Abstract

We retrospectively evaluated 20 y of using predator-deterrent fences to mitigate unusually high nest predation for the critically endangered Attwater's prairie-chicken Tympanuchus cupido attwateri at the Attwater Prairie Chicken National Wildlife Refuge. Refuge staff constructed fences of 0.9-m-high, 0.32–0.64-cm mesh hardware cloth, with 15.2-m-long sides, and staked at the bottom to discourage predators from going under the fence. Workers placed fences around nests at a mean of 3.2 d of incubation. Eighty-two percent of fenced nests were successful vs. 12% for unfenced nests. Daily survival rate increased from 0.9159 for unfenced nests in 1997–2011 to 0.9916 for fenced nests during 2000–2019. Fencing did not increase abandonment or reduce the proportion of eggs that hatched in successful nests. After 2012, we reduced fence sides from 15.2 m to 7.6 m in length. Larger and smaller fences were equally effective with respect to daily survival rate and proportion of abandoned nests. The median proportion of eggs that hatched from successful nests was 6% higher for larger fences vs. smaller fences, but this difference was not statistically significant. Predator-deterrent fences substantially increased Attwater's prairie-chicken nesting success in this study, and may represent a viable management strategy for increasing nesting success for other populations of ground-nesting birds with high conservation value.

Introduction

The Attwater's prairie-chicken Tympanuchus cupido attwateri is a grouse endemic to coastal prairie grasslands of Texas and southwest Louisiana. Populations disappeared from Louisiana circa 1919 (Lehmann 1965; Oberholser 1974), and those in Texas have been in steep decline for more than 80 y despite increasingly intensive management efforts (Lehmann 1941; U.S. Fish and Wildlife Service [USFWS] 2010; Morrow et al. 2015). The Secretary of the Interior listed the Attwater's prairie-chicken as endangered with extinction in 1967 (U.S. Endangered Species Preservation Act of 1966; Udall 1967). Habitat loss and degradation resulting from urban and industrial expansion, agriculture, overgrazing, invasion of prairies by native and nonnative woody species, and invasion by nonnative fauna (e.g., red imported fire ants Solenopsis invicta; Morrow et al. 1996, 2015; USFWS 2010) have driven declines. Populations declined to near extinction during the 1990s, and have remained there despite implementation of intensive recovery actions including habitat restoration, captive breeding and release, and research to identify factors limiting populations (Figure 1). Approximately 100 Attwater's prairie-chickens currently exist in wild populations. Captively reared individuals have supplemented remaining populations beginning with a pilot release of males in 1995, followed by release of both sexes since 1996 (USFWS 2010). Apparent nesting success for released females at the Attwater Prairie Chicken National Wildlife Refuge (APCNWR) during the first 3 y following implementation of the release program averaged 13.3%, and ranged from 0% in 1998 (n = 4) and 1999 (n = 6) to 40% in 1997 (n = 5). Peterson and Silvy (1996) reported >2× higher apparent nest success in a summary of historical Attwater's prairie-chicken data collected during 1937–1985 (mean 32.2% ± 4.31 SE; n = 143). Poor breeding success is a commonly reported malady for captive-bred animals (e.g., Parish and Sotherton 2007; Buner et al. 2011).

Figure 1.

Attwater's prairie-chicken Tympanuchus cupido attwateri population trend in Texas, 1970–present.

Figure 1.

Attwater's prairie-chicken Tympanuchus cupido attwateri population trend in Texas, 1970–present.

Destruction of nests by predators such as striped skunks Mephitis mephitis, opossums Didelphis virginiana, raccoons Procyon lotor, coyotes Canis latrans, and Texas rat snakes Elaphe obsoleta lindheimeri is the most commonly reported cause of nest failure for Attwater's prairie-chickens. Nest flooding following heavy rainfall is also a relatively common occurrence due to lack of topographic relief in coastal prairie habitats occupied by Attwater's prairie-chickens (Lehmann 1941; Jurries 1979; USFWS 2010). Traditional mitigation of nest predation generally consists of a combination of management practices including improvement of habitat quality, reduction of predator abundance, and predator exclusion (Greenwood and Sovada 1996). While maintenance of high-quality nesting habitat is often the most cost-effective strategy for improving nest success of most species on a large scale, predator management may be warranted for intensively managed populations of economically important species (e.g., for hunting, photography, viewing) or extremely vulnerable species like Attwater's prairie-chicken (Schroeder and Baydack 2001). Predator exclusion is generally more palatable than predator removal from a public perspective, but can be costly and difficult to effectively implement without adversely impacting behavior or survival of the protected species (Smith et al. 2011).

