Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Action: Release translocated/captive-bred mammals into area with artificial refuges/breeding sites Terrestrial Mammal Conservation

Key messages

Read our guidance on Key messages before continuing



  • Abundance (5 studies): Two of three studies (two replicated, two controlled) in Spain and the USA found that translocation release sites with artificial burrows provided had higher abundances of European rabbits and densities of California ground squirrels compared to those without. The other study found that abundance of European rabbits following translocation was similar with and without artificial burrows provided. A replicated, controlled study in the USA found that after translocating black-tailed prairie dogs to areas with artificial burrows, colonies increased in size. A before-and-after study in Spain found that translocating European rabbits into areas with artificial refuges to supplement existing populations did not alter rabbit abundance, although two of three populations persisted for at least three years.
  • Reproductive success (4 studies): Three studies in Australia, Ireland and the UK found that released captive-bred sugar gliders, most translocated female red squirrels and some translocated pine martens provided with nest boxes and supplementary food reproduced. A study of 12 translocation projects in Slovakia, the Czech Republic and Poland found that translocated European ground squirrels released initially into enclosures or burrows with retention caps reproduced after release, whereas those without enclosures or burrows dispersed from release sites.
  • Survival (9 studies): Five of eight studies in Australia, the USA, UK, Ireland and South Africa found that at release sites with artificial refuges, and in some cases food provided, a population of captive-bred sugar gliders survived at least three years, two of three populations of red-tailed phascogales survived for more than four years, most translocated black bears survived at least one year and over half translocated red squirrels and pine martens survived 8-12 months. Three studies found that at release sites with artificial refuges, food and in one case water provided, no translocated red squirrels survived more than five months, all translocated rock hyraxes died within three months and most translocated Tipton and Heermann’s kangaroo rat spp. died within five days. A randomised, replicated, controlled study in Hungary found that translocated European ground squirrels released into plugged artificial burrows had higher recapture rates than those released into unplugged artificial burrows.


  • Use (2 studies): Two studies in Australia found that released captive-bred sugar gliders used artificial nest boxes provided.
  • Behaviour change (1 study): A replicated, before-and-after study in the USA found that translocated Utah prairie dogs released into an area with artificial burrows, after the control of native predators, tended to leave the release site and spent more time being vigilant than before.

Supporting evidence from individual studies


A study in 1979–1981 of a young planted native forest reserve in Victoria, Australia (Suckling & Macfarlane 1983) found that a population of released, captive-bred sugar gliders Petaurus breviceps provided with artificial nest boxes and supplementary food survived, bred and used the nest boxes. In the third year after releases began, 37 individuals were recorded. Seven animals had been wild-born in the year after release and six females >2 years old showed signs of having reproduced. Occupation by sugar gliders or signs of previous occupation were recorded in 30 of 38 boxes, all three terra-cotta pipes and in 10 of 14 artificial hollow limbs. On a 130-ha island of planted native forest (trees ≤17 years old), 72 sugar gliders were released in January or February of 1979 (26 individuals), 1980 (34 individuals) and 1981 (12 individuals). Seventy boxes, pipes or hollowed limbs (dimensions not provided) were installed on trees, 3–7 m above the ground. Supplementary food was provided at release points during winters of 1979 and 1980. Gliders and artificial nest boxes were surveyed in May 1981.


A study in 1993–1994 on a forested peninsula in Dorset, UK (Kenward & Hodder 1998) found that none of the translocated red squirrels Sciurus vulgaris provided with nest boxes, supplementary food and water (in and once released from pre-release pens) survived over five months after release. Out of 14 translocated red squirrels, 11 (79%) survived over one week, three (21%) survived >3 months and none survived >4.5 months. At least half of the 14 squirrels were killed by mammalian predators. Intact carcasses that were examined showed signs of weightloss and stress (see original paper for details). Between October and November 1993, fourteen wild-born red squirrels were released into an 80-ha forest dominated by Scots pine Pinus sylvestris. The forest had no red squirrels but had introduced grey squirrels Sciurus carolinensis. Capture and release sites were similar habitats. Squirrels were transported in wooden nest boxes filled with dry hay. Squirrels were placed with their nest boxes into 1.5 × 1.5 × 1.5 m weldmesh pens surrounded by electric fencing for 3–6 days before release. Squirrels were kept individually except for 2 males who shared a pen. Supplementary food comprised a mixture of seeds, nuts and fruit on trays and in feed hoppers. After release, squirrels continued to have access to food, water and nest boxes inside the pens and outside (20-100 m away). All squirrels were radio-tagged and located 1–3 times/day, for 10–20 days after release and thereafter every 1–2 days.


