Data and code from: Song complexity in suboscine birds: Evolutionary drivers and ecological constraints
Data files
Apr 14, 2026 version files 3.28 MB
-
B_-_Species_level_data.csv
145.68 KB
-
C_-_Song_level_data.csv
692.69 KB
-
Data_S1_unformatted.xlsx
1.14 MB
-
Data_S1.xlsx
1.19 MB
-
Figure_2.R
4.06 KB
-
Figure_3.R
2.34 KB
-
Figure_4_and_S3__Table_S2-S3.R
7.54 KB
-
Figure_S1.R
4.16 KB
-
Figure_S2.R
1.78 KB
-
PGLS_main_models_song_level.R
3.40 KB
-
PGLS_main_models_species_level.R
8.92 KB
-
README.md
29.98 KB
-
T400F_AOS_Clements_sppnames.tre
53.32 KB
Abstract
Acoustic signal complexity varies widely in animals from single notes to highly sophisticated vocal displays. In birds, vocal complexity can evolve as an honest signal of individual quality driven by sexual selection. However, this hypothesis is rarely explored in conjunction with alternative drivers, including competition for ecological resources (social selection) and intra-group communication, both of which may favour increased signal complexity. Using Bayesian phylogenetic models, we test whether these alternative mechanisms predict the complexity of innate songs in 1,288 species of suboscine passerine birds, while accounting for ecological constraints on sound production, transmission and detection. We found that overall song complexity was reduced by sexual selection (estimated from mating systems), and declined with body size and vegetation density. Conversely, note count and song length increased in territorial species, particularly those using song to defend year-round territories during the non-breeding season. These findings challenge the common assumption that sexual selection is the main driver of increased signal complexity, and highlight the role of social selection via territorial competition as a factor increasing the temporal complexity of songs. Our results suggest that signal complexity depends on social, cultural and ecological contexts, reflecting a combination of multiple inter-related drivers and constraints.
Dataset DOI: 10.5061/dryad.6m905qgdb
Description of the data and file structure
Data and code to reproduce the analyses and figures presented in the manuscript ''Song complexity in suboscine birds: evolutionary drivers and ecological constraints'' (Proceedings B).
Data_S1.xlsx is the full dataset, including:
- the descriptions of all variables used in this study (A - metadata; also provided below),
- the species-level song complexity data and trait data (B - Species level data; also presented in csv format for use in analysis; B_-_Species_level_data.csv),
- the song-level complexity data (C - Song level data; also presented in csv format for use in analysis; C_-_Song_level_data.csv),
- relevant references (D - References; also provided below).
Metadata
| Variable | Description | Source |
| Species | Scientific names for all suboscine species, following names used in the latest species-level phylogeny for suboscines (Harvey et al. 2020) to enable phylogenetic analyses | Harvey et al. 2020 |
| Singing | Yes = species producing advertising songs | Billerman et al. 2022 |
| No = species not producing advertising songs | ||
| unknown = species not known to produce advertising songs or not, due to the lack of literature information and any song recordings | ||
| Signal type | 0 = species normally giving single-note songs | Billerman et al. 2022; xeno-canto.org |
| 1 = species normally giving multi-note songs | ||
| other values indicate that the species give both single- and multi-note songs, higher values mean the species more often produce multi-note songs | ||
| NA = assigned when the singing status of a species is unknown. Note that for unknown species, NA is also assigned to all measurements related to song structure to reflect the unknown status | ||
| No of songs analysed | The total number of songs sampled and analysed for the species | NA |
| Response variables | ||
| Note count | The total number of notes in the song | This study |
| Number of note types | The number of unique note types in the song | |
| Peak frequency SD | The standard deviation of peak frequency of all notes in the song. Peak frequency (Hz) is the frequency where the note reaches its highest amplitude. For one-note songs, the value is set to zero to reflect the absence of inter-note variation. | |
| Bandwidth SD | The standard deviation of bandwidth of all notes in the song. Bandwidth (Hz) is the difference between the two frequency limits spanning 5% and 95% of the energy of each note. For one-note songs, the value is set to zero to reflect the absence of inter-note variation. | |
| Note slope SD | The standard deviation of note slope of all notes in the song. Note slope is the flatness of the note shape, calculated as bandwidth/note length. For one-note songs, the value is set to zero to reflect the absence of inter-note variation. | |
| Frequency variation | The cumulative variation of peak frequency, bandwidth and slope of all notes in the song. For one-note songs, the value is set to zero to reflect the absence of inter-note variation. | |
| Note length SD | The standard deviation of note length of all notes in the song. Note length (ms) is the duration between the begin time and end time of the note. For one-note songs, the value is set to zero to reflect the absence of inter-note variation. | |
| Song duration | The duration (s) between the begin time of the first note and the end time of the last note. For one-note songs, song duration equals to the note length | |
| Note rate | The number of notes given per second. For normal songs, note rate = note count / song duration; for one-note songs, note rate = the number of songs given per second, calculated as the number of songs divided by the duration of all songs present in the recording | |
