Hummingbirds and other lightweight bird species are challenging to track because they have limited capacity to carry devices for data-logging. We present a simple and customizable three-loop ‘backpack’ harness for studying hummingbird migration and movement, with step-by-step instructions for harness construction and attachment. The harness has negligible weight and cost (< $0.50 USD/each), is easy for a single person to make and apply in the field, and it requires no complicated setup or equipment. We have field-tested this harness on 74 Giant Hummingbirds (Patagona gigas) with three different types of tracking devices (geolocators, GPS tags, and satellite transmitters) in Chile and Peru from 2017–2020. Based on recaptures to date, we report that harnesses last at least two years, even under high UV-light conditions. We found no evidence of adverse effects of the harnesses on birds after one to two years and apparent survival of geolocator-tracked Giant Hummingbirds was in line with published estimates for other hummingbird species. This harness method is a practical and effective option for mounting ultra-light devices on hummingbirds, and it can be readily modified for other species with short legs, prominent sternal keels, and long wings.

We designed a harness for tracking hummingbirds and field-tested this harness on 74 Giant Hummingbirds (Patagona gigas) with three different types of tracking devices (geolocators, GPS tags, and satellite transmitters) in Chile and Peru from 2017–2020. All collected data are original. We additionally conducted a literature review of temperate and tropical hummingbird survival and capture probability and used capture-mark-recapture models (Cormack-Jolly-Seber models) to estimate survival and capture probabilities of geolocator-tracked Giant Hummingbirds. Source code for reproducing analyses can be found at https://github.com/jlwilliamson/Hummingbird_Harness_Methods.

See README.txt file for description of all files included in this data publication and see fully-annotated code on GitHub for further details. All data files are also stored on GitHub for completeness of repository. We provide .csv files, which were read into R for analysis, and we additionally provide corresponding .xlsx files (see .xlsx files for metadata and descriptions; tab 2 in each file).