A study in 1996 of a forest in Victoria, Australia (Wood & Wallis 1998) found that nest boxes were used by a population of released captive-bred sugar gliders Petaurus breviceps. Twenty out of 67 nest boxes were occupied by sugar gliders. Additionally, 18 boxes were occupied by feral honeybees Apis mellifera, a potential competitor for use of boxes. Boxes used by sugar gliders were positioned higher (average 4.5 m) than boxes used by honeybees (average 3.5 m). The site was formerly logged and had subsequently been replanted. Sixty-seven boxes were inspected in July 1996. Boxes had been installed, and captive-bred sugar gliders released in 1979–1982. Boxes were 10–27 l in capacity. Fifty-three boxes were positioned 3–5 m above ground. Seven were >5 m high and seven were <3m high, including three that had fallen to the ground.


A replicated, controlled study in 1999–2003 on a grassland site in Montana, USA (Dullum et al. 2005) found that after translocating black-tailed prairie dogs Cynomys ludovicianus in social groups to areas with artificial burrows, colonies increased in size over four years. Six colonies receiving translocated prairie dogs grew more in area over four years (total growth 72 ha, 924% of pre-translocation area) than did 20 similar-sized colonies that did not receive translocated prairie dogs (total growth 27 ha, 93% increase). Two active colonies (with existing prairie dog populations at the start of the study) that each received 120 prairie dogs increased more over four years (total increase 37 ha, 971% of pre-translocation area) than did two active colonies each receiving 60 prairie dogs (total growth 31 ha, 768%). An inactive colony that received no prairie dogs remained inactive. In June–July 1999, prairie dogs were released into pre-existing burrows (up to eight prairie dogs/burrow) or drilled holes (8 cm diameter × 60 cm deep, 45° below horizontal, up to two prairie dogs/hole, 30 holes/site). Colony size was measured four years later. Nine experimental colonies, three each occupying areas of 0 ha (inactive), 0.1–2.0 ha and 2.0–6.6 ha, were studied. In each size class, translocations to the three colonies were of 0, 60 and 120 prairie dogs. Growth-rates of 20 non-supplemented colonies were also monitored.


A study in 2000–2003 in temperate forest in a wildlife refuge in Arkansas, USA (Wear et al. 2005) found that most translocated black bears Ursus americanus released into man-made dens survived at least one year after release. The first-year post-release survival rate for translocated adult female bears was 62%. For those surviving >1 year after release, second-year survival was 91%. The first-year survival rate of translocated cubs was 75%. Of eight documented adult female mortalities, at least three were due to poaching. Four bears returned to their capture site. In March 2000–April 2002, twenty-three wild adult female black bears and their 54 cubs were captured in White River National Wildlife Refuge and released, 160 km away, into man-made dens at Felsenthal National Wildlife Refuge. Radio-telemetry was used track bears and gather movement data weekly, through to January 2003.


A controlled study in 1999–2002 in a shrubland site in Huelva, Spain (Cabezas & Moreno 2007) found that providing artificial warrens to translocated European rabbits Oryctolagus cuniculus did not increase their abundance relative to those translocated without provision of artificial warrens. Over the three-year study, average rabbit pellet density in translocation plots where warrens were provided (4.4 pellets/m2) was not significantly different to that in plots where warrens were not provided (5.0 pellets/m2). The study was conducted in four 4-ha square plots (1–6 km apart) in Doñana National Park. Eight artificial warrens, with internal galleries and multiple entrances, were built in each of two plots. Two batches of rabbits, each totalling 64–67 animals, were translocated into each of two plots (one with and one without warrens) each winter from 1999–2000 to 2001–2002. Translocation plots were switched after the first winter, such that translocations in the second and third winter were into plots where no translocations were made in the first winter. Between September 1999 and November 2002, rabbit abundance was estimated every two months by counting the number of pellets in 33 fixed-position 0.5-m diameter sampling points/plot. Wild rabbits were present in all plots prior to translocations beginning.


