Behavioral Ecology Vol. 13 No. 1: 65-69
© 2002 International Society for Behavioral Ecology
The development of coordinated singing in cooperatively displaying long-tailed manakins
a Department of Biology, University of Northern Iowa, Cedar Falls, IA 50614, USA b Department of Zoology and Physiology, University of Wyoming, Laramie, WY 82071-3166, USA c Department of Computer Science, University of Northern Iowa, Cedar Falls, IA 50614, USA
Address correspondence to J.M. Trainer. E-mail: jill.trainer{at}uni.edu .
Received 14 July 2000; revised 15 February 2001; accepted 20 March 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Long-tailed manakins (Chiroxiphia linearis) have a puzzling social system in which teams of two males display cooperatively in dispersed lek arenas, but only the alpha partner mates with visiting females. One benefit of performing as a nonmating partner might be to gain experience as an "apprentice" to improve the performance of the complex duet song and joint dance. We examined the relationship between the age of singers and two measures of singing performance: song variability and sound frequency matching. Singing performance improved with age; variability in four song characteristics of males less than 3 years old was greater than that in their older partners, and frequency matching increased with the age of the younger partner. Randomization tests of song samples from seven well-established teams showed that males did not track the song-to-song variation in their partners' singing. Another randomization test showed that frequency matching by these teams was higher than that of randomly paired partners. We considered three alternative hypotheses for the congruent songs: (1) short-term accommodation to the partner's song; (2) active choice of partners with similar intrinsic frequencies; and (3) long-term development of congruent song through either practice or song copying. Our results and evidence from long-term monitoring of banded birds best support the hypothesis that frequency matching develops over several years during the complex and protracted process of partner formation. Nonmating males may benefit from increasing their competence at display, eventually enjoying increased mating success when they inherit display sites from older males.
Key words: age, Chiroxiphia, cooperation, learning, lek, manakin, Pipridae, sexual selection, song, suboscine, vocal development.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Long-tailed manakins (Chiroxiphia linearis) have an unusual cooperative courtship display in which two males perform a song and dance in dispersed lek arenas (Bradbury, 1981
) that we call "perch-zones," visited by lone
females or females in small groups (Foster,
1977,
1985). Each lek contingent
(defined as the males associated with a perch-zone) comprises a linear
dominance hierarchy of three to 15 males of varying ages. The two top-ranking
males (alpha and beta) perform the great majority of the duet song and dance
courtship displays. Only alpha males mate with visiting females, while the
subordinate, beta partners may spend as much as 4 to 10 years in the
supporting role, waiting for the opportunity to move up in rank and become
potential copulators (McDonald,
1993a). To understand this puzzling example of cooperation, we
need to know how the beta male benefits from the partnership. McDonald and
Potts (1994) ruled out an
indirect inclusive fitness benefit by showing that relatedness among display
partners is not higher than the ambient level in the population. Foster
(1981) suggested that one of
the direct, though delayed, benefits of performing as a beta male might be to
gain experience as an "apprentice" to improve the performance of
the complex duet song and joint dance. Subsequent studies showed that females
tended to visit and mate with males that are highly competent performers of
the song and dance displays (McDonald,
1989b; Trainer and McDonald,
1995). Therefore, the potential exists for beta males to increase
their eventual reproductive success by "apprenticing," providing
that their display performance improves during their tenure as a noncopulating
partner. Until now, we lacked information on the development of the manakin
courtship display.
Courtship display in long-tailed manakins is hierarchical, with duet song
acting as a long-range attractant for visitation, and dual-male dance acting
to induce females that have chosen to visit a perch to remain for copulations.
Only one team of males displays in a given perch-zone at any one time
(McDonald, 1989a). In unison,
two males (usually the alpha and beta) perform the duet song display
(phonetically resembling the word toledo)
(Figure 1), which advertises
their presence and attracts females to visit the perch-zone
(Trainer and McDonald, 1993).