Although refuge officials implemented active habitat management and predator removal on APCNWR in an attempt to mitigate the poor apparent nesting success observed during 1997–1999, we elected to also evaluate the use of nest predator exclosures around individual nests due to the precarious nature of Attwater's prairie-chicken populations. Sargeant et al. (1974) reported on fences used to protect sharp-tailed grouse Tympanuchus phasianellus nests, and J.E.T. demonstrated that incubating female greater prairie-chickens Tympanuchus cupido pinnatus in Minnesota would accept predator-deterrent fencing around their nests (J.E. Toepfer, Society of Tympanuchus Cupido Pinnatus, Ltd., personal observation). Further, of 18 Minnesota greater prairie-chicken nests that were fenced in 1998–1999, 78% were successful compared to 54% for 59 unfenced nests (J.E. Toepfer, personal observation). These observations demonstrated proof of concept for protecting individual nests of the endangered Attwater's prairie-chicken.

Wildlife managers have used fences to protect avian nests of various species at spatial scales ranging from relatively large blocks of habitat containing multiple nests to small areas encompassing single nests (Jimenez et al. 2001; Smith et al. 2011). Fences around individual nests have been shown to be effective for increasing nest success (Smith et al. 2011), but we were concerned that disturbance around the nest associated with fence construction and maintenance could result in abandonment by the incubating female. We were also concerned that disturbance or difficulty in traversing the fence following incubation recesses could cause incubating females to be away from the nest for extended periods resulting in reduced proportion of eggs that hatch. Therefore, our objectives for this paper were to 1) describe predator-deterrent fences constructed around individual Attwater's prairie-chicken nests and 2) evaluate the efficacy of these fences with respect to nest success, egg hatchability, and nest abandonment. Because Attwater's prairie-chicken populations were so perilously close to extinction in the wild and preliminary results of using predator-deterrent fences were so convincing, we proceeded with implementation of this technique before we collected enough data to complete a formal reference-treatment analysis. Therefore, many of the data presented in this paper are retrospective in nature.

Methods

Evaluation of predator-deterrent fences for Attwater's prairie-chickens was conducted at the 4,265-ha APCNWR near Eagle Lake, Texas (29°40′N, 96°16′W) from 2000 to 2019, but we also included data from unfenced nests beginning in 1997 for comparison. The APCNWR is located in the Gulf Coast Prairies and Marshes ecoregion of Texas (Hatch et al. 1990), and supports prairie grasslands managed to provide Attwater's prairie-chicken habitat using patch burning (combination of prescribed fire and grazing; Fuhlendorf and Engle 2001) and chemical treatment of invading brush. Texas Wildlife Services personnel removed known Attwater's prairie-chicken nest predators including striped skunks, opossums, and raccoons each year, February–May 1997–2019, although we do not know the proportion of total predators removed. The area surrounding the refuge was predominantly in rice production or heavily grazed rangelands with various amounts of native and nonnative brush encroachment. Attwater's prairie-chickens would occasionally use areas adjacent to APCNWR for nesting, particularly fallow rice fields.

We released most Attwater's prairie-chicken females at APCNWR from captivity and equipped them with necklace- or poncho-style radio transmitters (<3% of body mass; various manufacturers) at the time of release. We released most as 8–12+-wk-old poults during July–October 1996–2018 after they spent 2 wk in prerelease acclimation pens at release sites. We located Attwater's prairie-chicken nests during late-March–mid-June in subsequent years using a handheld Yagi antenna after lack of daily movement by radioed females suggested that they had initiated incubation.

In 2000, we chose four of seven Attwater's prairie-chicken nests at random for protection with fences, and left three unfenced for comparison. Thereafter (2002–2019), refuge workers fenced most nests except for rare cases when logistical constraints precluded fence construction. Fence construction material was 0.32-cm (0.125-in) or 0.64-cm) hardware cloth. Fences were approximately square with either 15.2-m (50-ft; 2000–2012) or 7.6-m (2012–2019) sides. We maintained sides in an upright position with 1.3–1.5-m pieces of metal rebar pushed or driven into the soil and attached to the hardware cloth with plastic cable ties. We generally constructed the 0.9-m-high fences immediately after locating nests, and usually without flushing incubating females (Figure 2). We secured fence bottoms with approximately 15-cm stakes as necessary (roughly every 30 cm) to close gaps at ground level that would allow access by snakes or other predators. Rigid galvanized wire stakes (3.7-mm) had one end bent into a J-shape to hold the fence down. Fences leaned toward the center of the nest because preliminary observations on fences surrounding Minnesota greater prairie-chicken nests revealed that females returning from incubation recesses would fly to the outside of the fence and then pace back and forth. Leaning the fence toward the center for 36–48 h until the female learned to traverse the fence minimized this problem (J.E. Toepfer, personal observation). We generally stood fences around Attwater's prairie-chicken nests upright 2 d after construction. The top 15 cm of the fence was bent roughly 90–135° outward approximately 4 d after construction to discourage climbing snakes (Figure 3). Initially, we used larger square fences with 15.2-m (50-ft) sides to minimize disturbance at the immediate nest site. Beginning in 2012, we reduced the length of each side by half to reduce the need for materials and time needed for construction. One to four individuals constructed large (15.2 m/side) fences, which took approximately two personnel hours to complete. Smaller fences (7.6 m/side) required approximately half the effort to construct compared to larger fences. Materials cost US $150–175 each for small fences, and approximately double that amount for large fences. We reused materials for multiple years.

