Study areas

Wild and pen-reared pheasants were released in two study areas (MiniCassia and Jefferson) privately owned in southern Idaho on the Snake River Plain within the range of pheasants in the state (Fig. 1). The study areas are relatively flat (1,260–1,450m a.s.l.) with an annual precipitation of 20–30 cm. Although the Snake River Plain was originally dominated by sagebrush Artemisia tridentata steppe vegetation, much of the area is now dominated by irrigated agricultural crops. Agricultural crops and habitat types include wheat Triticum aestivum, alfalfa Medicago sativa, sugar beet Beta vulgaris, grass pastures, grass road ditches, isolated patches of shrub-land, and narrow riparian zones with trees and berry producing shrubs.

Release areas were chosen by local advisory groups, appointed by state legislators to represent the interests of the local sport-hunting public. Release areas were based on the hunting experience of the advisory groups and, in their opinion, areas currently having adequate habitat but too low pheasant numbers.

Figure 1.

Study areas in Idaho where pheasants were released and monitored in 2000–2001. Shading denotes pheasant distributions.

Wild pheasant translocation

Wild pheasants were captured in Oregon (Malheur National Wildlife Refuge) with mist nets and walk-in baited traps (Nohrenberg 1999) and in California (Sacramento Valley) by night-lighting (Giesen et al. 1982, Wakkinen et al. 1992). During 2001, birds were only captured by night lighting and only in California.

Prior to transport to release sites, we equipped the pheasants with battery-powered 14-g necklace radio transmitters (Riley & Fistler 1992), programmed with 4-hour mortality sensors (Advanced Telemetry Systems, Isanti, MN 55040). Only pheasants weighing > 700 g were radio-marked to keep transmitters ≤2% of body mass (Kenward 1987).

Pen-reared pheasant releases

In both years, 10-month-old pen-reared stock was purchased (US $9.50/bird) from local game farms. During 2000, birds were held overnight, fitted with aluminum bands and radio-collars, nasal blinders removed, and released the next day. During 2001, we radio-marked pheasants three days prior to release allowing them to acclimate to the radios at the game farm before being released. The same style of radio transmitter was used on pen-reared as on wild pheasants.

Telemetry and survival

We monitored radio-marked birds from the ground with 3-element yagi antennas twice per week in 2000 and 3–5 times per week in 2001 from day of release to 1 October each year. We determined cause of death by characteristics presented by Einarsen (1956). We defined time of death as the mid-point between the first mortality signal and the last live contact. We estimated survival using the Kaplan-Meier staggered entry method (Pollock et al. 1989) with the computer program STAGKAM (Kulowiec 1988). Multiple comparisons among survival estimates were protected by significant χ ² tests with program CONTRAST (Hines & Sauer 1989, Sauer & Williams 1989). Comparisons of body mass were made using the GLM procedure and Tukey's Studentized Range Test (HSD) for multiple comparisons. Overlap of 95% confidence intervals (Fleis 1981) was used to test for differences in the proportions of birds dying over time. All values are presented with 95% confidence intervals (value ± 95% CI).

Pheasant production

We recorded number of eggs hatched immediately after termination of the nesting attempt. Additionally, in 2001, we determined clutch size by flushing the female from the nest during the last week of incubation. We considered a nest attempt successful if at least one egg hatched. Brood survival was determined by counting flushed chicks at four and eight weeks post-hatch (Nohrenberg 1999) in 2001. We used overlap of 95% confidence intervals (Fleis 1981) to test differences in proportions of females surviving to the nesting season (1 May), nesting rates and nest success. Nesting rate was defined as the proportion of hens attempting to nest of those alive at the beginning of the nesting season.

Predator control and abundance

After establishing baseline vital rates for translocated pheasant stocks in 2000, losses to predation were deemed unacceptably high and an adaptive management approach was implemented in late March 2001 through July 2001 by contracting two trappers, one for each study area. Trappers were given a list of target species and trapped within a 4.8-km (72 km²) radius of the release sites. Just prior to pheasant releases, trappers concentrated on mammalian predators within 1.6 km of release sites. Padded leg-hold traps (size #1.5, #3) were provided for trapping near residences to avoid harming domestic pets (Olsen et al. 1988, Onderka et al. 1990, Hubert et al. 1997). Unpadded steel jawed traps, snares and conibear traps (size #120) were used in remote areas. Walk-in circular live traps (Alsager et al. 1972) were used to remove corvids after pheasants were released. No other avian predators were removed.

We informed landowners of the predator control project and trappers were restricted to those granting permission. All captured predators were disposed of according to procedures established by the Idaho Department of Fish and Game. House cats without identification collars were euthanized and collared cats were released. Trappers recorded their time spent trapping, number of trap nights, and number and species of predators trapped for 2-week periods throughout the trapping season (March-July).

