Introduction

Kalij pheasants (Lophura leucomelana) are members of a large Asian genus, the chicken-like gallopheasants. Originally from Southeast Asia, the species was introduced to Hawaii Is. in 1962 where it quickly expanded into numerous island environments, including Hawaii National Park (Hawaii Audubon Society et al. 1997). In spite of their large abundance in the National Park and island-wide, little has been learned about Kalij ecology and their impact on native Hawaiian ecosystems. Research efforts began in 1984 by Lewin who documented the species establishment and spread on Hawaii Is (Scott et al. 1986).

The current study of Kalij breeding biology and population dynamics in HAVO's Kipuka Puaulu (Bird Park) began in 2000. Data gathered indicates that Kalij are territorial and long-term monogamous (Pratt 2001). Kalij also exhibit cooperative breeding with delayed dispersal of breeding-age sons. In contrast, hatch-year daughters disperse before the start of the next breeding season; yet it is unknown where or how far they disperse. In 2001, a live-trapping and banding component was developed to help identify family groups and population dynamics over time. In addition, increased efforts were made to census the local population and to observe Kalij social behavior and dietary habits. The 2003 field season was a continuation of the trapping, surveying, and observational efforts initiated in 2001. This paper reports findings from behavioral and demographic research performed on Kalij pheasants from January 6 through April 1, 2003 in Kipuka Puaulu.

Methods

2.1 Trapping

Details of trap development and methodology are discussed in Aruch et al. 2002. Kalij were captured using three trap types: (1) open-door traps, (2) turkey-traps, and (3) hand-nets. Open-door traps consisted of 100 x 50 mm wire-mesh fencing in 1.25 x 4 m sections, bent into a semicircular cage. A 40 cm opening in the semicircle was left as an entrance. From above, a trap resembled the letter "C", with an approximate diameter of 1.5 meters. Four attached, 1 m bamboo poles, were used to secure each trap to the ground. Tent stakes, sticks, and vines were used on some traps for additional support. A top-cover of plastic netting was then fastened to the wire fencing. The ground inside and around the trap was cleared of leaves and debris and a cracked corn and wheat feed mixture was then placed in and around the trap. Clearing the ground appeared to make bait easier to recognize and encouraged birds to forage within the trap. When attempting to trap larger groups of unbanded birds, two open-door traps were placed next to one another with their openings facing each other, separated approximately 40 cm apart. From above, this trap design looked like a figure-8 with a gap in the center. Turkey-traps (or box traps) were constructed of 1.25 x 2.5 m PVC rectangular frames with sides of woven plastic poultry netting. Traps were 5-sided, containing four vertical side-panels and a roof. One side panel was hinged to the horizontal PVC bar that formed part of the roof, creating a top-swinging door. The door was held open with a long, upright-placed stick that when pulled by an attached string, caused the door to swing shut. Lastly, we used hand-nets made of a rounded aluminum frame attached to a 1 m long handle.

Traps were introduced into locations where targeted family groups were frequently observed and where neighboring groups were rarely seen. This minimized the chance of clashes and injuries between different group members in and around traps. Once the bird(s) had entered a trap, it was quickly approached from the entrance side and captured by hand or with the aid of a hand-net. Birds were then transferred to bird-bags for weighing. Birds were fitted with two bands, one on each leg: a numbered aluminum band and plastic color band (Haggie Engraving Co.). Selection of bands and banding procedures are discussed in Aruch et al. 2002. Band number, band combination, weight, crest color, body molt, presence or absence of lesions, and tail, wing, tarsus, spur, and bill measurements were recorded for all previously uncaptured Kalij. For recaptured birds with missing bands, only band number and combination was recorded. Birds were also brachial bled and feather samples were collected.

2.2 Surveying

Surveys were conducted twice per month from January through March and once in early April. Surveys occurred between 7 AM and 1300 PM, lasting approximately 4 hours per survey. The survey route followed an established transect-grid containing flagged stations separated 25 m apart, covering a total area 0.09 km 2 (300m x 300m) (Figure 1). Surveys began at either the southeast or northwest corner of the grid. Surveyors walked slowly along east-west transect lines, stopping for two-minutes at even-numbered stations (approximately 50 m apart) listening and watching for pheasants. Upon reaching the last station of each transect line, surveyors would cut north or south, depending on the direction of the survey, until reaching the neighboring east-west transect line, 25 m away. The procedure was continued throughout the length of the grid. Location of pheasants relative to nearest flagged station, band combination, sex, group association, and behavior were recorded for each detection.

2.3 Behavioral Observations

Location and observed behavior of Kalij was recorded for all detections during the study. Time, sex, age, crest color, band combinations (if banded), and group number were also documented for all unique observations. To avoid repetition, the same individual was recorded twice only when it was certain to have left the area and returned or had not been seen for approximately 30 minutes. In cases where Kalij followed observers, only the original location and time of detections were recorded.