As soon as a female arrives in the vicinity of a display perch, the males
begin to perform a joint, backwards leapfrog dance. Copulations occur only
after one or more lengthy bouts of dancing
(McDonald, 1989b). Hence,
singing plays an important role in the first, long-range step of mate choice:
selecting which perch-zone to visit. Our previous work showed that teams of
males whose individual toledo components were well matched in sound frequency
received more visits from females (Trainer
and McDonald, 1995).
|
Acquisition of social status involves plumage sequences and interactions
among males that develop over a period of as much as 10 years. Three distinct
predefinitive plumages precede the definitive plumage, acquired at 4 years of
age. McDonald (1993a) provided
evidence that the predefinitive plumages function as honest status signals
directed at other young males in queuing for dominance rank in lek
contingents. Predefinitive and young definitive males interact with members of
two to six different lek contingents simultaneously, traveling among
perch-zones to interact with various partners at each. The number of lek
contingents with which a male associates decreases as his status increases.
Eventually, some males work their way up to the alpha and beta positions,
comprising the principal display team. By the time a male reaches beta status
at approximately 7 years of age, he usually interacts at no more than two
perch-zones. Males attain alpha status at approximately 10 years of age, and
almost always interact and display at only one perch-zone. Thus, alpha-beta
partnerships take many years to consolidate, gradually becoming more stable as
advancing males display with fewer and fewer partners. New opportunities arise
when higher-ranking males disappear, which occurs at the rate of nine to 23%
per year for definitive males. Because of the high survival rates of
top-ranking males, only approximately 8% of first-year males will actually
become alpha males with the potential of copulating
(McDonald, 1989a). Thus, lek
contingents are fluid, with memberships (especially at the lower ranks)
changing from year to year, and the network of social interactions is arguably
as complex as that found in any non-human society.
The long and complex route to acquisition of social status in lek
contingents entails deferring breeding for 6 to 10 years
(McDonald, 1993a;
McDonald and Potts, 1994).
Young males have almost no chance of breeding, since females appear not to be
attracted to songs in which a male in predefinitive plumage participates, and
rarely to watch dances involving predefinitive males. Even after a male
attains definitive plumage, females will mate only with the older, dominant
alpha. Deferring reproduction for such a long period of time would select for
long-term behavioral strategies that increase a male's personal reproductive
success, the benefits of which do not accrue until at least 6 years of age
(McDonald, 1993a;
McDonald and Potts, 1994).
The development of display competence may be a long-term strategy of great
potential importance for reproductive success in male long-tailed manakins.
Singing competence in this species is unique because it involves an unusual
dual phenotype in which frequency matching is dependent on the performance of
two individuals. Well-executed performance of the song confers a more
immediate benefit on the alpha male, which has a chance to mate with females
attracted to the dual song. Nevertheless, beta males tend to inherit the alpha
position when the older males disappears, and females exhibit site fidelity,
continuing to visit favorite perch-zones even after a change in alpha male
(McDonald and Potts, 1994).
Thus, beta males also benefit, albeit in the long run, from performing
frequency matched songs with their partners. The development of singing
competence may be a consequence of prolonged partnership formation that
benefits males over the long term.
Here we investigate development of the singing display used to attract females to perches where the dance display and copulation occur. We quantify singing competence using two parameters. First, we estimate the singing consistency of each male from the generalized variance in four song variables describing his contribution. We consider males with less variability in their singing to be more competent singers. Second, we develop an index of frequency matching (IFM) as the congruence between partners in the frequencies of three song elements. We use this quantitative information to test whether frequency matching and singing consistency improve with age. In addition, we determine whether the songs of established alpha-beta teams of definitive males are more congruent than expected by chance.
We test three alternative hypotheses for the high congruence of songs by partners in well-established teams: (1) a short-term accommodation hypothesis under which partners "track" each other's song contributions during each round of singing; (2) an active choice of partner hypothesis, whereby males gravitate toward other males with similar intrinsic frequencies as preferred singing partners; and (3) a long-term developmental hypothesis, under which partners converge on congruent song by practice or song copying.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The study area, in Monteverde, Costa Rica (10° 18' N, 84° 48' W), is 80 ha of premontane tropical moist forest at an elevation of 1300 m. We recorded toledo songs as teams of two males sang in the subcanopy, usually at a distance of 8 to 10 m. Duetting males perch 10 to 15 cm apart, making it easier to age both partners, and in many cases to identify them individually by their color bands. Using a database of age-specific information for color-banded birds, we aged males using the plumage sequences described in McDonald (1993a
).
