S42.2: Population recovery techniques in North America using captive-reared cranes

C. 7712 Midge St. NE, Albuquerque, New Mexico, 87109, USA, fax 505 821 3823; e-mail jclewis@unm.edu

The 15 species of cranes are among the world's most threatened groups of birds; as many as 11 may be globally threatened (Meine & Archibald 1996a, b). Captive populations are essential to effective conservation programs for threatened species. Captive populations support field conservation efforts through research, education, financial support, and provide the individuals for release and reintroduction (Meine and Archibald 1996b). This paper discusses progress in the use of captive-reared cranes for recovery programs within North America.

Several terms need explanation. Parent-rearing refers to chicks reared in the pen of their parents or foster parents until they are about fledging age. Hand-rearing, as used in this paper, refers to cranes cared for by caretakers who are not costumed, talking is not excluded from the rearing area, and taxidermy mounts, taped brood calls, and live imprinting or social models may or may not be used. Partial isolation-rearing refers to limiting human contact with cranes and utilising some imprinting cranes or models. Costume-rearing of cranes refers to the technique whereby cranes are reared in individual pens by caretakers in crane costumes, talking is not permitted in the rearing area, uncostumed humans are visible only during stressful activities, taxidermy mounts of a crane head are used to teach the chicks to feed, a taxidermy mount of adult crane in brood posture with a heat lamp overhead provides heat for the chick, brood calls are played by tape recorder and imitated by caretakers, and live cranes are penned next to young chicks as imprint and socialisation models. Group-rearing refers to raising several chicks of similar age (plus or minus a few days) together. In the group-rearing practised to date, the caretakers were not costumed. Continuous access to food and lots of exercise pursuing natural foods outside the pen helps minimise sibling aggression that might cause death or injury. Utilising the following instinct, the chicks were trained to follow the caretaker to field locations where they experienced environments similar to those at the release site. It is assumed that cranes reared in groups will be sexually imprinted on their own species, but this has not been confirmed.

A soft release is one in which the captive-reared cranes are penned at a release site for several days or weeks, fed, and allowed to acclimate. A hard or abrupt release is when the captive-reared cranes are released immediately upon arrival at the release site. Optimum chick rearing and behaviour management techniques for cranes that will be released in the wild are distinctly different from the techniques used for individuals that will remain in captivity for show or as breeders (see chapters 5, 6, and 11d in Ellis et al. 1997). It is not possible to discuss all these aspects in a brief paper. Anyone planning to rear cranes for wild release should contact aviculturists and field biologists with experience in such techniques and gain from their insight.

Hyde (1968) hatched and raised five Greater Sandhill Cranes Grus canadensis tabida under bantam hens in Oregon, in partial isolation from humans, where the captive cranes could view wild cranes in wetlands. They were released at fledging and joined wild cranes. At least one migrated south and returned in spring.

Florida Sandhill Cranes G. c. pratensis were released in Florida from 1971 to 1976 in a test of the suitability of captive-reared cranes to survive in the wild (Nesbitt 1979). Fourteen of these, hand-reared at Patuxent Wildlife Research Center (Patuxent), Maryland, were abruptly-released at age 5 months in September, but survived less than 4 months. Hand-reared cranes, when released abruptly, survive poorly. The only parent-reared crane in the sample seemed best suited for wild release because it survived three years.

From 1975 to 1988, the U.S. Fish and Wildlife Service and Canadian Wildlife Service collaborated on a cross-fostering experiment to establish a second migratory population of Whooping Cranes Grus americana. Two hundred and eighty nine Whooping Crane eggs, taken from the wild population and captive production, were substituted for Sandhill Crane eggs in wild nests in southeastern Idaho. The Sandhill Crane foster parents raised the Whooping Crane chicks and led them on migration from Idaho to central New Mexico. The experiment was discontinued because of high mortality and because none of the Whooping Cranes paired (U.S. Fish and Wildlife Service 1994). The cross-fostered Whooping Cranes appeared to be socially and sexually imprinted on the foster parent species ( Nakayama 1970, Mahan & Simmers 1992).

In an effort to augment a declining population of Greater Sandhill Cranes in British Columbia, 17 cranes partially isolation-reared in 1980, were soft-released in April 1981 on the summering area. Thirteen apparently migrated in October and at least 7 (41 %) returned to the summering area in spring of 1982 (Leach 1987). The soft release and partial isolation-rearing seemed to benefit survival.

Eleven Greater Sandhill Cranes parent-reared at Patuxent were shipped in 1980 to southeastern Idaho for release. The 5 one-year-olds, 2 two-year-olds, and 4 three-year-olds, were released abruptly (Drewien et al. 1982). Research goals included assessing survival when captive cranes are released directly to the wild without a prerelease period of acclimation at the release site. Seven survived to migrate south in October and the minimum survival to one year was one bird (8%). The authors concluded that younger cranes would be better candidates for release programs and an acclimation period at the release site might increase survival.

