Confirmation: captive-bred galliforms are maladapted

One of the more surprising findings of this review is that supplementations continue to be made and experiments conducted using unparented captive-bred galliforms, despite the overwhelming evidence that such animals are seriously maladapted for life in the wild, surviving very poorly and, if they live long enough, also breeding very poorly (e.g. Hill and Robertson Reference Hill and Robertson1988, Dowell Reference Dowell1992, Parish and Sotherton Reference Parish and Sotherton2007). If further evidence is needed, several recent papers, dealing with both reintroductions and introductions (after establishing that the habitats and conditions were sufficiently similar to justify the attempt), provide it.

• • It took the release of 17,000 Common Pheasants Phasianus colchicus over 12 years (1968–1980) to establish a population in Texas, USA (Sokos et al. Reference Sokos, Birtsas and Tsachalidis2008).

• • A supplementation of Common Pheasants in Idaho, USA, compared the performance of translocated wild birds and released pen-reared birds, and found ‘wild female pheasants were seven times more likely to survive to 1 October, 10 times more likely to survive to the nesting season, eight times more productive, and one-third as expensive per egg hatched than [sic] pen-reared females’ (Musil and Connelly Reference Musil and Connelly2009).

• • In a study of translocated wild and commercially reared Grey Partridges Perdix perdix released into a new area in the Czech Republic ‘none of the commercially reared birds survived in the wild until the end of the nesting period, and none produced a fledged brood’, demonstrating ‘the uselessness of releasing adult commercially reared partridges in an effort to establish viable populations’ (Rymešová et al. Reference Rymešová, Tomášek and Šálek2013).

• • Fostering propensity in Red-legged Partridges Alectoris rufa is much higher in wild than captive-bred birds (suggesting ‘fostering is strongly related to parental care behaviour’), and an experiment to study adoption in wild and captive-bred pairs already with their own broods had to be discontinued because captive-bred pairs simply failed to breed (Sánchez-García et al. Reference Sánchez-García, Alonso, Pérez, Rodríguez and Gaudioso2011).

• • The state of Virginia, USA, spent 30 years (1930–1960) working on releases of and release techniques for Wild Turkey Meleagris gallopavo from game farms before abandoning the programme as a total failure; Pennsylvania persisted for another 20 years, releasing over 200,000 birds ‘with little to no evidence that any had established populations’ (which was achieved instead by translocations) (Hughes and Lee Reference Hughes and Lee2015).

• • Between 1961 and 1971 some 10,000 Red Junglefowl Gallus gallus, derived from 117 pure individuals imported from India, were introduced into the south-eastern USA, but no descendants survive today (Condon et al. Reference Condon, Brisbin and Chandler2019).

Causes and consequences of maladaptation

Filial imprinting in chickens takes place from hatching for about three days; the primary object of imprinting is the mother, but to a lesser degree it also involves co-hatched siblings such that, in the absence of a mother, chicks will imprint on each other (Nicol Reference Nicol2015). Early social experience promotes normal reproductive behaviour (Wood-Gush Reference Wood-Gush1958). It follows that galliform chicks deprived of the opportunity to bond naturally with their mother must be significantly disadvantaged in later life. Videos in cages at Paignton Zoo, UK, revealed that Vietnam Pheasant chicks stay within 1 m of their mother for 80% of their first month of life and for almost 50% of their second (J. Gregson in litt. 2019).

