Secretive avian species can be difficult to detect. For such species, environmental DNA (eDNA) monitoring could improve our ability to estimate occupancy and population trends. There are often challenges associated with efforts to establish eDNA as an effective monitoring tool for species of conservation concern because it is difficult to hold nonmodel organisms in captivity to complete validation tests. To address this issue, we tracked four king rails Rallus elegans in Ohio, USA, in 2020 with radio telemetry to determine whether our eDNA monitoring approach could indicate king rail presence where these individuals occurred. Also, we collected eDNA near locations surveyed by human observers within potential habitat where king rail status was unknown. The eDNA surveys indicated king rail presence at one of five locations where we knew that an individual king rail had occurred within 30 h and at one of four locations where king rail presence was unknown but feasible. At these latter locations, human observers detected no king rails, although in 2022 human observers using traditional survey methods detected a king rail near the same location where our eDNA detection occurred in 2020. This study provides a preliminary, although informative, king rail eDNA monitoring methodology that we validated with radio telemetry data. We suggest ways to improve our approach to increase king rail detection rates and to add confidence in results.

Nongame and rarely hunted waterbird species declined by more than 20% in abundance as a broad group between 1970 and 2017 in North America (Rosenberg et al. 2019). Simultaneously, hunted waterbird species (predominantly waterfowl) that use similar habitat increased by more than 50% in abundance (Rosenberg et al. 2019). This disparity may be a result of management actions that historically focused upon and primarily benefited waterfowl species (Eddleman et al. 1988; Bolen 2000). Some waterbirds, such as secretive marsh birds, present a challenge to monitor because of their cryptic behaviors (Conway 2011) and in some cases due to their low abundances (Monfils et al. 2020). Methods that effectively monitor secretive marsh bird populations could help wetland managers to better evaluate and therefore justify the implementation of management activities intended to benefit this group of birds.

Currently, the Standardized North American Marsh Bird Monitoring (SNAMBM) protocol guides the primary approach used to monitor secretive marsh bird populations (Conway 2011) and entails human observers broadcasting the calls of focal species. This methodology has several limitations, including observer bias (Faanes and Bystrak 1981), limited numbers of qualified observers, short duration of surveys, and possible negative effects on birds due to playback of conspecific vocalizations (Harris and Haskell 2013). Especially for rarely occurring species, human observers may frequently fail to detect individuals that are present during a survey. For example, surveyors using the SNAMBM methodology failed to detect all 13 king rails Rallus elegans, a secretive marsh bird of conservation concern, detected by camera traps (Shirkey et al. 2017).

Populations of king rails have declined since the 1960s (Pardieck et al. 2020). Low detection probability and occupancy rates for king rails, especially at northern latitudes, may have also led to a paucity of detections using the SNAMBM methodology (e.g., Monfils et al. 2020; Brewer et al. 2023b; Kane et al. 2023). The use of environmental DNA (eDNA) as a monitoring tool could improve secretive marsh bird monitoring efficacy, especially for rarely occurring species. This method detects some aquatic species (e.g., Mahon et al. 2013; Lafferty et al. 2018; Yang et al. 2023; Sieber et al. 2024) and provides relatively precise temporal resolution of species occurrence due to the short persistence of eDNA in water (∼2 wk; Pilliod et al. 2014). Rodgers and Mock (2015) demonstrated detection of eDNA at water sources where terrestrial mammals have drunk, despite the relative brevity of this activity. Therefore, using eDNA to monitor bird species—including secretive marsh birds—that spend considerable amounts of time in water warrants further investigation. At least three studies document use of eDNA to detect secretive marsh birds, namely, black rail Laterallus jamaicensis (Neice and McRae 2021; Feist et al. 2022) and Ridgway’s rail Rallus obsoletus (Guan et al. 2023). However, species shed DNA at different rates (Allan et al. 2021) and environmental conditions—such as temperature and light—influence eDNA persistence (Allan et al. 2021). Thus, there is a need for further methodological development to optimize the use of eDNA as a monitoring tool for secretive marsh birds. Holding a focal species in captivity is often a good first step toward confirming the efficacy of a particular eDNA approach. For threatened or endangered species such as king rails (in many northern states; Cooper 2008), this step is less feasible. Alternatively, radio telemetry could be used to confirm the presence of a focal species in a natural setting while an eDNA monitoring methodology is evaluated there.

Our goal was to develop and test an initial approach for using eDNA to detect the presence of king rails in a region where this species is of conservation concern. We hypothesized that our approach would detect king rail eDNA where we knew that radio-tagged individuals occurred. We also sought to determine whether SNAMBM methodology and eDNA monitoring produced the same results pertaining to king rail presence where king rail status was unknown but feasible.

Fieldwork occurred in 2020 from May to July in Ohio, Indiana, and Michigan, USA, between the latitudes of 40.97 to 42.43° and the longitudes of −82.92 to −87.51° (Figure 1). This region has spring and summer temperatures ranging from ∼8 to 28°C. Fieldwork occurred exclusively in impounded wetlands dominated by robust herbaceous emergent vegetation, including primarily species in the genera Typha, Scirpus, and Phragmites. This study occurred on both public property (Pickerel Creek Wildlife Area, Ohio; Willow Slough Fish and Wildlife Area, Indiana; Waterloo State Recreation Area, Michigan) and private property (Winous Point Marsh Conservancy, Ohio). These properties probably experienced different levels of human activity. Logistics, availability of resources, and the desire to locate our negative control site far from where king rails were known to occur required conducting this study at multiple sites throughout the Midwest (Figure 1).

Figure 1.

