Captive breeding cannot sustain migratory Asian houbara Chlamydotis macqueenii without hunting controls

Abstract

To evaluate the potential contribution of captive breeding to the conservation of exploited migratory Asian houbara Chlamydotis macqueenii, we estimated release numbers required to stabilise a population in a hunting concession (14,300 km 2), under scenarios of local licensed hunting and flyway-scale protection. We developed a population model, initially 2350 adult females, re-sampling parameters measured through fieldwork and satellite telemetry, over 1000 iterations. With current flyway-scale unregulated harvest, and without any licensed hunting in the concession, populations declined at 9.4% year −1 (95% CI: −18.9 to 0% year −1); in this scenario a precautionary approach (85% probability λ ≥ 1.0) to population stabilisation required releasing 3100 captive-bred females year −1 (131% of initial wild numbers). A precautionary approach to sustainable hunting of 100 females year −1 required releasing 3600 females year −1 (153% of initial wild numbers); but if interventions reduced flyway-scale hunting/trapping mortality by 60% or 80%, sustaining this quota required releasing 900 or 400 females year −1 , 38% and 17% of initial wild numbers, respectively. Parameter uncertainty increased precautionary numbers for release, but even with reduced precaution (50% probability λ ≥ 1.0), sustainable hunting of 100 females year −1 required annual releases of 2200 females (94% wild) without other measures, but 300 (13%) or no (0%) females under scenarios of a 60% or 80% reduction in flyway-scale hunting/trapping. Captive breeding cannot alone sustain migrant populations of wild C. macqueenii because it risks replacement and domestication. Trade and exploitation must be restricted to avoid either extinction or domestication. For exploited populations, supplementation by captive breeding should be used with caution.

Full text

Contents lists available at ScienceDirect
Biological Conservation
journal homepage: www.elsevier.com/locate/biocon
Captive breeding cannot sustain migratory Asian houbara Chlamydotis
macqueenii without hunting controls
P.M. Dolman
a,⁎
, N.J. Collar
b,c
, R.J. Burnside
a
a
School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK
b
BirdLife International, Pembroke Street, Cambridge CB2 3QZ, UK
c
School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK
ARTICLE INFO
Keywords:
Arab falconry
Central Asian flyway
Population reinforcement
Population supplementation
Sustainable exploitation
Sustainable hunting
ABSTRACT
To evaluate the potential contribution of captive breeding to the conservation of exploited migratory Asian
houbara Chlamydotis macqueenii, we estimated release numbers required to stabilise a population in a hunting
concession (14,300 km
2
), under scenarios of local licensed hunting and flyway-scale protection. We developed a
population model, initially 2350 adult females, re-sampling parameters measured through fieldwork and sa-
tellite telemetry, over 1000 iterations. With current flyway-scale unregulated harvest, and without any licensed
hunting in the concession, populations declined at 9.4% year
−1
(95% CI: −18.9 to 0% year
−1
); in this scenario a
precautionary approach (85% probability λ≥ 1.0) to population stabilisation required releasing 3100 captive-
bred females year
−1
(131% of initial wild numbers). A precautionary approach to sustainable hunting of
100 females year
−1
required releasing 3600 females year
−1
(153% of initial wild numbers); but if interventions
reduced flyway-scale hunting/trapping mortality by 60% or 80%, sustaining this quota required releasing 900 or
400 females year
−1
, 38% and 17% of initial wild numbers, respectively. Parameter uncertainty increased pre-
cautionary numbers for release, but even with reduced precaution (50% probability λ≥ 1.0), sustainable
hunting of 100 females year
−1
required annual releases of 2200 females (94% wild) without other measures, but
300 (13%) or no (0%) females under scenarios of a 60% or 80% reduction in flyway-scale hunting/trapping.
Captive breeding cannot alone sustain migrant populations of wild C. macqueenii because it risks replacement
and domestication. Trade and exploitation must be restricted to avoid either extinction or domestication. For
exploited populations, supplementation by captive breeding should be used with caution.
1. Introduction
Captive breeding is increasingly used to re-establish or reinforce
species' populations, but its effectiveness needs rigorous assessment
(Converse et al., 2013a;IUCN/SSC, 2013;Taylor et al., 2017). In-
escapable behavioural and genetic risks (Frankham, 2008;Snyder et al.,
1996) become acute where large-scale releases reinforce exploited wild
populations (Laikre et al., 2010). An understanding is therefore essen-
tial of the levels of release, relative to wild numbers, needed to render
harvest sustainable (Burnside et al., 2016).
Despite best practices to minimise loss of genetic diversity
(Frankham, 2008;Williams and Hoffman, 2009), long-term ex situ re-
gimes inevitably select across multiple traits (Snyder et al., 1996).
Traits can diverge rapidly (Christie et al., 2016) affecting fecundity and
linked traits (Bagliacca et al., 2004;Chargé et al., 2014;Lacy et al.,
2013), immune-genetics (Worley et al., 2010), digestive morphology
(Moore and Battley, 2006), temperament (Robertson et al., 2017), and
intrinsic and learnt behaviours including anti-predator responses
(Mathews et al., 2005;McPhee, 2004;Rantanen et al., 2010). Despite
higher mortality and lower productivity in captive-bred individuals,
post-release selection may not purge maladaptive traits, which can in-
trogress into wild populations (Ford, 2002;Sanchez-Donoso et al.,
2014;Söderquist et al., 2017), and wild-born descendants of captive-
bred parents may exhibit substantially lower fitness than those of wild
parents (Araki et al., 2009).
The primary cause of the Asian houbara Chlamydotis macqueenii
Gray being listed as globally threatened (IUCN status Vulnerable) is
hunting, which occurs across the semi-arid desert flyway linking China
and Central Asia to wintering grounds in Iran and Pakistan, and in-
volves Arab falconers and local communities (often using firearms)
compounded by live-trapping for falcon-training (BirdLife
International, 2018). Asian houbara has suffered population losses and
https://doi.org/10.1016/j.biocon.2018.10.001
Received 4 June 2018; Received in revised form 14 September 2018; Accepted 1 October 2018
⁎
Corresponding author.
E-mail addresses: P.Dolman@uea.ac.uk (P.M. Dolman), Nigel.Collar@birdlife.org (N.J. Collar), R.Burnside@uea.ac.uk (R.J. Burnside).
Biological Conservation xxx (xxxx) xxx–xxx
0006-3207/ Crown Copyright © 2018 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Please cite this article as: Dolman, P.M., Biological Conservation, https://doi.org/10.1016/j.biocon.2018.10.001

Table 1
Parameters of migratory Asian houbara Chlamydotis macqueenii population model, showing sample sizes. Breeding components (probability of nesting, clutch size,
nest success, hatching probability, chick survival to fledging) were measured using nest monitoring, and satellite and UHF tracking of adult females with broods.
Subsequent survival was measured using satellite telemetry. Distributions of clutch size and first-egg date were defined by mean and SD; other parameters by the
mean and SE.
Demographic parameter (units) N, details of replication Parameter estimate
Probability of breeding (starting incubation)
after spring return
52 returns of 23 satellite-tagged adult females to the
breeding grounds (2012–2017); mean estimated from
binomial GLM treating arrivals as independent events:
outcome 0/1, did-not-breed, bred.
0.981 ± 0.019 SE
First-egg date of first nesting attempt (Jday) 51 first-egg dates from 23 satellite-tagged adult females
returning to the breeding grounds (2012–2017); mean and
variance sampled across individuals and years.