Each of 14 males in seven teams of definitive males that formed the crux of
the recording effort was color-banded—we will call these the
"focal" teams. The partnership patterns, singing, courting, and
mating behavior of these focal teams were well documented (McDonald,
1989a,b,
1993a,b;
McDonald and Potts, 1994;
Trainer and McDonald, 1995).
The focal teams received nearly all of the copulations and most of the visits
by females over the 2 years of the present study. In addition, we recorded 15
partner combinations that contained a banded or unbanded definitive male and
an unbanded predefinitive male. The 15 predefinitive males in these
combinations were evenly divided among three age classes (1, 2, and 3 years
old). Each individual male is represented in only one of the 22 partner
combinations described above. For each team we analyzed 10 consecutive songs from a single song bout recorded during June or July of the breeding season. We tested whether a consecutive sample was representative of a team's singing over these months by comparing the consecutive sample at perch-zone Z with a temporal sample of the same team over several days. Due to the difficulty of obtaining song samples in which the color-bands of both males were identified during a recording session, this was the only pair of focal males for which repeated samples existed. The frequency characteristics of 10 consecutive songs at perch-zone Z did not differ significantly from a sample of seven songs by the same team, each recorded on a different day between 12 June and 3 July (MANOVA; Wilks's Lambda = 0.54; F(6,10) = 1.42; p =.30). Visual inspection of sonagrams of two teams recorded on two or three different days showed their songs also varied little over time. We assumed, therefore, that samples of consecutive songs were representative of the rather stereotyped singing throughout the duration of the breeding season.
We recorded songs with a Sony WM D6C cassette recorder and a Sennheiser ME 80 directional microphone, and analyzed them using a Kay Elemetrics Model 5500 Sonagraph with a gray scale printer. The two males of a team sing almost identical song contributions of approximately 0.6 s duration, with one male commencing approximately 0.1 s after the first male begins (Figure 1). We were able to identify the singer of each contribution to the toledo songs based on slight differences in the frequency characteristics of partners' contributions. Additionally, one of the two males would sometimes give a small number of solitary toledos, allowing us to verify that male's contribution to the duet songs.
We used the techniques developed by Trainer and McDonald
(1995) to measure songs
quantitatively. In sonagrams, each male's contribution appears roughly
J-shaped (Figure 1). For each
male, we measured the minimum frequencies occurring at two nodes of high
amplitude in the bottom of the J (A and B of
Figure 1), the maximum
frequency at the top of the J (C of Figure
1), and the time interval between the onset of the song and the
stem of the J (D of Figure 1).
To determine how consistently each male sang his contribution to the duet
song, we calculated the generalized variance in the above four song parameters
for a sample of 10 songs from each male. This multivariate measure was the log
of the determinant of the variance/covariance matrix of the four parameters
(Sneath and Sokal, 1973). To
give each parameter approximately equivalent weight, we multiplied the time
interval measurements (ms) by 10 so that their magnitude would be comparable
to those of the frequency measurements (Hz). Using a two-way ANOVA of age
class and partner status, we compared the generalized variance in the song
parameters of males of different ages.
For each song we measured an index of frequency matching (IFM). We calculated the IFM by taking the sum of the disparities between the two males' song contribution in the three minimum and maximum frequency variables. We then subtracted this sum from 200, so that the magnitude of the IFM would increase as the disparities in frequencies decreased. Using a repeated-measures, one-way ANOVA with teams nested within age class, we compared the IFM among plumage classes to see if it increased with the age of the younger partner.
