Genetic evaluation of the captive breeding program of the Persian wild ass

UNCORRECTED PROOF

Genetic evaluation of the captive breeding program of thePersian wild ass

R. K. Nielsen1, C. Pertoldi1,2 & V. Loeschcke1

1 Department of Ecology and Genetics, Institute of Biological Science, University of Aarhus, Ny Munkegade, Aarhus C, Denmark.

2 Department of Landscape Ecology, National Environmental Research Institute, Kalø, Rønde, Denmark.

Keywords

conservation; microsatellites; genetic

structure; relatedness; effective population

size; bottleneck.

Correspondence

Rikke Kruse Nielsen, Department of

Ecology and Genetics, Institute of Biological

Science, University of Aarhus, Ny

Munkegade, Building 1540, DK-8000

Aarhus C, Denmark. Fax: +45 894 22722

Email: [email protected]

Received 11 September 2006; accepted

28 November 2006

doi:10.1111/j.1469-7998.2007.00294.x

Abstract

During the last century, many species have become endangered and conservation

in terms of captive breeding has been crucial for their survival. Classical manage-

ment of captive species is based on recorded genealogies. However, if pedigrees are

incomplete or inaccurate, it can bias the interpretation of the results obtained from

analyses based on such data. In this investigation, 12 microsatellite loci were

investigated to evaluate the studbook information of the critically endangered

Persian wild ass Equus hemionus onager. Relatedness and inbreeding coefficients

were calculated in order to compare the same coefficients estimated from the

recorded studbook. A significant correlation between coefficients obtained by

microsatellites and the studbook validates the recorded studbook as a reasonable

tool for future genetic management. Furthermore, a Bayesian-based method

divided the captive onager population into four subgroups that indicate departure

from random mating, and thus minor rotation of animals between zoos. Lastly,

analyses for inferring past demographic changes revealed a gradual population

decline and inbreeding over several generations. This may indicate a low genetic

load in captive onagers as a consequence of some degree of purging. Consequently,

the risk of inbreeding depression should currently be minimal in the captive

breeding program. Therefore, it is recommended to increase the connectivity

between the four subgroups of onagers in order to reduce the risk of demographic

and genetic stochasticity. This study underlines the importance of using molecular

markers to evaluate genetic management of captive breeding programs.

Introduction

Many species have become endangered in the last century

and require active management to ensure their survival

(Olech & Perzanowski, 2002; Wisely, McDonald & Buskirk,

2003; Wilson et al., 2005). Genetic variation is a primary

component of adaptive evolution, and its loss or reduction

will decrease the long-term survival probability of popula-

tions. Therefore, maintaining genetic variation has been a

major goal in captive breeding programs (Reed & Frank-

ham, 2003). Minimizing kinship, which involves choosing

individuals with the lowest mean kinship to be parents of

subsequent generations, has been a strategy used in various

breeding programs to maximize the retention of genetic

variation. Minimizing kinship reduces the overall level of

relatedness and maximizes founder representation in captive

populations, and additionally minimizes the expression of

deleterious alleles in inbred animals (Montgomery et al.,

1997).

Deleterious alleles are expressed through consanguineous

mating and through random genetic drift in small popula-

tions in which deleterious alleles can be fixed (Hedrick &

Kalinowski, 2000). This increases the risk of inbreeding

depression in populations, which can reduce individual

fitness (Seymour et al., 2001). Removal of deleterious alleles

by natural selection (purging) has become increasingly

interesting in breeding programs, especially in cases when

the number of founders is low and inbreeding unavoidable.

However, as deliberate inbreeding to purge deleterious

alleles causes a further reduction in fitness, ‘maximum

avoidance of inbreeding’ is the currently accepted breeding

strategy for eliminating inbreeding depression in captive

species (Hedrick & Kalinowski, 2000).

In many breeding programs, the complete pedigrees are

often unknown. When the ancestry of founders is unspeci-

fied, they are assumed to be non-inbred and unrelated,

referred to as ‘founder assumption’ (Russello & Amato,

2004). This may lead to an underestimation of relatedness

within the population and result in incorrect calculations of

mean kinship and inbreeding coefficients that conservation

decisions normally rely on (Russello & Amato, 2004).

However, Willis (2001) showed that it is more appropriate

to underestimate than overestimate relatedness in unpedi-

greed populations in order to maintain genetic variation.

