
support. Special thanks also to the Reserve manager Paulo Chaves and
two anonymous reviewers.
LITeRATuRe CITed
Amstrup, s. C.; Mcdonald, T. L. and Manly, b. F. J. (2005). (eds.)
Handbook of Capture-Recapture Analysis. Princeton University
Press, Princeton.
bayne, e. M. and hobson, K. A. (2002). Annual survival of adult
American Redstarts and Ovenbird in the southern boreal forest.
Wilson Bull., 114:358‑367.
boscolo, d.; Metzger, J. P. and vielliard, J. M. e. (2006). Efficiency
of playback for assessing the occurrence of five bird species
in Brazilian Atlantic Forest fragments. An. Acad. Bras. Ciênc.,
78:629‑644.
buckland, s. T.; Anderson, d. R.; burnham, K. P.; Laake, J. L.;
borchers, d. L. and Thomas, L. (2004). Introduction to Distance
Sampling, Oxford Press, Oxford. Oxford Press, Oxford.
dabelsteen, T. and McGregor, P. K. (1996). Dynamic acoustic
communication and interactive playback. Pp. 398‑408. In: D. E.
Kroodsma and E. H. Miller (eds.), Ecology and Evolution of Acoustic
Communication in Birds. Cornell Univ. Press, Ithaca, New York.
Ficken, M. s. and Ficken, R. W. (1970). Responses of Four Warbler
Species to Playback of their Two Song Types. Auk, 87:296‑304.
hansen, b. T. and slagsvold, T. (2003). Rival imprinting:
interspecifically cross‑fostered tits defend their territories against
heterospecific intruders. An. Behav., 65:1117‑1123.
Kearns, G. d.; Kwartin, n. b.; brinker, d. F. and haramis, G.
C. (1998). Digital playback and improved trap design enhances
capture of migrant Soras and Virginia Rails. J. Field Ornithol.,
69(3):466‑473.
Kroodsma, d. (2005). The Singing Life of Birds: the Art and Science of
Listening to Birdsong. Houghton Mifflin Company, Boston.
Lima, A. M. X. and Roper, J. J. (in review). Population dynamics
of the Black‑cheeked Gnateater (Conopophaga melanops,
Conopophagidae): facultatively temporary territories and
monogamy. J. Tropical Ecol.
Martin, d. J. (1980). Response by Male Fox Sparrows to Broadcast of
Particular Conspecific Songs. Wilson Bull., 92:21‑32.
Richards, d. G. (1981). Estimation of Distance of Singing
Conspecifics by the Carolina Wren. Auk, 98:127‑133.
Rigelski, d. J. and Moldenhauer, R. R. (1996). Discrimination
Between Regional Song Forms in the Northern Parula. Wilson
Bull., 108:335‑341
Roper, J. J. (2003). Nest‑sites influence predation differently at
natural and experimental nests. Ornitol. Neotrop., 14:1‑14.
Roper, J. J. (2005). Try and try again: Nest Predation Favors
Persistence in a Tropical Bird. Ornitol. Neotrop., 16:253‑262.
simpson, b. s. (1984). Tests of Habituation to Song Repertoires by
Carolina Wrens. Auk, 101:244‑254.
verner, J. and Milligan, M. M. (1971). Responses of Male White‑
crowned Sparrows to Playback of Recorded Songs. Condor,
73:56‑64.
yahner, R. h. and Ross, b. d. (1995). Seasonal response of Wood
Thrushes to taped‑playback songs. Wilson Bull., 107:739‑741.
4 The use of playbacks can influence encounters with birds: an experiment
André Magnani Xavier de Lima and James Joseph Roper
Revista Brasileira de Ornitologia, ##(#), ####