92.4 Jday ± 6.8 SD
Clutch size in first nesting attempt (number
of eggs)
41 known first clutches (23 individuals) 2012–2017, from
Koshkin et al. (2016a) supplemented by this study.
3.32 eggs ± 0.65 SD
Clutch size reduction from first to second
attempt
38 re-nesting events (25 individuals) 2013–2017. − 0.56 eggs ± 0.15 SE (median = −1)
Date-dependent daily nest survival
probability, clutch.dsr
d
249 nests monitored over 3025 exposure-days over 6 years
(2012–2017); clutch.dsr
d
related to Jday; for details see Fig.
A.1 and Table A.1.
clutch.dsr
d
= 1 − inverse logit[5.604642 (SE 0.89) − 0.08789 (SE
0.039726) ∗(Jday − 68) + 0.000836 (SE 0.000424) ∗(Jday − 68)
2
]
Mean nest success (23 days incubation) = 52.3% ± 3.6% SE
Hatching probability egg
−1
(accounting for
partial predation, infertile eggs and
embryo death)
339 eggs laid in 110 successful nests (2012–2017) hatched
294 chicks; hatching probability estimated in a known-fate
binomial trials Generalised Linear Model (GLM).
0.855 ± 0.019 SE
Date-dependent re-nesting probability,
Renest
date
Outcome following 99 nest failures (with known failure
day) of 36 satellite-tagged females where it was known
whether subsequent incubation was initiated (65 re-
nesting events).
Renest
date
= Inverse logit[43.5389 (SE 9.43) − 4.0366 (SE
0.88) ∗√day of failure]
Mean = 0.657 ± 0.048 SE
Re-nesting interval (days) 65 paired nest failure and re-nesting events of satellite-
tagged females, with both dates known.
8.57 days ± 0.30 SE
Chick survival probability from hatching to
fledging (35 days), S
chick
40 broods (initially comprising 104 hatched chicks) from
24 satellite-tagged females that successfully hatched one
or more clutches during 2013–2017; broods relocated at
34–38 days.
0.500 ± 0.049 SE
Wild juvenile survival probability from
fledging to 01 October (mean 116 days),
S
juv.oct
37 satellite-tracked juveniles fitted (2013–2017) at
fledging (35 days) monitored to October. Survival from
fledging to October, S
juv.oct
, was estimated as:
S
juv.oct
=juv.dsr
fd (days)
where juv.dsr
fd
, is fledging-date-
specific daily survival probability and days the brood-
specific number of exposure days from fledging to 01
October; details in Fig. A.3 and Tables A.3, A.4.
juv.dsr
fd
= (1 − inverse logit[−19.59017 (SE 5.298) + 0.08341 (SE
0.031) ∗fd])
Mean = 0.9965 ± 0.0012 SE
S
juv.oct
= 0.7113 ± 0.0766 SE
Wild first-winter survival probability from 01
October to 31 March (183 days),
excluding hunting in the concession,
S
juv.wtr
27 juveniles satellite-tracked (2013–2017) beyond
October.
0.3704 ± 0.0929 SE
Wild adult summer (01 April to 01 October)
survival probability, S
adult.su
58 satellite-tracked wild adults (15 adult males, 35 adult
females, plus 8 adults originally caught as juveniles)
monitored for 135 over-summer periods.
0.9037 ± 0.0254 SE
Wild adult winter (01 October to 31 March)
survival probability, S
adult.wtr
52 satellite-tracked adults (both sexes) monitored for 123
over-winter periods.
0.8926 ± 0.0282 SE
Captive-bred survival probability from
release to 01 October, S
cbr.oct
Total of 75 captive-bred released individuals monitored by
satellite transmitters, 65 from Burnside et al. (2016) plus
10 released April 2015.
0.560 ± 0.057 SE
Captive-bred released, first-winter survival
probability from 01 October to 31 March,
S
cbr.wtr
42 captive-bred released individuals, alive at October,
monitored by satellite transmitters, 38 from Burnside et al.
(2016) plus 4 released April 2015.
0.231 ± 0.068 SE
Proportion of winter mortality attributed to
hunting/trapping (excluding licensed
hunting in the concession)
13 wild adult winter, 17 wild first-winter, 30 captive-bred
first-winter and 13 wild adult summer mortalities,
respectively, from (Burnside et al., in press).
0.538 ± 0.138 SE; 0.529 ± 0.121 SE; 0.233 ± 0.078 SE;
0.231 ± 0.117 SE, respectively
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
2

range contraction across its southern resident range in the Middle East,
Arabia and Pakistan, as well as substantial declines in its migratory
Central Asian populations (Tourenq et al., 2004). To date, the main
conservation response has been supplementation by large-scale releases
of captive-bred birds (Burnside et al., 2016;IFHC, 2016). In the closely
related African houbara C. undulata, however, selection in captivity has
affected a range of physiological, reproductive and behavioural traits
(Chargé et al., 2014), with unknown or unreported consequences.
Captive-bred C. macqueenii released in Uzbekistan migrated shorter
distances than wild counterparts (Burnside et al., 2017), but whether
selection for docility, changes to immune-genetics and loss of learnt
behaviours reduce fitness in the wild is unknown. Crucially, the release
levels required to stabilise residual exploited populations remain un-
studied.
Using a population model parameterised by extensive fieldwork and
satellite telemetry, we estimate the current trajectory, and the potential
for supplementation to stabilise it, in a population of migratory C.
macqueenii that is targeted for small-scale harvest within a large
hunting concession but is also exposed to exploitation elsewhere along
its flyway. We assess whether the wild population possesses a har-
vestable surplus or, conversely, needs captive-bred supplementation or
other mitigations within the flyway to stabilise it. We then calculate the
additional releases required to offset licensed hunting, first under cur-
rent conditions, then under management scenarios that limit hunting
and trapping on the flyway and wintering grounds.
2. Materials and methods
2.1. Study area and population
We studied a migratory Asian houbara population in 14,300 km
2
of
the southern Kyzylkum Desert, Bukhara province, Uzbekistan
(39.34–40.56°N 62.21–65.20°E). The area supports c. 0.14 male hou-
bara km
−2
(Koshkin et al., 2016b) and is managed as a hunting con-
cession, licensed by the State Committee for Nature Conservation. A
captive-breeding facility, founded with locally sourced birds, supports
annual releases (mean 364 year
−1
during 2011–2017). Limited hunting
(with falcons) occurs in October when, although released birds are still
present, some adults have already left the area, with numbers supple-
mented by migratory individuals from other breeding populations, in-
cluding Kazakhstan and China (Combreau et al., 2011). Thus, while our
modelling assumes a closed population, captive-bred individuals re-
leased to compensate for licensed harvest within a hunting concession
may bolster local numbers but not offset the depletion of populations
breeding elsewhere in the flyway.