We collected eDNA samples in 2020 at three different sites (A, B, and C) in our midwestern USA study area (bottom right), where we tested an approach for detecting king rail Rallus elegans environmental DNA (eDNA). Site A: We collected eDNA samples between 18 June and 21 July from king rail present and status unknown locations at Winous Point Marsh Conservancy and neighboring Pickerel Creek Wildlife Area, Ohio. We separated locations in different categories at site A by more than 1 km. In the panels containing satellite imagery, we present homing points (red dots) and home ranges for each bird that we tracked (A1, A2, A3, A4). Pink circles surround homing points where we collected eDNA, but where we did not detect king rail eDNA. A blue circle surrounds the homing point (in panel A3) where we detected a king rail via eDNA. Note that we sampled bird A3 for eDNA twice and that this bird concentrated space use in a smaller area than did the other birds. Site B: We collected eDNA samples from environmental negative control locations on 5 July at Waterloo State Recreation Area, Michigan. Site C: We collected eDNA samples from king rail status unknown locations on 15 June at Willow Slough Fish and Wildlife Area, Indiana. Robert W. Hines drew the king rail, which we did not manipulate.

Figure 1.

We collected eDNA samples in 2020 at three different sites (A, B, and C) in our midwestern USA study area (bottom right), where we tested an approach for detecting king rail Rallus elegans environmental DNA (eDNA). Site A: We collected eDNA samples between 18 June and 21 July from king rail present and status unknown locations at Winous Point Marsh Conservancy and neighboring Pickerel Creek Wildlife Area, Ohio. We separated locations in different categories at site A by more than 1 km. In the panels containing satellite imagery, we present homing points (red dots) and home ranges for each bird that we tracked (A1, A2, A3, A4). Pink circles surround homing points where we collected eDNA, but where we did not detect king rail eDNA. A blue circle surrounds the homing point (in panel A3) where we detected a king rail via eDNA. Note that we sampled bird A3 for eDNA twice and that this bird concentrated space use in a smaller area than did the other birds. Site B: We collected eDNA samples from environmental negative control locations on 5 July at Waterloo State Recreation Area, Michigan. Site C: We collected eDNA samples from king rail status unknown locations on 15 June at Willow Slough Fish and Wildlife Area, Indiana. Robert W. Hines drew the king rail, which we did not manipulate.

Close modal

Trapping and tracking

In May and June 2020, we caught four king rails in northwestern Ohio by using walk-in traps, following methods described by Shirkey et al. (2017). We attached very high frequency radio transmitters (model A1050, weight 2.4 g; ATS-Isanti, Minnesota) to all captured king rails and collected blood from two individuals. We used blood to extract DNA for primer design and eDNA method validation, as described in the laboratory analyses section. We waited 24 h or longer after releasing individuals to allow them time to resume normal activity before tracking began. Before tracking, we practiced estimating distances to transmitters placed in the field to improve accuracy of the king rail homing locations that we later marked. During each tracking session, we tracked the focal king rail to where that individual naturally occurred. To do so, we used radio telemetry equipment (R2000 and R4000, three-element folding Yagi antenna; ATS-Isanti), biangulated the individual’s location from approximately 20 m away, and used a global positioning system unit to record a single homing location. This homing location was our best estimate of where the focal bird was before we disturbed it by approaching. We recorded all homing locations 24 h or longer apart, one to three times per week, to ensure independence. As described in Brewer et al. (2023a), we used kernel density estimation to calculate home-range size for each bird. We collected eDNA samples at a subset of homing locations, at locations where SNAMBM surveys occurred in potential king rail habitat, and at locations assumed to be unsuitable for king rails. Additional detail about sampling locations is presented below.

eDNA sampling

The same individual filtered water samples to collect eDNA in 2020 from 15 June to 21 July, during the king rail breeding season, at all sampling locations (Figure 1). The observer used a single-use, 60-mL syringe to pump water associated with each sample through a Sterivex filter (pore size, 0.22 μm) to collect eDNA (Spens et al. 2017). Water where king rails occurred was shallow (<20 cm) and difficult to collect without sediment and plant material impeding the water-filtering process. To address this challenge, the observer first collected water in a disposable cup. Next, the observer poured water from that cup into another cup covered by a fresh section of stainless steel-woven wire mesh (0.13-mm holes) to exclude larger debris from entering. The observer then used the cup filled (∼400 mL) with water devoid of large debris to fill the syringe. During this process, the observer refilled the cup as necessary (which depended on how much water could be pumped through the associated Sterivex filter). The observer noted the volume of water (in milliliters) that the syringe pushed (using multiple draws) through each Sterivex filter (mean ± SE, 297 ± 25.7 mL; range, 85–620 mL) before the filter became clogged and no more water could be pushed through. To avoid cross-contamination, the observer completed this filtering process on-site. We did not collect field blanks to test for equipment contamination because we used single-use/disposable equipment. The observer immediately filled each Sterivex capsule with Longmire’s solution (Longmire et al. 1997), wrapped the capsule in parafilm, and stored it at room temperature (see Renshaw et al. 2015). Finally, the observer sprayed boots/waders with bleach solution before traveling to new locations.