Figure 2.

Predator-deterrent fence surrounding an Attwater's prairie-chicken Tympanuchus cupido attwateri nest on April 17, 2020, at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas.

Figure 2.

Predator-deterrent fence surrounding an Attwater's prairie-chicken Tympanuchus cupido attwateri nest on April 17, 2020, at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas.

Figure 3.

Predator-deterrent fence surrounding an Attwater's prairie-chicken Tympanuchus cupido attwateri nest on April 21, 2020, at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas. We bent the top of the fence outward to discourage climbing predators like Texas rat snakes Elaphe obsoleta lindheimeri from crawling over the top of the fence.

Figure 3.

Predator-deterrent fence surrounding an Attwater's prairie-chicken Tympanuchus cupido attwateri nest on April 21, 2020, at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas. We bent the top of the fence outward to discourage climbing predators like Texas rat snakes Elaphe obsoleta lindheimeri from crawling over the top of the fence.

We monitored incubating females remotely by telemetry on a daily basis during incubation. We immediately inspected nests to assess their status if females were not at their nests during midday. We visited nest sites once during incubation, usually when females were away from the nest during morning or evening feeding recesses, to determine clutch size and collect various other data, including egg weights and measurements. After 2000, we treated the area within an approximately 15–20-m radius of most nests with Extinguish Plus™ or Amdro Pro™ to suppress red imported fire ants, which are threats to chicks, especially during the hatching process. We estimated projected hatch dates by either onset of incubation as indicated by telemetry movement data of the female, or by comparison of egg weights to estimated fresh egg weights as described by Hoyt (1979) and Burnham (1983). We generally found that the latter method provided more reliable estimates of hatch because it was not uncommon for females still laying eggs to be located at the nest site on multiple days while we were still attempting to determine onset of incubation. In those cases, we would incorrectly assume that incubation had started.

Beginning several days before expected hatch, we checked nests daily by direct observation or close-range telemetry of the incubating female to determine if she was still on the nest. Females generally leave the nest with their chicks within a few hours of hatch (Lehmann 1941). Once we confirmed hatch, we opened the fence on at least two sides to allow the female and chicks to exit (Figure 4). We collected hatched eggshells and any unhatched eggs at that time unless it was not possible to do so without disturbing the hen and chicks. In that case, we collected shells and unhatched eggs the following day. Because the hen sometimes crushed hatched eggshells during the hatching process, we determined the number of hatched eggs by subtracting any unhatched eggs from the known clutch size. We then calculated the proportion of eggs that hatched within a nest as the ratio of the number hatched eggs to known clutch size.

Figure 4.

Opening the predator-deterrent fence at an Attwater's prairie-chicken Tympanuchus cupido attwateri nest at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, after hatch on June 1, 2005, to allow exit by the female and chicks.

Figure 4.

Opening the predator-deterrent fence at an Attwater's prairie-chicken Tympanuchus cupido attwateri nest at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, after hatch on June 1, 2005, to allow exit by the female and chicks.

We included only nests of known fate in this study. We also excluded nest losses attributed to weather (e.g., heavy rain, nest flooding) because predator-deterrent fences would have no effect in those cases. We excluded all nests for 2016 (n = 26) because catastrophic flooding and wet conditions affected most nests. Several hens that year continued to incubate nonviable eggs which had likely been flooded. We considered nests successful if at least one egg hatched. We determined causes of nest failure by inspecting remaining shells and other sign at the nest site (Darrow 1938; Rearden 1951).

Data analysis

We conducted analyses using R 4.0.0 (R Core Team 2020). Because of small expected values in contingency tables, we used Fisher's exact test to evaluate whether nest fencing was associated with nest success in 2000, and nest abandonment across years (Daniel 1978; R function fisher.test). In addition to apparent nest success, we also calculated daily survival rate for fenced and unfenced nests using the RMark package (Laake 2013) within Program MARK 8.2 (White and Burnham 1999). We parameterized nest survival models to include 88 encounter occasions representing the interval from the earliest nest found (March 27) to latest active (June 22).