| Explanatory variables | ||
| Sexual selection score | 0 = Strictly monogamous | Barber et al. 2024 |
| 1 = Predominantly monogamous with infrequent (<5%) EPP in males | ||
| 2 = Monogamous with frequent (5-20%) EPP in males | ||
| 3 = Socially polygamous or EPP >20% in either males and females (i.e., including sex role reversed species) | ||
| 4 = Polygamous with direct competition between individuals (e.g., displaying in leks or arenas) | ||
| Territoriality | 0 = Non-territorial; species that never defend territories, including those that defend very small areas around nest sites, or species where males defend song or display posts only | Tobias et al. 2016 |
| 1 = Seasonal or weak territoriality; species that either defend territories only in the breeding season, or that habitually join mixed flocks with extensive or poorly defined spatial ranges (including those that may defend their position in such flocks) | ||
| 2 = Year-round territoriality; species that defend territories all year, including year-round pair and group territories, and migrantory species that use song to defend territories on both the breeding and non-breeding grounds | ||
| Sociality | Social group size; the maximum number of individuals aggregating and interacting as a pair or group for prolonged periods. In seasonal breeders, the maximum group size is two (paired) individuals; in tropical species with prolonged natal philopatry, maximum group size may increase to 3-4 individuals or more if offspring typically remain on the natal territory for many months. | This study |
| Mass | Body mass given as species average (including both male and female body mass) | Tobias et al. 2022 |
| Relative beak size | Beak size relative to the overall body size, calculated as the relative beak volume: beak culmen length × beak width × beak depth / body mass. Beak culmen length is the length from the tip of the beak to the base of the skull; beak width is the width of the beak at the anterior edge of the nostrils; beak depth is the depth of the beak at the anterior edge of the nostrils. Raw beak measurements and mass data were obtained from AVONET (Tobias et al. 2022). | Tobias et al. 2022; this study |
| Habitat density | 1 = Species primarily lives in desert, grassland, open water, low shrubs, rocky habitats, seashores, cities; also applies to species living mainly on top of forest canopy (i.e. mostly in the open) | Tobias et al. 2022 |
| 2 = Species primarily lives in open shrubland, scattered bushes, parkland, low dry or deciduous forest, thorn forest. | ||
| 3 = Species primarily lives in the lower, middle or upper storey (canopy) of forest, or in dense thickets, dense shrubland etc.; does not apply to species living on top of vegetation or forest canopies (i.e. habitually signalling in an environment uncluttered by foliage) | ||
| Species richness | The number of species (suboscine/ passerine/all birds) which share more than 1% of breeding and resident ranges with the focal species | Harvey et al. 2020 |
| Average peak frequency | The median peak frequency (Hz) of all notes in the song | |
| Average note length | The median note length (ms) of all notes in the song | |
| Other variables | ||
| Behavioural data certainty (of sexual selection, territoriality, sociality) | A = high levels of certainty based on direct evidence published in primary and secondary literature | Barber et al. 2024; Tobias et al. 2016; this study |
| B = moderate to high levels of certainty based on inference and reasonable circumstantial evidence, including reliable field observations suggestive of particular social behaviour, as well as inference from closely related species | ||
| C = low levels of certainty due to conflicting evidence (this includes conflicting published observations and cases where inference from closely related species is complicated by high variation in social behaviours) | ||
| D = no direct or indirect evidence, including inference from closely related species | ||
| Song source | The source of each song file included in this study (e.g. xeno-canto.org) | NA |
| Catalog number | The accession number of each song file or ID number used in the corresponding source | NA |
Data_S1_unformatted.xlsx - unformatted version of Data_S1.xlsx.
Code/software
R code for reproducing all model results, including species level (PGLS_main_models_species_level.R) and song level (PGLS_main_models_song_level.R) models, figures and tables is provided in the .R files with corresponding file names.
T400F_AOS_Clements_sppnames.tre: the phylogenetic tree of suboscines used in this study (Harvey et al. 2020).
References
| Barber RA, Yang J, Yang C, Barker O, Janicke T, Tobias JA. 2024 Climate and ecology predict latitudinal trends in sexual selection inferred from avian mating systems. PLoS Biol. 22, e3002856. (doi:10.1371/JOURNAL.PBIO.3002856) |
| Billerman SM, Keeney BK, Rodewald PG, Schulenberg TS, editors. 2022 Birds of the world. Cornell Lab of Ornithology, Ithaca, NY, USA. See https://birdsoftheworld.org/. |
| Harvey MG et al. 2020 The evolution of a tropical biodiversity hotspot. Science 370, 1343–1348. (doi:10.1126/science.aaz6970) |
| Tobias JA, Cornwallis C, Derryberry E. et al. 2014 Species coexistence and the dynamics of phenotypic evolution in adaptive radiation. Nature 506, 359–363. (doi:10.1038/nature12874) |
| Tobias JA, Sheard C, Seddon N, Meade A, Cotton AJ, Nakagawa S. 2016 Territoriality, social bonds, and the evolution of communal signaling in birds. Front. Ecol. Evol. 4, 74. (doi:10.3389/fevo.2016.00074) |
| Tobias JA et al. 2022 AVONET: morphological, ecological and geographical data for all birds. Ecol. Lett. 25, 581–597. (doi:10.1111/ele.13898) |
