---

Introduction

Radio telemetry is commonly used to estimate habitat use, survival, and productivity of ducks; nevertheless, few studies have attempted to evaluate the possible effects of radio packages on waterfowl. Most investigators using telemetry on dabbling ducks noticed no obvious effects of radio packages and thus assumed no major effects on the measured characteristics (Ball et al. 1975, Talent et al. 1983, Cowardin et al. 1985, Duncan 1986).

A few studies, however, reported effects of radio packages on duck behavior or condition. Schladweiler (1969) used a 2-loop harness to attach breast-mounted transmitters to game-farm mallards released as 6-week-old ducklings. He reported evidence of transmitter-induced mass loss, but detected no significant effect on mortality. Gilmer et al. (1974) used a similar breast-mounted package on wild free-ranging mallards and wood ducks (Aix sponsa) during the breeding season. They reported a possible decrease in survival, an increase in preening, and some feather wear and skin abrasion. Greenwood and Sargeant (1973) put back-mounted transmitters with Dwyer (1972) harnesses on captive mallards and blue-winged teal (Anas discors) in winter. They documented mass loss, feather wear, skin irritation, preoccupation with radio packages, and an aversion to swimming. More recently, studies were conducted on wild free-ranging mallards during the breeding season in Minnesota and North Dakota (Pietz et al. 1993) and Alberta (Rotella et al. 1993). Pietz et al. (1993) observed that females equipped with back-mounted, Dwyer-harness transmitters fed less and rested and preened more than unmarked birds. Rotella et al. (1993) reported that females equipped with back-mounted Dwyer-harness transmitters initiated fewer nests and devoted fewer days to egg laying and incubation than females equipped with implanted transmitters. All investigators warned that subtle effects of radio packages on bird behavior or condition could bias findings and lead to erroneous interpretations of data.

We investigated the direct effects of radio packages on reproductive biology of captive mallards. Our objectives were to evaluate whether there were effects of 3 types of radio packages on number of nesting attempts, nesting interval, clutch size, egg mass, and female mass; and to determine retention time and antenna durability of each radio package type. A radio package refers to harness, transmitter, and antenna. Nesting interval refers to number of days between radio package attachment and initiation of first clutch or interval between removal of a clutch and initiation of a successive clutch.

We appreciate help from T. R. Boschee, M. N. Miller, J. D. Moodie, P. J. Pietz, B. L. Reed, and G. L. Reed. R. E. Ingstad provided experimental mallards. G. L. Pearson provided veterinarian services. J. E. Austin contributed to initial study design. W. E. Newton provided valuable advice on data analyses. C. A. Calheim and D. Van Asperen maintained experimental pond facilities. J. E. Austin, J. F. Bergen, G. L. Krapu, P. J. Pietz, and J. K. Ringelman provided helpful comments on earlier drafts of this manuscript. Partial funding was provided by Ducks Unlimited through the Institute for Wetland and Waterfowl Research. Use of trade names in this document does not imply endorsement of products by Ducks Unlimited or the U.S. Fish and Wildlife Service. We followed the American Ornithologists' Union (1988) guidelines for use of wild birds in research, and our procedures were approved by the Northern Prairie Wildlife Research Center (NPWRC) Animal Care and Use Committee.

Study Area

Our study was conducted on 16 experimental ponds at NPWRC, approximately 3 km southeast of Jamestown, North Dakota. Ponds and adjacent upland areas were individually fenced and covered to exclude predators (Swanson et al. 1986). Each pond had an approximately 225-m² flat bottom. Water levels in ponds were monitored daily and maintained at 15-35 cm by constantly flowing well water, providing a surface area of approximately 285 m². Shorelines sloped upward 5 m from the bottom to a 2-m flat shoulder around each pond; approximately 550 m² of natural vegetation was available for nesting on the slope and shoulder. We divided each pond and adjacent upland in half with a 1.2-m-high wire fence and provided a covered nest box in each half. Ponds were drained independently for 6 hours every 1-3 weeks to reestablish water seals. Drinking water was available in shallow pans during the draw-down period.

Plant communities varied slightly among ponds and adjacent upland. Upland vegetation consisted of smooth brome (Bromus inermis), quackgrass (Agropyron repens), Kentucky bluegrass (Poa pratensis), absinth (Artemisia absinthium), leafy spurge (Euphorbia esula), and Canada thistle (Cirsium arvense). Wetland vegetation consisted mainly of tall mannagrass (Glyceria grandis), sloughgrass (Beckmannia syzigachne), and reed-canary grass (Phalaris arundinacea). Varying amounts of cattail (Typha spp.) and smartweed (Polygonum spp.) were also present in some ponds.

Methods

We used wild-strain mallards hatched from eggs collected in North Dakota. Birds were obtained in early April 1991 from Ingstad Ranches, Valley City, North Dakota, where they had been held in an outdoor pen since hatching. Winter diet at Ingstad Ranches was corn and wheat screenings. Females were 3 years old and males were 2-4 years old. Birds were free-flying when acquired but were wing-clipped at the beginning of the study. At NPWRC, all birds were initially held for approximately 2 weeks in 10 small (2.1 × 11.1 × 2.3 m) outdoor pens (6-10 birds/pen; equal sex ratio). Some birds were acquired as mated pairs; others paired during the first 2 weeks at NPWRC. Pairs were determined by their behavior (Lorenz 1971:22-42). We determined the mass of all females when acquired, at time of treatment, and at the end of the study. Change in mass was calculated as mass at end of study minus mass at time of treatment.