McPhee and McPhee (Reference McPhee and McPhee2012) observed that ‘captive environments can relax selective pressures on various traits, resulting in significantly more trait variance in captive-bred versus wild populations’. These traits include ‘fecundity, immune-genetics, digestive morphology, temperament and intrinsic and learnt behaviours including anti-predator responses’ (Dolman et al. Reference Dolman, Collar and Burnside2018). From the evidence furnished by the many studies of galliform introductions, reintroductions and supplementations to date, it is anti-predator responses that are the most decisive in determining the fate of releasees. Study after study has testified to the poor survival rates of captive-bred birds (all references in the preceding section plus, e.g., Leopold Reference Leopold1944, Robertson Reference Robertson1988, Dowell Reference Dowell1992, Priddel and Wheeler Reference Priddel and Wheeler1994, Putaala and Hissa Reference Putaala and Hissa1998, Buner et al. Reference Buner, Browne and Aebischer2011, Gaudioso et al. Reference Gaudioso, Pérez, Sánchez-García, Armenteros, Lomillos and Alonso2011b, Gruychev Reference Gruychev2014, Mihaylov et al. Reference Mihaylov, Gruychev and Stoyanov2014, Sanchez-Donoso et al. Reference Sanchez-Donoso, Rodríguez-Teijeiro, Quintanilla, Jiménez-Blasco, Sardà-Palomera, Nadal, Puigcerver and Vilà2014, Carter Reference Carter2015, Merta et al. Reference Merta, Kobielski, Krzywiński, Theuerkauf and Gula2015 and multiple citations in each). However, the degree to which anti-predator responses and other traits are inherited (genetic) or developmental (ontogenetic) is not clear, and the apparent naïvety of birds towards predators may reflect such things as reduced feeding ability and/or resistance to disease, attributable to the lack of parental superintendence. One study showed that captive-bred Grey Partridges Perdix perdix occupy field margins, which are rich in both food and predators and hence represent an ecological trap that wild birds avoid, apparently through cultural transmission (Rantanen et al. Reference Rantanen, Buner, Riordan, Sotherton and Macdonald2010). Lack of the opportunity for mother–offspring cultural transmission is equally believed to explain the failure of captive-bred Wild Turkeys to survive in the USA (Hughes and Lee Reference Hughes and Lee2015). Research has increasingly focused on finding techniques to compensate for this behavioural deficit in pen-reared birds, but recent work on inherited tameness in junglefowl finds a correlation with reduced brain size (Agnvall et al. Reference Agnvall, Bélteky and Jensen2017), a condition that may be relatively difficult to reverse.

Recent research has also highlighted the role of stress in curtailing survival rates in released captive-bred animals (Dickens et al. Reference Dickens, Delehanty and Romero2010). While the ‘unfearfulness’ (lack of vigilance and awareness) resulting in predation might be considered the consequence of lack of stress, any kind of new environment, making unprepared-for demands on an animal’s stamina and resourcefulness, will itself elevate levels of stress: having to find food and shelter in unfamiliar contexts potentially represents a shock to an individual, one effect of which may be reduced cognitive ability, including predator identification (Teixeira et al. Reference Teixeira, Azevedo, Mendl, Cipreste and Young2007). One cause of the Red Junglefowl’s failure to become established in the south-eastern USA (see above) was rapid long-distance dispersal, but birds were also reported to move, where possible, into denser vegetation (Condon et al. Reference Condon, Brisbin and Chandler2019); rapid dispersal was probably therefore a highly stressed response to the absence of appropriate cover. Captive-bred Grey Partridges were more stressed by unpredictable food supply—which is far more likely a circumstance in the wild than in captivity—than wild ones, and survived less (Homberger et al. Reference Homberger, Jenni-Eiermann, Roulin and Jenni2013, Reference Homberger, Jenni, Duplain, Lanz and Schaub2014). However, evidence is mounting that even in the confinement of a cage an animal can be unduly stressed by multiple factors including, simply, the lack of behavioural stimuli (see next section).

Meanwhile, it is predictable that changes in the traits listed by Dolman et al. (Reference Dolman, Collar and Burnside2018) (above) will become more pronounced with number of generations, as a result of (unintended) artificial selection and cultural transmission. Since cases are rare where this is an issue, it is not a mainline topic in reintroduction biology (it is not mentioned in WPA 2009 or IUCN/SSC 2013). However, recent research has begun to indicate that a serious effect exists. In a rodent species McPhee (Reference McPhee2003) found that ‘the more generations a population has been in captivity, the less likely an individual is to take cover after seeing a predator’, while a review of 90 papers simply concluded that ‘the best approach to minimize genetic adaptation is to reduce the number of generations that a species spends in captivity’ (Williams and Hoffman Reference Williams and Hoffman2009).

Such testimony is of alarming relevance to the case of the Vietnam Pheasant, and is supported by evidence documented nearly 40 years ago and cited by Ridley (Reference Ridley1986): ‘Pielowski (Reference Pielowski1981) found that pheasants were substantially changed merely by being kept in captivity for 20 generations. Their chances of surviving in the wild were reduced to about half of those of identically reared chicks from the eggs of wild birds’.

Given that the Vietnam Pheasant has passed through at least 35 generations (Hennache et al. Reference Hennache, Mahood, Eames and Randi2012), any project to return the species to the wild must accept the obligation to research this problem and provide the fullest set of management procedures and practices to offset and neutralise it.

Managing maladaptation, 1: environmental enrichment

A review of the evidence relating to the release of captive animals found that ‘a lack of complexity in the developmental environment of an individual negatively impacts its neural development, with wide-ranging implications for adult behavior and survival’ (Reading et al. Reference Reading, Miller and Shepherdson2013). Consequently, it was recommended that when animals are being bred and kept for release into the wild the conditions in their cages should mimic the release site habitat as closely as possible, since the benefits include reducing the chance of dispersal, increasing physical fitness, and promoting locomotory, anti-predator, foraging and social skills.