In 1982, two chicks were hatched and puppet-reared from two Greater Sandhill Crane eggs found in an abandoned nest in Wisconsin (Archibald & Archibald 1992). They were reared in partial isolation from humans, and at 4 to 5 weeks of age were confined in a pen in a marsh frequented by the parent pair. They were eventually released in the vicinity but a strong bond never developed with the adult pair. The chicks disappeared 30 August but one was seen on the Florida wintering grounds that winter, and subsequent winters, and in 1990 had paired and produced a chick 5 km from the release site. The combination of partial isolation-rearing, on-site rearing, and soft release seemed beneficial.

Twenty-one Sandhill Cranes parent-reared at Patuxent were shipped to southeastern Idaho in June 1984 (Bizeau et al. 1987). The experiment was designed to test the fitness of parent-reared cranes to survive in the wild and migrate. These one-year-old (19) and two-year-old (2) cranes were held 4 to 6 days in an acclimation pen before release in summering areas occupied by wild Sandhill Cranes. Cohort sizes were 8, 8, and 5. Penmates formed their own social groups and at release did not integrate well with wild cranes. Nine survived (43 %) to migrate with wild Sandhill Cranes and four may have survived their first winter (2 were seen at New Mexico wintering sites). The primary cause of mortality at the release site was predation by Golden Eagles (Aquila chrysaetos). The authors suggested that in future releases of captive-reared cranes into a migratory situation the birds be released singly into soft release pens spaced at least 2 km apart. A wild crane might be confined with the captive-reared bird to facilitate a social bond that would benefit the captive-reared bird when both were released to the wild. Four to six days did not seem sufficient for acclimation to the release site.

Twenty-seven parent-reared juvenile Greater Sandhill Cranes were transferred in 1986 and 1987 from Patuxent to Florida (Nesbitt, S.A. unpubl. An evaluation of the Florida sandhill crane population of peninsular Florida and its potential to support a population of non-migratory Whooping Cranes. Report to Florida Game and Fresh Water Fish Commission. Steve Nesbitt, Florida Game and Freshwater Fish Commission, 4005 South Main, Gainesville, Florida 32601). The study objective was to determine if subadult Greater Sandhill Cranes produced in captivity, and soft-released, would migrate north with wild Greater Sandhill Cranes that winter in Florida. They were penned at the release site for 4 to 6 weeks, then released. They did not migrate so it appeared that captive-reared Whooping Cranes could also be released in Florida and would not migrate. Survival to one year after release was 15 individuals (56 %) and it appeared the soft-release and parent-rearing benefited survival.

Ten Greater Sandhill Cranes were hand-reared at International Crane Foundation (ICF), Baraboo, Wisconsin in 1985 using a moveable crane puppet, crane costume, and vocalising brood models as substitute parents (Horwich et al. 1992). At 8 weeks of age they were moved to Necedah National Wildlife Refuge, Wisconsin, where they were penned by a wetland. A crane-costumed caretaker led them daily around a marsh to introduce them to the wild environment and foods. At 3.5 months of age the five survivors were released to the wild. Despite costume-rearing they initially had little fear of humans. Human aversion conditioning included chasing the cranes, loud yelling, and conducting medical examinations without a costume. Within a month post-release they were associating regularly with wild cranes and four (80%) migrated with wild cranes and returned to Wisconsin the following spring. This combination of costume-rearing, soft release, partial on-site rearing, and release near wild conspecifics was very successful in promoting adaptation to the wild and survival.

In 1988, seven Greater Sandhill Cranes were costume-reared at Welder Wildlife Refuge (WWR), in typical crane wintering habitat in southern Texas (Nagendran 1992). The research objective was to see if cranes could be induced to learn migration routes from wintering conspecifics. Because of severe drought conditions the three surviving cranes were moved to Laguna Atascosa National Wildlife Refuge, coastal Texas, in January 1989. In March they initiated northward migration with wild cranes but stopped near Rosebud, Texas, after going 600 km. They were recaptured and moved by truck 1,100 km north to the Platte River, Nebraska, and released near wild cranes. The research cranes then migrated north at least another 160 km on 21 April 1989 before their transmitter signals were no longer audible. On 22 May 1989 they were found 72 km north of Rosebud, Texas, having returned south. In June a Bobcat Lynx rufus killed one of the cranes 32 km from WWR and the two survivors returned to WWR in August. These cranes showed a strong affinity for their natal area, and ability to navigate long distances across which they were transported by truck. The results indicate that releases should be made on the breeding grounds, not at wintering areas, if the objective is to establish a migratory population (Nagendran 1992).