The observer collected three eDNA samples within 20 m of the associated homing location (a region that the focal king rail likely used), less than 30 h (mean ± SD, 20.4 ± 8.4 h) after recording the location. Thus, each sampling event consisted of three sampling points associated with the homing location to which we tracked the focal bird. We did not collect eDNA samples at all homing locations; we used some homing locations to indicate how often the focal bird had been in the sampled area before collection of eDNA samples (see below). We sampled homing locations for one bird twice (on 2 July and on 21 July; i.e., we used two different homing locations for the same individual). The observer placed all three sampling points associated with each eDNA sampling event at a water–vegetation interface, where king rails forage (Eddleman et al. 1988). During this process, the observer collected a sample at the exact homing location whenever water–vegetation interface occurred, and water could be filtered, there. For each sampling event associated with a homing location (n = 5), the observer collected one sample within 5 m of the homing location and ensured that the placement of the other two samples was random with the constraint that they were 10–20 m from each other and from the homing location to increase the likelihood that the area where the bird had occurred was well represented. We collected these samples (five sets of three; n = 15) at Winous Point Marsh Conservancy/Pickerel Creek Wildlife Area (Figure 1) and categorized them as king rail present (KP). For each KP sampling event, we noted how many times the focal bird had been recorded during prior homing events within 50 m of the homing location during the previous 2 wk (Pilliod et al. 2014). We also noted how much time elapsed between determination of the homing location and collection of associated eDNA samples.

In addition, the observer collected eDNA samples at four locations (three samples at three locations, two samples at one location; n = 11) within potential, but not confirmed, king rail habitat separated by greater than or equal to 250 m. We conducted secretive marsh bird surveys as described by Brewer et al. (2023b) at these locations. Surveys followed the SNAMBM protocol (Conway 2011) and occurred at Willow Slough Fish and Wildlife Area and Winous Point Marsh Conservancy (Figure 1). We categorized samples from these locations as king rail status unknown (KU). The observer collected these KU eDNA samples near each SNAMBM survey location as described relative to homing locations (for the KP samples).

Finally, the observer collected eDNA at Waterloo State Recreation Area (Figure 1) at locations where we presumed that king rails did not occur and categorized them as environmental negative control (ENC). These samples came from two public lakes (two samples separated by ≥10 m within each lake; n = 4) not connected to marshes by flowing water and devoid of king rail habitat. We considered these sites, which respectively consisted of a public beach and a boat launch, to lack king rail habitat because there was no shallow water (<20 cm) interspersed with herbaceous emergent vegetation that could provide food or cover (Cooper 2008; Brewer et al. 2023a). We also used eBird (2021) to confirm that no king rails had been reported within 20 km of the ENC sites during the previous 30 y.

Laboratory analyses

We used droplet digital polymerase chain reaction (ddPCR) to analyze collected water samples for the presence of king rail eDNA. Our methodology followed an active surveillance (single-species) approach involving radio telemetry, similar to previous studies (e.g., Lafferty et al. 2018). Our ddPCR approach used relatively new technology (Pinheiro et al. 2012) that is more sensitive (Nathan et al. 2014; Doi et al. 2015; Simmons et al. 2015) and requires different procedures than real-time PCR (quantitative PCR). First, we selected several candidate DNA sequences that were potentially specific to king rails by searching GenBank, a database produced by the National Center for Biotechnology Information (Sayers et al. 2020). Specifically, we chose nucleotide sequences that minimized similarity to nontarget rallids that shared habitat with king rails. These nontarget species were the congeneric Virginia rail Rallus limicola and the confamilial sora Porzana carolina. Soras are closely related genetically (Kirchman et al. 2021) to the other rallids that occurred in the study area, namely, the American coot Fulica americana and the common gallinule Gallinula galeata. Next, we created candidate primer sets (Table S1, Supplemental Material) via the National Center for Biotechnology Information’s Primer-BLAST program to target the identified sequences. We set PCR product size at 125–300 bp, specified ‘nr’ for database, specified ‘birds’ for organism, and otherwise used default parameters. Finally, we used DNA extracted from target and nontarget blood samples (see next paragraph) to test all candidate primer sets and to identify a single king rail-specific primer set (Table S1) for subsequent analyses. We used the same equipment to complete this step and to analyze eDNA samples (see below for equipment and primer set details). We confirmed that all rallids and other wading/secretive marsh birds that were likely to occur in the same area as king rails in our study area (Table S2, Supplemental Material) had sequences represented in GenBank for the cytochrome c oxidase subunit 1 gene. Thus, we checked these species via Primer-BLAST for nontarget amplification regarding the cytochrome c oxidase subunit 1 gene, which included the sequence that our chosen primer set targeted. Our repeated SNAMBM surveys indicated that other secretive marsh birds were not common (Table S2) at the KP or KU locations. Because we used a targeted approach meant to detect presence of king rails, distantly related species that were not present in our study system posed only a minimal risk of false positive detections.

Before use during analyses, we stored blood samples in ethanol at room temperature for a mean of 84 d after collection (range, 28–132 d). We extracted DNA from two blood samples that we collected from different king rails in our study. We also extracted DNA from blood samples collected in Illinois, USA, by a collaborator to represent both nontarget rail species (sora and Virginia rail, three blood samples for each). We used a DNeasy blood and tissue kit (Qiagen) to extract and concentrate DNA from the blood samples and stored the eluted DNA at −80°C until we could analyze the field samples in the laboratory.

We stored collected eDNA samples in Sterivex capsules at room temperature for a mean of 105.5 d (range, 88–123 d) before eDNA extraction. Wegleitner et al. (2015) found that even after this period, preserved eDNA samples remain stable. Using a DNeasy blood and tissue kit (Qiagen), we extracted eDNA from the capsules following the protocol of Spens et al. (2017), with the following modifications. We changed the incubation time for the eDNA pellets and filter columns, both potential sources of eDNA, to 6 h at 56°C given that this amount of time improves extraction efficiency (A.R.M., unpublished data; Resh and Mahon 2019). We inverted samples twice during incubation and, after completing the final extraction steps, stored eluted eDNA at −80°C.