We compared proportion of eggs hatched in successful fenced and unfenced nests with a one-tailed Mann–Whitney test (Daniel 1978; R function wilcox.test) to determine whether fencing decreased hatchability. We also used a Mann–Whitney test to evaluate whether fence size (7.6 vs. 15.2 m/side) affected the proportion of eggs that hatched. Except for the preliminary evaluation of fences on nest success in 2000, nests in the unfenced sample included those from 1997 to 1999 (all unfenced), unfenced nests from 2000 (selected at random), and the few nests that were not fenced after 2000 because of logistical constraints.

Results

A total of 317 Attwater's prairie-chicken nests were located on the Attwater Prairie Chicken National Wildlife Refuge (n = 302) or surrounding privately-owned properties (n = 15) from 1997 to 2019. Of these, we excluded 52 from our study because we attributed nest failure to weather (10.1%), we found the hen dead when we located the nest (2.5%), we collected eggs for the Attwater's prairie-chicken captive rearing program (1.6%), we applied other experimental treatments to the nest (1.3%), or nests were already destroyed when found (0.6%; Table S1, Supplemental Material; Data S1 and S2, Supplemental Material). In addition, we excluded one nest containing nonviable eggs because no males were known to occur within >3.2 km of the nest.

All Attwater's prairie-chicken nests chosen at random for fence protection in 2000 were successful, and all of the unfenced nests in that year failed (Fisher's exact test P = 0.03). Based on these preliminary observations, we placed fences around 240 of 250 nests included in our study in subsequent years (Table S1, Supplemental Material; Data S1 and S2, Supplemental Material). We placed fences around nests at an estimated mean 3.2 d ± 0.3 SE (n = 228) of incubation. Daily survival rates for fenced nests across years (2000–2019) was 0.9916 (95% CI 0.9886–0.9938; n = 240) compared with 0.9159 (95% CI 0.8755–0.9440; n = 25) for unfenced nests (Figure 5). Apparent success for fenced Attwater's prairie-chicken nests was 82% vs. 12% for unfenced nests (Table 1). We observed equal effectiveness for the larger fences (daily survival rate = 0.9914; 95% CI 0.9876–0.9941; n = 154) we initially used for Attwater's prairie-chicken nests during 2000–2012 compared with smaller fences (daily survival rate = 0.9918; 95% CI 0.9862–0.9951; n = 86) used during 2012–2019. Excluding weather, the most common causes of nest failure for both fenced and unfenced nests were snake predation, mammalian predation, and abandonment. In addition, four (1.7%) fenced nests failed when females were killed by predators away from the nest, presumably during feeding recesses (Table 1). Fourteen (5.8%) nests failed during the first few days after fence placement while the nest fence was leaning toward the center to facilitate the female learning to traverse the fence. Of these 14 failures, which represent 33% of fenced nest losses, the female abandoned six nests and predators destroyed eight (five predated by snakes, two by mammals, one hen killed away from the nest site).

Figure 5.

Estimated daily survival and 95% confidence intervals for Attwater's prairie-chicken Tympanuchus cupido attwateri nests on the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas. Nests in the fenced group were surrounded by large (15.2 m/side; 2000–2012) or small (7.6 m/side; 2012–2019) predator-deterrent fences. We constructed fences of 0.32- or 0.64-cm hardware cloth, 0.9 m high, approximately square in shape, and were centered on the nest. We placed fences around nests at an estimated mean 3.2 ± 0.3 SE d of incubation. Daily survival rates for Fenced nests was 0.9916 (95% CI 0.9886–0.9938; n = 240) compared with 0.9159 (95% CI 0.8755–0.9440; n = 25) for unfenced nests.

Figure 5.

Estimated daily survival and 95% confidence intervals for Attwater's prairie-chicken Tympanuchus cupido attwateri nests on the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas. Nests in the fenced group were surrounded by large (15.2 m/side; 2000–2012) or small (7.6 m/side; 2012–2019) predator-deterrent fences. We constructed fences of 0.32- or 0.64-cm hardware cloth, 0.9 m high, approximately square in shape, and were centered on the nest. We placed fences around nests at an estimated mean 3.2 ± 0.3 SE d of incubation. Daily survival rates for Fenced nests was 0.9916 (95% CI 0.9886–0.9938; n = 240) compared with 0.9159 (95% CI 0.8755–0.9440; n = 25) for unfenced nests.

Table 1.