2.2. Parameterisation of adult female re-nesting model
Breeding productivity, F, was estimated from an individual-based
re-nesting model (following Dolman et al., 2015) coded in R. Propensity
to nest and first clutch dates were determined from returning adult
satellite-tagged females (2013–2017). Clutch size, nest outcomes and
hatching rates were measured from wild nests (2012–2017) monitored
using protocols in Koshkin et al. (2016a). Incubating females were
confirmed as wild at 74% of nests, and since few remaining nesters
were likely to have been captive-bred (95% probability ≤5 nests: Ap-
pendix A) the likelihood of confounding effects was considered negli-
gible. Nests were treated as independent observations since nest pre-
dation is spatially unpredictable (Koshkin et al., 2016a) and most were
from different individuals. Nest success was date-dependent, declining
mid-season before partially recovering (Appendix A and Fig. A.1) in
response to the phenology of a key nest predator (Koshkin et al.,
2016a). Following nest failure, date-dependent re-nesting probability,
re-nesting interval and clutch size were determined from satellite-
tagged females (Appendix A and Fig. A.2; sample sizes in Table 1). As
negative effects on reproduction are largely associated with neck collars
(Bodey et al., 2017), breeding parameters derived using backpack-
mounted telemetry were not adjusted.
2.3. Age-specific survival
Chick survival (over 35 days) to fledging, S
chick
, was measured for
replicate broods of satellite-tracked adult females (some tracked over
multiple years: Table 1) that had hatched at least one egg. Early long-
distance flights indicated complete brood loss; other females were re-
located 34–38 days post-hatch and the numbers of surviving chicks
were recorded (details in Appendix A). S
chick
was estimated using a
binomial Generalised Linear Model (GLM) and was independent of
hatching date (Table A.2). Juvenile survival from fledging to October,
S
juv.oct
, was examined by satellite telemetry of individuals captured at
34–38 days (details in Appendix A). The fledging-date-specific daily
survival rate was estimated using a binomial trials GLM (which con-
siders the number of repeated trials, here exposure days to October,
leading to the outcome of each sample) and declined for later broods,
such that earlier-fledged juveniles were more likely to survive to Oc-
tober (Fig. A.3 and Tables A.3, A.4).
A binomial GLM was used to estimate subsequent first-winter sur-
vival of wild satellite-tagged juveniles that survived to October.
Survival of satellite-tagged wild adults (following protocols in Burnside
et al., 2017) considered two periods: winter (including outward and
return migration), and summer (including settlement, breeding, post-
breeding dispersal and moult). Survival of captive-bred houbara to 01
October was similar for yearlings released in spring (March–May) and
young-of-the-year released in August (Burnside et al., 2016). To ex-
amine underlying demography, when analysing survival probabilities
individuals killed by licensed hunting within the concession were coded
as surviving up to that date; hunting quotas were then applied in po-
pulation management scenarios (see below).
Transmitters may reduce avian survival, particularly when back-
pack-mounted on species with flapping flight, using tags that exceed 1%
of body mass, although mean effect size is small (Bodey et al., 2017). As
we cannot exclude satellite-tag effects on houbara survival, we adjusted
all transmitter-derived survival parameters (captive-bred released, wild
juvenile from fledging to October, first winter, adult summer and adult
winter) by the mean effect size on avian survival from meta-analysis
(Bodey et al., 2017). Mean effect sizes reported as the Fisher-trans-
formed correlation coefficient (Zr = −0.064) were converted to Co-
hen's d, a measure of the difference between means (d= −0.128),
following Nakagawa and Cuthill (2007). All mean survival parameters
were increased by this proportion of the parameter's standard error
(SE), as an estimate of the pooled SD of the original and adjusted
parameter.
2.4. Population model
Age-specific female cohorts were modelled, coded in R, assuming:
equal sex ratio at hatching (Dolman et al., 2015), breeding un-
constrained by male availability, and maximum age 20 years (Preston
et al., 2015). Table 1 summarises demographic parameters. The initial
distribution of age-classes was determined from a Leslie matrix model
using mean demographic parameters, which stabilised at 35 years. Po-
pulation trajectories over 20 generations were examined using a po-
pulation model, applied to an initial population of 2350 adult females,
estimated from spring male point count data (Appendix B). At each of
1000 iterations, each demographic parameter was randomly sampled:
for survival rates from a beta distribution bound by limits of 0 and 1,
defined by the mean and its SE; for clutch size and incubation start-date
from a normal distribution defined by the mean and observed standard
deviation (SD); and for other parameters from a normal distribution
defined by the mean and SE. Parameter estimates pooled data across
individuals and years, and iteration-specific resampling combined both
inter-annual variability and empirical sampling error. Individual
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
3

probabilistic events (probability of breeding, daily nest survival during
incubation, egg hatching probability, date-dependent re-nesting prob-
ability after clutch loss, and individual chick, juvenile and adult sur-
vival) were assessed as random binomial trials against the iteration-
specific parameter threshold.
As few one-year-old females attempt to breed (Azar et al., 2018;
Maloney, 2003), we assumed first breeding at two years. We further
assumed that subsequent breeding performance was independent of
age; however, as breeding propensity and success increase up to three
years (Bacon et al., 2017;Maloney, 2003) the contribution of second-
year birds to productivity was slightly over-estimated. Reproductive
senescence was not incorporated. For captive-bred C. undulata, re-
productive senescence is negligible, with only subtle reductions in chick
hatching by 18 years (Preston et al., 2015) and, for released birds, in
clutch size and egg volume after eight years (Bacon et al., 2017). Fur-
thermore, our monitoring of wild nests and broods sampled perfor-
mance across the current female age structure.
For computational efficiency, a single matrix of potential breeding
outcomes was populated by running the re-nesting model for 10,000
independent female breeding seasons. Within the re-nesting model,
individual probability of breeding was assessed and, if breeding, the
first-egg date of the first nesting attempt was determined by randomly
sampling the observed distribution (Table 1). Initial clutch size was
sampled from the distribution of known first attempts (rounded to an
integer, bounded 1–5), incremented by 1.5 days for each additional egg
(Saint Jalme and van Heezik, 1996) to determine incubation start-date.
As the primary cause of nest failure is predation (Koshkin et al., 2016a),
which is considered spatially unpredictable, nest success was sampled
independently for each attempt (rather than individual). A vector of
date-specific daily nest survival probabilities (clutch.dsr
d
) and their SE
was generated from a binomial trials model of monitored nests. For
each nesting attempt, the beta distribution of the first day's clutch.dsr
d
was sampled to determine a nest-attempt-specific adjustment that was
applied (as a Z score relative to the date-dependent SE) to each sub-
sequent clutch.dsr
d
that formed the basis of iterative daily binomial
trials, until the clutch failed or hatched. Females hatching eggs were
assumed not to re-nest during that season. Whether a female re-nested
after failure was determined by a binomial trial performed on the date-
dependent re-nesting probability, sampling the re-nesting interval to
determine the subsequent first-egg date, sampling the distribution of
known second clutches to determine the re-nesting clutch size, and then
modelling success as above. The re-nesting model yielded a matrix of
outcomes for each female-season.
The population model randomly sampled the matrix of nesting
outcomes for each breeding-age female in each breeding season. For
each successful nest, the matrix provided clutch size and hatching date;
the iteration-specific hatching probability was then applied (by bino-
mial trial for each egg) to determine chick numbers. Chick survival to
fledging was tested by an individual binomial trial, against the itera-
tion-specific value of S
chick
. For each juvenile, survival from fledging to
October was determined by an iteration-specific adjustment (by Z
score) of the relevant fledge-date-dependent juv.dsr
fd
, with daily bino-
mial trials repeated according to the number of exposure days to 01
October.
Following breeding, adult summer survival was tested by individual
binomial trials against the iteration-dependent probability, together
with surviving juveniles providing total numbers at the October hunt.