We used a QX200 ddPCR system (Bio-Rad, Hercules, California) to determine that the primer set that targeted a 164-bp segment of the cytochrome c oxidase subunit 1 gene was king rail specific and therefore appropriate for subsequent analyses. This primer set produced greater than or equal to 444 copies per microliter for both king rail blood samples and less than or equal to 0.7 copy per microliter for all Virginia rail and sora blood samples analyzed (n = 5; Table S1). The minimal amplification produced by this primer set for nontarget blood samples was less than the threshold that we used to indicate king rail presence based on eDNA samples taken from the field (see below). The forward primer for our chosen primer set was 5′-CACCGCCGTCCTCCTATTAC-3′ and the reverse primer was 5′-TGTAGACTTCTGGGTGACCG-3′. We determined target eDNA concentrations, and thus focal species presence, as in Hauger et al. (2020) and as detailed below.

For each sample, we prepared a reaction with 1,000 nM of the forward and reverse primers, 1× ddPCR Supermix for SYBR Green (Bio-Rad), 2.5 μL of DNA, and water for a total volume of 25 μL. We partitioned each sample into thousands of nanodroplets (i.e., replicate droplets; range: 12,984–16,687) by using the QX200 autodroplet generator (Bio-Rad) and then amplified them in a PCR cycler under the following conditions: 95°C for 10 min, 45 cycles at 94°C for 30 s and 60°C for 1 min, followed by 98°C for 10 min. For each PCR trial, we included two no-target controls (sora and Virginia rail DNA extracted from blood; 5–25 ng/μL of genomic DNA), a no-template control (sterile water), and a positive control (king rail DNA extracted from blood). Doing so allowed us to confirm a lack of contamination and that the PCR was successful. We analyzed droplet concentrations with QuantaSoft 1.7 software (Bio-Rad) and used the automatic threshold to identify positive droplets. We used the number of positive droplets to statistically estimate the target eDNA concentration in each sample (Pinheiro et al. 2012). We confirmed that positive detection droplets occurred in positive controls and field samples but were completely or virtually absent in no-target and no-template controls. Thus, we considered our approach valid and concluded that nontarget eDNA would not result in false positive detections. We observed no issues with droplet rain; thus, we attempted no further optimization of lab methods.

Statistical analysis

We compared target eDNA concentrations from each KP and KU sample to the mean target eDNA concentration for ENC samples. Specifically, we considered each sample with a greater number of target eDNA copies per microliter than was included in the 95% confidence interval for ENC samples to indicate king rail presence. In other words, if a sample had an eDNA concentration that was statistically greater than the mean of the ENC sample concentrations, we considered that sample to indicate king rail presence. We chose the 95% confidence interval based on the statistical convention of using P = 0.05 as the acceptable threshold for committing a Type 1 error. If a sampling event contained a single positive sample, we considered a king rail to be present and considered multiple positive samples to indicate increased evidence of king rail presence where the sampling event occurred.

The 95% confidence interval for mean target eDNA copies per microliter (hereafter, target copies) for ENC samples (n = 4) was 0.35–1.9. Of our 26 KP and KU eDNA samples, 3 (11.5%) met the criteria for positive king rail detection (i.e., exceeded 1.9 target copies; Figure 2; Table S3, Supplemental Material). Two of these samples occurred at a KP location (there were 15 KP samples total), were generated from the same sampling event, and had a target copy count of 514 and 3, respectively. The other sample that indicated king rail presence occurred at a KU location (there were 11 KU samples total) and had 3.8 target copies. Overall, eDNA indicated king rail presence for one of five KP sampling events and for one of four KU sampling events (Figure 2). We did not detect king rails via the SNAMBM surveys that occurred at KU locations. We report the number of within-sample, replicate droplets that we used to estimate eDNA concentrations for each sample (Table S3).

Figure 2.

Target environmental DNA (eDNA) copies per microliter at different location categories visited from June to July 2020 in our midwestern USA study area, where we tested an approach for detecting king rail Rallus elegans eDNA. Location categories are as follows: king rail present (king rail), king rail absent (negative control), and king rail status unknown (unknown). Dots represent samples that are color coded based on sampling event within each location category. The region above the horizontal dashed line (>1.9 copies) contains samples that we considered king rail detections. Note the scale break.

Figure 2.

Target environmental DNA (eDNA) copies per microliter at different location categories visited from June to July 2020 in our midwestern USA study area, where we tested an approach for detecting king rail Rallus elegans eDNA. Location categories are as follows: king rail present (king rail), king rail absent (negative control), and king rail status unknown (unknown). Dots represent samples that are color coded based on sampling event within each location category. The region above the horizontal dashed line (>1.9 copies) contains samples that we considered king rail detections. Note the scale break.

Close modal

The relatively low detection rate that we observed, along with our small sample size, underscores that this study is a starting point rather than a definitive protocol for monitoring the presence of king rail and similar species by using eDNA. Numerous methodological and environmental factors may have influenced our results. For example, the volume of water pumped through filters may influence findings, although our results suggest that a wide range of volumes can result in a detection. The 95% confidence interval for mean amount of water pumped through the 30 Sterivex filters used during this study (one for each sample) was 246.7–347.3 mL. The three samples that indicated king rail detections came from a filtered water volume of 210, 330, and 555 mL, resulting in 3.8, 3, and 514 target copies, respectively. Five samples where we did not detect king rails, but where they were known to be present, had greater than or equal to 400 mL pumped through their respective Sterivex filters.