From 1988-1990, 38 Greater Sandhill Cranes were costume-reared and released in Upper Michigan in research to develop a reintroduction technique for Whooping Cranes (Urbanek & Bookhout 1992). Similar to the research of Horwich et al. (1992) they were costume-reared in the field at the release site. Wild Sandhill Cranes were confined as imprinting models adjacent to the chick pens. The chicks were acclimated to the wild in a release pen and taken on marsh walks. They were encouraged to migrate by releasing them in small groups near wild cranes at the time for southward migration. The resulting minimum survival rate of 84% one year after release exceeded that of any other study, and the minimum return rate to Upper Michigan was 74%. Again, the combination of rearing cranes at the release site, costume-rearing, exposing them to wild foods and the environment, and release near wild conspecifics, proved very effective.

The Mississippi Sandhill Crane (G. c. pulla), a CITES Appendix I subspecies, considered critically endangered by IUCN (Meine & Archibald 1996), is found on and near the 19,300 acre Mississippi Sandhill Crane National Wildlife Refuge in Jackson County, Mississippi. From the 1940s to 1980s the population was estimated at 25 to 60. Some of the research, described previously in this paper, provided clues to rearing and release techniques that would be most successful in rebuilding the Mississippi Sandhill Crane population so it could be self-sustaining. It is the largest captive-reared crane reintroduction to date with 274 released. Releases began in 1981 using parent-reared juveniles from Patuxent. Two to 13 juveniles were soft released in pens annually from 1981 to 1989 but the wild population probably never exceeded 65 (Ellis et al. 1992). Release pen size varied from .5 to 2 ha. Integration to the wild and survival was poor in the first two years. Survival and integration into the wild flock were markedly improved by increasing release cohort formation time and the pre-release acclimation period to one month each. The one-year survival rate for these early releases averaged 62% (Ellis et al. op cit.). By 1988 over half the wild flock consisted of cranes reared at Patuxent and some had successfully reproduced (Valentine and Hereford 1997).

Captive production was accelerated 1989 to 1992 and intensive research compared the survival rate of parent-reared versus costume-reared cranes. There were no statistically significant differences in survival between costume-reared and parent-reared birds, although costume-reared showed slightly better survival. Survival was better in mixed cohorts than in cohorts containing only parent-reared or only costume-reared birds. The estimated one-year survival from the day of release, for all birds (n=132) used in the analysis, was 80%. Costume-reared cranes survived at least as well as parent-reared and survival was improved by mixing costume-reared and parent-reared in the release cohorts (Ellis et al. at press). One possible explanation for the improved survival of parent-reared when placed in release cohorts with costume-reared is the greater wariness of costume-reared cranes. Parent-reared cranes are accustomed to the approach of vehicles and uncostumed humans, consequently they are less wary of humans and predators. But after release, parent-reared cranes seemed to learn from and mimic the more-wary response of the costume-reared flock members.

Fewer birds have been released annually since 1992 because the release goal has been to maintain the population at 90 to 120 cranes until the population becomes self-sustaining and carrying capacity increases with continuing habitat restoration. The population peaked at 142 in December 1992 and since then has varied between 118 and 96 with captive-reared making up over 80% (Ellis et al. at press). The nesting population did not increase until release cranes began breeding (Valentine & Hereford 1997). Between 1965 and 1986 the number of pairs nesting annually ranged from 3 to 8.

After 1986 there was a gradual but irregular increase in the number of pairs nesting annually. Release cranes nested within a few kilometres of their release site. To promote colonisation of new areas, the release pens must be located nearby. Twelve new nesting territories were established after 1984, and 11 of these were occupied by pairs containing one or more release cranes (Valentine & Hereford op cit.). Twenty pairs hatched 18 chicks in 1997. The Mississippi Sandhill Crane population remains endangered, but an intensive habitat management program, and soft releases of combined cohorts of parent-reared and costume-reared stock, show promise of ensuring the long-term survival of this subspecies.

With the failure of the cross-fostering technique (described previously), crane conservationists were left with no tested technique for reestablishing a migratory population. Consequently, the U.S. and Canadian Whooping Crane Recovery Teams decided to reestablish a nonmigratory wild population. Such a sedentary population still existed in southern Louisiana in the 1940s but was decimated during a hurricane. The last surviving Whooping Crane in this population was captured in 1949. Like the nonmigratory Sandhill Crane population of the southeastern United States, nonmigratory Whooping Crane populations may have extended from Louisiana to Florida. Early naturalists reported Whooping Cranes living year round and breeding in Florida. However, the only specimen record from Florida was collected in winter. Central Florida was identified as the most suitable habitat and the area where the human population was most supportive of the reintroduction.