Fate of Attwater's prairie-chicken Tympanuchus cupido attwateri nests on the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, during 1997–2019. Approximately square fences constructed of 0.32-cm or 0.64-cm mesh hardware cloth, 0.9 m high, and with either 7.6 or 15.2 m side length surrounded fenced nests. We included only nests of known fate. We excluded a total of 52 nests from analyses: nests impacted by weather including all nests from 2016 (n =32); nests where the female was found dead when the nest was located (n = 8); nests from which we collected eggs for the captive breeding program (n = 5); nests that received other experimental treatments (n = 4); nests that were already destroyed when found (n = 2); nests where eggs were infertile because no male was in the area (n =1). We determined causes of nest failure by inspecting remaining shells and other sign at the nest site (Darrow 1938; Rearden 1951).

Fate of Attwater's prairie-chicken Tympanuchus cupido attwateri nests on the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, during 1997–2019. Approximately square fences constructed of 0.32-cm or 0.64-cm mesh hardware cloth, 0.9 m high, and with either 7.6 or 15.2 m side length surrounded fenced nests. We included only nests of known fate. We excluded a total of 52 nests from analyses: nests impacted by weather including all nests from 2016 (n =32); nests where the female was found dead when the nest was located (n = 8); nests from which we collected eggs for the captive breeding program (n = 5); nests that received other experimental treatments (n = 4); nests that were already destroyed when found (n = 2); nests where eggs were infertile because no male was in the area (n =1). We determined causes of nest failure by inspecting remaining shells and other sign at the nest site (Darrow 1938; Rearden 1951).
Fate of Attwater's prairie-chicken Tympanuchus cupido attwateri nests on the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, during 1997–2019. Approximately square fences constructed of 0.32-cm or 0.64-cm mesh hardware cloth, 0.9 m high, and with either 7.6 or 15.2 m side length surrounded fenced nests. We included only nests of known fate. We excluded a total of 52 nests from analyses: nests impacted by weather including all nests from 2016 (n =32); nests where the female was found dead when the nest was located (n = 8); nests from which we collected eggs for the captive breeding program (n = 5); nests that received other experimental treatments (n = 4); nests that were already destroyed when found (n = 2); nests where eggs were infertile because no male was in the area (n =1). We determined causes of nest failure by inspecting remaining shells and other sign at the nest site (Darrow 1938; Rearden 1951).

We observed no difference (Fisher's exact test, P = 1) in abandonment for fenced (5.8%; n = 14) nests in our study compared to unfenced nests (4.0%; n = 1) nests, nor did placing a fence decrease the proportion of eggs that hatched in successful nests (n =196, median = 91% for fenced; n = 3, median = 64% for unfenced; W = 97; P1-sided > 0.98). The median proportion of eggs that hatched was 86% and 92% for nests protected by small and large fences, respectively, but this difference was not statistically significant (W = 5076.5; P = 0.10). We observed no association (Fisher's exact test, P = 1) between abandonment and fence size.

While nest fences substantially improved nest survival as evidenced by apparent success and daily survival rate (Table 1; Figure 5), fences directly resulted in nest failures in a small number of cases. On one occasion, the female apparently struck one of the rebar supporting stakes that extended a short distance above the fence, resulting in injury to her crop. The female survived this injury and the nest survived the entire incubation period, but leakage of crop material contaminated eggs during incubation resulting in death of all embryos. Thereafter, we made corrections to ensure that the tops of all stakes were placed below fence level. In addition, incorrect estimate of hatch resulted in the entrapment and loss of one brood. These two cases represented <1% of the 240 fenced nests.

Discussion

Even though our study was retrospective in nature rather than a balanced statistical evaluation of predator-deterrent fences for mitigating predation of Attwater's prairie-chicken nests, data continue to support our preliminary observations of this technique's effectiveness after 20 y of use (Figure 5). The 82% apparent nest success we observed for fenced Attwater's prairie-chicken nests was 6.8× that of unfenced nests in our study, and 2.6× the historical average reported by Peterson and Silvy (1996). Other studies have also demonstrated substantial increases in nest success using fences at the individual nest level for various species representing several avian orders (e.g., Smith et al. 2011; Ragheb et al. 2019).

