---

Study Areas and Methods

Field Sites and Exclosure Designs

The 3 predator exclosures used in this study were located in waterfowl production areas near Kulm and Jamestown, North Dakota, and Detroit Lakes, Minnesota. Each exclosure encompassed about 25 ha of dense nesting cover (Duebbert et al. 1981). The exclosures were designed to exclude red fox (Vulpes vulpes), striped skunk (Mephitis mephitis), raccoon (Procyon lotor), badger (Taxidea taxus), and coyote (Canis latrans).

Exclosure fences were 1.8 m high and were supported by wooden posts at 5-m intervals. Atop each post, a 38-cm steel rod extended outward at a 45° angle. Wire mesh was attached to these rods to create an overhang. Electrified wires were attached at the top of the fence and 1.2 m and 10 cm above the ground along the outside of the fence. The bottom 0.6 m of the fence was 5- × 8-cm vinyl-coated wire mesh; the remaining fence was 2.5- × 3.8-cm wire mesh. Soil sterilant was applied to a 1-m wide band on each side of the fence to retard plant growth.

Before the 1989 nesting season, we installed 8 pairs of ground-level openings (design 1) in all 3 exclosures (Table 1). Exits were placed on each side of all 4 corner posts, with paired exits at the midpoints of each wall (200-310 m from corner exits). Inside the exclosure, leads were installed to help direct females to exits. Each lead was made of a rectangular piece of 2.5-cm hardware cloth placed at right angles to the fence, between the paired openings. One end of the lead was attached to the exclosure fence and the other end to a steel rod driven into ground. To keep out predators, each opening was bisected by a clear plexiglass flap (2.5 cm wide) that hung down on the outside of the opening and moved only when pushed from inside the exclosure. An electrified wire was attached to the outside surface of the flap.

By early June 1989, it was apparent that these exits were inadequate; some radiomarked female mallards with broods spent up to 27 hours along fences before we flushed the females over the fence. In late June 1989, we replaced design-1 exits in the Kulm exclosure with 32 pairs of much larger openings (design 2, Table 1). Four pairs of openings were in corners, the other 28 were evenly spaced around the perimeter of the exclosure (55-66 m apart). Because of their large size, design-2 exits were covered with wire mesh nightly to exclude predators. During the remainder of the 1989 season, we monitored radiomarked gadwall broods from nests in the Jamestown (design 1) and Kulm (design 2) exclosures and measured their exit times.

During the 1990-1991 nesting seasons, we tested 2 other exit designs (designs 3 and 4, Table 1) at all 3 exclosures. With these designs, we hoped to achieve brood-exit times similar to those with design 2 while reducing potential predator access and eliminating the need to cover openings nightly. The new exits were narrower than those in design 2, and the 28 paired exits along the perimeter were replaced by 24 single exits (65-78 m apart). To deter predators, the openings were protected on the outside by protruding metal rods and electrified wires (Fig. 1). In the Kulm exclosure, we installed hardware-cloth leads on alternate sides of adjacent exits. We monitored radiomarked gadwall and mallard broods to measure the time they required to exit the exclosure with leads (design 3) and the 2 exclosures without leads (design 4).

Fig. 1. Exterior view of ground-level brood exit installed in predator exclosures (designs 3 and 4) showing electrified wires were strung at the ends of steel rods that protruded 14 cm from the frame of the opening (design by R. J. Greenwood, NPWRC, Jamestown, N.D.).

Data Collection

We located nests inside exclosures by dragging a chain between 2 vehicles to flush females (Higgins et al. 1969). We attached 1.5- to 2.5-g radio transmitters (Krapu and Luna 1991) to gadwall and mallard ducklings at the nest site. Transmitters were glued and sutured to backs of 1-3 ducklings/brood. We followed American Ornithologists' Union (1988) guidelines for use of wild birds in research, and our procedures were approved by the NPWRC Animal Care and Use Committee. Broods were monitored continuously from the time they left the nest until they exited the exclosure. We recorded the time when each female and brood first encountered the fence. A brood's exit time was defined as time elapsed between encountering the fence and exiting the exclosure.