A study in 2005–2007 in a mixed conifer forest in Galway, Ireland (Poole & Lawton 2009) found that over half of translocated red squirrels Sciurus vulgaris provided with nest boxes and supplementary food (in and once released from holding pens) survived over eight months after release and most females reproduced during that period. At least 10 out of 19 (53%) translocated squirrels survived over eight months post-release and five out of nine translocated females (56%) were lactating 5-7 months after release. In August 2006, seven juvenile squirrels were caught. At least one squirrel was still alive at the release location two years after the original release. Two squirrels died while in the release pen or shortly afterwards. Another four squirrels died 1-2 months after release. Nineteen squirrels were translocated to a nature reserve (19 ha) in the middle of a 789-ha commercial pine plantation, 112 km from the capture site. Individuals were marked, radio-tagged and kept on average for 46 days in one of two pre-release enclosures (3.6 × 3.6 × 3.9 m high). Enclosures contained branches, platforms, nest boxes, and supplementary feeders (containing nuts, maize, seeds and fruit). Supplementary food (50/50 peanut/maize mix) was provided in six feeders in the nature reserve until July 2006. Twenty nest boxes were also provided. Squirrels were radio-tracked in September and November 2005 and February and May 2006, and were trapped in February, May and August 2006 and observed once in October 2007.


A study in 2005–2006 at rocky outcrops on a reserve in KwaZulu-Natal Province, South Africa (Wimberger et al. 2009) found that translocated rock hyraxes Procavia capensis that were provided with an artificial refuge and food after release in a social group, having been held in captivity, all died (or were presumed to have died) within 87 days of release. Eighty-seven days after the release of 17 hyraxes, none could be relocated. In July 2005, ten adult hyraxes were caught in baited mammal traps (900 × 310 × 320 mm) in an area where they were abundant, and held in captivity for 16 months, during which time three died. The remaining seven were released in November 2006, along with the eight juveniles and two pups born to them in captivity, to a 656-ha reserve where the species was nearly extinct. For four months prior to release, the group was housed together in an outdoor cage (5.9 × 2.5 × 3.2 m). Hyraxes were released into a hay-filled hutch which was left in place for several months, and were provided with cabbage for one week after release. Hyraxes were monitored by direct observations and by walking regular transects, daily for the first week but decreasing to monthly by the end of the study.


A study in 2001 in a grassland and shrubland site in California, USA (Germano 2010) found that most Tipton kangaroo rats Dipodomys nitratoides nitratoides and Heermann’s kangaroo rats Dipodomys heermanni ssp. translocated into artificial burrows provided with supplementary food died within five days of release. All four Tipton kangaroo rats were predated within five days of translocation, and only one out of seven Heermann’s kangaroo rats survived over 45 days. Three Heermann’s kangaroo rats were predated, two died as a result of aggression from other Heermann’s kangaroo rats, and the fate of one was unknown. In September 2001, four juvenile Tipton kangaroo rats and three Heermann’s kangaroo rats were captured and held in captivity for two months before release at a protected site in November. In December 2001, a further four Heermann’s kangaroo rats were caught and translocated to the same site. All 11 animals were fitted with a radio-transmitter and ear tags, and monitored for seven days in captivity prior to release. The release site was already occupied by Heermann’s kangaroo rats. Animals were released into individual artificial burrows (two 90-cm-long cardboard tubes with a chamber about 30 cm below the surface), dug 10–15 m apart and provided with seeds. Burrows were plugged with paper towels until dusk. Animals were radio-tracked every 1–8 days for 18–45 days after release.


A randomised, replicated, controlled study in 2000 in a grassland site in central Hungary (Gedeon et al. 2011) found that translocated European ground squirrels Spermophilus citellus released into plugged artificial burrows had higher recapture rates than did ground squirrels released into unplugged artificial burrows. From four to 10 days after release, a higher proportion of ground squirrels released into plugged artificial burrows were recaptured (19 out of 60, 32%) than squirrels released into unplugged artificial burrows (6 out of 57, 11%). The highest recapture rate came from the group released into plugged burrows in the morning (15 out of 30). From 22–24 April 2000, one hundred and seventeen wild-caught European ground squirrels were translocated to a fenced 40-ha protected grassland. Four 40 × 40-m grid cells were established, each containing vertical, artificial burrows (50 cm long, 4.5 cm diameter) spaced 4.5 m apart. Sixty animals were released into burrows plugged with wood caps (from which they could only exit by digging out) across two grid cells and 57 into unplugged artificial burrows in the other two grid cells. One individual was released/burrow. Approximately half the squirrels were released in the afternoon on the day of capture. Animals to be released in the morning were kept in individual wire cages (10 × 10 × 40 cm) for one night and provided with fresh apple slices prior to release. From 28 April–2 May, squirrels were recaptured with snares to record retention.


