Data from: Seed preference is only weakly linked to seed-type-specific feeding performance in a songbird

Andries, Tim1 ; Müller, Wendt1 ; Van Wassenbergh, Sam1

Research facility: University of Antwerp

Published Mar 22, 2024 on Dryad. https://doi.org/10.5061/dryad.9w0vt4bpc

Data files

Mar 22, 2024 version files 39.42 KB

analysis.zip

11.23 KB
foodpref_raw.csv

8.38 KB
foodpref_summary.csv

15.44 KB
README.md

4.37 KB

Abstract

The dehusking of seeds by granivorous songbirds is a complex process that requires fast, coordinated and sensory-feedback-controlled movements of beak and tongue. Hence, efficient seed handling requires a high degree of sensorimotoric skill and behavioural flexibility, since seeds vary considerably in size, shape and husk structure. To deal with this variability, individuals might specialise on specific seed types, which could result in greater seed handling efficiency of the preferred seed type, but lower efficiency for other seed types. To test this, we assessed seed preferences of canaries (Serinus canaria) through food choice experiments and related these to data of feeding performance, seed handling skills and beak kinematics during feeding on small, spindle-shaped canary seeds and larger, spheroid-shaped hemp seeds. We found great variety in seed preferences among individuals: some had no clear preference, while others almost exclusively fed on hemp seeds, or even prioritized novel seed types (millet seed). Surprisingly, we only observed few and weak effects of seed preference on feeding efficiency. This suggests that either the ability to handle seeds efficiently can be readily applied across various seed types, or alternatively, it may indicate that achieving high levels of seed-specific handling skills does not require extensive practice.

README: Seed preference is only weakly linked to seed-type-specific feeding performance in a songbird

https://doi.org/10.5061/dryad.9w0vt4bpc

This dataset contains all the necessary data for reproducing the results of the research, as described in the manuscript with the name: 'Seed preference is only weakly linked to seed-type-specific feeding performance in a songbird'.

The data includes seed preference data obtained from manual video analysis and data on feeding performance, skills and beak kinematics obtained from Andries et al. (2023).

Description of the data and file structure

This dataset includes 2 .csv files with measurement data and a .zip file named 'analysis' which contains an R-file with the code for statistical analysis and associated .txt files (see below; code/software).

General terminology (this is a list of terms that is used across most of the files):

bird: identification number of the individual
sex: sex of the individual (M = male, F = female)
birth_year: Year of birth of the individual
age_group: Indicates whether the individual is an adult (adult) or juvenile (young)
weight: Body mass of the individual prior to the experiment (in g)
NA: no data available because the individual refused to feed on the respective seed type.

foodpref_raw.csv: This file contains the count data of seeds selected by the individuals, extracted from the video files. Aside from the general terminology, this file contains 9 columns, one for each seed present in the food choice experiment, with the raw counts and a second set of 9 columns with Manly's selectivity index for each seed, indicated as 'seed_name'% (which is in our experimental setup equivalent to the proportion of seeds selected in relation to all seeds selected).

foodpref_summary.csv: This file contains the seed preference variables of interest and the metrics on feeding performance, feeding skills and beak kinematics (taken from Anonymous 2023) during feeding on canary seed and hemp seed.

Terminology:

seed: indicates whether total_duration, success_rate, splitting_seed and frequency are for feeding on either canary seed (canseed) or hemp seed (hemp).
levins_ind: Levin's Niche Breath Index, calculated in R using raw count data (see code/software). Measure for degree of generalism, here in terms of seed selection (number between 0 and 1)
hemp: Manly's Selectivity Index for hemp seed, measure for preference of hemp seed (number between 0 and 1)
millet: Manly's Selectivity Index for millet seed, measure for preference of millet seed. Includes both white and red millet (number between 0 and 1)
total_duration: Average duration of the entire feeding act (in s)
success_rate: Proportion of seed husking attempts that are successful (in %)
splitting_seed: Proportion of successful seed husking attempts where the seed is being split in two halves (in %)
frequency: Average opening-closing frequency of the beak (in Hz)

analysis.zip: contains the following files used for statistical analyses:

food_final.R: R-file with code of the statistical analyses and graph generation
levins_ind.txt: Text file containing raw count data, ready to use in the R-script to calculate Levin's Niche Breadth Index.
food_alt.txt: Text file containing Manly's Selectivity Indexes in a different format, ready to use in the R-script to generate graphs of seed selection.
food.txt: Text file containing the data from foodpref_summary, ready to use in the R-script.

Sharing/Access information

Raw video data is too large to deposit, but can be accessed upon request by contacting the corresponding author Tim Andries (tim.andries@uantwerpen.be)

Data on feeding performance, seed handling skills and beak kinematics is taken from Andries et al. (2023), which can be accessed via:

https://doi.org/10.5061/dryad.1jwstqk09

Code/Software

All statistical analyses were conducted in R (version 4.2.1).

This dataset includes an R-file with the full script for all analyses and generation of graphs. Additionally, .txt-files of the relevant data are included in this dataset, ready-to-use in the R-file.

The required packages are: 'MicroNiche' for calculating Levin's Niche Breadth Index, 'lme4' and 'lmerTest' for linear mixed modelling and 'ggplot2' for graph generation.