We gathered information by direct observation and time-lapse photography on behavior of females with ducklings at fences and exits. Observations were made through telescopes or binoculars from vehicles parked near exclosures. Film records (10 frames/min) were obtained with movie cameras on tripods inside the Jamestown exclosure near 4 exits in 1989 and 10 exits in 1990. Cameras were positioned 1 m from the fence and 8 m from an exit. If we observed or filmed a female along the fence, we noted her posture and locomotor behavior, whether she probed the fence with her bill, and whether she paced between exits. We documented changes of direction over greater distances with radio telemetry and direct observation. We used triangulation of radio signals and direct observation to determine approximate travel paths broods followed between the nest and the first wetland they reached.

Results

Brood Exit Times

The mean and variance of exit times for gadwall broods were lower from the exclosure with larger, more numerous openings (design 2) than from the exclosure with fewer, smaller openings (design 1, Table 2). All 9 gadwall broods exited quickly (≤1 hr) from the design-2 exclosure. Some broods exited quickly from the design-1 exclosure, but others failed to exit for several hours.

Despite the apparent success of design 2, we did not consider it a practical option for managers. Designs 3 and 4 provided a compromise between designs 1 and 2. The mean exit times for gadwalls from exclosures with designs 3 and 4 in 1990-1991 were between the means for designs 1 and 2 (Table 2). However, as with design 1, some female gadwalls required several hours to exit exclosures with designs 3 and 4 (e.g., 4 of 34 exit times were >4 hr).

Ranges of gadwall exit times from the exclosure with leads (design 3) and without leads (design 4) overlapped considerably (Table 2). Ranges in mallard exit times from exclosures with and without leads also overlapped. No trends were apparent; the mean exit time for mallards was lower from the exclosure with leads, but the mean for gadwalls was lower from the exclosure without leads.

Behavior of Females with Broods

For a number of reasons, sample sizes of females differed for each aspect of behavior we examined. The time and distance over which females could be monitored varied with observer and bird location, topography, and vegetation. Some behaviors (e.g., probing the fence) could be assessed only for females on which adequate observational records were obtained. Other behaviors (e.g., passing exits, changing direction) sometimes could be assessed from direct observation, film records, or radio telemetry. For each behavior examined, different criteria had to be met in deciding which females' records provided adequate information.

Of 55 females monitored as they left exclosures with their broods, 53 (96%) used ground-level exits and 2 flew over the fences. Females with broods first encountered the fence where it intersected their line of travel from the nest to a nearby wetland. When the fence blocked their paths, females typically ran or walked along it until they discovered an exit. Their postures while traveling along the fence varied from fully alert, with head and neck stretched high, to crouched, with head and neck extended low to the ground. While searching for an exit, 78% of 28 females observed at fences were seen probing the fence with their bills. Some females also stuck their heads and necks through the mesh and pushed against it with their bodies.

Eighty-five percent of 39 females paced back and forth between exits; 68% of 56 females changed directions while moving along the fences. Some females ran or walked along the fence for several hundred meters before stopping or back-tracking. Among 42 females monitored in 1990-1991, 33% traveled ≥200 m, 17% traveled ≥400 m, and 5% traveled ≥800 m along exclosure perimeters.

Of 60 females monitored during 1989-1991, 72% passed exits along the fences. Leads did not prevent females from passing exits. At least 63% of 30 females passed exits with no leads (design 4); however, 67% of 15 females passed exits with leads (design 3). Observations of female behavior at leads revealed different reactions. Some females stopped and changed directions when they encountered leads; others walked around leads and proceeded.

Exit-site Characteristics

The location of wetlands near the exclosures seemed to influence which exit was used. At the Jamestown study site, 2 wetlands were <100 m from the exclosure. In 1989-1991, 32 of 33 females led their broods to these wetlands. All 33 females used exits on the sides of the exclosure adjacent to the wetlands. Exits were used along only 39% of the exclosure perimeter.

Exits in exclosure corners also were used frequently. In 1990-91, 30% of 53 broods used corner exits although these constituted only 14% of the available exit sites and 25% of the total number of openings.