The goal is to establish a population of 25 or more breeding pairs by year 2020. Techniques successfully used to supplement the Mississippi Sandhill Crane population were duplicated in the Whooping Crane releases. The soft releases began in February 1993 using 1992 parent-reared and costume-reared cranes from Patuxent and from ICF. Release groups contained from 5 to 14 individuals (Nesbitt et al. 1997). They were quarantined for 60 days before shipment to Florida and for 2 to 4 weeks in Florida in the release pen. All releases were within the Kissimmee Prairie of central Florida but the first releases were centered at Three Lakes Wildlife Management Area. The release pen was 30 m x 120 m and 3 m high. First year survival rates were 0.36 for 1993, 0.32 for 1994, and 0.47 for 1995. Despite the existence of some Bobcat control measures, such predation was the only known source of mortality in the first three years. Cranes from the first two years of releases tended to roost on dry ground and to use heavily vegetated areas, making them vulnerable to Bobcat.

Changes were then made at the rearing facilities, such as mowing vegetation inside pens, to better prepare the birds for release. Standing water was provided in most pens to encourage water roosting, and when first socialised with their cohorts the cranes were housed in pens with an open, unobstructed view to habituate them to open, low vegetation. At ICF, cranes were taken on 'marsh walks' with costumed caretakers to expose the cranes to food items and feeding sites similar to the release area, and encouraged to take short flights to promote flight ability prior to release. At the release site, larger expanses were mowed or control-burned to remove suitable Bobcat habitat.

In 1994, project personnel also started using smaller temporary pens, called satellite pens, for the last period of acclimation. The cranes were held in large pens for the initial acclimation period, then moved into satellite pens which were placed in optimum crane habitats. Since 1994, most satellite pens have been placed on large private cattle ranches where grazing and pasture management reduces the suitable Bobcat habitat, and prairie wetlands provide good roosting and feeding sites. As a consequence of the changes in captive management and release techniques, first year survival has increased. Of the 151 Whooping Cranes released to date (May 1998), 60 (42 %) survive. Five young pairs held territories and built nests in 1998 but no eggs were laid. Annual releases will continue for the foreseeable future and efforts will continue to diminish the losses to Bobcat predation, remains the primary mortality factor.

After the cross-fostering project was discontinued the U.S. Fish and Wildlife Service and Canadian Wildlife Service evaluated potential uses for the surviving cross-fostered Whooping Cranes. One proposal was to test whether these adults could be used as 'guides' to adopt and introduce young captive-reared Whooping Cranes to the wild. The objective was to have chicks imprint on a cross-fostered adult(s) and learn survival skills and migration from the adult (Drewien et al. 1997). Two wild adult cross-fostered Whooping Cranes, a male and female, were captured on the wintering grounds in January 1993, transported to Idaho, and placed in a pen. In June, four costume-reared Whooping Crane chicks, 14 to 21 days old, were placed in enclosures next to the adults' pen. Chicks were placed in the adults' pen when they were 24-47 days old and remained there until they fledged in late August. Then three chicks and their foster parents were moved to a pen at Grays Lake National Wildlife Refuge, held overnight, and the pen door opened in the morning. Although adults and chicks appeared to develop bonds in captivity, adults left the chicks upon release, and about two months later the adults migrated without the chicks. The chicks did not socialise with any Sandhill Crane or Whooping Cranes after their release although many wild cranes were present in the vicinity. One technique proposed for introducing captive-reared juvenile Whooping Cranes has been to release them into nesting or staging areas of Sandhill Cranes with the expectation that they will learn survival skills and migration patterns from the Sandhill Cranes. The observations of Drewien et al. (1997) in Idaho and Nesbitt et al. (1997) in Florida showed that juvenile captive-reared Whooping Cranes did not readily join flocks of wild Sandhill Cranes. The Sandhill Cranes often harassed or chased juvenile Whooping Cranes. Based on these observations, releasing juvenile Whooping Cranes on summer or fall staging areas, where they are expected to join flocks of wild Sandhill Cranes, will likely meet with little success (Drewien et al. 1997.)

Lishman et al. (1997) evaluated the potential for use of ultra-light aircraft to lead birds in migration. Initial research utilised captive-reared Canada Geese (Branta canadensis), leading them from a natal area in Ontario, along a preselected migration route, to a wintering site. Bill Lishman was the first to train geese to follow ultralight aircraft. Eighty-six geese were led 640 to 1,312 km on southbound migrations in three years. Of 51 geese led to Virginia or South Carolina in 1993 and 1994, 46 (90%) returned unassisted to Ontario. In 1995, 30 geese were transported by truck from Ontario into New York and then led by aircraft to South Carolina; only 50% returned to Ontario the following spring. Another 16 geese were transported in a truck from Ontario to Virginia, stopping overnight at several locations where the geese were allowed to fly. None of the 16 returned to Ontario the following spring. None of the (1993 to 1995) geese were known to return to the Virginia or South Carolina wintering sites in subsequent winters. Geese transported by truck had difficulty navigating back to their natal area, but those led by aircraft were able to return to their natal area in spring without human assistance.

