Data from: Hummingbird feeders can provide eDNA for detection of nectar-feeding bats
Data files
Feb 02, 2026 version files 484.81 KB
Abstract
In the face of ongoing anthropogenic pressures on global biodiversity, effective monitoring strategies are essential for understanding species distributions and guiding conservation. Leptonycteris nivalis, L. yerbabuenae, and Choeronycteris mexicana are nectar-feeding bats of conservation concern that occur in the southwestern United States and Mexico. Leptonycteris nivalis and L. yerbabuenae are morphologically similar and difficult to differentiate in the field, making eDNA a particularly relevant monitoring tool. Since previous studies have shown that eDNA detection of these species is possible when swabs are taken from their dietary flowers, we sought to determine whether they can be detected from artificial feeders, which are commonly used on residential properties to attract hummingbirds, and are a known supplemental food source for L. yerbabuenae. Between 2023 and 2024, citizen scientists (n = 12) in Arizona and New Mexico took 306 swabs of hummingbird feeders, which we tested with an existing qPCR assay for L. nivalis and newly-developed qPCR assays for L. yerbabuenae and C. mexicana. Leptonycteris yerbabuenae and C. mexicana showed the highest number of detections (300 and 274 swabs, respectively). Previously only known to occur in the U.S. in Texas and New Mexico, we detected L. nivalis near Portal, Arizona, within 50 km of documented foraging range in western New Mexico. Detections of this endangered species suggest its range extends beyond Hidalgo County, New Mexico and into neighboring Cochise County, Arizona, highlighting a need for increased surveillance of this species. Our work with artificial nectar feeders expands the eDNA detection method for nectar-feeding bats to the human-wildlife interface, and shows that citizen science can be successfully used for eDNA surveys. Such methods provide an alternative to mist netting or acoustics, which can help monitor occupancy and clarify range extensions. This work corroborates other studies illustrating that eDNA can effectively detect and monitor terrestrial species.
https://doi.org/10.5061/dryad.905qfttw7
Description of the data and file structure
The data was collected and analyzed by Northern Arizona University's Bat Ecology and Genetics Lab and by Bat Conservation International. Detailed methods are available in the associated publication.
HummingbirdfeedereDNA_TotalDetections.csv
- Target: Name of assay run; LENIS3 = Leptonycteris nivalis, LEYES5 = L. yerbabuenae, CHME14 = Choeronycteris mexicana
- Sample detections: Total number of samples in which the target species was detected
- Total screened: Total number of samples screened; this number includes extraction (reagent) blanks
HummingbirdfeedereDNA_DetectionTable-Location.csv
- Site: Citizen scientists' residence or sample collection site; coded for privacy
- State: State where residence is located
- County: County where residence is located
- City: City where residence is located
- LENIS3: Total number of swabs collected at each residence in which L. nivalis was detected
- LEYES5: Total number of swabs collected at each residence in which L. yerbabuenae was detected
- CHME14: Total number of swabs collected at each residence in which C. mexicana was detected
- Sample size: Total number of swabs screened from each residence
HummingbirdfeedereDNA_DetectionTable-Sample.csv
- Target: Name of assay run; LENIS3 = Leptonycteris nivalis, LEYES5 = L. yerbabuenae, CHME14 = Choeronycteris mexicana
- Sample: Unique identifying name of each sample
- Positive: TRUE/FALSE; "TRUE" indicates the target species was detected in at least one replicates; "FALSE" indicates the target was not detected in any replicates
- qPCR replicates with detections: Number of replicates run for a sample in which the target was detected
- Max. qPCR replicates: Total number of qPCR replicates run for each sample; all samples were originally run in triplicate. Any number higher than 3 indicates a sample was re-run
- Run: Batch of samples run together
- Site: Citizen scientists' residence or sample collection site; coded for privacy
- Sample type: DNA source; either a hummingbird feeder swab ("Swab - eDNA") or an extraction (reagent) blank ("Extraction blank")
- Collection date: Date a sample was collected; in cases where no specific date is known, collection year is provided
- Country: Country where residence is located
- State: State where residence is located
- County: County where residence is located
- City: City where residence is located
HummingbirdfeedereDNA_qPCRData.csv
- Sample: Unique identifying name of each sample
- Site: Citizen scientists' residence or sample collection site; coded for privacy
- Target: Name of assay run; LENIS3 = Leptonycteris nivalis, LEYES5 = L. yerbabuenae, CHME14 = Choeronycteris mexicana
- Detection: TRUE/FALSE; Indicates whether the target species amplified in a sample
- Replicate: qPCR replicate 1-3
- Well: Well in a 384-well plate; columns 1-24, rows A-P
- Task: Unknown, Extraction blank, Postive Control, Negative Control, or Standard
- Dyes: Dye associated with the fluorescent probe; target species = FAM, Internal Postive Control = VIC
- Cq: Cycle of quantification; NA indicates no amplification
- Quantity: Starting quantity of DNA; calculated by QuantStudio 7 Pro Real-Time PCR System
- LOQ: Limit of quantification; specific to each assay
- Amp Status: Amp/No amp; Indicates whether amplification occurred or not; calculated automatically by QuantStudio 7 Pro Real-Time PCR System
- Flag IPC: TRUE/FALSE; Internal positive control; TRUE indicates inhibition; FALSE indicates no inhibition
- Late Cq: TRUE/FALSE; Amplification occurred after the y-intercept
- Run: Batch of samples run together
HummingbirdfeedereDNA_RunMetrics.csv
-
Run: Batch of samples run together
-
Target: Name of assay run; LENIS3 = Leptonycteris nivalis, LEYES5 = L. yerbabuenae, CHME14 = Choeronycteris mexicana
-
Slope: Estimated from standard curve; ideally between -3.2 and -3.5
-
Intercept: Number of cycles at which 1 copy of DNA would be detected; ideally between 34 and 41
-
R^2: Likelihood of any given sample to occur along the standard curve; ideally close to 1
-
Efficiency %: A measure of whether the reaction doubled with each cycle; ideally between 90 % and 110 %