Individual over-winter survival to spring (31 March) was tested against
the iteration-specific, age-dependent probability and survivors then
incremented in age, providing spring breeding numbers.
2.5. Demography
Population trajectories were considered in terms of mean popula-
tion growth rate (λ, the linear regression of log population size across
years 1–20). The value of each demographic parameter (breeding
productivity, first-winter, adult winter and adult summer survival) at
which λ= 1 (holding other parameters at their mean) was predicted
from a series of GLMs (with normal error) which related iteration-
specific λto the iteration-specific parameter value.
2.6. Population management/mitigation scenarios
To examine the magnitude of releases required to stabilise the
exploited population within the hunting concession, we considered
scenarios of releasing between 0 and 4000 captive-bred female hou-
bara year
−1
, under different levels of regulated hunting, ranging from 0
to 200 females year
−1
. For captive-bred C. undulata, susceptibility to
hunting did not vary between sexes or with time since release
(Hardouin et al., 2015). We treated wild adult, wild juvenile and cap-
tive-bred released females as equally susceptible to hunting, with their
mortality dependent on the ratio of the hunting quota to total female
numbers alive in October. Subsequent individual winter survival was
tested by a binomial trial against the iteration-specific, cohort-specific
survival probabilities and survivors incremented in age next spring. We
assumed (perhaps generously: see Discussion) that, after their first year,
captive-bred released birds performed as wild birds.
For each release and hunting scenario, we examined release num-
bers required to stabilise the population (λ≥ 1). We examined two
scenarios: an ‘average approach’ based on the mean λ, thus giving a
50% probability of stability or increase, and a more rigorous ‘precau-
tionary approach’ based on the lower 15% confidence interval of the
mean λ(85% probability λ≥ 1).
Despite extensive ecological study (Koshkin et al., 2016a;Koshkin
et al., 2016b;Koshkin et al., 2014), no opportunity exists to enhance in
situ breeding productivity to offset mortality. We reject predator con-
trol across breeding landscapes because: (1) the key predator, Varanus
griseus, is vulnerable and protected (Anon, 2009); (2) poison-bait is
indiscriminate; (3) disrupting predator/prey dynamics risks unforeseen
consequences; (4) predator reduction may affect selection on wild
houbara traits (e.g. Guilherme et al., 2018); and (5) no evidence exists
that anthropogenic factors inflate predator numbers in areas remote
from agriculture or settlements.
Actions to limit illegal trapping and sustainably regulate hunting
have long been advocated as conservation measures (CMS, 2005;
Combreau et al., 2005) and may reduce dependence on captive-bred
supplementation. We therefore examined the demographic con-
sequences of scenarios that reduced anthropogenic mortality. The
proportion of mortality in each cohort (wild adults in summer, wild
adults in winter, wild juvenile first-winter and captive-bred birds first-
winter) attributable to hunting/trapping, was estimated from PTT lo-
cation and engineering data (Burnside et al., in press). A similar pro-
portion of wild adult winter and wild first-winter mortalities was at-
tributed to hunting/trapping, at 53% (Table 1), less than an earlier
estimate of 74% of over-winter mortality attributable to hunting for
adult-sized migrant Asian houbara (Combreau et al., 2001). Hunting/
trapping pressure may have decreased, or alternatively more hunters or
trappers may now simply remove and discard rather than destroy a
PTT, in which case anthropogenic mortality will be underestimated. We
therefore consider our estimates of the proportion of total winter
mortality attributed to hunting/trapping, and thus the extent to which
limiting this offers a reduction in captive-bred numbers needed for re-
lease, as conservative. Fewer (23%) captive-bred first-winter mor-
talities were attributed to hunting/trapping (Burnside et al., in press).
Of wild adult summer mortalities, 23% were attributed to hunting/
trapping (Burnside et al., in press). Using this information, we in-
crementally reduced hunting/trapping mortality according to:
= +S S mort red P S. . ( )·(1 )
j 2 j 1 j 1 j 1. . . .
(1)
where S
j.1
and S
j.2
are the original and improved winter or summer
survival probabilities of cohort j,mort.red the proportionate reduction
in hunting/trapping (by 0.1 increments) and P
j.1
the proportion of
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
4

mortality attributable to hunting/trapping.
3. Results
3.1. Demographic performance
3.1.1. Productivity
Nesting propensity was high, and nest success relatively high (52%
overall), with most eggs in successful nests hatching (Table 1). Mean
annual productivity, estimated from the re-nesting model for a starting
population of 2056 breeding-age (≥2 years) females and applying date-
independent chick mortality and date-dependent juvenile mortality,
was 0.334 ± 0.037 SD female recruits in October per breeding-age
female (Table A.5).
3.1.2. First-winter and adult survival
The mean number of wild female yearling recruits returning the
following year was 0.132 ± 0.016 SD (breeding-age female)
−1
(Table
A.5). First-winter survival of released captive-bred houbara did not
differ (t
67
= 1.235, P = 0.221) from that of wild first-winter birds
(Table 1). Adult annual survival (excluding licensed hunting within the
concession; estimated for 58 individuals over 132 individual-years
2011–2018) was 0.8030 ± 0.0346 SE, and similar between winter and
summer periods (Table 1).
3.1.3. Population trend
The wild population, examined in the absence of licensed hunting in
the concession or release of captive-bred birds (Fig. 1), was estimated to
be declining at 9.4% year
−1
(mean λ, 0.906, 95% CI 0.811–1.00), with
only 2.7% of population iterations increasing. The trajectory was si-
milar without adjustment of survival parameters to compensate for
potential transmitter effects (mean λ= 0.904, 0.818–0.985 CI; 0.9%
iterations increasing). To achieve population stability (λ= 1.0) re-
quired large increases in demographic parameters: for productivity, to
0.599 (95% CI 0.583–0.616) female recruits in October per breeding-
age female, substantially beyond the upper quartile of current pro-
ductivity (Fig. 2a); for first-winter or adult winter survival, to 0.70
(95% CI 0.69–0.72) and 0.996 (95% CI 0.992–1.0) respectively, again
substantially beyond their current upper quartiles (Fig. 2b, c). Stability
could not be achieved solely through increased summer survival
(to > 1.0) (Fig. 2d).
3.2. Scenarios for population stabilisation and creating a huntable surplus
With no licensed hunting occurring in the concession, stabilising
population trajectories required, on average (50% probability λ≥ 1.0),
annual releases of 1700 females (Fig. 3a), 72% of the initial wild female
numbers. Under this ‘average approach’, compensating for an annual
hunting quota of 100 females within the concession would require
annually releasing 2200 females, approaching (94%) initial wild num-
bers, while a quota of 200 females required annually releasing 2700
females, exceeding (115%) wild numbers. However, a precautionary
approach (85% probability λ≥ 1.0) required annually releasing 3100
females (132%) with no hunting, and 3600 (153%) or > 4000
(> 170%) females for sustainable annual quotas of 100 or 200 females
respectively. Total numbers of houbara hunted or released would, as-
suming an equal sex ratio and no differential hunting susceptibility, be
double these levels.
In scenarios where the hunting/trapping component of winter
mortality was reduced by 30%, a quota of 100 females could, on
average, be sustained by annually releasing 1100 females, half those
required without wider regulation but still 47% relative to wild num-
bers (Fig. 3b); or under a precautionary approach by annually releasing
2200 females (94% wild numbers). If winter hunting/trapping mor-
tality were reduced by 60%, a quota of 100 females could, on average,
be sustained by annually releasing 500 females (21%), while a pre-
cautionary approach required annually releasing 1200 females (33% of
those required without wider regulation but still 51% of wild numbers).