In 1995, Lishman et al. (1997) also costume-reared Sandhill Cranes in Ontario. Three chicks fledged and were trained to follow the costumed pilot and ultralight aircraft on local flights for up to 64 km from the rearing site. No attempt was made to lead the cranes on migration. In 1997, Sandhill Cranes were costume-reared at Patuxent and shipped to Ontario for flight training. Thirteen were led in migration 24 October to 13 November to Virginia. Seven Sandhill Cranes flew with the plane and six were transported by truck much of the distance. During winter the cranes were penned at night but were free to roam a sanctuary during the day. In late March, 1998, six of the seven cranes which migrated south with the aircraft, returned to Ontario on their own initiative, landing about 99 km west of their rearing site. This is the first documentation of captive-reared cranes returning to their natal area without the potential influence of wild cranes following the same migration route. After wandering about Ontario for two weeks, interacting with humans, they were led by aircraft back to the rearing site. Unfortunately the cranes had an affinity for people that may have resulted because the rearing protocol of strict isolation from humans was not followed (Stehn, T. unpublished report. Whooping Crane recovery activities, December, 1997 - April, 1998. 16 pp.).

Kent Clegg was the first individual to train captive-reared Sandhill Cranes to follow an aircraft (Clegg et al. 1997). In 1994 he group-reared six Sandhill Cranes in Idaho and trained them to follow an ultralight aircraft in flights in a 50 km radius of his ranch. In 1995 he group-reared 12 Sandhill Cranes in an experiment to see if they could be led along a 1,200 km migration route and if, after release on a wintering area, they would integrate with wild cranes and migrate north in spring to their natal area without assistance from humans. Training the cranes to follow capitalised on the following behaviour of chicks. The cranes responded to Clegg's imitation of a brood call and followed him as he walked, drove an all terrain vehicle, or piloted an ultralight. They were led daily to fields and ditch banks on the ranch where they fed in grain and alfalfa fields on seeds, insects, and Earthworms (Lumbricus sp.). They were penned at night but allowed to roam the ranch without accompanying humans part of each day, providing exposure to habitat, foods, and predators they might encounter after release.

Eleven greater Sandhill Cranes were led in October 1995 on the 1,204 km migration from Idaho to central New Mexico which required 11 days (Clegg et al. 1997). Aircraft and cranes flew early and late in the day when the air was calm and the birds were placed in portable pens at night. Daily distances flown varied from 43 to 217 km depending on wind conditions, temperature, terrain, fuel supplies, Golden Eagle attacks, stamina of the cranes, and diminishing daylight. Mean flight altitude was about 300 m. The highest mountain pass traversed was 2,590 m above sea level. Average flight speed was 52 km per hour with occasional bursts of up to 70 km per hour. The typical morning flight lasted 1 to 2 hours. Hazards encountered included mountainous terrain, turbulent air, winter storms, and two cranes were killed in flight by Golden Eagles.

Two days after their abrupt release in New Mexico, the survivors were associating with and imitating the behaviour of wild cranes (Clegg et al. 1997). Two were killed by Coyotes, two by hunters, and one disappeared. The four survivors migrated north with wild Sandhill Cranes and two spent the summer within 53 km of their natal area. Three returned to New Mexico the following winter, thus, a minimum of 3 of 9 (33%) survived one year after release.

In 1996 the experiment was repeated in an effort to reduce mortality. Eight Sandhill Cranes were led in migration. None were known lost to Golden Eagles, hunters or Coyotes. A key to the survival of these cranes, despite the abrupt release, is believed to be their daily experience afield while they were being reared and intensive efforts of biologists to modify any inappropriate behaviour following release. For example, the research cranes were led to an appropriate water roosting site where they could associate with wild conspecifics the first evening after release. Using foods and crane decoys the research cranes were familiar with, they were attracted to safe feeding sites where they could find abundant natural foods. The research cranes were encouraged to associate with and imitate the behaviour patterns of wild cranes. Within two days after their release at the New Mexico wintering site, two research cranes apparently joined wild cranes migrating to Mexico and were not seen again. The six remaining research cranes survived the winter in New Mexico. Five of the 1996 and three 1995 research cranes were identified as they migrated north through Colorado in spring of 1997. Two 1995 and four 1996 research cranes were found in the summer of 1997. All were utilising typical crane habitat, associated with wild cranes, and summered in locations normal for wild cranes hatched in southeastern Idaho.