Results

3.1 Trapping

A total of 49 unbanded Kalij (26 male: 23 female and 29 adult: 13 juvenile: 7 unknown) from 17 of 23 different family groups were captured and banded from January 22 - March 31, 2002 (Table 1). Three additional birds, which were banded in 2001 but had lost one of their two bands, were refitted with a replacement band. Blood and feather samples were also collected from 48 of the total 52 birds banded. Two people were required to band and bleed each pheasant: one to hold the bird and one to apply bands, take and record measurements, and obtain feather and blood samples. Birds were held with one hand clutching the base of the wings and the other securing the legs.

Of the three trap methods, 35 of the 52 Kalij banded were caught using the open-door traps (67.3 percent). Turkey-traps were also effective, capturing 15 of the 52 individuals (28.9 percent). Only two Kalij were caught using hand-nets (3.8 percent). Capture attempts with hand-nets were made only for individuals that would not readily enter other traps introduced to them but would advance extremely close to the person handing out bait. A disadvantage of using hand-nets was that birds that excessively struggled in them tended to lose high numbers of feathers.

Four birds escaped from open-door traps, either by initially forcing the trap off the ground and escaping under it or by flying out just as researchers approached the trap. Three of those four birds were recaptured and banded. Also, two individuals escaped when attempting to remove them from bird-bags. One of these two birds was recaptured. To minimize future loss of trapped birds, traps should be firmly anchored to the ground. In some areas of the study site, the soil is rocky and dense with roots and can be difficult to secure. Use additional stakes and other surrounding objects to secure traps before attempting to use them. Also, removing birds from bags from inside the trap reduces the chances of them permanently escaping.

3.2 Baiting and Habituating Kalij

Capturing and banding Kalij requires both habituating them to human presence and pre-baiting traps. Time and effort taken to capture birds varied in the study, presumably due to differences in the level of comfort both groups and individuals had towards people and traps. Aruch et al. 2001 found Kalij to be extremely wary of foreign objects. In some cases, Kalij were wary of traps but readily approached researchers. In other cases, the opposite was true. Trap method also influenced the amount of time it took to capture birds. Although open-door and turkey-traps had similar average times for capturing birds (6.8 and 6 days respectively), open-door traps caught individuals over a broader range of days (1 to 18 days, N=35).

We spent an average of 4.5 actual days, not calendar days, pre-bating birds before introducing them to traps. Once exposed to traps, it took an average of 6.8 days to capture Kalij (min = 1 day, max = 18 days). Collectively, Kalij in the study required an average of 10.7 total days (min = 2 days, max = 22 days) of baiting and habituating to be caught. It is worth noting that juveniles and females tended to exhibit greater hesitation upon entering traps. It is unclear whether this observed behavior is due to differences in anxiety towards traps between sexes and age classes or whether females and juveniles are simply scared of bullying by adult males within the confined trap space. In some cases however, perhaps due to “trap naivety,” juveniles readily entered traps. Likewise, it appeared that groups that had never been attempted were less wary of traps than those groups who had been attempted before and had some banded members. Family groups that had been trapped previously were on average caught within 11.3 (N=9) days from initial baiting while previously attempted groups were caught within 10.8 days (N=8). In addition, it was also observed that among groups that had been trapped in 2001, remaining unbanded individuals were the most reluctant to investigate traps.

3.3 Survey and Observational Data

Survey data, including the number of individual Kalij and family groups discovered within Kipuka Puaulu Park has not been analyzed. However, between January 7 and April 1, 2911 unique visual detections from at least 23 family groups were recorded within the study site.

Discussion

4.1 Interesting Observations

A primary concern over the Kalij pheasant in Hawaii National Park is its ecological impact on native ecosystems. Due to their wide-spread abundance, effects on seed dispersal, seed and invertebrate predation, and browsing of rare native seedlings may be substantial. As such, observing Kalij dietary habits and foraging behavior is an important aspect of the project. Several interesting foraging observations were made during the study, including Kalij feeding on native plants, insects, skinks, and fungi. Kalij were also observed digging up nickel-sized, fern root-ball structures, and eating them whole. Several Kalij also had plant seeds attached to their feathers, especially Pisonia brunoniana (Table 2). However, observations of Kalij foraging on forest-food resources were rare. The infrequency of these behaviors was due in large part to the feeding of birds. Given the choice between searching for food in the natural environment and receiving generous hand-outs, Kalij almost exclusively chose the latter. Thus, our manipulation of food resource likely altered their foraging behavior. As such, our observations of Kalij foraging behavior and dietary habits are likely different than those occurring in more natural, undisturbed conditions.