We were concerned that presence of fences and the associated activity required to construct and monitor them might lead to increased nest abandonment. However, we observed no statistical difference associated with fences in the relative number of Attwater's prairie-chicken females that abandoned their nests compared to unfenced nests. Of the fenced nests that were abandoned, females abandoned 8 of 14 (57%) within the first week after the fence was built, possibly due to disturbance around the nest or the inability of the female to negotiate the fence. The remaining abandonments occurred in mid to late incubation and >2 wk after fences were placed. Females abandoned five of six in the final few days before projected hatch; we observed pipped eggs in at least two of these nests. Two of the six females abandoned nests ∼1 wk after we captured them at nest sites to administer treatment for a heavy feather louse (Order Phthiraptera) infestation. Since these six females had successfully negotiated the fence for 2+ wk of incubation and embryos were viable up to the point of abandonment, it is unlikely that fences contributed directly to abandonment of these nests, although disturbance associated with daily monitoring for hatch may have played a role. At the first indication of abandonment we usually peremptorily salvaged eggs from nests that we thought the female had deserted; we used these eggs in the Attwater's prairie-chicken captive breeding program. We have since learned that some females can spend extended time away from the nest and still hatch chicks, especially late in the nesting season when temperatures become summer-like (late May–mid-June). It is possible that we incorrectly concluded abandonment was imminent and salvaged the eggs from two nests like this. Regardless, the abandonment rate we observed for both the fenced and unfenced groups was consistent with previous work on nesting Attwater's prairie-chickens, which reported 0–15.8% abandonment of unfenced nests (mean = 6.3%; n = 127; Lehmann 1941; Horkel 1979; Lutz 1979; Lawrence 1982; Morrow 1986).

We were also concerned that egg hatchability might suffer if hens had difficulty negotiating fences after incubation recesses. However, we saw no convincing evidence of adverse effects of fencing on the proportion of eggs that hatched within a successful nest. Even though a slightly decreased proportion of eggs hatched from nests protected by small vs. large fences, the median (86%) for small fences was comparable to the 87% average hatchability reported by Peterson and Silvy (1996) for historic Attwater's prairie-chicken data.

Placement of predator-deterrent fences requires that the female and chicks are able to exit the fence when hatch occurs. One (0.4%) Attwater's prairie-chicken brood perished after being unable to exit the fence perimeter when the nest unexpectedly hatched due to inaccurate estimates of hatch date. In this case, we used methods described by Hoyt (1979) and Burnham (1983) to project hatch dates. These methods rely on changes in egg weight compared to estimated fresh egg weight to determine stage of incubation. While we have found that these methods provide better estimates of hatch than telemetry in many cases, we have observed that unusual precipitation patterns (abnormally wet or dry) alter moisture loss from the egg compared to more normal conditions, occasionally resulting in skewed hatch date projections. At that time, we were checking nests 3–5 d in advance of the projected hatch date, but this nest hatched outside that window. While it is important that the monitoring timeframe for determining hatch be as narrow as possible to minimize potential disturbance at the nest site, we have since extended the monitoring window to begin 1 wk before projected hatch.

After accounting for abandonment and other losses, a total of 9–15 (3.8–6.2%) of Attwater's prairie-chicken nests may have resulted in failure due to nest predator fences depending on whether all nest abandonments were associated with presence of the fence. Either way, the proportion of nests negatively impacted by predator-deterrent fences was below abandonment rates reported by others for unfenced Attwater's prairie-chicken nests (Lehmann 1941; Horkel 1979; Lutz 1979; Lawrence 1982; Morrow 1986). The substantial increase in nest survival for remaining fenced nests far outweighed the possible negative consequences we observed.

Management Implications

We have demonstrated that the use of predator-deterrent fences around individual nests is a viable strategy for increasing Attwater's prairie-chicken nesting success. However, a relatively precise estimate of projected hatch is required to minimize the amount of disturbance associated with checking for hatch and to ensure that managers open fences in a timely fashion when hatch occurs. We concur with Estelle et al. (1996) that a thorough knowledge of the protected species including their behaviors, response to disturbance, and potential predators along with an evaluation of possible effects (positive and negative) of nest-predator fences is essential to successful implementation of this management technique. Consistent with Liebig's law of the minimum as applied to wildlife populations, an increased number of chicks leaving the nest does not guarantee additional recruits if nesting success is not limiting the population. While the nest success we observed for fenced Attwater's prairie-chicken nests was considerably higher than averages for unfenced nests we observed or those reported in the literature, it is important to evaluate increases in nest success resulting from management actions like predator-deterrent fences within the context of the whole life cycle. Such an evaluation was beyond the scope of this study. Finally, we believe the expense and amount of effort required to implement predator-deterrent fencing as we described could only be justified for species of extreme conservation concern.

Supplemental Material

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.

Data S1. Attwater's prairie-chicken Tympanuchus cupido attwateri nest survival dataset (n = 265) used to retrospectively evaluate efficacy of predator-deterrent fences as a management tool for increasing nesting success at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, 1997–2019. We excluded a total of 52 nests from analyses: nests impacted by weather including all nests from 2016 (n = 32); nests where the female was found dead when the nest was located (n = 8); nests from which we collected eggs for the captive breeding program (n = 5); nests that received other experimental treatments (n = 4); nests that were already destroyed when found (n = 2); nests where eggs were infertile because no male was in the area (n =1). We determined causes of nest failure by inspecting remaining shells and other sign at the nest site (Darrow 1938; Rearden 1951).