In 2001, we used roadside spotlight counts at night and scent station surveys on a 40-km route in each area immediately surrounding the release site to assess predator activity and effectiveness of predator control. Roadside spotlight counts were conducted in June-July and involved driving a vehicle 20 km/hour while an observer scanned with a 1-million candle power spotlight and binoculars. Scent surveys (Roughton & Sweeny 1982, Travaini et al. 1996, Sargeant et al. 1998) were conducted with 20 stations per study area during the first week of August, after predator removal ceased. Stations were monitored for four evenings of exposure.

Translocation

Wild pheasants were passively released in both years. Padded lids (to avoid scalping) of the transport crates were pulled open simultaneously with a string by concealed observers and birds walked or flushed from the site without disturbance. From 29 February-31 March 2000, 92 wild pheasants (71 females and 21 males) were released within three days of capture (Table 1). On 28 March 2001, 51 wild pheasants (41 females and 10 males) were released six days after capture. Wild birds were held in captivity longer in 2001 because results of mandatory blood tests were delayed. All wild birds had to test negative for disease before legally released into Idaho. We radio-marked 74% of wild birds in 2000, and 98% in 2001.

Table 1.

Number of wild and pen-reared pheasants translocated into Idaho, USA. M = males and F = females. Number of birds with radio-transmitters are given in parentheses.

Pen-raised birds were actively released by pouring them out of the transport crates, because they tend to remain in the crates even if opened. On 6–7 April 2000, 159 female and 30 male pen-reared pheasants were released in MiniCassia and Jefferson study areas (see Table 1). On 16 March (MiniCassia) and 30 March (Jefferson) 2001, 900 female and 200 male pheasants were released. We equipped 30% of the pen-reared birds in 2000 with radios, and 5% in 2001. Wild females were significantly lighter than all other birds, with no detectable effect of year on body mass (Table 2). For males, we detected no effect of release stock or year on body mass.

Predator control

Because of frozen ground conditions in MiniCassia, predator control could not begin earlier than two days before release of pen-reared birds and 14 days before that of wild birds in 2001. In Jefferson, predator control started 11 days and 9 days before pen-reared and wild releases, respectively. Trapping in MiniCassia removed 271 black-billed magpie Pica pica, 35 striped skunk Mephitis mephitis, 13 coyote Canas latrans, five mink Mustela vison, two red fox Vulpes fulva, two badger Taxidea taxus, and one feral cat Felis domesticus for 2,358 trap nights costing US $3,119. In Jefferson, predator trapping removed 189 black-billed magpie, 53 striped skunk, 33 red fox, 11 feral cat, and two coyote for 1,526 trap nights costing US $5,441. Predator trapping cost US $13.87/animal removed and US $2.20/trap night. Only the contracted labor was included in the estimate, not the cost of traps.

Table 2.

Mean body mass of pheasants released into Idaho, USA. GLM F statistic = 40.0, P < 0.001, multiple comparisons (Tukey's Studentized Range Test) with same letter are similar at 0.05 level of significance.

Survival

From time of release to 1 October, wild female pheasants had greater 6-month survival than pen-reared females during both years (Table 3). Wild males survived better than pen-reared males in 2001, but not in 2000. Wild male pheasants had greater overall survival in 2001 than 2000, but there were no differences between years within the same stock and gender for wild females or for pen-reared females and males.

A higher proportion of pen-reared female pheasants died during the first seven days post-release in 2000 than in 2001 (0.59 ± 0.15 vs 0.22 ± 0.14). Moreover, the proportion of females that died during these seven days in 2000 was greater among pen-reared than wild hens (0.06 ± 0.08). Proportionally more wild female mortalities than pen-reared occurred > 30 days post-release in 2000 (0.88 ± 0.11 vs 0.14 ± 0.10), but not in 2001 (0.68 ± 0.19 vs 0.36 ± 0.16). All pen-reared male mortalities (100%, N = 17) occurred within 30 days post-release and most (50–86%) occurred within seven days of release. No other differences for males could be detected due to low sample sizes. Survival of pen-reared females appeared to decline more rapidly during the first two months post-release for the year prior to predator removal (2000) than after predator removal (2001), but survival approached the same estimate for both years by 1 October (Fig. 2).

Table 3.

Translocated pheasant survival (1 March-1 October) and comparisons for pre- (2000) and post-predator (2001) removal in Idaho, U.S.A.

Figure 2.

Survivorship functions and causes of mortality by month for wild and pen-reared female pheasants released during 2000 and 2001 in Idaho, USA.