Oyster shell, grit, and 16% protein commercial ration (Zip Duck and Goose Grower, Zip Feed Mills, Sioux Falls, S.D.) were provided ad libitum during the study; feed was the same as during previous nesting seasons at Ingstad Ranches. Wheat screenings were provided ad libitum during the first month that birds were on the experimental ponds. Seed shrimp (Ostracoda), fairy shrimp (Eubranchiopoda), and mosquito larvae and pupae (Culicidae) were present in all ponds. Nesting ducks consume invertebrates (Meyer and Swanson 1982, Swanson et al. 1985), but we did not quantify invertebrate use or abundance.

Because we expected large differences among ponds, we randomly assigned mallard pairs (32) to one of 4 treatments, each pair to 1 side of a pond in a randomized, unbalanced, incomplete block design. Treatments consisted of 4-g (4.05 ± 0.22 g [ ± SD]) radio packages attached with sutures and glue (Wheeler 1991), 10-g (9.89 ± 0.59 g) radio packages attached with Dwyer harnesses, 18-g (17.54 ± 0.37 g) radio packages attached with Dwyer harnesses, and controls that had no radio packages attached. Radio packages were non-operational and were obtained from Advanced Telemetry Systems (Isanti, Minn.) and Holohil Systems, Ltd. (Woodlawn, Ont.). Radio packages equalled 0.40, 0.99, and 1.75% of average body mass of mallard females at time of treatment for the 4-g, 10-g, and 18-g radio packages, respectively. All radios were attached to mallards on 17 or 18 April by the senior author. Mallards were hooded during radio package attachment; control females were hooded and handled as if radio packages were attached.

Four-g sutured radio packages were 24 mm long × 12 mm wide × 7 mm high. The transmitter was attached dorsally between the wings and adjacent to the backbone with 2 sterile nylon nonabsorbable sutures (2-0 Supramid, S. Jackson, Inc., Alexandria, Va.). A 244-mm antenna extended from the posterior end of the transmitter and lay flat along the bird's back. Each suture was threaded laterally through holes in the back or front of the transmitter and through the skin. Connective tissue or muscle were not included in the sutures. Individual sutures were tied off with 2 square knots, which were secured with cyanoacrylate glue (Quick Gel Super Glue™). Glue also was applied to the underside of the transmitter and bonded to back feathers. Birds were not anesthetized for this procedure. After radio packages were attached, birds preened both transmitter and antenna under feathers.

Ten-g harness transmitters were 21 mm long × 28 mm wide × 6 mm high, and 18-g harness transmitters were 30 mm long × 23 mm wide × 16 mm high. Both types had a 220-mm antenna that extended from the posterior end through a small spring (12 mm long × 4 mm diam). Antennas angled 10° upward from the plane of the 10-g, and 45° in 18-g transmitters. Harness transmitters were attached following published procedures (Dwyer 1972). Body and neck loops were tightened so a finger (approx. 1.5 cm diam) could fit between each loop and the bird's keel. Both loops were preened under feathers by the investigator. The loose end of harness tubing was tied off with 2 overhand knots and secured with polyvinyl chloride (PVC) cement. To prevent loops from changing size, PVC cement was applied where the harness passed through the transmitter housing. In addition, a square knot (using excess harness tubing) was tied and cemented at the base of the neck loop just in front of the transmitter.

We observed each mallard pair daily to determine nesting status and retention of transmitter, but we did not quantify their behaviors. When nesting was suspected, nest boxes were examined and upland cover was searched. Nests were revisited every 2 days to count eggs. We removed clutches and determined the mass of eggs 2 days after the last egg of the clutch was laid. Females then were allowed to renest. We repeated these procedures until 1 August, 10 days after the last clutch was completed.

We analyzed treatment effects on number of clutches and body mass with analysis of variance (ANOVA) blocked by pond. Treatment effects on clutch size, nesting interval, and average egg mass per clutch were analyzed with a split-plot ANOVA, split by clutch number. We included clutches in these analyses only if represented by data from all treatments. Statistical treatment of the data followed methods described in Milliken and Johnson (1984) with the SAS General Linear Model procedure (SAS Inst. Inc. 1987). Significance levels were set at 0.05. We estimated power of F-tests following procedures in Kirk (1982:142).

Results

Data from 29 mallard females were available for analyses (Table 1). Two females lost radio packages before nesting, and 1 pair was removed 7 days into the study after the drake died. Sample sizes decreased with successive clutches because some radio packages fell off and some females stopped nesting.