A study of 12 translocation projects in 1989–2010 in 14 grassland sites in Slovakia, the Czech Republic and Poland (Matějů et al. 2012) found that translocated European ground squirrels Spermophilus citellus released initially into enclosures or burrows with retention caps (‘soft-release’) reproduced on site after release, but individuals released without an initial preadaptive period or support after release (‘hard-release’) dispersed from release sites. Translocations in which at least 23 individuals/season were released into enclosures or capped abandoned/artificial burrows led to reproduction (results reflect statistical model outcomes). However, animals released without initial containment did not settle at release sites. The study analysed data from 12 projects, involving release of ground squirrels at 14 sites. Around 2,500 grounds squirrels were released (4–1,057 individuals/project; 4–136 individuals/release season). Animals were ‘soft-released’ in eleven projects, ‘hard-released’ in two and combined hard and soft-released in one project. Three releases involved both captive-bred and wild-bred individuals. The remainder were of wild-bred translocated animals.


A replicated, site comparison study in 2008–2012 in 32 shrubland sites in Andalusia, Spain (Guerrero-Casado et al. 2013) found that release sites with shelter and artificial warrens provided had higher abundances of European rabbits Oryctolagus cuniculus following translocation. There were more rabbit latrines at sites where artificial warrens and wooden branches were provided (1.6–7.1 latrines/km) than at sites where they were not provided (0.3–3.4 latrines/km), although the size of the effect was less when scrub coverage was high (see original paper for details). In 2008–2009, between 75 and 90 rabbits/ha were released inside 32 fenced plots (0.5–7.7 ha). Artificial warrens and wooden branches were added within a 500-m radius of some plots and, in some sites, scrubland was cleared to create pasture (number of plots/treatment not stated). Twelve plots had no wooden branches or artificial warrens (wooden pallets covered with stones, branches and earth) added. From the end of the 2009 breeding season, small gates on fences were opened and the rabbits could disperse into adjacent areas. Relative rabbit abundance was estimated by latrine counts, in four 500-m transects outside each plot, in the summers of 2008–2009 and 2012. Scrub cover was classified as low (0-30% coverage), medium (30-60%) and high (>60%).


A replicated, before-and-after study in 2010–2011 in two grassland sites in Utah, USA (Curtis et al. 2014) found that translocated Utah prairie dogs Cynomys parvidens released into an area with artificial burrows after the control of native predators tended to leave the release site and spent more time being vigilant than before. Only 50 out of 779 (6%) were still present at the release sites two months after release. After translocation in both family groups and groups of unrelated individuals, prairie dogs spent more time being vigilant (48%) than they had done before translocation (22%). In July 2010 and 2011, prairie dogs (379 and 400) were caught on a golf course using baited Tomahawk wire box-traps. Individuals were marked with hair dye and ear tags and released the same day at two sites with artificial burrow systems, with up to 10 animals/burrow. Each site had four release areas at least 200 m apart, each containing five burrows, 4 m apart. Each burrow consisted of a 30 × 45 × 30 cm box, buried 1.8m deep, and with two entrances (10-cm diameter and 4-m long) made from plastic tubing. Extra holes were left in the box and tubing to allow burrow expansion. Burrow entrances were protected from predators by mesh cages. At each site, two release areas were used for family groups and two were used for non-related groups. Predator removal of coyote Canis latrans and badgers Taxidea taxus was conducted for several weeks before and after prairie dog release. In September 2010 and 2011, prairie dogs were trapped, using 100 traps/site, during two sessions of four days each to determine numbers remaining at the site.


A before-and-after study in 2004–2007 in three mixed pasture and scrubland sites in southwest Spain (Guil et al. 2014) found that translocating European rabbits Oryctolagus cuniculus into areas with artificial refuges to supplement existing populations did not alter rabbit abundance, though populations persisted at two of three sites for at least three years. Three years after artificial warrens were built and rabbits were released, rabbit abundance was not significantly different to that before warrens were built (no data reported). In two of three sites, the rabbit population persisted for at least three years, but at one site no rabbits were seen three years after release. In 2004, at three sites, 20–72 artificial warren tubes were installed. In autumn 2004, wild translocated rabbits were released at each site and, in autumn 2005, more rabbits were released at two of the sites. In total, 150–387 rabbits were released at each site. Rabbit presence was detected at two of the sites before releases of translocated animals. In June–September of 2004–2007, rabbit droppings were counted along 10–12 transects, each 500 m long.