Concerted action to reduce winter hunting/trapping mortality by 80%
would allow sustainable hunting of 100 females by annually releasing
only 200 (on average) or 700 (precautionary) females, 9% and 30% of
initial wild numbers respectively.
If both summer and winter hunting/trapping mortality were re-
duced similarly by 30%, a quota of 100 females could be sustained by
annually releasing 900 (on average) or 1900 (precautionary) females
(Fig. 3c), 38% and 81% of initial wild numbers respectively. If annual
hunting/trapping mortality were reduced by 60%, a quota of 100 fe-
males could be sustained by annually releasing 300 (on average, 13%)
or 900 (precautionary, 38%) females. However, if annual hunting/
trapping mortality were reduced by 80% a sustainable hunt would re-
quire no (0%) or 400 (17%) females under average and precautionary
scenarios, respectively.
4. Discussion
Population modelling indicated the C. macqueenii population in the
Kyzylkum of Uzbekistan is declining by 9.4% year
−1
(95% CI:
−18.9% year
−1
to 0% year
−1
), despite strong breeding productivity,
due to unsustainable mortality. With no harvestable surplus, population
stabilisation required a reduction in mortality and/or the release of
captive-bred birds; licensed hunting within the concession then re-
quired additional releases. However, if captive breeding and release
was the sole conservation measure, under a precautionary approach
(85% probability of stabilising the population) numbers that must be
released annually exceed the wild population: 132% of initial wild
numbers in the absence of licensed hunting within the concession;
153% or > 170% with a modest hunting quota of 100 or 200 females
respectively (200 or 400 houbara). Unless action to regulate hunting
and limit trapping across the flyway was extremely effective (e.g. re-
ducing annual hunting/trapping mortality by 80%), annual numbers of
captive-bred releases required for population stabilisation remained
substantial relative to wild numbers.
Fig. 1. Demography of Asian houbara Chlamydotis macqueenii populations,
showing the number of females in spring (yearlings plus adults) over 20 years,
and the geometric mean trajectory (continuous line) and SD (dashed lines)
across 500 model iterations (pale lines).
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
5

4.1. Wild performance
Nesting propensity of satellite-tagged females was high (98%),
while overall nest success (52%) and hatching probability (86%) were
similar to those of wild houbara in China (59% and 84% respectively,
Combreau et al., 2002). Wild chick survival probability to five weeks
(50%) exceeded that of wild chicks in China (39% to eight weeks,
Combreau et al., 2002), and wild juvenile survival was high from
fledging to 01 October (71%). Thus predicted population declines
cannot be attributed to poor productivity, and population stabilisation
would require an unfeasible increase above current levels of recruit-
ment.
Wild adult annual mortality (19.7%) was below the 28.3%
(18.6–40.4% CI) for adult-sized houbara (possibly including some first-
winter birds) migrating between China and Pakistan, 1994–2000, when
mortality was 11.2 times higher in winter than the breeding season
(Combreau et al., 2001). Although wild first-winter mortality (63%)
was greater, the proportion attributable to hunting/trapping was si-
milar to that of wild adults. However, in the absence of hunting/
trapping mortality, wild first-winter mortality may be < 45%. Although
houbara remain at risk on migration and wintering grounds in Pakistan,
Afghanistan and Iran, adult summer mortality (including unlicensed
hunting and powerline collisions: Burnside et al., in press;Burnside
et al., 2015) was similar to that in winter. Conservation measures
should therefore address threats of infrastructure and hunting on
breeding grounds as well as the widely recognised threat (CMS, 2005;
Combreau et al., 2005) of unsustainable harvest in wintering areas.
We consider these results robust, although further parameter re-
finement may narrow confidence limits and improve predictions. In two
aspects, our modelling was conservative and thus may have under-es-
timated numbers of releases needed to stabilise populations. First, it
will have slightly over-estimated productivity by not allowing for lower
output of two-year-old females and senescent birds. Second, it assumed
that, after surviving their first winter, captive-bred birds performed as
well as wild adults (see below). Conversely, any transmitter impacts
may over-estimate numbers of releases required. However, we adjusted
survival estimates upward by the mean transmitter effect from recent
meta-analysis (Bodey et al., 2017). Moreover, given that transmitter
Fig. 2. Mean rate of wild Asian houbara Chlamydotis macqueenii population change (λ) across 5000 model iterations, related to (a) productivity: female recruits in
October per breeding-age female; (b) wild first-winter survival; (c) adult over-winter survival and (d) adult over-summer survival. Vertical lines show median (solid
line, x axis value) and inter-quartile range (shaded, dashed lines) of parameters re-sampled across iterations, and the fit (red solid line) and 95% CI (red dashed lines)
of a GLM relating λ to recruitment or survival. The vertical dashed line and x axis value show the parameter value for which λ =1.0 (horizontal dashed line). (For
interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
6

0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
1.05
1.1
1.15
1.2
0 (0)
300 (13)
600 (25)
900 (38)
1200 (51)
1500 (64)
1800 (76)
2100 (89)
2400 (102)
2700 (115)
3000 (127)
3300 (140)
3600 (153)
3900 (166)
Hunt quota
(females)
0
50
100
150
200
lambda
Number of captive-bred females released (% of wild population)
(a)
0 (0)
300 (13)
600 (25)
900 (38)
1200 (51)
1500 (64)
1800 (76)
2100 (89)
2400 (102)
2700 (115)
3000 (127)
3300 (140)
3600 (153)
3900 (166)
Winter
hunting/trapping
reduction
(quota 100 females)
0
0.2
0.4
0.6
0.8
1
lambda
(b)
0 (0)
300 (13)
600 (25)
900 (38)
1200 (51)
1500 (64)
1800 (76)
2100 (89)
2400 (102)
2700 (115)
3000 (127)
3300 (140)
3600 (153)
3900 (166)
Summer & Winter
hunting/trapping
reduction
(quota 100 females)
0
0.2
0.4
0.6
0.8
1
lambda
(c)
Fig. 3. Mean and 95% CI of population growth rate (λ), across 1000 iterations for each release/hunt scenario, shown in relation to (a) numbers of female captive-bred houbara released year
−1
for different annual hunting
quotas (numbers of females hunted), and (b, c), maintaining an annual quota of 100 females, but with proportionate reductions (0 to 1.0) in (b) adult and first-winter hunting/trapping mortality, and in (c) adult winter,
first-winter and adult summer hunting/trapping mortality. Assuming an equal sex ratio, total numbers released and hunted will be twice the values shown.
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
7

effects are more detrimental for nesting propensity than survival
(Barron et al., 2010), the high (98%) nesting propensity of returning
satellite-tagged adult females suggests transmitters minimally affected
condition. Indeed, where protocols have been tested and expertly
adapted to bird species, satellite transmitters commonly have minimal
effects (e.g. Ashbrook et al., 2016;Sergio et al., 2015). Nevertheless, a
concern is whether transmitters affected juvenile or first-winter survival
if not properly fitted to allow for growth. Accepting that development,
growth and thus harness adjustments may differ for active wild birds, 4
captive-bred juveniles, fitted with transmitters at c.35 days using
identical protocols to tagged wild juveniles, and kept over winter to the
following spring, showed normal growth and mobility, no lesions and
no signs of musculoskeletal problems.