Short-term accommodation
We performed randomization tests to determine whether alpha and beta males
in the focal teams tracked one another's songs
(Manly, 1997;
Smith et al., 1990). The song
contributions of each partner in a team showed slight variation from one song
to the next. If partners tracked one another, then their song contributions
should have varied together. Therefore, the frequency matching in the observed
sample of 10 songs was expected to be higher than that in artificial samples
containing rearranged duet songs. For example, an alpha male's contribution to
the first song might be compared to the beta male's contribution to the fifth
song in an artificial sample. For each of seven focal teams, we used an Ada
program on a PC to generate 999 random permutations of rearranged songs and
calculated the mean IFM for these artificial samples. We compared the mean IFM
of the actual sample to the distribution of the randomly generated mean IFMs.
The significance level is equal to the number of artificial samples with a
mean IFM equal to or greater than the mean IFM of the actual sample, divided
by 1000 (the number of random samples plus the actual sample;
Manly, 1997).
Nonrandom frequency matching by partners
We tested the hypothesis that the IFM of the focal teams was higher than
expected by chance by performing a randomization test. We generated randomly
formed assemblages, each containing seven artificial teams. We formed
artificial teams by repairing alpha males with non-partner beta males. We
calculated an IFM for each artificial team using the means of the three song
variables in the samples for each male in a team. The assemblages were random
permutations of the artificially paired teams. We generated 999 random
permutations, and calculated the mean IFM for each permuted assemblage of
seven artificial teams. We performed a significance test by comparing the mean
IFM of the seven focal teams with the distribution of mean IFMs of the random
assemblages. The significance level is equal to the number of mean IFMs that
are equal to or greater than the observed IFM divided by 1000 (the number of
random assemblages plus the one focal assemblage;
Manly, 1997).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The song contributions of predefinitive, non-alpha partners were significantly more variable than those of their definitive, alpha partners (two-way ANOVA; F = 6.93; p =.012; Figure 2). However, the differences in variability among predefinitive males of different age classes were not significant (F = 0.88; p =.462). We consider more variable singing to indicate less competent performance.
|
ANOVA showed that the IFM differed significantly among four age classes to which the non-alpha partner belonged, and increased with the age of the predefinitive partner (one-way ANOVA; F = 103.10; p =.0001; Figure 3). The differences among teams of males within age classes were also significant (F = 39.14; p =.00001). Tukey's Studentized Range Test revealed that all comparisons were significant at the.05 level.
|
.05). There were five teams in age classes 1, 2, and 3,
and seven teams in age class 4.We found no evidence that members of a team track one another's songs. The mean IFMs of the seven focal teams' 10 field-recorded songs could not be distinguished from those in randomized samples of rearranged songs. The p-values for the comparisons between the seven focal teams' observed simultaneous samples versus randomly matched non-simultaneous samples were.49,.08,.56,.21,.91,.14, and.67.
The randomization test showed that frequency matching by the seven focal teams was significantly higher than that of the artificially constructed, random assemblages. The mean IFM of the seven focal teams was greater than all but 33 of the values in the distribution of IFMs for the 999 artificial assemblages (p =.033; Figure 4).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our results clearly demonstrate (1) that singing competence, including consistent singing and frequency matching, increased as a function of the age of the subordinate partner and (2) that frequency matching by established teams was greater than the random expectation. Because frequency-matched songs attract more females (Trainer and McDonald, 1995
), any process that improves frequency matching
would enhance the mating success of males. We consider three alternative
hypotheses that could account for our results on change in performance with
age, and the increased congruence of partnerships: short-term accommodation
between singing partners, active choice of partners with similar intrinsic
frequencies, and long-term development of frequency matching.
Short-term accommodation hypothesis
We found no evidence that congruence in partners' songs results from males
making short-term accommodations to their partners' songs, adjusting the sound
frequencies as they sing. If this were true, then partners would be expected
to track subtle variation in one another's songs, matching the frequencies of
each song. That did not occur. The randomization tests showed that song
contributions that actually co-occurred were no more similar than randomly
paired, non-simultaneous contributions. Observations of color-banded males
provide further counter-evidence. Under the short-term accommodation
hypothesis, males would also be expected to change the frequency
characteristics of their songs when singing with different partners. A
color-banded alpha male sang with two different partners at the same perch on
the same day. Inspection of sonograms and comparisons of frequency
measurements showed that the alpha male did not alter his song contribution to
match the frequencies of the two different partners. Similarly, a color-banded
beta male sang with two different alpha males at different perch-zones one
week apart, but did not alter his song contribution. Long-tailed manakin males
do not, therefore, appear to make short-term accommodations to produce
congruence with partners' songs.