Radio telemetry has been previously used to confirm that eDNA can detect the presence of a species (e.g., a shark; Lafferty et al. 2018). Our study shows how this approach could be applied in a natural setting to secretive marsh birds, a taxon known to be declining in our study area (Tozer 2016). Furthermore, radio telemetry allowed us to preliminarily evaluate how the amount of time a bird spent near a sampling site related to detectability. Our eDNA samples indicated king rail presence for one sampling event where radio telemetry also confirmed king rail presence. During the 2 wk before eDNA collection for this sampling event, two of the four homing locations for the radio-tagged king rail were within 50 m of the eDNA sampling locations (Table 1). This bird, a female based on genetic testing (Brewer et al. 2023a), concentrated space use in a relatively small area compared with other individuals included in this study (95% kernel density estimation home range of 5 ha vs. 5.8, 6.1, and 15.8 ha; Figure 1). The second eDNA sampling event in the vicinity of this bird, which was 19 d after the first sampling event, was within 50 m of all four of this bird’s homing locations during the previous 2 wk, but did not result in a detection. King rail detections via eDNA did not occur where the other three radio-tagged birds occurred; we documented the radio-tagged birds associated with these sampling events within 50 m of the sampling event just once during the previous 2 wk when we marked three, four, and five homing locations, respectively (Table 1). Time between when the homing location had been marked and when eDNA sampling occurred ranged from 6 to 27 h (Table 1). For the instance of likely eDNA detection by two samples during a sampling event, the focal bird had been documented at the eDNA sampling area 25 h before sampling occurred.

Table 1.

An overview of sampling activity, which occurred in July 2020, at and in the vicinity of each king rail Rallus elegans present location in our midwestern USA study area, where we tested an approach for detecting king rail environmental DNA (eDNA). Note that “Bird A3_Detected” denotes the sampling event when detection of the focal bird was likely via eDNA on 2 July. “Bird A3_NotDetected” denotes a separate sampling event, on 21 July, when there was no detection of king rail eDNA where that bird occurred. Also indicated is the sex of each bird based on genetic analysis. We marked homing points by using radio telemetry to document locations of each focal bird during the 2 wk before each eDNA sampling event occurred.

An overview of sampling activity, which occurred in July 2020, at and in the vicinity of each king rail Rallus elegans present location in our midwestern USA study area, where we tested an approach for detecting king rail environmental DNA (eDNA). Note that “Bird A3_Detected” denotes the sampling event when detection of the focal bird was likely via eDNA on 2 July. “Bird A3_NotDetected” denotes a separate sampling event, on 21 July, when there was no detection of king rail eDNA where that bird occurred. Also indicated is the sex of each bird based on genetic analysis. We marked homing points by using radio telemetry to document locations of each focal bird during the 2 wk before each eDNA sampling event occurred.
An overview of sampling activity, which occurred in July 2020, at and in the vicinity of each king rail Rallus elegans present location in our midwestern USA study area, where we tested an approach for detecting king rail environmental DNA (eDNA). Note that “Bird A3_Detected” denotes the sampling event when detection of the focal bird was likely via eDNA on 2 July. “Bird A3_NotDetected” denotes a separate sampling event, on 21 July, when there was no detection of king rail eDNA where that bird occurred. Also indicated is the sex of each bird based on genetic analysis. We marked homing points by using radio telemetry to document locations of each focal bird during the 2 wk before each eDNA sampling event occurred.

The nondetections in our study where king rails had recently been present could have occurred for numerous reasons. Species shed DNA at different rates (e.g., Allan et al. 2021), and king rails may shed less DNA than other species. Environmental factors, such as temperature, may have contributed to rapid decay of king rail eDNA such that when we collected samples little or no target eDNA remained (Allan et al. 2021). Water currents—although not apparent to us—may have also had an impact on our results by transporting target eDNA from sampling locations (Stoeckle et al. 2017).

The random nature of sampling likely also affected detection success. For example, the sample that contained an exceptional amount of target copies (514) may have incidentally occurred near to where a king rail had recently been. Although no king rail (or other) fecal matter was visible in the filtered water, it may have been present. Alternatively, the sample could have occurred very near (e.g., within 1 cm) to where a king rail had been very recently (e.g., <10 min). Another sample from that sampling event also indicated king rail presence, providing further support for the accuracy of that detection. If we collected more samples during each sampling event, there would have likely been an increase in the number of king rail detections.

Investigators may wish to conduct analytical steps that we could not perform due to monetary limitations. For example, practitioners could carefully assess the limit of detection for our primers to evaluate sensitivity as well as to consider impacts on occupancy modeling and false negative rates (Hunter et al. 2017). If false positive detections are of particular concern to stakeholders, future investigators may wish to retest all positive samples to confirm the presence of target eDNA (Hauger et al. 2020), Sanger sequence a subset of positive samples to confirm target accuracy (e.g., Feist et al. 2022; Porco et al. 2022), and/or collect more ENC samples. If false negatives are a concern, collection of more samples from each location and analysis of each sample multiple times could minimize the likelihood of failing to detect target eDNA (Feist et al. 2022).

The primer set that we used is best suited for use in our study area. For example, it likely would not differentiate between king rail and the closely related clapper rail Rallus crepitans and so would not be ideal for use where those two species coexist (in tidal marshes). Further validation of our chosen primer set, such as expanded testing of nontarget tissue samples, would add confidence to subsequent king rail eDNA surveys that may use our assay. The development of additional king rail primer sets could facilitate the use of multiple assays for each sample to increase detection rates and, if multiple lines of evidence suggest presence of target DNA, to provide increased confidence in results (Feist et al. 2022; Guan et al. 2023). Future investigators may also be able to improve results by optimizing their ddPCR protocols (Kokkoris et al. 2021; Porco et al. 2022) or by developing a probe for increased specificity (Brys et al. 2023).