Support for this position came in a study in which the introduction of perches to the cages of captive-bred Common Pheasants had the dramatic effect of allowing birds to reduce aggression, improve body development and spatial memory, roost more securely and, overall, survive better (Whiteside et al. Reference Whiteside, Sage and Madden2016). In the light of this it is perhaps salutary to recall that, when liberated into wild habitat, artificially reared Cheer Pheasants Catreus wallichii roosted on the ground at night and were heavily predated (Garson et al. Reference Garson, Young and Kaul1992).

Managing maladaptation, 2: soft releases

The value of ‘soft’ releases (where animals are conditioned to the release environment by being caged within it for varying periods of time) is well recognised by WPA (2009) and IUCN/SSC (2013). Evidence relating to galliforms was furnished by Lockwood et al. (Reference Lockwood, Griffin, Morrow, Randel and Silvy2005), but more recent research has further demonstrated the value of habituating birds to the habitat into which they are to be released.

The duration of acclimatisation within the soft-release enclosure positively influenced post-release survival of Red-billed Curassows, and a period of at least 47 days was consequently recommended (Bernardo et al. Reference Bernardo, Lloyd, Bayly and Galetti2011). Nine Cabot’s Tragopans Tragopan caboti kept for over 50 days in a soft-release enclosure in the species’s montane habitat survived far better than 11 birds held there for only three days (86% vs 20% after 50 days), and they selected habitats more typical of wild birds (Liu et al. Reference Liu, Li, Lloyd, Xia, Zhang and Zheng2016). In Grey Partridges ‘longer acclimatisation time at the release site was related to lower levels of corticosterone [associated with passive and therefore sociable behaviour] and partly enhanced survival after release’ (Homberger Reference Homberger2014).

The failure of the huge project to introduce Red Junglefowl to the south-eastern USA (see above) has largely been attributed to the use of ‘hard’ (i.e. instant) releases (Condon et al. Reference Condon, Brisbin and Chandler2019). The fact that some birds were found in or seen heading toward dense vegetation (as mentioned above), when matched with the finding in Liu et al. (Reference Liu, Li, Lloyd, Xia, Zhang and Zheng2016) that soft-released birds selected more appropriate habitat, strongly implies that establishing the release site amidst dense cover is likely to help releasees in terms of stress minimisation and survival capacity. Armstrong and Wittmer (Reference Armstrong and Wittmer2011) emphasised that ‘the quality of the release site is a particularly important factor affecting the benefits gained from large releases.’

Soft releases may include the provision of food and shelter for released birds. This can help anchor birds to the release site and prevent rapid, perhaps panic-driven dispersal. In the case of the White-winged Guan Penelope albipennis, productivity has bluntly been attributed to food supplementation: ‘since 2001 there have been 50 wild White-winged Guan chicks born from reintroduced birds at CPCA… due to food availability’ (Angulo Reference Angulo2008).

Managing maladaptation, 3: parent-rearing

Buner and Schaub (Reference Buner and Schaub2008) found that the survival of captive-bred but parent-reared Grey Partridges was lower than in translocated wild individuals and captive-bred birds fostered to wild parents, attributing the difference to wild birds being more familiar with predators. Nevertheless, several recent studies have demonstrated that there is some value in parent-rearing galliforms intended for release.

Breeding stock (25 males, 25 females) of Red-legged Partridges was obtained from a game farm, introduced in autumn into a 4,000 m² pen enclosing grass and shrubs, and allowed to mate freely (a strategy already shown to benefit females); a year later the 42 offspring were caught at night, released into a 7.5 km² protected area alongside a cohort of 36 artificially reared birds, and after six months achieved a re-sighting rate of 22.6% against 0 for the ‘artificial’ stock (Santilli et al. Reference Santilli, Galardi and Bagliacca2012). Similarly, parent-reared Red-legged Partridges greatly outperformed intensively reared birds in terms of survival and behaviour, and came moderately close to figures achieved by wild birds (Pérez et al. 2012). Western Capercaillies Tetrao urogallus reared in semi-liberty by their captive mother and released next to her cage survived far better (males 549 vs 253 days, females 293 vs 56 days) than chicks reared in captivity and released in the absence of their mother, suggesting that ‘mother-assisted rearing and release reduces the mortality of capercaillie chicks because they are assisted by anti-predator behaviour of their mother and explore the new environment in a similar way as chicks hatched in the wild’ (Merta et al. Reference Merta, Kobielski, Krzywiński, Theuerkauf and Gula2015).