One way to avoid this bias is to distinguish between “natural” observational data and observational data collected during the capture component. We propose that the capture component and associated observations occur exclusively during the fall through winter while the recording of “natural” foraging behavior--when no baiting or trapping occurs--take place in the spring through summer. This would provide valuable information on both (1) the effects of baiting and trapping on Kalij behavior and (2) on natural pheasant foraging and social behavior.

It was observed that several juvenile daughters had dispersed from their family groups in late March. Two juvenile females of the Bully group were last observed with their family on March 12, while all other group members had been seen consistently (5 to 8 times each) throughout the remainder of the study. Additionally, two juvenile females from the H3 group, last seen on March 28, failed to be observed with the rest of their family group on the final three consecutive sightings in late March. These are perhaps early indications of considerable demographic changes about to occur within this population.

4.2 Turkey-traps vs. Open-door Traps

Both turkey-traps and open-door traps proved effective at capturing Kalij. We found that the two trap types are best used in conjunction with one another, whereby open-door traps are used for small groups of 2 to 3 birds and turkey-traps are used for large or entire unbanded groups. Because of their small stature and scant amount of material, open-door traps take less time for the pheasants to get accustomed to. Additional advantages to open-door traps are that they are cheap and simple to make and that they are relatively small and easy to transport through the forest. Unfortunately, their small size can also pose a problem in that they can only accommodate 1-3 birds at a time. Thus, unbanded groups larger than three would have to be caught over multiple trapping attempts. Witnessing a family member caught in a trap appeared to discourage or delay other group members from approaching traps in the future. As mentioned earlier, certain individuals, mostly juveniles and females will not enter open-door traps if one or more adults already occupy the trap. This makes capturing even a single, intimidated individual difficult.

On the other hand, the turkey-trap's large size and volume make it ideal for large groups with socially wary individuals. Although their imposing size discouraged Kalij from initially entering them, most group members eventually did enter turkey-traps. Fifteen of 16 individuals (93.7 percent) attempted from three separate groups were caught in the turkey-traps. One drawback to using turkey-traps is that they are potentially injurious to birds. In an attempt to escape, Kalij would repeatedly poke their heads against the mesh side panels, occasionally inflicting abrasions on the head. Future construction of turkey-traps should consider using finer mesh material that Kalij cannot fit their heads through.

Interestingly, turkey and open-door traps took approximately the same amount of days, on average, to capture individual Kalij (6 days compared to 6.8 days, respectively). However, the number of attempted capture days varied widely among birds caught with the open-door traps (1 to 18 days, N=35). For example, the Southside and the H3 group both had all four of its unbanded members caught during the study. Although three Southside members were readily caught after introduction to the open-door trap (3, 5, and 5 days respectively), it took eleven days to capture the last individual in the entire group. Thus, the average time taken to capture birds in the group was 6 days. In comparison, no individuals from the H3 group were caught within the first 5 days, but all 4 individuals of the group were captured on day 6. In spite of the broader time frame of which it took to capture the Southside group, the H3 and Southside groups averaged the same amount of time to capture birds.

4.3 Multiple Traps Open at Once

One important key to our success in banding a high number of individuals was having several traps at different territories open and baited at the same time. This allowed several groups exposure to traps at once, habituating them simultaneously. When attempting to capture multiple individuals from a group, it is important to exercise patience and wait for all or several unbanded birds to occupy traps at once. Again, the act of approaching and handling Kalij in traps is stressful to Kalij and deters on-looking birds from readily entering traps in the future.

A final suggestion is to experiment with attracting pheasants to bait without human presence. Perhaps test the feasibility of trapping Kalij through leaving trails of food that lead to baited traps. Diminishing habituation to humans will likely reduce panhandling and other "unnatural" behaviors of the population. Understanding impacts of Kalij on Hawaiian ecosystems requires learning how natural Kalij populations interact with the environment. This objective fails to be realized if we only learn the habits of a relatively small population of birds that are consistently exposed to humans, and furthermore largely manipulated by that exposure.

Acknowledgements

Thank you to USGS-BRD for providing research funding and to Hawaii National Park for land use permission. We would like to especially thank Darcy Hu, Taryn Jarvis, and Kristin Wilson for assistance with field work and Rick Camp, Marcos Gorresson, and Sam Aruch for logistical support and expertise. Mahalo.

Literature Cited

Aruch, S. N., T. K. Pratt, and J. Vetter. 2002. Methods for trapping and banding Kalij Pheasants. Hawaii Conservation Conference.

Pratt, T. K. 2001. Social monogamy in the Kalij Pheasant: skewed sex ratios, parental care, and cooperative breeding. American Ornithologists' Union annual meeting.

Scott, J. M., S. Mountainspring, F. L. Ramsey, and C. B. Kepler. 1986. Forest bird communities of the Hawaiian Islands: their dynamics, ecology, and conservation. Studies in Avian Biology 9.