Most of the mortality was caused by predation of pen-reared (99%) and wild (82%) released female pheasants (see Fig. 2). Predator-specific mortality could not be determined for the majority (54%) of documented female pheasant deaths, therefore, cause-specific mortality rates were not computed. At least 26% of the mortalities were caused by mammalian predators, 12% by avian predation, 4% from natural causes, and 4% were human caused.

Nesting

The proportion of radio-marked pen-reared female pheasants surviving to the nesting season (1 May) was not different before predator control (0.18 ± 0.11, N = 49) and after (0.40 ± 0.15, N = 40). Survival to the nesting season for wild females also remained constant with 0.89 ± 0.08 (N = 62) before and 0.83 ± 0.12 (N = 40) after predator control. Both years combined, more wild females (0.86 ± 0.07, N = 102) survived to the nesting season than pen-reared (0.28 ± 0.09, N = 89) birds. However, perhaps due to small sample sizes, we could not detect a change in nesting rate or survival to produce chicks for either pen-reared or wild females before and after predator control. For both years combined, nesting rates were similar for pen-reared (0.28 ± 0.09, N = 25) and wild (0.23 ± 0.09, N = 88) released pheasants, but, a smaller proportion of pen-reared (0.01 ± 0.02, N = 89) than wild (0.10 ± 0.06, N = 102) radio-marked females survived to produce chicks. We could not detect differences in nest success and average clutch sizes due to small samples sizes of nests (seven for pen-reared and 20 for wild nests for both years combined).

Predator abundance

The goal of monitoring predators was to document their presence during the later part of the trapping period and after trapping was concluded. Spotlight counts were conducted during the last two months of predator trapping, June–July, 2001. House cat presence in MiniCassia increased from 0.13/km–0.99/km during June through July. Detection of striped skunk in Jefferson increased from 0.08/km–0.39/km for the same period. Scent station surveys were conducted within seven days of the end of the trapping season. In MiniCassia, we detected attendance by house cats (0.36 visits/station), domestic dogs Canis familiarus (0.13 visits/station), unknown canids (0.03 visits/station), and striped skunks (0.02 visits/station). Scent station attendance in Jefferson included house cats (0.08 visits/station), coyotes (0.06 visits/station) and red fox (0.01 visits/station). Of all the predators removed in 2001, 11% of the black-billed magpies, 7% of the striped skunks, and 3% of the red fox were removed during the last two weeks of trapping in July. These predator surveys and trapping records clearly indicate that a variety of predators were still relatively abundant throughout the summer in 2001, and continued to cause pheasant mortality (see Fig. 2) despite the implementation of predator control.

Pheasant costs

We did not determine cost estimates for wild pheasant trapping in 2000, but in 2001, an estimated US $53/bird was needed to capture, transport and release wild pheasants (D. Musil, unpubl. data). Pen-reared pheasants cost US $9.50/bird to be released. Using production estimates from radio-marked pheasants and the cost of release in spring, it cost US $170/pen-reared egg hatched and US $62/wild egg hatched. Estimates for brood survival were too limited to estimate cost/bird available for the fall harvest.

Translocation

In Idaho, pen-reared pheasants released into the same habitats as translocated wild pheasants had significantly lower survival. In fact, wild females were seven times more likely to survive from spring to 1 October than pen-reared females. In 2001, we acclimated pen-reared pheasants to radio-collars and reduced transportation time, but did not detect an increase in overall survival. Our study provided similar survival results to those of Leif (1994) for radio-marked pen-reared (0.08 ± 0.05) and wild (0.55 ± 0.12) females released in South Dakota in early April and monitored until October. Similarly, Anderson (1964) had 0.28 game-farm and 0.52 wild pheasant survival from winter to May, while Wilson et al. (1992) reported 0.26–0.42 survival (February-June) for translocated wild female pheasants. Brittas et al. (1992) also reported significantly higher survival for translocated wild female pheasants (0.74) than pen-reared (0.52) for three months post-release (May-July).

High mortality commonly occurs immediately after release as birds become acclimated (Hessler et al. 1970, Wilson et al. 1992) and this finding is supported by our data, especially for pen-reared stock. Translocated wild game birds may be more vulnerable to predation immediately after release due to increased movements while searching for adequate habitat (Kurzejeski & Root 1988, Musil et al. 1993). Burger (1964) attributed heavy initial mortality of game-farm stock to ‘release shock’ rendering pheasants more vulnerable to predation. Pen-reared pheasants may be more susceptible to predation, because they are unfamiliar with wild foods, thus increasing their foraging time and reducing predator avoidance behavior.