Our eDNA results suggest a king rail was present in 2020 at Willow Slough Fish and Wildlife Area, where no king rails were detected by SNAMBM surveys at any survey location (n = 13) during 2018–2021 (Audubon Great Lakes and Indiana Department of Natural Resources, unpublished data). This underscores the potential usefulness of further developing eDNA monitoring techniques meant to detect the presence of king rails and other secretive species that may frequently be missed by traditional survey methods. In 2022, the SNAMBM survey effort detected a single king rail at Willow Slough Fish and Wildlife Area for the first time and that detection occurred at the same survey location where our eDNA detection occurred in 2020 (bird detected ∼40 m to the north of our eDNA sampling location). In this case, it appears that eDNA indicated occurrence of a king rail within a marsh impoundment, and wetland complex, years before annually occurring SNAMBM surveys did. This demonstrates how SNAMBM and eDNA monitoring approaches could be used complementarily at the same location to identify where focal species occur and so inform holistic management actions.

Our study provides a starting point for those interested in monitoring king rails by using eDNA. We have demonstrated how coupling radio telemetry with a preliminary eDNA monitoring methodology can establish feasibility of eDNA as a monitoring tool for a focal species. Given the potential of eDNA as a novel and useful monitoring tool, other cryptic waterbirds of conservation concern may benefit from similar studies.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Reference S1.Cooper T. 2008. King rail (Rallus elegans) conservation plan version 1. U.S. Fish and Wildlife Service, Fort Snelling, Minnesota.

Available: govinfo.gov/content/pkg/GOVPUB-I49-PURL-LPS108314/pdf/GOVPUB-I49-PURL-LPS108314.pdf.

Reference S2. Monfils MJ, Hayes DB, Al-Saffar M, Soulliere GJ, Pierce R. 2020. Marsh bird population estimates to inform conservation decisions in the Midwest. Michigan Natural Features Inventory Report 2020-03, Lansing.

Available: mnfi.anr.msu.edu/reports/MNFI-Report-2020-03.pdf.

Table S1. Primer sets—which amplify portions of three different genes—tested for specificity to king rail Rallus elegans, the species for which we tested an approach to detect environmental DNA in 2020 in our midwestern USA study area. Pos. conc. (positive concentration) represents the mean target copy concentration (copies per microliter) for the king rail DNA samples (n = 2). Neg. conc. (negative concentration) represents the mean target copy concentration for the nontarget DNA samples, including Virginia rail Rallus limicola and sora Porzana carolina (n = 5). Note that we chose the primer set (bolded) with the greatest positive concentration-to-negative concentration ratio for subsequent analyses. The sequence used to generate this chosen primer set, AY666315.1, is highly similar to another king rail sequence, JQ176092.1, that occurred in GenBank.

Available: https://doi.org/10.3996/JFWM-23-043.S1 (14 KB DOC)

Table S2. Our study occurred in the midwestern portion of the USA, where we tested an approach for detecting king rail Rallus elegans environmental DNA (eDNA) at locations sampled in June and July 2020. This table overviews species that we confirmed were represented in GenBank regarding the cytochrome c oxidase subunit 1 gene, but for which we did not have a tissue sample to test our chosen primer set. Thus, we created our chosen primer set with specificity to king rails while considering these nontarget species based on in silico testing, although we did not confirm that tissue samples from them would not result in amplification. In each column, we note whether we detected these species during the 2 wk before each eDNA sampling event using the Standardized North American Marsh Bird Monitoring (SNAMBM) protocol. For each eDNA sampling location, “N” represents the number of SNAMBM surveys that occurred during the 2 wk before eDNA sampling. “Yes” indicates detection of a species and “No” indicates no detection at the corresponding eDNA sampling location. “N/A” indicates a nonfocal species for our SNAMBM surveys, which would not have been noted even if detected by the surveyor. We detected king rail eDNA during the first sampling event for bird A3 and during the unknown (Unk.) 1 sampling event (both bolded). No SNAMBM surveys occurred during the 2 wk before the second eDNA sampling event for bird A3, which did not result in a king rail detection.

Available: https://doi.org/10.3996/JFWM-23-043.S2 (15 KB DOC)

Table S3. An overview of the number of king rail Rallus elegans target copies per microliter (“Copies”) detected at each location visited in June and July 2020 in our midwestern USA study area, where we tested an approach for detecting king rail environmental DNA (eDNA). In addition, the “Total Droplets” column indicates how many nanodroplets we used during droplet digital polymerase chain reaction analysis. “Positive Droplets” indicates how many of these nanodroplets were positives, which produced our estimate of eDNA concentration in each sample. The “Detected?” column indicates whether we considered each sample a detection.

Available: https://doi.org/10.3996/JFWM-23-043.S3 (19 KB DOC)

We appreciate funding support provided by the American Ornithological Society, the Indiana Audubon Society, and the Upper Mississippi/Great Lakes Joint Venture (grant award F19AP00330). The Earth and Ecosystem Science PhD program at Central Michigan University supported this research. We also received support from Central Michigan University’s Department of Biology, College of Science and Engineering, and Office of Research and Graduate Studies. L. Durden, as well as two anonymous reviewers and the Associate Editor, provided helpful comments and A.M.V. Fournier provided samples. M. Schoof, along with personnel from Audubon Great Lakes and Winous Point Marsh Conservancy, conducted marsh bird surveys. The National Audubon Society and Indiana Department of Natural Resources contributed data. Approval for this research was via Central Michigan University’s Institutional Animal Care and Use Committee protocol 19-21, Michigan (permit TE216), Ohio (permit 24-002), and the U.S. Federal Government (permit 24149). This is contribution number 205 of the Central Michigan University Institute for Great Lakes Research.