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S1 (22 KB CSV).

Data S2. Description of variables contained in the Attwater's prairie-chicken Tympanuchus cupido attwateri nest survival dataset (Data S1, Supplemental Material) used to retrospectively evaluate efficacy of predator-deterrent fences at the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, 1997–2019.

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S2 (24 KB DOC).

Table S1. Attwater's prairie-chicken Tympanuchus cupido attwateri nests (n = 317) on or near the Attwater Prairie Chicken National Wildlife Refuge, Colorado County, Texas, 1997–2019. Approximately square fences constructed of 0.32-cm or 0.64-cm mesh hardware cloth, 0.9 m high, and with either 7.6 or 15.2 m side length surrounded fenced nests. We included only nests of known fate. We excluded a total of 52 nests from evaluation of fence efficacy: nests impacted by weather including all nests from 2016 (n =32); nests where the female was found dead when the nest was located (n = 8); nests from which we collected eggs for the captive breeding program (n = 5); nests that received other experimental treatments (n = 4); nests that were already destroyed when found (n = 2); nests where eggs were infertile because no male was in the area (n =1).

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S3 (26 KB XLSX).

Reference S1. Jurries RW. 1979. Attwater's prairie chicken. Austin: Texas Parks and Wildlife Department, F. A. Series No. 18.

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S4 (2.34 MB PDF).

Reference S2.Laake, JL. 2013. RMark: an R interface for analysis of capture-recapture data with MARK. Seattle, Washington: National Marine Fisheries Service, Alaska Fisheries Science Center Processed Report 2013-01.

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S5 (319 KB PDF); also available at https://www.researchgate.net/publication/267509042_RMark_an_R_Interface_for_analysis_of_cpture-recapture_data_with_MARK.

Reference S3.[USFWS] U.S. Fish and Wildlife Service. 2010. Attwater's prairie-chicken recovery plan, 2nd revision. Albuquerque, New Mexico.

Found at DOI: https://doi.org/10.3996/112019-JFWM-099.S6 (2.86 MB PDF); also available at https://ecos.fws.gov/ecp0/profile/speciesProfile?sId=7259.

Acknowledgments

We thank staff and interns at the Attwater Prairie Chicken National Wildlife Refuge for assistance with construction of fences and data collection. The Society of Tympanuchus Cupido Pinnatus, Ltd., the Minnesota Prairie Chicken Society, and the U.S. Fish and Wildlife Service provided logistical and financial support. Funders had no influence on the content of the manuscript, and did not require approval prior to publication. This research was authorized by permits from the U.S. Fish and Wildlife Service (TE051839), Texas Parks and Wildlife Department (SPR-0491-384), and the Minnesota Department of Natural Resources. We appreciate comments provided by G. Huschele, J. Mueller, D. Sherer, two anonymous reviewers and the Associate Editor that greatly improved earlier versions of this manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.