Application of polymorphic molecular markers has allowed

‘black holes’ in pedigrees to be eliminated and to infer

J Z O 2 9 4 B Dispatch: 19.1.07 Journal: JZO CE: Chandrika

Journal Name Manuscript No. Author Received: No. of pages: 9 TE: Rathna/Suresh

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 1

Journal of Zoology. Print ISSN 0952-8369

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

relatedness of individuals with previously unknown ancestry

(Jones et al., 2002; Wilson et al., 2005).

One species that threatened to go extinct is the Persian

wild ass Equus hemionus onager. Equus hemionus comprises

six subspecies including onagers that are categorized on the

IUCNRed List: The critically endangered Asiatic wild asses

(E. h. onager and Equus hemionus kulan); the Syrian wild ass

Equus hemionus hemippus categorized as extinct; the Indian

wild ass Equus hemionus khur classified as endangered; the

Mongolian wild ass Equus hemionus hemionus; and the Gobi

khulan Equus hemionus luteus as vulnerable (Feh et al., 2002;

IUCN Red List, 2004). In historical time, the distribution

of E. hemionus ranged from Turkey to northern China, and

from Kazakhstan to Saudi Arabia and India. Today,

E. hemionus are found restricted to a few places in China,

Mongolia, Turkmenistan, Kazakhstan, India and Iran (Feh

et al., 2002).

Wild onagers are now restricted to only two geographi-

cally isolated populations in Iran consisting of c. 100 and

500 individuals, respectively (Feh et al., 2002; Tatin et al.,

2003). The major threats to onagers are poaching, but also

habitat destruction, competition from domestic animals and

disturbance during mating and breeding periods are affect-

ing their survival. In 1954, a captive population of onagers

was founded and today it comprises c. 62 individuals in

Europe and worldwide 106 individuals, when the American

captive population is included. Genetic research is essential

to improve the captive breeding program and to plan future

management strategies to secure survival of the species.

This investigation only included individuals from the

European captive population. The main aims were (1) to

determine relatedness and inbreeding coefficients by means

of microsatellites in order to compare the same estimates

calculated from the recorded studbook, (2) to assess the

genetic structure in the captive population of onagers in

order to validate or invalidate random mating in the breed-

ing program, (3) to reconstruct the demographic history of

the onager population in order to examine whether the

observed genetic structure was due to recent demographic

changes and/or a more ancient event. Finally, the informa-

tion revealed was used to discuss the best genetic manage-

ment strategy for the captive onager population.

Methods

Study population and sample collection

Tissue or blood samples were collected from 60 onagers (21

males and 39 females) kept in captivity in 12 different

institutions taking part in the European Endangered Species

breeding Program (EEP) of onagers. The 12 institutions and

number of samples were: Germany: Tierpark Hagenbeck

(n=5), Zoologischer Garten Augsburg (n=5), Wilhelma

Zoologisch-botanischer Garten Stuttgart (n=4) and Zool-

ogischer Garten Koln (n=3); Switzerland: Werner Stamm –

Stiftung zur Erhaltung seltener Einhufer, Oberwil (n=7);

France: Reserve Africaine de Sigean (n=11), Parc Zoologi-

que de Lunaret, Montpellier (n=10) and Parc Zoologique

de Paris (n=1); the Netherlands: Rotterdam Zoo (n=5)

and Zoo Parc Overloon (n=4); and England: Chester Zoo

(n=1) and Whipsnade Wild Animal Park (n=4). The

samples consisted primarily of tissue collected with a biopsy

needle and a Daninject injection rifle (model JM Special).

Only four of the 60 samples were blood samples. Tissue

samples were preserved in 60% alcohol and together with

blood samples stored at �20 1C until use.

Microsatellite genotyping procedure

DNA extraction was performed with the standard CTAB

procedure (Doyle & Doyle, 1987). A total of 12 horse

microsatellite markers were applied to examine the geno-

types of the onagers by cross-species PCR amplification.