Allan
EA,
Zhang
WG,
Zhang,
Lavery
AC,
Govindarajan
AF.
2021
.
Environmental DNA shedding and decay rates from diverse animal forms and thermal regimes
.
Environmental DNA
3
:
492
514
.
Bolen
EG.
2000
.
Waterfowl management: yesterday and tomorrow
.
Journal of Wildlife Management
64
:
323
335
.
Brewer
DE,
Gehring
TM,
Garcia
MM,
Shirkey
BT,
Simpson
JW,
Fournier
AMV.
2023a
.
King rail (Rallus elegans) home range and microhabitat characteristics in western Lake Erie coastal marshes
.
Ecology and Evolution
13
:
e10043
.
Brewer
DB,
Gehring
TM,
Shirkey
BT,
Simpson
JW
.
2023b
.
King rail (Rallus elegans) response to audio playback: implications for population estimation, monitoring methodology, and trapping approach
.
Journal of Field Ornithology
94
:
7
.
Brys
R,
Halfmaerten
D,
Everts
T,
Van Driessche
C,
Neyrinck
S.
2023
.
Combining multiple markers significantly increases the sensitivity and precision of eDNA‐based single‐species analyses
.
Environmental DNA
5
:
1065
1077
.
Conway
CJ.
2011
.
Standardized North American marsh bird monitoring protocol
.
Waterbirds
34
:
319
346
.
Cooper
T.
2008
.
King rail (Rallus elegans) conservation plan version 1
.
U.S. Fish and Wildlife Service, Fort Snelling
,
Minnesota
(see Supplemental Material, Reference S1).
Doi
H,
Takahara
T,
Minamoto
T,
Matsuhashi
S,
Uchii
K,
Yamanaka
H.
2015
.
Droplet digital polymerase chain reaction (PCR) outperforms real-time PCR in the detection of environmental DNA from an invasive fish species
.
Environmental Science & Technology
49
:
5601
5608
.
eBird
.
2021
.
eBird: an online database of bird distribution and abundance
. Available: http://www.ebird.org (August 2024)
Eddleman
WR,
Knopf
FL,
Meanley
B,
Reid
FA,
Zembal
R.
1988
.
Conservation of North American rallids
.
Wilson Bulletin
100
:
458
475
.
Faanes
CA,
Bystrak
D.
1981
.
The role of observer bias in the North American Breeding Bird Survey
.
Studies in Avian Biology
6
:
353
359
.
Feist
SM,
Guan
X,
Malmfeldt
MP,
Lance
RF.
2022
.
Two novel qPCR assays to enhance black rail (Laterallus jamaicensis) eDNA surveys in the United States
.
Conservation Genetics Resources
14
:
321
329
.
Guan
X,
Lance
RF,
Feist
SM.
2023
.
Environmental DNA (eDNA) assays for the detection of Ridgway’s rail (Rallus obsoletus) in the United States
.
Conservation Genetics Resources
15
:
179
185
.
Harris
JBC,
Haskell
DG.
2013
.
Simulated birdwatchers’ playback affects the behavior of two tropical birds
.
PLoS One
8
:
e77902
.
Hauger
AM,
Hollis-Etter
KM,
Etter
DR,
Roloff
GJ,
Mahon
AR.
2020
.
Use of environmental DNA (eDNA) in streams to detect feral swine (Sus scrofa)
.
PeerJ
8
:
e8287
.
Hunter
ME,
Doazio
RM,
Butterfield
JSS,
Meigs-Friend
G,
Nico
LG,
Ferrante
JA.
2017
.
Detection limits of quantitative and digital PCR assays and their influence in presence-absence surveys of environmental DNA
.
Molecular Ecology Resources
17
:
221
229
.
Kane
ME,
Brewer
DE,
Gehring
TM,
Shirkey
BT,
Pangle
KL,
Uzarski
DG,
Picciuto
MA,
Simpson
JW.
2023
.
King rail (Rallus elegans) presence in the midwestern United States is predicted by local‐scale factors and avian community
.
Ecology and Evolution
13
:
e10732
.
Kirchman
JJ,
Rotzel McInerney
N,
Giarla
TC,
Olson
SL,
Slikas
E,
Fleischer
RC.
2021
.
Phylogeny based on ultra-conserved elements clarifies the evolution of rails and allies (Ralloidea) and is the basis for a revised classification
.
Ornithology
138
:
ukab042
.
Kokkoris
V,
Vukicevich
E,
Richards
A,
Thomsen
C,
Hart
MM.
2021
.
Challenges using droplet digital PCR for environmental samples
.
Applied Microbiology
1
:
74
88
.
Lafferty
KD,
Benesh
KC,
Mahon
AR,
Jerde
CL,
Lowe
CG.
2018
.
Detecting southern California’s white sharks with environmental DNA
.
Frontiers in Marine Science
5
:
355
.
Longmire
JL,
Maltbie
M,
Baker
RJ.
1997
.
Use of ‘lysis buffer’ in DNA isolation and its implication for museum collections
.
Museum of Texas Tech University
163
:
1
3
.
Mahon
AR,
Jerde
CL,
Galaska
M,
Bergner
JL,
Chadderton
WL,
Lodge
DM,
Hunter
ME,
Nico
LG.
2013
.
Validation of eDNA surveillance sensitivity for detection of Asian carps in controlled and field experiments
.
PloS One
8
:
1
6
.
Monfils
MJ,
Hayes
DB,
Al-Saffar
M,
Soulliere
GJ,
Pierce
R.
2020
.
Marsh bird population estimates to inform conservation decisions in the Midwest
.
Michigan Natural Features Inventory Report 2020-03, Lansing
(see Supplemental Material, Reference S2).
Nathan
LM,
Simmons
M,
Wegleitner
BJ,
Jerde
CL,
Mahon
AR.
2014
.