Managing maladaptation, 4: dietary enrichment

High-protein low-fibre diets commonly fed to commercially reared galliforms do not prepare them for the high-fibre dietary conditions that tend to apply in the wild (Parish and Sotherton Reference Parish and Sotherton2007). Exposing Northern Bobwhite Colinus virginianus chicks to insects, in part to allow them to broaden their foraging skills and in part to add a nutritional component considered vital to feather growth (and hence to thermoregulation and flight), has been presumed but not (yet) proved to enhance post-release survival rates (Gall et al. Reference Gall, Kuvlesky, Gee, Brennan, Palmer, Burger and Pruden2000).

Improved post-release survival was, however, achieved in Common Pheasants through their prior exposure to more naturalistic diets, which meant that they kept better watch because they took less time to forage, handled live prey faster, relied less on supplementary feed and developed different gut morphologies; and therefore reduced the risk of predation by their vigilance, fed themselves better, and coped better with natural forage once supplementary feeding ceased (Whiteside et al. Reference Whiteside, Sage and Madden2015). In Western Capercaillies, the lower diversity of bacterial microbiota in the caeca produced by a captive diet ‘may be responsible for the high mortality of captive birds released into nature’ (Wienemann et al. Reference Wienemann, Schmitt-Wagner, Meuser, Segelbacher, Schink, Brune and Berthold2011), and captive ducks developed relatively short, light intestines that were judged to compromise digestive efficiency and hence survival prospects in releasees, a condition considered remediable by high fibre and greater diversity in the diet (Moore and Battley Reference Moore and Battley2006).

Managing maladaptation, 5: anti-predator training

Anti-predator training prior to the release into the wild of captive-bred animals has become a widespread practice in reintroduction biology (review in Griffin et al. Reference Griffin, Blumstein and Evans2000), where it is also labelled ‘environmental enrichment’ (Roberts et al. Reference Roberts, Taylor and Garcia de Leaniz2011). It has been used occasionally in some galliform programmes: as early as 1975 a Northern Bobwhite release project included the harassment of potential releasees by humans, dogs and hawks, but found that too few birds could be trained in that way (Carpenter et al. Reference Carpenter, Gabel and Goodwin1991). More recently captive-bred adult Red-legged Partridges trained by caged birds calling and reacting to model aerial predators survived six times longer than untrained birds (105.2 vs 17.8 days), although when the training shifted to terrestrial (non-avian) predators the panic caused fatalities and work was suspended (Gaudioso et al. Reference Gaudioso, Sánchez-García, Pérez, Rodríguez, Armenteros and Alonso2011a). Visits by predators to release cages at night (see next section) are assumed to stress releasees (Keiter and Ruzicka Reference Keiter and Ruzicka2017), but presumably may still benefit the birds by familiarising them with the shape, smell and sound of hostile animals (and are perhaps less stressful than—albeit not necessarily a substitute for—training sessions). This dilemma requires further evaluation, along with the need to train birds to fear humans.

Training (calling) animals should preferably be of the same species as the trainees, but this is not always necessary; and training should coincide with the period of maximum learning receptiveness—evidently before six months (see below)—in young birds.

Managing maladaptation, 6: anti-predation stratagems

Unnaturally high concentrations of predators attracted to the vicinity of release pens (containing potential prey for up to 50 days) obviously prejudice the survival of releasees (Robertson Reference Robertson1988, Dowell Reference Dowell1992, Parish and Sotherton Reference Parish and Sotherton2007). One stratagem to overcome this problem is to deploy multiple smaller release stations within the release site. Gortázar et al. (Reference Gortázar, Villafuerte and Martín2000) released one cohort of Red-legged Partridges at a single central site and a second the following year at multiple sites around the centre: although both groups suffered 25–34% mortality in the first 72 hours after release, 37% of birds at dispersed sites survived the first month as against 6% at the central site. Eitniear (Reference Eitniear2010), seeking to reintroduce Montezuma Quail Cyrtonyx montezumae to an area of Texas, set up night-cameras around release pens to record what mammalian predators came to the area; live-trapped and translocated as many of these predators as possible; and placed seven male quails in a small cage within the larger release enclosure to lure the releasees back for food and roosting each night.

Straightforward predator control was practised by Angulo (Reference Angulo2006, Reference Angulo and Soorae2011) with White-winged Guans. A technique apparently untried with galliform reintroductions is the deployment of top predator odours such as tiger faecal extract to deter potential meso-predators from entering the area (Murray et al. Reference Murray, Burns and Davy2006, Garvey et al. Reference Garvey, Glen and Pech2016).