References

Buner
FD,
Browne
SJ,
Aebischer
NJ.
Experimental assessment of release methods for the re-establishment of a red-listed galliform, the grey partridge (Perdix perdix)
.
Biological Conservation
144
:
593
601
.
Burnham
W.
1983
.
Artificial incubation of falcon eggs
.
Journal of Wildlife Management
47
:
158
168
.
Daniel
WW.
1978
.
Applied nonparametric statistics
.
Boston, Massachusetts
:
Houghton Mifflin
.
Darrow
R.
1938
.
Possibilities of recognizing the evidence of predation and the species involved in the remains of grouse nests found destroyed
.
Proceedings of the North American Wildlife and Natural Resources Conference
3
:
834
838
.
Estelle
VB,
Mabee
TJ,
Farmer
AH.
1996
.
Effectiveness of predator exclosures for pectoral sandpiper nests in Alaska
.
Journal of Field Ornithology
67
:
447
452
.
Fuhlendorf
SD,
Engle
DM.
2001
.
Restoring heterogeneity on rangelands: ecosystem management based on evolutionary grazing patterns
.
BioScience
51
:
625
632
.
Greenwood
RJ,
Sovada
MA.
1996
.
Prairie duck populations and rates of nest predation in birds
.
Pages
31
42
in
KG,
Wadsworth
McCabe
RE,
editors.
Issues and problems in predation management to enhance avian recruitment
.
Logan
:
Utah State University, Berryman Institute
.
Hatch
SL,
Gandhi
KN,
Brown
LE.
1990
.
Checklist of the vascular plants of Texas
.
College Station
:
Texas Agricultural Experiment Station, MP-1655
.
Horkel
JD.
1979
.
Cover and space requirements of Attwater's prairie chicken (Tympanuchus cupido attwateri) in Refugio County, Texas. Doctoral dissertation
.
College Station
:
Texas A&M University
.
Hoyt
DF.
1979
.
Practical methods of estimating volume and fresh egg weight of bird eggs
.
The Auk
96
:
73
77
.
Jimenez
JE,
Conover
MR,
Messmer
TA.
2001
.
Exclusionary methods to reduce predation on ground-nesting birds and their nests
.
Logan
:
Utah State University, Berryman Institute Publication 20
.
Jurries
RW.
1979
.
Attwater's prairie chicken
.
Austin
:
Texas Parks and Wildlife
Department, F. A. Series No. 18 (see Supplemental Material, Reference S1).
Laake
JL.
2013
.
RMark: an R interface for analysis of capture-recapture data with MARK
.
Seattle, Washington: National Marine Fisheries Service, Alaska Fisheries Science Center Processed Report 2013-01
(see Supplemental Material, Reference S2).
Lawrence
JS.
1982
.
Effect of predator reduction on the reproductive success of Attwater's prairie chicken. Master's thesis
.
College Station
:
Texas A&M University
.
Lehmann
VW.
1941
.
Attwater's prairie chicken, its life history and management
.
North American Fauna
57
:
1
65
.
Lehmann
VW.
1965
.
Fire in the range of Attwater's prairie chicken
.
Proceedings Tall Timbers Fire Ecology Conference
4
:
127
143
.
Lutz
RS.
1979
.
The response of Attwater's prairie chicken to petroleum development. Master's thesis
.
College Station
:
Texas A&M University
.
Morrow
ME.
1986
.
Ecology of Attwater's prairie chicken in relation to land management practices on the Attwater Prairie Chicken National Wildlife Refuge. Doctoral dissertation
.
College Station
:
Texas A&M University
.
Morrow
ME,
Adamcik
RS,
Friday
JD,
McKinney
LB.
1996
.
Factors affecting Attwater's prairie-chicken decline on the Attwater Prairie Chicken National Wildlife Refuge
.
Wildlife Society Bulletin
24
:
593
601
.
Morrow
ME,
Chester
RE,
Lehnen
SE,
Drees
BM,
Toepfer
JE.
2015
.
Indirect effects of red imported fire ants on Attwater's prairie-chicken brood survival
.
Journal of Wildlife Management
79
:
896
906
.
Parish
DM,
Sotherton
NW.
2007
.
The fate of released captive-reared grey partridges Perdix perdix: implications for reintroduction programmes
.
Wildlife Biology
13
:
140
149
.
Peterson
MJ,
Silvy
NJ.
1996
.
Reproductive stages limiting productivity of the endangered Attwater's prairie chicken
.
Conservation Biology
10
:
1264
1276
.
Oberholser,
H.
1974
.
The bird life of Texas
.
Austin
:
University of Texas Press
.
R Core Team.
2020
.
R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria
.
Ragheb
ELH,
Miller
KE,
Leone
EH.
2019
.
Exclosure fences around nests of imperiled Florida grasshopper sparrows reduce rates of predation by mammals
.
Journal of Field Ornithology
90
:
309
324
.
Rearden
JD.
1951
.
Identification of waterfowl nest predators
.
Journal of Wildlife Management
15
:
386
395
.
Sargeant
AB,
Kruse
AD,
Afton
AD.
1974
.
Use of small fences to protect ground bird nests from mammalian predators
.
Prairie Naturalist
6
:
60
63
.
Schroeder
MA,
Baydack
RK.
2001
.
Predation and the management of prairie grouse
.
Wildlife Society Bulletin
29
:
24
32
.
Smith
RK,
Pullin
AS,
Stewart
GB,
Sutherland
WJ.
2011
.
Is nest predator exclusion an effective strategy for enhancing bird populations?
Biological Conservation
144
:
1
10
.
Udall
SL.
1967
.
Native fish and wildlife endangered species. Federal Register 32:4001
.
Available at
: .
U.S. Endangered Species Preservation Act of
1966
,
Pub. L. No. 89-669, 80 Stat. 926 (Oct. 15, 1966).
[USFWS] U.S. Fish and Wildlife Service.
2010
.
Attwater's prairie-chicken recovery plan. 2nd revision. Albuquerque, New Mexico
(see Supplemental Material, Reference S3).
White
GC,
Burnham
KP.
1999
.
Program MARK: Survival estimation from populations of marked animals
.
Bird Study
46
Supplement
,
120
138
.

Author notes

Citation: Morrow ME, Toepfer JE. 2020. Use of predator-deterrent fences to increase Attwater's prairie-chicken nest success. Journal of Fish and Wildlife Management 11(2):455–462; e1944-687X. https://doi.org/10.3996/112019-JFWM-099

Competing Interests

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.

Deceased