In 1997, Kent Clegg became the first person to train Whooping Cranes to follow an ultralight aircraft and to lead them on migration (Clegg and Lewis. Unpubl. 1997-1998 progress report on use of ultralight aircraft for introducing migratory crane populations. 15 pp.) Research objectives were to determine if captive-reared Whooping Cranes could be taught to follow an ultralight aircraft along a migration route and, after release on a wintering area, would integrate with wild cranes and migrate north in spring to the vicinity of their natal area without human assistance. The Whooping Cranes were hatched and reared for two weeks at Patuxent to utilise their avicultural and veterinary expertise. They were reared in two groups without the presence of taxidermy mounts or crane costumes. Then they were flown to Idaho and raised at the Clegg Ranch. Rearing techniques for the Whooping Cranes were similar to those for Sandhill Cranes in previous summers. Studies also continued with greater Sandhill Cranes which were reared separate from the Whooping Cranes to avoid improper sexual imprinting. After the Sandhill Cranes fledged they were allowed to join with the Whooping Cranes during the daytime. The association of the Whooping Cranes with the Sandhill Cranes was designed to increase the likelihood of association between the Whooping Cranes and wild Sandhill Cranes on the wintering grounds and, thereby, to accelerate the Whooping Cranes adjustment to the wild.

Four Whooping Cranes and eight Sandhill Cranes were led in migration from Idaho in mid October (Clegg and Lewis op cit.). The 1,137 km migration took 9 days. Hazards encountered en route included mountainous terrain and attacks by Golden Eagles. One Whooping Crane was injured by a Golden Eagle. One Sandhill Crane stuck its feet through the propeller guard in flight and was fatally injured. Average flight speed was 52.5 km per hour at an average altitude of 305 m. Daily flight times and distances travelled were less than that of migrating wild cranes (Melvin and Temple 1982) which flew from 30 to 462 miles in daily flights that lasted 1 to over 10 hours. Within 2 to 4 days after release at the wintering site, the research cranes were associating with and imitating the behaviour of wild cranes. As a group the juvenile Whooping Cranes were dominant to wild adult Sandhill Cranes and readily associated with them. On the wintering grounds one Sandhill Crane was killed by hunters, one Whooping Crane was killed by a Bobcat, and another by a Coyote. The 8 surviving cranes migrated north in March and April of 1998.

The techniques of group rearing in the field, migration using ultralight aircraft, and abrupt introduction to the wild among wild cranes seemed suitable and show promise for improved use in reintroducing a migratory population where other wild crane species exist to help the transition to wild behaviour. Although some disappointing losses of project cranes were experienced, such losses may be diminished in the future as techniques are refined.

Research on another possible technique for establishing a migratory population of Sandhill Cranes, teaching the cranes to fly following a truck, was initiated in 1995 (Ellis et al. 1997). The cranes were reared at Patuxent until the youngest was 74 days old and then transported to northern Arizona within former Sandhill Crane breeding range. Ten cranes were partially isolation-reared but the caretakers did not wear crane costumes. In northern Arizona they were trained for 5 weeks to fly behind a truck and then migrated 600 km (495 of which were flown) to spend winter on the Arizona-Mexico border. The problems experienced during migration were collisions with power lines (3, 1 fatal), Golden Eagle attacks, and overheating as the cranes crossed a desert. They were penned on a refuge during winter and released at a wetland in spring but failed to initiate northward migration. They were placed in a truck, transported to northern Arizona, and released. During the summer they became too tame and were removed from the wild. In 1996, 12 Sandhill Cranes were led in migration flying behind a truck from northern to southwestern Arizona and released abruptly with wild cranes. Eight returned unassisted to northern Arizona in spring and seven returned to southwestern Arizona in fall (a minimum of 58% survived one year after release). Release of the trucked cranes in the vicinity of wild cranes at the winter site benefited survival and apparently stimulated appropriate migration in spring.

Captive-reared Sandhill Cranes seem to have an innate behaviour to roost at night in shallow water after release, even if they did not have access to such water during captivity. In contrast, Whooping Cranes reared in pens without access to shallow water tended to choose inappropriate upland roosting sites when they were released to the wild (Nesbitt et al. 1997). Rearing cranes on site (Horwich et al. 1992, Urbanek and Bookhout 1992) or providing diversified opportunities for them to experience semi-wild environments before release, such as marsh walks or spending the day afield (Archibald and Archibald 1992, Wellington et al. 1996, Clegg et al. 1997, Nesbitt et al. 1997), seemed to improve survival and adaptability to the wild. Cranes less than one year old seem to adapt more readily to the wild and survive longer than older released cranes that were more accustomed to captive living (Drewien et al. 1982). Cranes reared in pens with rank, tall vegetation, trees for shade, and tennis netting screening the sides, after release were more likely to choose rank, dense, or tree-shaded habitats for loafing and feeding, and consequently were vulnerable to predators (Nesbitt et al. 1997). Pens should have open vistas and low vegetation if the cranes are to be released in the wild. Intuitively, it seems advantageous to provide chicks some deliberate negative exposure to uncostumed humans and Dogs so the cranes are not overly tame or naive when released. Such training may include mock attacks on the chicks (Wellington et al. 1996). Cranes have been reared with costumed dummies or crane decoys and the cranes continue to be attracted to them when they are released to the wild. The decoy or dummy is then moved to an appropriate roosting or feeding site as a means of manipulating the cranes behaviour. The cranes are colour-banded and equipped with radio transmitters so they can be monitored following release.

