Individuality obscures geographic song variation in two passerine sister species with contrasting distributions and movement ecology

Data and sound files underlying the study

Oñate-Casado J., Pradeep Sakhalkar S., Porteš M., Beran V., Pérez-Granados C., Traba J., Barrero A., Petrusek A., Petrusková T. (2025): Individuality obscures geographic song variation in two passerine sister species with contrasting distributions and movement ecology. Ibis, doi: 10.1111/ibi.13429

Data are provided in either Microsoft Excels (XLSX) format or, in case of large datasets, as comma separated text files (CSV) with dot (".") as a decimal separator. If relevant, XLSX files contain a sheet "legend" with description of the columns.

Relevant data files and archives are provided separately for species, BP denotes Berthelot's Pipit (Anthus berthelotii), TP denotes Tawny Pipit (Anthus campestris). Studied localities are occasionally abbreviated by two-letter codes, and the male origin and species is also indicated in their codes, as follows:

Tawny Pipits (Ac): 
Agios Mamas, Greece (AM, AcAM in male code)
Altos de Barahona, Spain (BR, AcB)
Gran Sasso, Italy (GS, AcGS)
Most, Czechia (CZ, AcCZ)
Mount Parnassus, Greece (MP, AcMP)
Rincón de Ademuz, Spain (AD, AcA)

Berthelot's Pipits (Ab)
Fuerteventura, Canaries (FV, AbF)
Tenerife, Canaries (TF, AbTE)
La Palma, Canaries (LP, AbLP)
Madeira (MA, AbM)
Porto Santo, Madeira (PS, AbPS)

Basic information about analysed birds:

Supplementary Table S1.xlsx

A XLSX table, identical to Supplementary Table S1 in the manuscript, contains male codes (used consistently in the remaining data files), geographic origin of birds (including coordinates of recording positions in WGS84 decimal degree format, with negative longitude values indicating the western hemisphere), recording year, number of song types in male's repertoire (1 or 2), and indication which males were used to assess intra-individual variation of song for Berthelot's Pipits, and songs of which males were used to create spectrograms in Fig. 2 of the study. 

===============================

Song dissimilarity matrices

BP_dissimilarity.csv
TP_dissimilarity.csv
BP_3song_dissimilarity.csv
Luscinia_DTW_settings_screenshot.png

Pairwise song dissimilarity, as generated by the Dynamic Time Warp analysis implemented in Luscinia 2.24.09.12.02 software (http://rflachlan.github.io/Luscinia/), with the following settings (as exported from Luscinia after the analysis):
Time: 5.0; Fundamental frequency: 1.0; Fundamental frequency change: 1.0; FF norm: 1.0; Weight by amplitude: true; Log frequencies: true; Normalise with sd: true; Interpolate time warping: true; Dynamic warping: 1; Use squared dissimilarities: false; Stitch syllables: 1; Compression factor: 0.25; Minimum number of points: 2; SD ratio: 1.0; SD ratio 2: 0.0; Alignment cost: 0.2; ArcTan transformation factor: 0.02; Stitch threshold: 100.0; Number of alignment points: 4*; Syllable repetition weighting: 0.0; Maximum warp: 100.0. [* The option selected in Luscinia prior to analysis was "Alignment points: Five points".] The screenshot of all selected options is provided in Luscinia_DTW_settings_screenshot.png.

BP_dissimilarity.csv and TP_dissimilarity.csv contain pairwise comparison of all song types analysed for Berthelot's and Tawny PIpits, respectively. BP_3song_dissimilarity.csv contains the dissimilarity values for a subset of 50 Berthelot's Pipits (25 from Fuerteventura and 25 from Tenerife), which were represented by three different instances of their song types to evaluate variation of song delivery by the same singing individual. 

Matrices are provided as a list of unique pairwise comparisons, including the self-comparisons (with zero dissimilarity values). 