Managing maladaptation, 7: numbers to release

Success rates in reintroductions correlate positively with number of animals involved; projects releasing over 100 individuals—not necessarily or desirably all at once—have been most successful, since populations below this number (and even above it, if the sexes are not in balance) are disproportionately susceptible to the influence of demographic events and the disruption of social and sexual interactions, including through Allee effects (sources in Faria et al. Reference Faria, van Oosterhuit and Cable2010, Armstrong and Wittmer Reference Armstrong and Wittmer2011, Tracy et al. Reference Tracy, Wallis, Efford and Jamieson2011). Moreover, compensation for persistent ‘behavioral variation’ (maladaptive behaviour resulting from generations in captivity, notably loss of fearfulness) requires an increase in the number of releasees, in one case calculated at 30–50% (McPhee and Silverman Reference McPhee and Silverman2004).

For European grouse a review found that ‘annual releases of at least 30 birds are needed for a period of more than 6 years, to reach a 50% probability for survival and reproduction of released birds’ (Seiler et al. Reference Seiler, Angelstam and Bergmann2000), i.e. at least 180 birds altogether. However, a 50% probability of success is an unacceptable threshold for a Critically Endangered species that might already have died out in the wild; a far higher number of releasees (below I arbitrarily assume double) would be necessary to achieve 95% or 99% probability.

Guidance exists (Canessa et al. Reference Canessa, Hunter, McFadden, Marantelli and McCarthy2014) on how to balance numbers released with numbers retained in captivity to preserve full genetic diversity. Guidance does not exist, however, on how birds should be released (for which many options exist, e.g. 180 birds over six years could involve 30 in each year, as mentioned above, or 80 in year 1 and 20 each in years 2–6; and the 80 in the first year could involve one, two or three batches, at various several-month intervals, etc.). Without full explanation, Lockwood et al. (Reference Lockwood, Griffin, Morrow, Randel and Silvy2005) concluded that future releases of Prairie Chickens Tympanuchus cupido should be multiple, ‘with the initial release containing the fewest number of birds’ but including ‘excess males’ as a priority.

Managing maladaptation, 8: choice and treatment of birds

Pre-release Red-billed Curassows suffered mortalities caused by aggressive individuals in release pens, leading to the suggestion that cameras should be used to identify dangerously dominant behaviour early and allow the problematic individuals to be isolated (Bernardo et al. Reference Bernardo, Lloyd, Bayly and Galetti2011). This recommendation did not imply that these individuals should not be released. However, ‘boldness’ and ‘risk-taking’ in captive-bred animals has been found to be a threat to the individuals themselves (Bremner-Harrison et al. Reference Bremner-Harrison, Prodohl and Elwood2004, Roberts et al. Reference Roberts, Taylor and Garcia de Leaniz2011), and in Grey Partridges survival increased with diminishing levels of both temerity and timidity (Homberger Reference Homberger2014). Clearly such individuals need to be either treated for their behaviour or removed from the reintroduction programme. Importantly for the many galliform birds which seasonally form social groups, which includes most Lophura pheasants (Hennache and Ottaviani Reference Hennache and Ottaviani2005), ‘coveys’ or social groups appear to serve to regulate corticosterone in individuals, such that ‘the substantial covey effects… suggest that strategic manipulation of the social group prior to release could prove useful to enhance re-introduction success’ (Homberger Reference Homberger2014). Releasing in coveys has proved relatively efficacious in Grey Partridge restocking (Buner et al. Reference Buner, Browne and Aebischer2011).

Little information appears to exist on the selection of individuals for release. In a White-winged Guan reintroduction, only first- or second-generation and parent-raised birds were selected, requiring (a) maximum distance in blood relationship between selected individuals, (b) 1:1 sex ratio, (c) optimal health status and (d) sexual maturity (Angulo Reference Angulo2006, Reference Angulo and Soorae2011). These criteria are clearly sound, but there is a strong case for varying the ages of birds by releasing immature individuals with their parents, since the former have a higher capacity to learn and adapt up to the age of six months, so that release at 4–5 months (allowing for 50 days in the release aviaries) might be expected to balance high levels of security with the opportunity to adjust to wild conditions (C. J. Nicol in litt. 2019). Learning by the releasers (i.e. the staff in charge) as well as the releasees allows adaptive management of the process, with the proportion of adults and juveniles determined by circumstance (Runge Reference Runge2013).