Duckling Mortalities

Thirty-six gadwall and 23 mallard females with broods were monitored as they tried to leave exclosures with exit designs 3 and 4. None of the females lost or abandoned their broods while trying to exit. Only 2 of 134 radiomarked ducklings (both mallards) were known to die while females tried to exit. One duckling was killed by a mink (Mustela vison); the other apparently died from exposure. In addition, 1 unmarked duckling from a radiomarked gadwall brood was killed by a northern harrier (Circus cyaneus) while their mother was pacing the fence.

Discussion

Predator exclosures with inadequate ground-level exits can be obstacles to females leading broods from the nest to water. Exclosures similar to ours but without ground-level exits were evaluated in Canada during 1991 (Trottier et al. 1994). Mallard broods took an average of 27 hours to exit those exclosures, approximately 7 times longer than the average for mallard broods exiting our exclosures (designs 3 and 4).

Some biologists and managers have suggested that brood delays may be related to the height of our exclosure fences. They believed that shorter fences may present less of a visual obstacle to females thus increasing females' likelihood of flying over fences. However, Sargeant et al. (1974) suggested that females were reluctant to fly over relatively short mesh fences. They constructed wire-mesh fences 0.6 m high and 30.5 m in circumference to protect individual nests of dabbling ducks and noted that ≥3 of 6 northern shoveler (Anas clypeata) females paced along the inside of these fences while leading their broods from the nest (A. D. Afton, Louisiana Coop. Fish and Wildl. Res. Unit, Baton Rouge, and A. B. Sargeant, NPWRC, Jamestown, N.D., pers. commun., 1991). Two ducklings from 1 brood were found dead along the fence; 2 other broods exited through the fences only after the female was flushed over the fence by an investigator several hours after they had left their nests.

Although our problem with brood delays involved mesh predator fences that completely surround areas of nesting cover, similar delays have been noted where open-ended fencing is used on peninsulas extending into marshes or lakes (as described by Lokemoen and Woodward 1993). In southcentral North Dakota, L. B. Albright (U.S. Fish and Wildl. Serv.[USFWS], Valley City, N.D., pers. commun., 1992) monitored an L-shaped wire-mesh fence used to protect 28 ha of nesting cover next to a large marsh. He saw dabbling duck broods in the landward corner of the fence nearly every time he visited the site and found the remains of several ducklings along the fence. In west central Minnesota, K. J. Brennan (USFWS, Fergus Falls, Minn., pers. commun., 1992) twice observed blue-winged teal (Anas discors) females with broods pacing along a wire-mesh fence used to cut off a peninsula of nesting cover. These broods had been at the fence at least a day and had created a path in the vegetation.

The longer broods are delayed along fences, the greater their chances of being detected by avian predators. Northern harriers frequently nested inside our exclosures; harriers and Swainson's hawks (Buteo swainsoni) often hunted in and around the exclosures and perched on the wooden fence posts. Trottier et al. (1994) observed raptors swooping near broods along the inside of exclosure fences on 4 occasions. Although only 1 of our radiomarked ducklings was killed by an avian predator along the fence, several ducklings were killed by hawks while traveling from the nest to the fence or on plowed fields immediately outside the fence. The strip of bare ground maintained along many predator fences may have helped females find exits more quickly but also made females and ducklings more conspicuous there.

Long delays at fences may increase the risk that females will abandon broods. In 1989, we flushed a radiomarked female mallard over an exclosure fence after she had spent >27 hours there; she did not return to her brood. Trottier et al. (1994) found that 3 of 18 females left and lost their broods at exclosure fences with no ground-level openings.

The inferential scope of findings presented in this paper is limited by the lack of replication of exclosures with different exit designs. The problem of brood delays, however, is likely to apply to most dabbling ducks nesting in exclosures with mesh fencing. Delays have been documented for northern pintail (Anas acuta) (P. J. Pietz, unpubl. data), mallard, gadwall, and blue-winged teal; for exclosures in North Dakota, Minnesota, and Canada; for closed and open-ended mesh fences; and for fences extending from 0.6 to 1.8 m above the ground.

Management Recommendations

Predator exclosures can increase waterfowl nest success where mammalian predation rates on nests are high. However, careful consideration must be given to exclosure design to ensure that hatched young can move to wetlands without long delays. Exclosures and other barriers with mesh fencing to the ground should have a system of ground-level exits for females with broods.