Active choice of partners hypothesis
Another possible cause of congruence in songs would be males choosing
partners with songs similar to their own. McDonald's
(1989a,
1993a) observations, however,
suggest that social history and patterns of male mortality in different lek
contingents are the most important determinants of which partners end up
together. Rather than commit to a single partner early on, males interact with
many partners in several different lek contingents, narrowing the range of
partners and lek contingents over a period of several years. This should allow
them to advance in the most opportune queue when a higher-ranking male
unpredictably disappears. Such a long, fluid process of establishing an
alpha-beta partnership is unlikely to result in partnerships between males
based primarily on active choice of those with intrinsically similar song
frequencies.
Long-term developmental hypothesis
The final alternative is a long-term developmental hypothesis, under which
males gradually accommodate to their partners' frequencies. Consistent with
this hypothesis is our observation that predefinitive males gained competence
at dual singing as they aged; variability in song structure decreased and
frequency matching increased with age in predefinitive males.
In addition, we found that the song contributions of definitive members of established teams were congruent; that is, they were more similar to each other than expected, given the song variation in the population. Under the developmental hypothesis, congruence may develop over time as males spend time singing together. Males may make long-term adjustments to more closely match their partners' songs. It seems most likely that the younger, beta male would match the alpha, but alphas matching betas and mutual convergence of songs would also produce congruence in partners' songs. If congruence develops over time, then frequency matching in a new team should increase over the years as the partnership develops. Again, this is supported by our observations that singing variability decreased with age and that frequency matching of teams increased with the age of the predefinitive partner.
Frequency matching could develop in long-tailed manakins either by imitation of adult males early in life or by practice, over time, as partners tune their song contributions to one another. The long tenure of subordinate males as non-copulators would provide ample opportunity to perfect song through practice. Direct observations of vocal development in color-banded long-tailed manakins are needed to discover how songs are initially acquired and how and at what ages they may become modified with practice as partnerships develop.
Whereas vocal learning is widespread and well documented in the songbirds
(Kroodsma, 1988;
Kroodsma and Baylis, 1982), its
occurrence has only recently been suggested in the closely related taxon, the
suboscines, which include manakins. Kroodsma (unpublished data) has found
perhaps the best evidence of vocal learning in a suboscine, the three-wattled
bell-bird, Procnias tricarunculata. Several young males in a contact
zone between two song dialects have been observed singing both dialect
variants. In addition, bellbird songs appear to change over time, similarly to
many songbirds that exhibit vocal learning. These observations contrast with
earlier studies of three species of flycatchers, the suboscines most closely
related to manakins, which showed that normal songs develop in socially
isolated young flycatcher males without imitation of a taped model song
(Kroodsma, 1984,
1989). Further studies of the
relatively poorly known suboscines may reveal other examples of song
learning.
Our long-term developmental hypothesis has implications for understanding
the selective advantages that maintain the unusual system of cooperation in
long-tailed manakins. Present success of the alpha partner is dependent on the
cooperation, since females will mate only after observing one or more
prolonged dual dance displays, and they avoid perch-zones where overt
aggression among males occurs in the presence of females
(McDonald, 1993a). Whereas the
benefits of cooperative display to the copulating male are clear and present,
the benefits to his partner are not.