The chosen microsatellite markers were from a genotyping

kit normally used for genotyping and parentage testing

of horses (StockMarkss, Applied Systems, http://www.

appliedbiosystems.com). The loci examined were: AHT4,

AHT5, ASB17, ASB23, HMS2, HMS3, HMS6, HMS7,

HTG4, HTG7, HTG10 and VHL20. The PCR reactions

for all 12 markers were carried out in a 6mL reaction

containing 0.6 mL reaction buffer (1.5mM MgCl2), 0.96mLdNTP (1.25mM of A, C, G and T, respectively), 0.5 mLprimer mix (forward/reverse), 0.06mL Taq polymerase (Am-

plitaq Gold Q1, 5UmL�1), topped up with distilled water to 5

and 1mL DNA template added. PCR conditions using a

9700 GeneAmp Q2machine were: an initial denaturation at

95 1C for 10min, followed by 30 cycles of denaturation at

95 1C for 30 s, annealing at 60 1C for 30 s and extension at

72 1C for 60 s. Cycling culminated with a 60-min extension

at 72 1C. The amplified loci were analyzed on an ABI 310

Genetic Analyzer Q3. All alleles were scored manually using the

program GENOTYPER version 2.5.2 (Applied Biosystem Q4).

Population genetic analysis

All microsatellite loci were tested for deviation from Hard-

y–Weinberg expectations (HWE) using the Markov chain

method in FSTAT version 2.9.3.2 (Goudet, 1995); http://

www2.unil.ch/popgen/softwares/fstat.htm. Overall Bonfer-

roni’s adjustments were used to correct for the effect of

multiple tests (Rice, 1989). FSTAT was also used to calcu-

late the inbreeding coefficient (FIS) and allelic richness (AR)

at each locus. The observed (HO) and expected heterozygos-

ity (HE) per locus were all estimated with GENECLASS

version 2.0 (Piry et al., 2004). Furthermore, we tested for the

presence of null-alleles with MICRO-CHECKER version

2.2.3. (Van Oosterhout et al., 2004).

Detection of population genetic structure

A Bayesian clustering procedure implemented in the soft-

ware STRUCTURE 2.0 was used to infer whether distinct

genetic subgroups were represented in the samples of 60

onagers originally assumed to be one single population

(Pritchard, Stephens & Donnelly, 2000). If the onager

population is composed of more differentiated subgroups,

the likelihood of a partition in several panmictic clusters

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London2

Conservation of onagers R. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

should depart from its expected distribution under the

hypothesis of no structure (a single panmictic population).

STRUCTURE 2.0 was run with five independent simula-

tions, each one of 1 000 000 iterations, following a burn-in

period of 100 000 iterations. series of independent runs with

K=1–8 were performed.

Genetic differentiation between distinct subgroups in the

onager population was characterized by estimating overall

and pairwise FST using GENEPOP on the web (http://

wbiomed.curtin.edu.au/genepop/). The significance levels

for overall and pairwise FST were determined after 10 000

permutations in FSTAT. Furthermore, for all distinct sub-

groups in the onager population, the microsatellite loci were

tested for deviation fromHWE, and also FIS and AR at each

locus were calculated as described above. The number of

private alleles (PA), HO and HE per locus were also esti-

mated for each subgroup with GENECLASS. For graphic

visualization of the genetic differences among individuals

and subgroups, factorial component analysis (FCA) using

the software GENETIX version 4.05 was run (Belkhir et al.,

2004); http://www.univ-montp2.fr/�genetix/genetix.htm

Relatedness and inbreeding coefficients

Relatedness between pairs of the 60 onagers was estimated in

terms of the relatedness coefficient, rxy (Queller & Good-

night, 1989), using the software IDENTIX (Belkhir, Castric

& Bonhomme, 2002); http://www.Univ-montp2.fr/�genetix/identix01.zip. Relatedness coefficients were determined

assuming that the 60 individuals constitute one single

panmictic population (non-structured population), and ad-

ditionally relatedness coefficients were estimated between

the individuals composing the subgroups determined by the

software STRUCTURE 2.0 (structured population). A

power analysis has been conducted and has revealed the

capacity of discriminating half-sibs at the 5% level, given the

genetic variability found. Furthermore, individual inbreed-

ing coefficients were calculated for a structured and non-

structured onager population. Relatedness and inbreeding

coefficients calculated from the microsatellite data were

compared with those calculated from the studbook for

captive onagers. Studbook estimates were calculated by the

software package, GENES, which is a program for analysis

and management of pedigrees (written by Robert C. Lacy,

available at: http://www.vortex9.org/genes.html). To test

for significant correlation between microsatellite and stud-

book estimates, a Pearson’s r-test was performed.