Quantifying environmental DNA signals for aquatic invasive species across multiple detection platforms
.
Environmental Science & Technology
48
:
12800
12806
.
Neice
AA,
McRae
SB.
2021
.
An eDNA diagnostic test to detect a rare, secretive marsh bird
.
Global Ecology and Conservation
27
:
e01529
.
Pardieck
KL,
Ziolkowski
DJ
Lutmerding
M,
Aponte
VI,
Hudson
MAR.
2020
.
North American Breeding Bird Survey dataset 1966–2019
:
U.S. Geological Survey data release
. Available: https://doi.org/10.5066/P9J6QUF6 (August 2024)
Pilliod
DS,
Goldberg
CS,
Arkle
RS,
Waits
LP.
2014
.
Factors influencing detection of eDNA from a stream-dwelling amphibian
.
Molecular Ecology Resources
14
:
109
116
.
Pinheiro
LB,
Coleman
VA,
Hindson
CM,
Herrmann
J,
Hindson
BJ,
Bhat
S,
Emslie
KR.
2012
.
Evaluation of a droplet digital polymerase chain reaction format for DNA copy number quantification
.
Analytical Chemistry
84
:
1003
1011
.
Porco
D,
Hermant
S,
Purnomo
CA,
Horn
M,
Marson
G,
Colling
G.
2022
.
Getting rid of ‘rain’ and ‘stars’: mitigating inhibition effects on ddPCR data analysis, the case study of the invasive crayfish Pacifastacus leniusculus in the streams of Luxembourg
.
PLoS One
17
:
e0275363
.
Renshaw
MA,
Olds
BP,
Jerde
CL,
McVeigh
MM,
Lodge
DM.
2015
.
The room temperature preservation of filtered environmental DNA samples and assimilation into a phenol–chloroform–isoamyl alcohol DNA extraction
.
Molecular Ecology Resources
15
:
168
176
.
Resh
CA,
Mahon
AR.
2019
. Improving the efficiency of DNA extraction from samples collected for environmental DNA surveillance.
Conference Proceeding: Integrative and Comparative Biology
.
Oxford University Press
,
Cary, NC, USA
.
Rodgers
TW,
Mock
KE.
2015
.
Drinking water as a source of environmental DNA for the detection of terrestrial wildlife species
.
Conservation Genetics Resources
7
:
693
696
.
Rosenberg
KV,
Dokter
AM,
Blancher
PJ,
Sauer
JR,
Smith
AC,
Smith
PA,
Stanton
JC,
Panjabi
A,
Helft
L,
Parr
M,
Marra
PP.
2019
.
Decline of the North American avifauna
.
Science
366
:
120
124
.
Sayers
EW,
Beck
J,
Bolton
EE,
Bourexis
D,
Brister
JR,
Canese
K,
Comeau
DC,
Funk
K,
Kim
S,
Klimke
W,
Marchler-Bauer
A,
Landrum
M,
Lathrop
S,
Lu
Z,
Madden
TL,
O’Leary
N,
Phan
L,
Rangwala
SH,
Schneider
VA,
Skripchenko
Y,
Wang
J,
Ye
J,
Trawick
BW,
Pruitt
KD,
Sherry
ST.
2020
.
Database resources of the National Center for Biotechnology Information
.
Nucleic Acids Research
49
:
D10
D17
.
Shirkey
BT,
Simpson
JW,
Picciuto
MA.
2017
.
King rail (Rallus elegans) trapping efficiency and detection techniques in southwestern Lake Erie coastal marshes, USA
.
Waterbirds
40
:
69
73
.
Sieber
N,
King
A,
Krieg
R,
Zenker
A,
Vorburger
C,
Hartikainen
H.
2024
.
Large-scale eDNA monitoring of multiple aquatic pathogens as a tool to provide risk maps for wildlife diseases
.
Environmental DNA
6
:
e427
.
Simmons
M,
Tucker
A,
Chadderton
WL,
Jerde
CL,
Mahon
AR.
2015
.
Active and passive environmental DNA surveillance of aquatic invasive species
.
Canadian Journal of Fisheries and Aquatic Sciences
73
:
76
83
.
Spens
J,
Evans
AR,
Halfmaerten
D,
Knudsen
SW,
Sengupta
ME,
Mak
SST,
Sigsgaard
EE,
Hellström
M.
2017
.
Comparison of capture and storage methods for aqueous macrobial eDNA using an optimized extraction protocol: advantage of enclosed filter
.
Methods in Ecology and Evolution
8
:
635
645
.
Stoeckle
BC,
Beggel
S,
Cerwenka
AF,
Motivans
E,
Kuehn
R,
Geist
J.
2017
.
A systematic approach to evaluate the influence of environmental conditions on eDNA detection success in aquatic ecosystems
.
PLoS One
12
:
e0189119
.
Tozer
DC.
2016
.
Marsh bird occupancy dynamics, trends, and conservation in the southern Great Lakes basin: 1996 to 2013
.
Journal of Great Lakes Research
42
:
136
145
.
Wegleitner
BJ,
Jerde
CL,
Tucker
A,
Chadderton
WL,
Mahon
AR.
2015
.
Long duration, room temperature preservation of filtered eDNA samples
.
Conservation Genetic Resources
7
:
789
791
.
Yang
J,
Zhang
L,
Mu
Y,
Zhang
X.
2023
.
Small changes make big progress: a more efficient eDNA monitoring method for freshwater fish
.
Environmental DNA
5
:
363
374
.

Author notes

Citation: Brewer DE, Mahon AR, Wirick MJ, Gehring TM. 2024. Using environmental DNA to detect radio-tagged king rails: initial validation results and needed advancements. Journal of Fish and Wildlife Management 15(2):xx-xx; e1944-687X. https://doi.org/10.3996/JFWM-23-043

The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.

Supplemental Material