If the objective is to establish a migratory population, the releases should occur at the breeding grounds, not at wintering sites (Nagendren 1992). Costume-reared and parent-reared cranes are statistically almost equally fit for adapting to and surviving in the wild (Ellis et al. at press). The survival of parent-reared cranes was increased by placing them in a release cohort with costume-reared cranes (Ellis et al. at press). But release cohorts comprised of two rearing methods seemed less cohesive than cohorts comprised of a single rearing method (Nesbitt et al. 1997). Release cohorts of 6 to 8 cranes seemed more cohesive after release than a cohort of 14 (Nesbitt et al. 1997). Portable 'satellite' soft release pens at the release site can be moved to take advantage of the optimum water, food, and activities of wild cranes (if present) and are less expensive than more permanent acclimation pens (Nesbitt et al. 1997). Agricultural foods may be provided adjacent to the acclimation pen before release to attract wild cranes to the vicinity and, thereby, to promote quicker association of captive-reared and wild cranes at the time of release. In some instances, commercial or agricultural foods are provided post release, until the release cranes appear to be fully adapted to natural foods. A soft release (Horwich et al. 1992, Urbanek & Bookhout 1992, Ellis et al. 1997, Nesbitt et al. 1997), that allows the cranes to acclimate to the local environment and foods, in most instances is preferable to a hard release (Drewien et al. 1982, Nesbitt 1979). Cranes tend to disperse greater distances and have lower survival rates when released abruptly (Drewien et al. 1982, Ellis et al. 1992). However, an abrupt release is feasible when the cranes have been reared with extensive field experience and their behaviour is manipulated at release so they quickly integrate with and mimic behaviour of wild cranes (Clegg et al. 1997, Clegg and Lewis op cit.). Whenever captive-reared cranes can be released in the vicinity of wild conspecifics or other wild crane species there is the possibility of improving survival of release cranes as they learn from the wild associates (Horwich et al. 1992, Urbanek & Bookhout 1992, Clegg et al. 1997, Stehn, unpubl. report).

The science and art of wild release of captive-reared cranes has made substantial progress but is still young, and there is much more to be learned. In the evolution of wild release of captive-reared cranes we see the following trends that benefit survival. Efforts to make the captive-rearing pen environment like the habitats where the cranes will be released, exposure of the chicks to field and marsh experiences outside their pen to prepare them for similar environments and foods at the release site, some efforts to rear chicks at the release site, increasing isolation from human-related audio and visual stimuli, availability of social and sexual models that will encourage appropriate imprinting, some negative conditioning to encourage fear of humans and canines, a period of acclimation at the release site before release, manipulation of behaviour after release to promote survival, management of habitats and predator populations to increase survival, and portable acclimation pens.

My thanks to Dave Ellis, and Scott Hereford who provided unpublished data. Claire Mirande, Scott Hereford, and Kent Clegg reviewed the manuscript. Special gratitude goes to George Archibald for inviting me to present this paper and for providing editorial assistance. My deepest appreciation to all those who have contributed hundreds of thousands of hours studying cranes in North American, each contributing to our present state of knowledge.

Archibald, K. & Archibald, G. 1992. Releasing puppet-reared Sandhill Cranes into the wild: A progress report. In: Proceedings 1988 North American Crane Workshop: 251-254. Florida Game Fresh Water Fish Comm. Nongame Wildl. Program Tech. Rep. 12.

Bizeau, E.G., Schumacher, T.V., Drewien, R.C. & Brown, W.M. 1987. An experimental release of captive-reared Greater Sandhill Cranes. In: Proceedings 1985 Crane Workshop: 78-88. Platte River Whooping Crane Habitat Maintenance Trust and U.S. Fish and Wildlife Service, Grand Island, Nebraska.

Clegg, K.R., Lewis, J.C. & Ellis, D.H. 1997. Use of ultralight aircraft for introducing migratory crane populations. Proceedings North American Crane Workshop 7: 105-113.

Drewien, R.C., Derrickson, S.R. & Bizeau, E.G. 1982. Experimental release of captive parent-reared Greater Sandhill Cranes at Grays Lake Refuge, Idaho. In: Proceedings 1981 Crane Workshop: 99-111. National Audubon Society, Tavernier, Florida.

Drewien, R.C., Munroe, W.L., Clegg, K.R. & Brown, W.M. 1997. Use of cross-fostered Whooping Cranes as guide birds. Proceedings North American Crane Workshop 7: 86-95

Ellis, D.H., Olsen, G.H., Gee, G.F., Nicolich, J.M., O'Malley, K.E., Nagendran, M., Hereford, S.G., Range, P, Harper, W.T., Ingram, R.P. & Smith, D.G. 1992. Techniques for rearing and releasing nonmigratory cranes: Lessons from the Mississippi Sandhill Crane program. Proceedings North American Crane Workshop 6: 135-141.