Other studies (Giudice & Ratti 2001) have shown average clutch sizes for first attempts of > 12 eggs/nest and later attempts were closer to our average. Mean clutch size for wild pheasants in our study (8.6 eggs/nest) was similar to that reported by Nohrenberg (1999) for resident pheasants in southern Idaho (8.8 eggs/nest) and may be the normal response to habitat conditions rather than the result of translocation.

Many studies have demonstrated the inability of game-farm stock to augment wild populations. In our study, the production of a successful nest required either 11 wild or 89 pen-reared females pheasants. Similarly, Haensly et al. (1985) found it required seven game-farm females released in spring to produce one rooster in fall while Wilson et al. (1992) suggested a February release of 150 translocated wild female pheasants was needed to have 50 survive to nest. Hill & Robertson (1988) determined that breeding success was 2–5 times greater for wild than hand-reared male pheasants. They also found wild females were four times more productive than hand-reared females and hand-reared females were three times more vulnerable to predation than wild birds during April-August. Hill & Robertson (1988) estimated that wild pheasants produced seven times more 12-week-old chicks than hand-reared females. Although Brittas et al. (1992) found no difference in breeding success between wild and pen-reared stock, they indicated that wild brood size was significantly larger 40 days after hatch than brood size of pen-reared females. Leif (1994) concluded, “Releasing pen-reared females in spring to augment wild ring-necked pheasant populations is not a practical management option in South Dakota” and our results indicate that his conclusion can be applied more broadly than just to South Dakota.

Similar results have been reported for translocation of game-farm birds of other gallinaceous species. Parish & Sotherton (2007) examined the fate of released grey partridge Perdix perdix in Scotland and concluded that birds released in autumn had poor over-winter survival. Hess et al. (2005) demonstrated that pen-reared Attwater's prairie chickens Tympanuchus cupido attwateri had deficiencies in flight endurance and predator avoidance compared to wild greater prairie chickens T. c. pinnatus. Snyder et al. (1996) stated that captive breeding should be considered a last resort for species recovery and not a long-term or prophylactic solution to problems. Finally, Svedarsky et al. (2000) recommended that translocations should only be conducted in areas with adequate habitat where problems that caused the original population decline have been corrected.

Predator control

We could not detect an increase in overall survival of wild female pheasant transplants as a result of our short-term and small-scale predator removal, but it may have improved survival of wild male pheasants. Predator control appeared to decrease mortality for pen-reared female pheasants early after release, but did not ultimately increase overall survival.

Several studies have measured effects of predator removal on pheasant vital rates. Chesness et al. (1968) concluded that predator removal, even though it did increase nest success of pheasants, was not economically feasible for increasing pheasant numbers over large agriculturally dominated areas. Similarly, Trautman et al. (1974) found that multi-species predator control substantially increased pheasant numbers, but, intensive control of only foxes showed little effect on pheasants. In contrast, Nohrenberg (1999) could not demonstrate significant increases in pheasant abundance after multi-species predator removal in southern Idaho and his April-July wild female survival (0.51 ± 0.27 with predator control vs 0.34 ± 0.24 without) was similar to our wild female survival with (0.54 ± 0.16) and without (0.54 ± 0.14) predator removal for April-July.

Management implications

Wildlife managers cannot expect comparable survival and production from pen-reared pheasants relative to wild stock. It is clearly more cost effective to release wild birds to supplement wild populations. In our study, the capture and release of a wild bird costs 5.6 times more than the purchase and release of a pen-reared bird. But the cost of producing a young bird in summer was about three times greater for pen-reared than for wild stock. Wild female pheasants were 8.1 times more productive (hatched nests/female) than pen-reared females. If the intention is strictly to increase hunting opportunity, releasing before the gun is a much more efficient and practical use of pen-reared stock. Artificially supplementing game birds for harvest was criticized by Leopold (1933:394), when he said, “A proper game policy seeks a happy medium between a game supply and that which deteriorates its quality or recreational value” and “The recreational value of a head of game is inverse to the artificiality of its origin.”

Predator control immediately before translocation appeared to slow the rate of mortality for pen-reared pheasants in our study, but did not increase overall survival for either pen-reared or wild translocated female pheasants. Predators were still relatively common after trapping ended and predation on pheasants still occurred, therefore, predator control was an extra added cost that was not effective in our study. If managers decide to use predator control, they must be committed to the cost of long term removal (Beasom 1974, Duebbert & Kantrud 1974) over large areas (Frey et al. 2003), and on multiple species (Chesness et al. 1968, Trautman et al. 1974) to be effective in reducing predator populations. The severity of the predator impacts (Balser et al. 1968), cost-benefit (Jimenez & Conover 2001) and social impacts (Rollins & Carroll 2001) must be determined by managers before predator control is implemented.