Skip to main content

Evidence of repertoire sharing and stability despite a high turnover rate in a duetting neotropical wren

Cite this dataset

Quirós-Guerrero, Esmeralda; Janeiro, Maria Joao; Cresswell, Will; Templeton, Christopher (2020). Evidence of repertoire sharing and stability despite a high turnover rate in a duetting neotropical wren [Dataset]. Dryad.


In songbirds, the spatial pattern of song sharing among individuals is influenced by the song learning and dispersal strategies within each species. In birds where females and males sing and create joint acoustic displays (duets), the processes defining the patterns of song sharing become more complex as there might be different selection pressures shaping the behaviour of each sex. To provide further insight into the vocal development and the dispersal strategy of duetting tropical species, we investigated the patterns of individual and pair repertoire sharing, as well as the stability of these repertoires, in a colour-marked population of riverside wrens, Cantorchilus semibadius, located in the Osa Peninsula, Costa Rica. Using data collected over a five-year period, we found considerable variation in the sharing levels of phrase and duet type repertoires among neighbouring individuals coupled with a general decline of repertoire sharing as distance increased between birds’ territories. These results are consistent with a pattern predicted in age-restricted learners that establish preferentially near their tutors. Furthermore, we found no evidence of individuals changing their phrase type repertoires over time, including after remating events. Duet type repertoires were also stable when pairs remained together. However, we observed a surprisingly high turnover rate. When individuals remated, even though the majority of the previous duet type repertoire remained, several new duet types were included. Taken together, our findings suggest that riverside wrens might create their individual repertoires by copying their same-sex parent and neighbouring individuals before dispersal. Additionally, we speculate that even though birds were able to create new duet types after changing partners, a substantial portion of their duet type repertoire might also be copied from their parents and neighbouring pairs during the initial critical period of song learning.


We conducted observations and recordings of 37 pairs (61 individuals) in 23 territories within a population of riverside wrens, Cantorchilus semibadius, for four years during a five-year period (2013, 2015-2017) in Osa Peninsula, Costa Rica. From our recordings, we isolated all phrase and duet types using Syrinx and created sound spectrograms and a library of phrase types for each individual and a library of duet types for each pair. The ID of each pair includes the name of the territory and the year when the pair was recorded in that territory. The duet types present in the repertoire of each pair are composed by a female phrase type and a male phrase type. The first female and male phrase types and the first duet type recorded were categorized as following: PF1, PM1, and PF1PM1 respectively.

To investigate the spatial patterns of song sharing, we performed the following analyses: first, we quantified repertoire sharing by using the song sharing index, = 2Ns/(2Ns a + b), in which Ns is the number of shared phrase/duet types between two individuals/pairs and a and b are the non-shared phrase/duet types of each individual/pair (McGregor and Krebs 1982). Then, we performed permutation-based Mantel tests, which assess the correlation between matrices of song sharing (song sharing coefficient, S) across territories and the corresponding matrices of the geographical distance between them. Distances were measured in meters between territories’ centres. We only included the three years in which we had 10 or more territories with complete or near complete repertoires of phrase and duet types (2015, 2016, 2017). 

To investigate repertoire stability in pairs that remained together across years and in individuals that changed mates, we used the Chao-1 index estimates of phrase and duet types richness (number of different phrase/duet types in a repertoire). Reference years and pairs were defined for comparison purposes, to test stability across years and across territories, respectively. For these references, asymptotic estimators of phrase/duet type richness (repertoire sizes), as well as the corresponding 95% confidence intervals, were obtained. For the remaining years/pairs, i.e. those compared to the reference years/pairs, the number of different phrase types and duet types recorded that were not recorded in the reference were registered. Additions to the estimated reference repertoire that exceeded the upper 95% confidence interval of the Chao index were considered to be evidence of a change in individuals and pairs repertoires and, therefore, lack of stability. We only included pairs for which we had nearly complete repertoires. As a result, 11 pairs were considered in the analysis for individuals that remained together in the same territory and 8 pairs were considered in the analysis for individuals that changed partners. 


Consejo Nacional de Humanidades, Ciencias y Tecnologías, Award: 381393

Fundação para a Ciência e Tecnologia, Award: SFRH/BD/96078/2013

NERC Environmental Bioinformatics Centre, Award: NE/J018694/1

Royal Society, Award: RG2012R2

M J Murdock Charitable Trust, Award: 2014199

Consejo Nacional de Humanidades, Ciencias y Tecnologías, Award: 327118