Skip to main content

Social partners and temperature jointly affect morning foraging activity of small birds in winter

Cite this dataset

Madsen, Anastasia; Vander Meiden, Laura; Shizuka, Daizaburo (2020). Social partners and temperature jointly affect morning foraging activity of small birds in winter [Dataset]. Dryad.


Daily foraging activity of small wintering birds is classically thought to be driven by the need to gather enough energy reserves to survive each night. A separate line of research has shown that sociality is a major driver in winter foraging activities in many species. Here, we use wintering birds as a study system to move towards an integrative understanding of the influence of energy requirements and sociality on foraging ecology. We used RFID-enabled feeders in Lincoln, Nebraska, USA in January-March 2019 to measure foraging activity in two species (downy woodpecker, Picoides pubescens, and white-breasted nuthatches, Sitta carolinensis). We analyzed the relationship between overnight temperature and morning foraging activity and found that lowest overnight temperature was negatively correlated with morning visitation at feeders. We then used a network approach to ask if flock associations explain similarity in birds’ foraging activity. In both species, individuals with stronger associations in a social network were more likely to share similar feeder activity, and an index of social partners’ activity explained foraging activity better than overnight temperature. This brings forth new questions about the interplay between individual response to temperature and social factors in shaping how small animals cope with harsh winter conditions.


The dataset was collected in three parts: 1) measurements and band numbers recorded during banding efforts, 2) a time series of visits detected at bird feeders and 3) weather measurements collected by a nearby weather station. The following is a revised, condensed version of the data collection methods reported in the manuscript. 

We caught songbirds using mist nets near bird feeders in Lincoln, Nebraska, USA. We banded all captured birds with aluminum leg bands distributed by the United States Fish and Wildlife Services (USFWS) and we placed RFID leg bands (Eccel Technology, Leicester UK) on downy woodpeckers (n=18) and white-breasted nuthatches (n=13). Birds were sexed by plumage and other morphological measurements were taken, including weight, culmen length, tarsus length, and wing length.

Next, we distributed 8 RFID feeders of uniform design over an area of approximately 150,866 m2, with a mean distance of approximately 287 m between feeders. Feeders were hung from trees and spaced as evenly as possible (i.e., given availability of suitable trees) to maximize coverage of the field site. Each feeder (New Generation(R) 23 inch feeder: Droll Yankee, Plainfield, CT) was equipped with an IBT EM4102 data logger board (Eccel Technology, Leicester, UK) to record RFID tag, date, and time when a bird visited the feeder. Data loggers were programmed to scan for RFID tags every ¼ second from 6:30 am-8:00 pm. We observed that the feeders could detect birds more than once during a single visit, therefore we condensed these data into discrete visits using an empirical cumulative distribution function. After 2 seconds, the density distribution of time delays exponentially decreased and we found it reasonable to accept that any detection of the same bird within 2 seconds was likely to be part of the same feeder visit. For a given bird, we collapsed consecutive detections ≤ 2 seconds apart into a single visit at the time of the first detection. These data are included as a master data file ("all_visits.dat") and were further transformed into the included data subsets for various analyses, as explained in the ReadMe text file. 

Last, we collected weather data from the Lincoln Municipal Airport (approximately 8.4 km from the study site) weather station through the Weather Underground website (, accessed 11 April 2019). We selected the lowest overnight temperature from this dataset to include in temperature response analyses. 

Usage notes

There are 16 files in all, including code scripts, data files, spreadsheet files, and a ReadMe file. We have included descriptions of the data files in the ReadMe file, and code scripts are annotated to allow navigation to re-create our analyses. The code scripts were written either for program R or Matlab, as explained in the ReadMe file. 


National Science Foundation, Award: NSF IOS-1750606

National Science Foundation, Award: NSF DGE-1735362

National Science Foundation, Award: NSF DGE-1545261