While ruling out indirect benefits, McDonald and Potts
(1994) found evidence for
three direct benefits of cooperative display to the subordinate males: an
older beta male may rarely copulate when the alpha male is absent temporarily,
subordinate males may eventually succeed to alpha status, and females usually
remain faithful to a perch-zone after alpha male turnover. Our results provide
support for a fourth benefit: subordinate males increase their competence at
display while they help to attract mates for their older partners. Since
well-performed displays increase the success of the partnership at attracting
females to the dance perch and stimulating females to mate, the beta male
whose singing improves with time should eventually enjoy increased mating
success when he inherits the perch-zone and its suite of site-faithful
females.
|
ACKNOWLEDGEMENTS |
|---|
We dedicate this article to the memory of Luis F. Baptista, whose friendship, encouragement, and inspiration continue to nurture those of us interested in bird song. This work was funded by the Graduate College, University of Northern Iowa and National Science Foundation grant BNS-8814038 to J.M.T. D.B.M. is grateful for long-term support from the Earthwatch Institute and its volunteers. Members of the Monteverde community helped in many ways, especially Joe and Jean Stuckey, who graciously permitted the use of their farm as a study site. We thank the people and government of Costa Rica for hospitality and provision of necessary permits. We are grateful to David Nutter for his assistance with the fieldwork, Mark Ecker for help with the statistical analysis, and Donald E. Kroodsma and an anonymous reviewer for comments on the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bradbury JW, 1981. The evolution of leks. In: Natural selection and social behavior: recent research and new theory (Alexander RD, Tinkle DW, eds). New York: Chiron; 138-169.
Foster MS, 1977. Odd couples in manakins: a study of social organization and cooperative breeding in Chiroxiphia caudata. Am Nat 111: 845-853.
Foster MS, 1981. Cooperative behavior and social organization of the swallow-tailed manakin (Chiroxiphia caudata). Behav Ecol Sociobiol 9: 167-177.
Foster MS, 1985. Pre-nesting cooperation in birds: another form of helping behavior. In: Neotropical ornithology (Foster MS, Morton ES, Ridgely RS, Buckley FG, eds). Ornith Monog 36: 817-828.
Kroodsma DE, 1984. Songs of the alder flycatcher (Empidonax alnorum) and willow flycatcher (Empidonax traillii) are innate. Auk 101: 13-24.
Kroodsma DE, 1988. Contrasting styles of song development and their consequences among the Passeriformes. In: Evolution and learning (Bolles RC, Beecher MD, eds). Mahwah, New Jersey: Erlbaum; 57-184.
Kroodsma DE, 1989. Male eastern phoebes (Sayornis phoebe, Tyrannidae, Passeriformes) fail to imitate songs. J Comp Psychol 103: 227-232.
Kroodsma DE, Baylis JR, 1982. A world survey of evidence for vocal learning. In: Acoustic communication in birds, vol. 2 (Kroodsma DE, Miller EH, eds). New York: Academic Press; 311-337.
Manly BGJ, 1997. Randomization, bootstrap, and Monte Carlo methods in biology. New York: Chapman & Hall.
McDonald DB, 1989a. Cooperation under sexual selection: age-graded changes in a lekking bird. Am Nat 134: 709-730.
McDonald DB, 1989b. Correlates of male mating success in a lekking bird with male-male cooperation. Anim Behav 37: 1007-1022.
McDonald DB, 1993a. Delayed plumage maturation and orderly queues for status: a manakin mannequin experiment. Ethology 94: 31-45.
McDonald DB, 1993b. Demographic consequences of sexual
selection in the long-tailed manakin. Behav Ecol
4: 297-309.
McDonald DB, Potts WK, 1994. Cooperative display and
relatedness among males in a lek-mating bird. Science
266: 1030-1032.
Smith EP, Pontasch KW, Cairns J Jr, 1990. Community similarity and the analysis of multispecies environmental data: a unified statistical approach. Wat Res 24: 507-514.
Sneath PHA, Sokal RR, 1973. Numerical taxonomy. San Francisco: WH Freeman.
Trainer JM, McDonald DB, 1993. Vocal repertoire of the long-tailed manakin and its relation to male-male cooperation. Condor 95: 769-781.
Trainer JM, McDonald DB, 1995. Singing performance, frequency matching and courtship success of long-tailed manakins (Chiroxiphia linearis). Behav Ecol Sociobiol 37: 249-254.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. L. Patricelli and A. H. Krakauer Tactical allocation of effort among multiple signals in sage grouse: an experiment with a robotic female Behav. Ecol., January 1, 2010; 21(1): 97 - 106. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||