Ellis, D.H., Gee, G.F. and Mirande, C.M. 1996. Cranes: Their biology, husbandry, and conservation. Department of the Interior, National Biological Service, Washington, D.C., and the International Crane Foundation, Baraboo, Wisconsin. 307 pp.

Ellis, D.H., Clauss, B., Watanabe, T., Mykut, R.C., Kinloch, M. & Ellis, C.H. 1997. Results of an experiment to lead cranes on migration behind motorized ground vehicles. Proceedings North American Crane Workshop 7: 114-122.

Ellis, D.H., Hereford, S.G., Olsen, G.H., Chisolm, T.D., Nicolich, J.M., Sullivan, K.A., Thomas, N.J., Nagendran, M. & Hatfield, J.S. In press. Post-release survival of Mississippi Sandhill Cranes: A comparison of rearing techniques. Journal of Wildlife Management.

Horwich, R.H., Wood, J. & Anderson, R. 1992. Release of Sandhill Crane chicks hand-reared with artificial stimuli. In: Proceedings 1988 North American Crane Workshop: 255-262. Florida Game Fresh Water Fish Comm. Nongame Wildl. Program Tech. Rep. 12. .

Hyde, D.O. 1968. Sandy. Dial Press, New York, New York. 214 pp.

Leach, B. 1987. the Sandhill Crane recovery project in the Lower Fraser Valley, British Columbia. In: Proc. 1983 International Crane Workshop: 415-423. International Crane Foundation, Baraboo, Wisconsin.

Lishman, W.A., Teets, T.L., Duff, J.W., Sladen, W.J.L., Shire, G.G., Goolsby, K.M., Bezner Kerr, W.A. & Urbanek, R.P. 1997. A reintroduction technique for migratory birds: Leading Canada Geese and isolation-reared Sandhill Cranes with ultralight aircraft. Proceedings North American Crane Workshop 7: 96-104.

Mahan, T.A. & Simmers, B.S. 1992. Social preference of four cross-foster reared Sandhill Cranes. Proceedings North American Crane Workshop 6: 43-49.

Meine, C. & Archibald, G.W. l996a. Ecology, status, and conservation. In: Ellis, D.H., Gee, G.F. & Mirande, C.M. (eds) Cranes: Their biology, husbandry, and conservation. National Biological Service, Washington, D.C. & International Crane Foundation, Baraboo, Wisconsin.

Meine, C. & Archibald, G.W. (eds) 1996b. The cranes:   Status survey and conservation action plan. IUCN, Gland, Switzerland, and Cambridge, U.K. 294 pp.

Melvin, S.M. & Temple, S.A. 1982. Migration ecology of Sandhill Cranes: A review.  In: Proceedings 1981 Crane Workshop. National Audubon Society, Tavernier, Florida:  73-87

Nagendran, M. 1992. Winter release of isolation-reared Greater Sandhill Cranes in south Texas. Proceedings North American Crane Workshop 6:131-134.

Nagendran, M., Urbanek, R.P. & Ellis, D.H. 1996. Reintroduction techniques. In: Ellis, D.H., Gee, G.F. & Mirande, C.M. (eds) Cranes: their biology, husbandry, and conservation. National Biological Service, Washington, D.C. & International Crane Foundation, Baraboo, Wisconsin.

Nakayama, T. 1970. Japonensis raised by foster parents. Animals and Zoos 22(2): 42-43.

Nesbitt, S.A. 1979. Notes on suitability of captive-reared Sandhill Cranes for release into the wild. In: Proceedings 1978 Crane Workshop: 85-88. Colorado State University Printing Service, Fort Collins.

Nesbitt, S.A., Folk, M.A., Scalding, M.A., Schmidt, DJ, Schwikert, S.A., Nicolich, J.M., Wellington, M., Lewis, J.C. & Logan, T.V. 1997. An experimental release of Whooping Cranes in Florida: The first three years. Proceedings North American Crane Workshop 7: 79-85.

Stehn, T. unpubl. report. Whooping Crane recovery activities April - November, 1997. 6 pp

Urbanek, R.P. & Bookhout, T.A. 1992. Development of an isolation-rearing/gentle release procedure for reintroducing migratory cranes. Proceedings North American Crane Workshop 6: 120-130.

Valentine, J.M., Jr. & Hereford, S.G. 1997. History of breeding pairs and nesting sites of the Mississippi Sandhill Crane. Proceedings North American Crane Workshop 7: 1-9.

Wellington, M., Burke, A., Nicolich, J.M. & O'Malley, K. 1996. Chick rearing. In: Ellis, D.H., Gee, G.F. & Mirande, C.M. (eds) Cranes: Their biology, husbandry, and conservation. National Biological Service, Washington, D.C. & International Crane Foundation, Baraboo, Wisconsin.