Inferring population demographic changes

The occurrence of bottleneck or expansion was tested with

two softwares. Recent population bottlenecks were inferred

for the structured population of onagers with BOTTLE-

NECK 1.2.02 (Cornuet & Luikart, 1996); http://www.

montpellier.inra.fr/URLB/bottleneck/bottleneck.html, as-

suming an infinite allele model (IAM), a stepwise mutation

model (SMM) or a two-phase model (TPM, with 95%

SMMs).

Detection of historical population decline or expansion

was conducted using MSVAR 1.3 (Storz, Beaumont &

Alberts, 2002). The program assumes a stepwise mutation

model, closed populations and that the populations are in

HWE. It estimates the posterior probability distribution of

several genealogical and demographic parameters using

Markov Chain Monte Carlo simulations based on the

observed distribution of microsatellite alleles and their

repeat numbers (Beaumont, 1999). Five independent chains

(a total of 10 000 points for each run) were run for the total

population of onagers. Density estimation, estimated mode

and highest probability density limits were carried out by

summing up the five independent chains for each analysis,

using the R script program. As MSVAR assumes popula-

tions in HWE, the analysis should have been performed on

the structured populations. However, similar results were

obtained when running MSVAR on subgroup A (n=24),

and therefore the total population (n=60) was used to

achieve higher statistical power. Still, the departure from

HWE in the total population may result in a small over-

estimation of the time since population decline or expan-

sion.

Results

Population genetic analysis

Highly significant departure from HWE was observed when

testing overall loci in the total population of onagers

(FIS=0.142, Po0.001). Significant deviation from HWE

was found at five loci (Po0.00417, Bonferroni’s corrected):

AHT5, HMS2, HTG7, HTG10 and VHL20. The level of

genetic variation in the 12 microsatellite loci for the total

population of onagers varied between 0.09 and 0.78, and

between 0.10 and 0.87 forHO andHE, respectively (Table 1).

There were two to ten alleles per locus with an average of

5.33 alleles per locus (Table 1). Finally, no significant

frequencies of null-alleles were found according to Brook-

field (Allele frequencies available on request).

Genetic structure

The Bayesian clustering procedure STRUCTURE 2.0 max-

imized the likelihood of the data with four genetically

distinct subgroups (K=4; Ln=�1438.5). The posterior

probability was estimated to be P(K=4)=0.99. The 60

onagers were assigned to the four inferred clusters at an

assignment level 40.864 (Table 2). For graphic illustration

of the assignment of the four subgroups in STRUCTURE

2.0, see Fig. 1. The FCA supported the findings of substruc-

ture in the onager population (Fig. 2).

The overall measure of genetic differentiation across all

loci and subgroups was significantly higher than zero

(FST=0.18, Po0.001). The pairwise FST between the sub-

groups were all highly significant and ranged from 0.10 to

0.35 (Table 3). The highest genetic differentiation was

estimated between subgroups C and D and the lowest

between subgroups A and B. Agreement with HWE was

Q5

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 3

Conservation of onagersR. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

Table 1 Summary of microsatellite data for the total population of captive onagers and the total population divided into subgroups A, B, C and D,

respectively (in later figures the subgroups A, B, C and D are visualized with red, blue, green and yellow colors, respectively)

Subgroup/

locus

A (red)

(n=24)

B (blue)

(n=14)

C (green)

(n=13)

D (yellow)

(n=9)

Mean

values

Total population

(n=60)