A study in 2006–2015 in three woodland and shrubland sites in Western Australia and Northern Territory, Australia (Short & Hide 2015) found that following release into areas with artificial refuges, two translocated populations of red-tailed phascogales Phascogale calura survived for more than four or five years, but one captive-bred population survived for less than a year. The two populations of phascogales established from wild-caught animals survived longer (4–5 years) than one population established from captive-bred animals (which had been kept in pre-release pens and given supplementary food; < 1 year). Authors suggest that the unsuccessful site may also have had a shortage of tree hollows for nesting. In July 2006 and January–February 2007, thirty-two captive-bred phascogales were released into a 26-ha fenced reserve (outside which feral cats were abundant) after spending either 10 days or over four months in a pre-release pen (3×6×2 or 4.5×3×2.2 m). Supplementary food was provided for one week after release. In April 2009 and June 2010, twenty-seven wild-caught phascogales were released into a 430-ha fenced reserve. In May 2010 and May 2011, thirty wild-caught phascogales were released into a 389-ha unfenced reserve, where poison baiting was used to control foxes Vulpes vulpes until 2012, but this was suspended due to a possible positive effect on feral cats Felis catus. Wild-caught animals had no pre-release pen or supplementary food. Nest boxes (11–35/site) were provided in every reserve. Phascogales were monitored after each release using radio-collaring or Elliott live traps, and through periodic monitoring of the nest boxes.


A replicated, controlled study in 2011–2014 of two areas of grassland and scrubland in southern California, USA (McCullough-Hennessy et al. 2016) found that where holes were drilled into the soil, densities of translocated California ground squirrels Otospermophilus beecheyi were higher than where no holes were drilled. Two years after management commenced, there were more squirrel burrows in drilled areas (43–124/subplot) than in areas that had not been drilled (11–122/subplot). Six plots each comprised a 0.8-ha circle, divided into three equal wedge-shaped subplots. Subplots were mown (in May, for two years, at 7.5–15 cm height, with cut material removed) and were either drilled with a soil auger (20 holes/subplot) or not drilled. The third subplot (data not presented here) was not mown and did not have holes drilled. Management commenced in 2011 (two plots) and 2012 (four plots). Squirrels were translocated into plots at a rate of 30–50/plot. Squirrel abundance was determined by counting squirrel burrows.


A study in 2015–2016 in a wooded mountain region in central Wales, UK (MacPherson 2017) found that some translocated pine martens Martes martes held in pre-release pens and then provided with supplementary food and nest boxes survived and bred in the first year after release. At least four out of 10 females that had been kept in pre-release pens survived and bred the year after release. Around 10–12 months after release, 14 out of 20 martens were alive and in good condition. Twelve were within 10 km of their release site. Six martens died in the first year, two had a fungal infection two weeks after release. Authors suggest this may have been due to damp conditions in November. From September–November 2015, twenty breeding age (>3-years-old) pine martens were caught in Scotland, health checked, microchipped and fitted with a radio-collar, and in some cases a GPS logger. Martens were transported overnight to Wales, and held in individual pre-release pens (3.6 × 2.3 × 2 m) for up to seven nights. Males’ pens were within 500 m of a female, but >2 km from the nearest male. Releases took place in autumn, and supplementary food was provided for 2–6 weeks after release (for as long as it continued to be taken). Den boxes were provided within 50 m of each release pen. Martens were radio-tracked until home-ranges were established, then located daily–weekly. Intensive tracking of females was carried out in March to locate breeding sites. Hair tubes and camera traps were used to monitor breeding success. A further 19 martens were released using the same procedure in September–October 2016.

Referenced papers

Please cite as:

Littlewood, N.A., Rocha, R., Smith, R.K., Martin, P.A., Lockhart, S.L., Schoonover, R.F., Wilman, E., Bladon, A.J., Sainsbury, K.A., Pimm S. and Sutherland, W.J. (2020) Terrestrial Mammal Conservation: Global Evidence for the Effects of Interventions for terrestrial mammals excluding bats and primates. Synopses of Conservation Evidence Series. University of Cambridge, Cambridge, UK.