===============================

Geographic distance matrices

BP_distance.csv
TP_distance.csv

Pairwise geographic distance matrices between recording points (coordinates provided in Supplementary Table S1) in meters, calculated from the coordinates of the recording position by the ‘distHaversine’ function from the geosphere R package (Hijmans 2024; https://github.com/rspatial/geosphere). The structure of the file and the arrangements of lines fully corresponds to song dissimilarity matrices described above (a list of unique pairwise comparisons, including the self-comparisons with zero distance). The distances between two song types of the same male were also set to 0. Note that for analyses of log-transformed distances, a formula log(distance+1) was used. 

===============================

Non-metric Multidimensional Scaling results

TP_nMDS_all_results.xlsx
BP_nMDS_all_results.xlsx

Results of non-metric Multidimensional Scaling implemented  implemented in Luscinia 2.24.09.12.02 software, based on the song dissimilarity values ontained from the Dynamic Time Warp analysis characterised above. In the first sheet, coordinates of all analysed song types in an ordination space are provided. In the second sheet, cumulative explained variation, eigenvalues and variation explained by each MDS dimension (up to 80% of total explained variation) are provided. (For algorithm implemented to quantify explained variation, see Luscinia documentation).

===============================

Dendrogram source files

BP_UPGMA_tree.tre
TP_UPGMA_tree.tre
BP_3song_UPGMA_tree.tre

UPGMA dendrograms generated from the song type pairwise dissimilarity matrices (described above), provided in the Newick tree format. Branches are labelled by the male code. In case of the tree characterising wthin-individual variation of song delivery, the labels contain both the male code and the name of the corresponding sound (WAV) file.

===============================

Song acoustic features

BP_song_features.xlsx
TP_song_features.xlsx

A dataset of 16 quantitative song features based on temporal and frequency characteristics, based on measurements within Luscinia, used to test for potential differentiaiton between studied pipit populations. These include both characteristics of the whole song (duration, number of elements and element rate, minimal and peak frequency, frequency bandwith) and some summary characterics of the elements within songs (minimal, maximal and mean values for duration, bandwith and peak frequency). Detailed explanation for all columns are provided in the "legend"sheet.

===============================

Source sound files

BP_WAVs_geograpic_scale.zip
TP_WAVs_geograpic_scale.zip

ZIP archives containing sound files (PCM format, 16 bit, mono, WAV file type) of representative instances of song types analysed in the study by Luscinia software. Each song type is represented once, as for analyses of among-individual and spatial variation. Songs of each studied species are included in separate ZIP files, their file names correspond to individual codes (as provided in Supplementary Table S1). Different song types of males that used two song types in their repertoire are further labelled by "(Song A)" and "(Song B)".

BP_WAVs_individual_scale.zip

A ZIP archive containing sound files (PCM format, 16 bit, mono, WAV file type) of three different instances of a song types sung by the same Berthelot's Pipit individuals,  analysed in the study to evaluate variation of song delivery by the same singing individual. 25 individuals from Fuerteventura and 25 males from Tenerife were analysed for this purpose. Files are arranges in folders labelled by individuals' codes, and file names correspond to original labels used in the Luscinia analyses.

===============================

Overview spectrograms of song types

BP_spectrograms.zip
TP_spectrograms.zip

ZIP archives containing spectrograms (PNG format, greyscale) visualizing the song types used in analyses of among-individual and spatial variation. These spectrograms are for overview only, as they were batch-generated by an R script, using "spectro" function of the R package "seewave", and their parameters were not adjusted individually to optimise singal to noise ratio. Before generating spectrogram, a band-pass filter was applied, removing noice below 2 kHz and over 8 kHz. 
Spectrograms from each studied species are included in separate ZIP files, Names of PNG files correspond to those of sound filedsdescribed above. Names refer to individual codes as provided in Supplementary Table S1; different song types of males that used two song types in their repertoire are further labelled by "(Song A)" and "(Song B)".