4.2. Captive-bred released performance
Initial survival of captive-bred houbara from release to October was
higher (56%) than for houbara bred in the UAE and released in Saudi
Arabia, Kazakhstan and Pakistan (to 3 months: 25%, 25%, 29% re-
spectively, IFHC, 2017). Houbara released in Central Asia face chal-
lenges in their first migration, through exposure to hunting and trap-
ping. Overall first-year survival was, unsurprisingly, lower (12.9% from
release to spring return) than that of non-migratory captive-bred hou-
bara released into protected areas in the UAE (48% for their first year,
Azar et al., 2016) and Morocco (44%, IFHC, 2016).
First-winter captive-bred birds were more susceptible to natural
mortality from starvation, disease or predation than wild juveniles
(Burnside et al., in press), potentially indicating poorer fitness in terms
of physiological condition, immunity to pathogens, anti-predator be-
haviours or foraging ability. Our estimates of captive-bred survival
were derived from cohorts reared in earlier years of the facility; sub-
sequent refinement of diet and rearing protocols may improve perfor-
mance. If post-release survival were improved by rearing and pre-re-
lease protocols such as exercise (Rymešová et al., 2013), or predator-
aversion training (Gaudioso et al., 2011;van Heezik et al., 1999),
numbers needed for release could perhaps be reduced.
Uncertainty in long-term demographic performance of released in-
dividuals is common when planning reintroduction or reinforcement
programmes (e.g. Converse et al., 2013b;Dolman et al., 2015). As noted
above, a critical uncertainty in our model concerns the subsequent
demographic performance of surviving captive-bred released birds that
returned to the breeding grounds. We assumed this matched that of
wild birds and, additionally, that their wild-hatched progeny were in-
distinguishable from wild individuals. However, even after reaching
3+ years, annual survival of captive-bred C. undulata (67%) is sub-
stantially below that of wild adults (89%, IFHC, 2016), suggesting
persistent effects of long-established captive-breeding programmes on
fitness. We lack robust data on returning captive-bred C. macqueenii due
to high first-winter mortality rates, but initial data suggest annual
survival (12 individuals, over 20 individual-years, 0.40 ± 0.1096 SE)
may be substantially less than for wild adults (0.8030 ± 0.0346). More
generally, captive-reared gamebirds released to reinforce wild popula-
tions commonly achieve significantly poorer productivity and survival
than wild counterparts (Buner et al., 2011;Hill and Robertson, 1988;
Madden et al., 2018;Rymešová et al., 2013). The conventional response
to this poorer performance is to release larger numbers of birds
(Robertson et al., 2017;Söderquist et al., 2017), the unintended con-
sequences of which we have reviewed in the Introduction.
4.3. Conservation of migratory Asian houbara
For migratory Asian houbara, the sheer scale of captive breeding
and release needed to stabilise exploited populations—approaching or
annually exceeding initial wild numbers—could itself become a desta-
bilising and unsustainable measure, owing to the known and pre-
dictable dangers from various adverse heritable or acquired traits (see
Introduction). Such scaling-up of supplementation would risk replace-
ment and domestication of wild stock rather than conservation re-
inforcement, with unknown population consequences. While our
modelling assumed a closed population, localised compensatory release
of captive-bred individuals may supplement local numbers but will not
offset hunting pressure on other populations during their migratory
stop-over in the hunting concession. Moreover, a flyway-wide pro-
gramme of captive breeding and release, currently under development,
poses even greater risks of wild replacement. Only by substantially
reducing mortality along the flyways can the decline of C. macqueenii be
halted and releases of captive-bred birds moderated to levels at which
the dangers to wild populations are minimised.
Conservation success can, therefore, only be achieved by interested
parties committing jointly to a scientifically informed, strictly observed
and fully integrated system of legal enforcement, quotas, concessions
and supplementations (CMS, 2005) to reverse the current downward
trajectory of the wild population and allow truly sustainable hunting.
Extending the supply of captive-bred birds for falcon training has the
potential to reduce markets for wild-caught houbara (Combreau et al.,
2005), although losses from powerline collisions (Burnside et al., 2015)
may be costly to mitigate. Considerable resources are already invested
by Gulf states in the captive breeding of houbara and in falcons, lo-
gistics for falconry and hunting expeditions, and the social and eco-
nomic support provided to host countries. Given the level of this in-
vestment, the move to a sustainably managed hunting regime should
represent not a threat to the deep tradition of falconry but the only
means of ensuring its long-term survival. Host countries and con-
servation agencies will surely find common cause in helping develop
and implement this now imperative strategy.
5. Conclusions
This study adds to the mounting evidence of potential long-term
risks to wild populations from supplementation and reinforcement
through the release of captive-bred animals. Where the fitness reduc-
tion of captive-bred relative to wild individuals is heritable, as has been
frequently demonstrated (e.g. Araki et al., 2009;Christie et al., 2016),
release of even slightly less-fit individuals may cause long-term reduc-
tion in the size, genetic diversity and viability of wild populations
(Lynch and O'Hely, 2001;Willoughby and Christie, 2018), particularly
where captive-born individuals are released for more than a few years
(Willoughby and Christie, 2018). The use of repeated annual supple-
mentation to reinforce exploited populations over the long term is,
therefore, a high-risk strategy. For hunted populations, a further un-
intended consequence of releasing captive-bred stock is that, instead of
reducing pressure on wild individuals through a dilution effect, per-
ceptions of high abundance and effective mitigation can lead to greater
hunting pressure on wild stock (Casas et al., 2016). Although it is im-
portant to develop strategies to reduce genetic adaptation in captivity
and to increase survival and performance after release, without ade-
quately addressing in situ threats, over-reliance on supplementation
risks domestication and replacement.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.biocon.2018.10.001.
Declaration of interest
The Ahmed bin Zayed Charitable Foundation funded this work as
part of a wider research collaboration of the Emirates Bird Breeding
Center for Conservation (EBBCC), BirdLife International and University
of East Anglia (UEA), UK. BirdLife sub-contracted UEA to undertake
independent and objective evaluation of strategies using fieldwork and
satellite telemetry; all analyses were conducted independently by UEA,
and other partners had no influence on the magnitude or direction of
findings. BirdLife co-authored the paper; EBBCC was informed of results
and discussed interpretation. The animal research was approved by the
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
8

UEA Animal Welfare and Ethical Review Body (AWERB), which is
governed by UK Government Home Office (2014) guidance on the
operation of the Animals Scientific Procedures Act (ASPA 1986; HMSO:
London) and the European Commission (2010) Directive 2010/63/EU
of the European Parliament and of the Council of 22 September 2010 on
the protection of animals used for scientific purposes (Official Journal
of the European Union L276:33-79).
Acknowledgements
We thank the State Committee for Nature Conservation of the
Republic of Uzbekistan for permission and the Emirates Bird Breeding
Center for Conservation for collaboration and logistical support. We
warmly thank Marta Acácio, Alex Brighten, Elizabeth Grayshon, João
Guilherme, David Hodkinson, Adam Formica, Ryan Irvine, Erasil
Khaitov, Maxim Koshkin, Judit Mateos-Herrero, Christopher Panter,
Philip Saunders, David Showler, Andrew Taylor, Elizabeth Taylor,
Vladimir Terentyev and Jonathan Willans for their contributions to data
collection. We are grateful to Doug Armstrong and two anonymous
reviewers for comments that improved an earlier version of the paper.