Mean number of clutches did not vary (F_3,6 = 0.62, P = 0.625, power = 0.12) among treatments. The average number of clutches for controls and females (n) that retained their radio packages for the duration of the study was 2.9 (n = 8, 0.4 SE) for controls, 3.0 (n = 1) for the bird with the sutured transmitter, 3.0 (n = 7, 0.04 SE) for birds with 10-g harness transmitters, and 3.5 (n = 8, 0.5 SE) for birds with 18-g harness transmitters. The maximum number of clutches was 5, produced by 2 birds with 18-g harness transmitters. Because no other birds nested 5 times, we did not include the fifth nesting attempt in the following analyses.

Average clutch size (Table 2) did not differ (P = 0.187) among treatments (Table 3), and we found no interaction between treatment and clutch number (P = 0.881). Average nesting interval (Table 4) did not differ (P = 0.854) among treatments, and there was no interaction between treatment and clutch number (P = 0.993). However, clutch size and nesting interval varied with clutch number (P < 0.001; Tables 3 and 4). Clutch sizes for all treatments generally decreased with consecutive clutches, whereas nesting intervals between clutches generally increased (Table 2). The initial nesting interval was larger than intervals between successive clutches. Average egg mass per clutch (Table 5) did not vary among treatments (P = 0.935) or among clutch number (P = 0.142), and we detected no interaction between treatment and clutch number (P = 0.268). Powers of the above tests were low (Tables 3, 4, and 5), indicating small sample sizes may have influenced our ability to detect differences among treatments.

Only 1 sutured transmitter remained attached for the entire study (106 days); at the end of the study it was loose and not bonded to any back feathers. Knots and suture material were still intact and showed no signs of wear. The average retention time for sutured transmitters was 43.5 days (SD = 34.8, range = 3-106 days). Sutured transmitters that fell off were not recovered, but scabs at the sutured area of 1 female examined several days after the radio package came off suggested that her sutures ripped out. All harness transmitters (n = 15) were retained for the entire study. In general, harnesses were not as tight as when attached, probably because birds had lost mass. Harness loops were still preened under feathers and seemed to be in the same position as when attached. Knots and cement were intact on all but 1 radio package; loops on this radio package changed size, but did not affect transmitter position.

Controls and birds with harness transmitters lost mass during the study, but the 1 bird that retained a sutured transmitter gained mass (Table 6). Mass change did not differ among the 4 treatments (F_3,6 = 0.65, P = 0.6114, power = 0.15). Following the methods of Milliken (1984), we analyzed the data to see if mass change was affected by total number of eggs laid or number of days between removal of last clutch and the end of the study (days since last clutch). For this analysis we excluded the 4-g suture treatment because n = 1, and combined the 10- and 18-g harness treatments because we detected no difference in mass change between them. Mass change between the combined harness treatments and control was not related to total eggs laid (F_2,5 = 0.01, P = 0.9938). Likewise, we detected no relationship between treatments in mass change and days since last clutch (F_2,5 = 0.24, P = 0.6380). Because of limited sample sizes we could not simultaneously test total eggs and days since last clutch on mass change.

The antenna of 1 sutured transmitter broke 62 days after attachment to the bird; the break was 45 mm from the base. Antennas of 2 18-g harness transmitters broke 26 and 58 days after attachment. Both antennas broke at the end of the spring, 12 mm from the base. No antennas broke off the 10-g harness transmitters; these radio packages had a dab of silicone where the antenna exited the spring. Daily observations revealed females with radio packages often aggressively tugged at the antenna and preened around the harness or transmitter. Control birds also were seen preening at times. Drakes were not observed pulling on radio packages of their mates.

At the end of the study, the 1 bird that retained a sutured transmitter showed no sign of feather wear or skin irritation beneath the transmitter. Scar tissue was present where sutures pierced the skin. Feather wear and skin irritation also were minimal under the 10- and 18-g transmitters. Four of the 15 birds with harnesses showed no feather wear or skin irritation under transmitters; 11 had slight to moderate feather wear under transmitters. Six of the 15 birds showed no sign of feather wear or skin irritation behind their wings; nine had slight to moderate feather loss behind wings from rubbing of the posterior harness loop.

Discussion

Our results suggest there was no influence on nesting from radio packages tested on captive female mallards, but these results are not directly applicable to wild, free-ranging mallards. Nesting in wild mallards is affected by a variety of factors, including food availability (Swanson et al. 1986) and habitat conditions (Pospahala et al. 1974, Cowardin et al. 1985). Birds in our study had unlimited food, experienced little environmental stress, were protected from predators, and expended no energy on flight. Other recent studies, however, indicated that wild female mallards with back-mounted radio packages attached with harnesses fed less and rested and preened more, initiated fewer nests and devoted fewer days to egg laying and incubation than females with other types of transmitters, or females with no transmitters (Pietz et al. 1993, Rotella et al. 1993); some effects could bias recruitment estimates (Cowardin et al. 1985).

Research Implications

We found, as did Rotella et al. (1993), that small transmitters attached with sutures are not currently a viable alternative to radio packages attached with harnesses because sutures do not provide reliable attachment. Improvements are needed in methods of attaching light-weight radio packages. Research also is needed to evaluate the impact of radio packages on wild mallards in other phases of their life cycles, and on other duck species.

---

---

---

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