Additional exit design tests are needed to improve detection and use of exits by females, and to ensure predator exclusion. Meanwhile, our design 4 should serve adequately. The openings are large enough to be seen by an adult mallard in an alert posture, yet constructed with enough safeguards to deter most predators. We found no evidence (e.g., from traps and track plots inside fences) that mammalian predators entered the exclosures through these exits. To minimize the potential for predator entry, exits should be kept closed except during the hatching period.

The number, placement, and spacing of ground-level exits should be determined to some extent by the exclosure site and design. The most important locations for exits are at corners because many females did not search for an exit until they reached a corner. Other important exit locations are on portions of fences nearest wetlands. For hens that search short sections of fence for long periods of time, the chances of finding an exit improve as the distance between exits decreases. The average distance between openings in our exclosures was about 70 m, too great to facilitate exodus by all females but sufficient to allow an average exit time of 2.3 hours ( for mallards, gadwalls, designs 3 and 4 combined). Increasing the number of exits in areas predicted to have high use should expedite departures of pacing females.

Attention also should be given to maintaining exit access and visibility for females with broods. One female moved her brood from the nest to an exclosure corner in <30 minutes, then spent 7 hours at the exit blocked by windblown kochia (Kochia scoparia). The strip of bare ground next to the fence seemed to serve as a travel lane for females and probably helped them find exits more quickly. At sites where visibility to avian predators is a concern, tops of fence posts could be modified to prevent raptors from perching there (Melvin et al. 1992, USFWS and U.S. Navy 1990:3-15, 3-16).

Summary

Predator exclosures to protect duck nesting areas often are constructed with wire-mesh fencing. During a study that involved 3 such predator exclosures, we noted that mesh fences delayed ducks leading broods from nests to wetlands. Females often paced along the inside of fences for hours rather than flying over fences. In 1989, we installed ground-level openings in our exclosure fences and measured the time it took females with radiomarked ducklings to exit. Mean exit time was lower (0.6 hr) for 9 gadwall broods in an exclosure with 32 pairs of 20- × 20-cm openings than for 6 gadwall broods (3.2 hr) in an exclosure with 8 pairs of 10- × 12-cm openings. In 1990-1991, we used exit designs with 32 single 22- × 16-cm openings in all 3 exclosures. Mean exit times were 1.5 hours for 34 gadwall broods and 3.9 hours for 16 mallard broods. Hardware-cloth leads directing females to openings did not reduce exit times. In 1990-1991, no broods were abandoned, and only 3 of 134 radiomarked ducklings died while the female tried to exit. Only 2 of 55 females left exclosures by flying over the fences. Ground-level openings reduced the time it took females with broods to leave exclosures and reduced potential exposure of ducklings to predators and adverse weather. We recommend installing ground-level exits in predator exclosures with mesh fences.

---

Acknowledgements

R. J. Greenwood, A. B. Sargeant, J. T. Lokemoen, and L. M. Cowardin jointly coordinated the construction and modification of predator exclosures with the authors. R. J. Greenwood and R. O. Woodward designed and built ground-level exits installed in predator exclosures. L. Sileo and M. D. Samuel (USFWS, Natl. Wildl. Health Res. Cent.) examined mallard duckling carcasses for causes of death. A host of dedicated technicians radiotagged and monitored duck broods and females; special thanks to C. R. Luna, R. T. Speer, D. W. Howerter, D. A. Brandt, L. A. Baker, C. P. Dwyer, E. D. Gilbert, R. J. Dusek, J. T. Huesby, G. G. Mack, W. L. Mulvaney, R. L. Sanders, D. J. Sausville, and S. J. Yu. J. T. Lokemoen and R. O. Woodward led nest-searching crews. A. D. Afton, L. B. Albright, K. J. Brennan, D. C. Duncan, D. W. Howerter, and G. C. Trottier kindly provided unpublished information from research conducted at other predator fences. We thank W. E. Newton for providing statistical advice and analyses. J. T. Lokemoen, A. B. Sargeant, R. J. Greenwood, D. H. Johnson, A. D. Kruse, M. R. Ryan, and an anonymous reviewer provided helpful comments on earlier drafts of this manuscript.

---

Literature Cited

---

---

---

NPWRC Home | Site Map | About Us | Staff | Search | Contact | Web Help | Copyright