References
Aebischer, N.J., 1999. Multi-way comparisons and generalized linear models of nest
success: extensions of the Mayfield method. Bird Study 46, 22–31.
Anon, 2009. The Red Data Book of the Republic of Uzbekistan Vol. II, Animals. Chinor
ENK, Tashkent.
Araki, H., Cooper, B., Blouin, M.S., 2009. Carry-over effect of captive breeding reduces
reproductive fitness of wild-born descendants in the wild. Biol. Lett. 5, 621–624.
Ashbrook, K., Taylor, A., Jane, L., Carter, I., Székely, T., 2016. Impacts of survival and
reproductive success on the long-term population viability of reintroduced great
bustards Otis tarda in the UK. Oryx 50, 583–592.
Azar, J.F., Rautureau, P., Lawrence, M., Calabuig, G., Hingrat, Y., 2016. Survival of re-
introduced Asian houbara in United Arab Emirates' reserves. J. Wildl. Manag. 80,
1031–1039.
Azar, J.F., Chalah, T., Rautureau, P., Lawrence, M., Hingrat, Y., 2018. Breeding success
and juvenile survival in a reintroduced captive-bred population of Asian houbara
bustards in the United Arab Emirates. Endanger. Species Res. 35, 59–70.
Bacon, L., Hingrat, Y., Robert, A., 2017. Evidence of reproductive senescence of released
individuals in a reinforced bird population. Biol. Conserv. 215, 288–295.
Bagliacca, M., Profumo, A., Ambrogi, C., Leotta, R., Paci, G., 2004. Egg-laying differences
in two grey partridge (Perdix perdix L.) lines subject to different breeding technology:
artificial egg hatch or mother egg hatch. Eur. J. Wildl. Res. 50, 133–136.
Barron, D.G., Brawn, J.D., Weatherhead, P.J., 2010. Meta-analysis of transmitter effects
on avian behaviour and ecology. Methods Ecol. Evol. 1, 180–187.
BirdLife International, 2018. Species factsheet: Chlamydotis macqueenii. Downloaded
from. http://www.birdlife.org (on 01/04/2018).
Bodey, T.W., Cleasby, I.R., Bell, F., Parr, N., Schultz, A., Votier, S.C., Bearhop, S., 2017. A
phylogenetically controlled meta-analysis of biologging device effects on birds: de-
leterious effects and a call for more standardized reporting of study data. Methods
Ecol. Evol. 9, 946–955.
Buner, F.D., Browne, S.J., Aebischer, N.J., 2011. Experimental assessment of release
methods for the re-establishment of a red-listed galliform, the grey partridge (Perdix
perdix). Biol. Conserv. 144, 593–601.
Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference. A
Practical Information Theoretical Approach, 2nd Edition. Springer, New York.
Burnside, R.J., Collar, N.J., Koshkin, M.A., Dolman, P.M., 2015. Avian powerline mor-
talities, including Asian houbara Chlamydotis macqueenii, on the Central Asian flyway
in Uzbekistan. Sandgrouse 37, 161–168.
Burnside, R.J., Collar, N.J., Scotland, K.M., Dolman, P.M., 2016. Survival rates of captive-
bred Asian houbara Chlamydotis macqueenii in a hunted migratory population. Ibis
158, 353–361.
Burnside, R.J., Collar, N.J., Dolman, P.M., 2017. Comparative migration strategies of wild
and captive-bred Asian houbara Chlamydotis macqueenii. Ibis 159, 374–389.
Burnside, R.J., Collar, N.J., Dolman, P.M., 2018. Mortality attributable to hunting,
trapping and powerlines in wild and captive-bred migratory Asian houbara
Chlamydotis macqueenii. (in press).
Casas, F., Arroyo, B., Viñuela, J., Guzmán, J.L., Mougeot, F., 2016. Are farm-reared red-
legged partridge releases increasing hunting pressure on wild breeding partridges in
central Spain? Eur. J. Wildl. Res. 62, 79–84.
Chargé, R., Sorci, G., Saint Jalme, M., Lesobre, L., Hingrat, Y., Lacroix, F., Teplitsky, C.,
2014. Does recognized genetic management in supportive breeding prevent genetic
changes in life-history traits? Evol. Appl. 7, 521–532.
Christie, M.R., Marine, M.L., Fox, S.E., French, R.A., Blouin, M.S., 2016. A single gen-
eration of domestication heritably alters the expression of hundreds of genes. Nat.
Commun. 7, 10676.
CMS, 2005. Meeting to Conclude the Agreement on the Conservation of the Asian
Houbara Bustard (Chlamydotis undulata macqueenii). UNEP/CMS/HB/5. Sixth
Revision December 2004: Annex 2 Draft Action Plan. UNEP, CMS (Convention on the
Conservation of Migratory Species of Wild Animals), Nairobi.
Combreau, O., Launay, F., Lawrence, M., 2001. An assessment of annual mortality rates in
adult-sized migrant houbara bustards (Chlamydotis [undulata]macqueenii). Anim.
Conserv. 4, 133–141.
Combreau, O., Qiao, J., Lawrence, M., Gao, X., Yao, J., Yang, W., Launay, F., 2002.
Breeding success in a houbara bustard Chlamydotis [undulata]macqueenii population
on the eastern fringe of the Jungar Basin, People's Republic of China. Ibis 144 (on-
line), E45–E56.
Combreau, O., Launay, F., Lawrence, M., 2005. Progress, challenges and perspective in
houbara bustard conservation in Asia. In: Bota, G., Morales, M.B., Mañosa, S.,
Camprodon, J. (Eds.), Ecology and Conservation of Steppe-land Birds. Lynx Editions,
Barcelona, pp. 307–325.
Combreau, O., Riou, S., Judas, J., Lawrence, M., Launay, F., 2011. Migratory pathways
and connectivity in Asian houbara bustards: evidence from 15 years of satellite
tracking. PLoS One 6, e20570.
Converse, S.J., Moore, C.T., Armstrong, D.P., 2013a. Demographics of reintroduced po-
pulations: estimation, modeling, and decision analysis. J. Wildl. Manag. 77,
1081–1093.
Converse, S.J., Moore, C.T., Folk, M.J., Runge, M.C., 2013b. A matter of tradeoffs: re-
introduction as a multiple objective decision. J. Wildl. Manag. 77, 1145–1156.
Dolman, P.M., Collar, N.J., Scotland, K.M., Burnside, R.J., 2015. Ark or park: the need to
predict relative effectiveness of ex situ and in situ conservation before attempting
captive breeding. J. Appl. Ecol. 52, 841–850.
Ford, M.J., 2002. Selection in captivity during supportive breeding may reduce fitness in
the wild. Conserv. Biol. 16, 815–825.
Frankham, R., 2008. Genetic adaptation to captivity in species conservation programs.
Mol. Ecol. 17, 325–333.
Gaudioso, V.R., Sánchez-García, C., Pérez, J.A., Rodríguez, P.L., Armenteros, J.A., Alonso,
M.E., 2011. Does early antipredator training increase the suitability of captive red-
legged partridges (Alectoris rufa) for releasing? Poult. Sci. 90, 1900–1908.
Guilherme, J.L., Burnside, R.J., Collar, N.J., Dolman, P.M., 2018. Consistent nest-site
selection across habitats increases fitness of Asian houbara. Auk 135, 192–205.
Hardouin, L.A., Legagneux, P., Hingrat, Y., Robert, A., 2015. Sex-specific dispersal re-
sponses to inbreeding and kinship. Anim. Behav. 105, 1–10.
van Heezik, Y., Seddon, P.J., Maloney, R.F., 1999. Helping reintroduced houbara bustards
avoid predation: effective anti-predator training and the predictive value of pre-re-
lease behaviour. Anim. Conserv. 2, 155–163.
Hill, D., Robertson, P., 1988. Breeding success of wild and hand-reared ring-necked
pheasants. J. Wildl. Manag. 52, 446–450.
IFHC, 2016. International Fund for Houbara Conservation Annual Report 2015–2016.
IFHC, Abu Dhabi.
IFHC, 2017. Internation Fund for Houbara Conservation Annual Report 2016–2017.
IFHC, Abu Dhabi.
IUCN/SSC, 2013. Guidelines for Reintroductions and Other Conservation Translocations.
Version 1.0. IUCN Species Survival Commission, Gland, Switzerland.
Koshkin, M.A., Collar, N.J., Dolman, P.M., 2014. Do sheep affect distribution and habitat
of Asian houbara Chlamydotis macqueenii? J. Arid Environ. 103, 53–62.
Koshkin, M., Burnside, R.J., Packman, C.E., Collar, N.J., Dolman, P.M., 2016a. Effects of
habitat and livestock on nest productivity of the Asian houbara Chlamydotis mac-
queenii in Bukhara Province, Uzbekistan. Eur. J. Wildl. Res. 62, 447–459.
Koshkin, M.A., Burnside, R.J., Collar, N.J., Guilherme, J.L., Showler, D.A., Dolman, P.M.,
2016b. Effects of habitat and land use on breeding season density of male Asian
houbara Chlamydotis macqueenii. J. Ornithol. 157, 811–823.
Lacy, R.C., Alaks, G., Walsh, A., 2013. Evolution of Peromyscus leucopus mice in response
to a captive environment. PLoS One 8, e72452.
Laikre, L., Schwartz, M.K., Waples, R.S., Ryman, N., 2010. Compromising genetic di-
versity in the wild: unmonitored large-scale release of plants and animals. Trends
Ecol. Evol. 25, 520–529.
Lynch, M., O'Hely, M., 2001. Captive breeding and the genetic fitness of natural popu-
lations. Conserv. Genet. 2, 363–378.
Madden, J.R., Hall, A., Whiteside, M.A., 2018. Why do many pheasants released in the UK
die, and how can we best reduce their natural mortality? Eur. J. Wildl. Res. 64.
Maloney, R.F., 2003. Survival, Breeding and Movements of Reintroduced Asiatic Houbara
(Chlamydotis [undulata]macqueenii) in Mahazat As-Sayd Reserve, Saudi Arabia (PhD
thesis). University of New England, New Zealand.
Mathews, F., Orros, M., McLaren, G., Gelling, M., Foster, R., 2005. Keeping fit on the ark:
assessing the suitability of captive-bred animals for release. Biol. Conserv. 121,
569–577.
McPhee, M.E., 2004. Generations in captivity increases behavioral variance: considera-
tions for captive breeding and reintroduction programs. Biol. Conserv. 115, 71–77.
Moore, S.J., Battley, P.F., 2006. Differences in the digestive organ morphology of captive
and wild Brown Teal Anas chlorotis and implications for releases. Bird Conserv. Int.
16, 253–264.
Nakagawa, S., Cuthill, I.C., 2007. Effect size, confidence interval and statistical sig-
nificance: a practical guide for biologists. Biol. Rev. 82, 591–605.
Preston, B.T., Saint Jalme, M., Hingrat, Y., Lacroix, F., Sorci, G., 2015. The sperm of aging
male bustards retards their offspring's development. Nat. Commun. 6, e6146.
Rantanen, E.M., Buner, F., Riordan, P., Sotherton, N., Macdonald, D.W., 2010. Vigilance,
time budgets and predation risk in reintroduced captive-bred grey partridges Perdix
perdix. Appl. Anim. Behav. Sci. 127, 43–50.
Robertson, P.A., Mill, A.C., Rushton, S.P., McKenzie, A.J., Sage, R.B., Aebischer, N.J.,
2017. Pheasant release in Great Britain: long-term and large-scale changes in the
survival of a managed bird. Eur. J. Wildl. Res. 63, 100.
Rymešová, D., Tomášek, O., Šálek, M., 2013. Differences in mortality rates, dispersal
distances and breeding success of commercially reared and wild grey partridges in
the Czech agricultural landscape. Eur. J. Wildl. Res. 59, 147–158.
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
9

Saint Jalme, M., van Heezik, Y., 1996. Propogation of the Houbara Bustard. Kegan Paul
International, London.
Sanchez-Donoso, I., Rodríguez-Teijeiro, J.D., Quintanilla, I., Jiménez-Blasco, I., Sardà-
Palomera, F., Nadal, J., Puigcerver, M., Vilà, C., 2014. Influence of game restocking
on the migratory behaviour of the common quail, Coturnix coturnix. Evol. Ecol. Res.
16, 493–504.
Sergio, F., Tavecchia, G., Tanferna, A., López Jiménez, L., Blas, J., De Stephanis, R.,
Marchant, T.A., Kumar, N., Hiraldo, F., 2015. No effect of satellite tagging on sur-
vival, recruitment, longevity, productivity and social dominance of a raptor, and the
provisioning and condition of its offspring. J. Appl. Ecol. 52, 1665–1675.
Snyder, N.F.R., Derrickson, S.R., Beissinger, S.R., Wiley, J.W., Smith, T.B., Toone, W.D.,
Miller, B., 1996. Limitations of captive breeding in endangered species recovery.
Conserv. Biol. 10, 338–348.
Söderquist, P., Elmberg, J., Gunnarsson, G., Thulin, C.G., Champagnon, J., Guillemain,
M., Kreisinger, J., Prins, H.H.T., Crooijmans, R.P.M.A., Kraus, R.H.S., 2017.
Admixture between released and wild game birds: a changing genetic landscape in
European mallards (Anas platyrhynchos). Eur. J. Wildl. Res. 63, 98.
Taylor, G., Canessa, S., Clarke, R.H., Ingwersen, D., Armstrong, D.P., Seddon, P.J., Ewen,
J.G., 2017. Is reintroduction biology an effective applied science? Trends Ecol. Evol.
32, 873–880.
Tourenq, C., Combreau, O., Pole, S.B., Lawrence, M., Ageyev, V.S., Karpov, A.A., Launay,
F., 2004. Monitoring of Asian houbara bustard Chlamydotis macqueenii populations in
Kazakhstan reveals dramatic decline. Oryx 38, 62–67.
Williams, S.E., Hoffman, E.A., 2009. Minimizing genetic adaptation in captive breeding
programs: a review. Biol. Conserv. 142, 2388–2400.
Willoughby, J.R., Christie, M.R., 2018. Long-term demographic and genetic effects of
releasing captive-born individuals into the wild. Conserv. Biol. https://doi.org/10.
1111/cobi.13217. Accepted Author Manuscript.
Worley, K., Collet, J., Spurgin, L.G., Cornwallis, C., Pizzari, T., Richardson, D.S., 2010.
MHC heterozygosity and survival in red junglefowl. Mol. Ecol. 19, 3064–3075.
P.M. Dolman et al. Biological Conservation xxx (xxxx) xxx–xxx
10