AHT4

No. of alleles 7 7 4 3 5.25 10

AR (PA) 5.78 (2) 6.23 (1) 3.69 (0) 3.00 (0) 4.68 9.98

HO 0.79 0.93 0.77 0.56 0.78

HE 0.80 0.83 0.67 0.63 0.87

AHT5

No. of alleles 5 5 4 2 4.00 7

AR (PA) 3.99 (2) 4.82 (1) 3.91 (0) 2.00 (0) 3.68 7.00

HO 0.67 0.77 0.77 0.33 0.66�

HE 0.64 0.78 0.73 0.29 0.76

ASB17

No. of alleles 2 2 2 2 2.00 2

AR (PA) 2.00 (0) 2.00 (0) 2.00 (0) 2.00 (0) 2.00 2.00

HO 0.35 0.57 0.54 0.44 0.46

HE 0.29 0.52 0.51 0.52 0.48

ASB23

No. of alleles 3 5 4 2 3.50 5

AR (PA) 2.78 (0) 4.21 (0) 3.98 (0) 2.00 (0) 3.24 5.00

HO 0.30 0.57 0.85 0.33 0.49

HE 0.47 0.48 0.75 0.43 0.58

HMS2

No. of alleles 2 2 1 1 1.50 2

AR (PA) 2.00 (0) 1.64 (0) 1.00 (0) 1.00 (0) 1.41 2.00

HO 0.21 0.07 – – 0.10�

HE 0.31 0.07 – – 0.15

HMS3

No. of alleles 1 3 2 1 1.75 3

AR (PA) (0) 2.63 (1) 1.69 (0) 1.00 (0) 1.58 3.00

HO – 0.29 0.08 – 0.09

HE – 0.32 0.08 – 0.10

HMS6

No. of alleles 2 2 2 2 2.00 2

AR (PA) 2.00 (0) 2.00 (0) 2.00 (0) 2.00 (0) 2.00 2.00

HO 0.25 0.29 0.54 0.56 0.37

HE 0.34 0.42 0.52 0.50 0.46

HMS7

No. of alleles 5 6 4 2 4.25 6

AR (PA) 4.71 (0) 5.29 (0) 3.67 (0) 2.00 (0) 3.92 6.00

HO 0.86 0.86 0.77 0.33 0.77

HE 0.76 0.68 0.65 0.29 0.80

HTG4

No. of alleles 5 3 2 2 3.00 5

AR (PA) 4.11 (2) 2.89 (0) 2.00 (0) 2.00 (0) 2.75 4.97

HO 0.67 0.43 0.23 0.44 0.48

HE 0.66 0.54 0.41 0.37 0.57

HTG7

No. of alleles 6 5 3 1 3.75 7

AR (PA) 5.23 (1) 4.49 (1) 2.69 (0) 1.00 (0) 3.35 6.98

HO 0.75 0.64 0.62 – 0.58�

HE 0.80 0.70 0.52 – 0.77

HTG10

No. of alleles 6 6 3 4 4.75 8

AR (PA) 4.95 (1) 5.27 (0) 2.91 (0) 4.00 (0) 4.28 8.00

HO 0.58 0.93 0.54 0.78 0.68�

HE 0.67 0.75 0.44 0.74 0.80

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London4

Conservation of onagers R. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

found when testing over all loci in all four subgroups

(Po0.00104, Bonferroni’s corrected). The same was found

for all single loci in each subgroup (Po0.00104). The mean

number of alleles detected at each locus for the subpopula-

tions varied between 1.5 (HMS2) and 5.25 (AHT4). The

mean AR per locus varied between 1.41 and 4.68 (Table 1).

The mean AR per subgroup was the highest for subgroup B

(3.57) and the lowest for subgroup D (2.17).

Eight PA were found at five loci in subgroup A, four PA

were found in four loci in subgroup B and two PA were

found in subgroup C (Table 1). The highest frequency of one

PA in subgroup A was 0.146 and in the other seven loci the

frequencies ranged between 0.021 and 0.063. In subgroup B,

the highest frequencies of two PA were 0.308 and 0.179,

respectively, and the two other PA showed lower frequencies

at 0.036 and 0.071. The two PA in subgroup C both

occurred at higher frequencies: 0.346 and 0.308, respec-

tively. The HE in the four subgroups varied between 0.36

and 0.57 and theHO varied between 0.37 and 0.60 (Table 1).

Relatedness and inbreeding coefficients

Pairwise relatedness coefficients calculated in the non-struc-

tured population of onagers were compared with the relat-

edness coefficient estimates from the studbook. A significant

correlation was found (r=0.52, Po0.05). A relatedness

matrix between males and females was constructed based

on relatedness coefficients calculated from the non-struc-

tured population (Supplementary Material Table S1). A

relatedness matrix between individuals of each structured

population (A, B, C and D) was constructed and had lower

values than estimates based on the non-structured onager

population (data not shown). Inbreeding coefficients esti-

mated from the structured and non-structured onager po-

pulation showed a significant correlation compared with the

inbreeding coefficients from the studbook (r=0.374,

Po0.05; r=0.365, Po0.05). The inbreeding coefficients

found from the structured population had values closer to

the estimates from the studbook (data not shown).

Table 1. Continued.

Subgroup/

locus

A (red)

(n=24)

B (blue)

(n=14)

C (green)

(n=13)

D (yellow)

(n=9)

Mean

values

Total population

(n=60)

VHL20

No. of alleles 5 5 4 4 4.50 7

AR (PA) 4.28 (0) 4.62 (0) 3.69 (2) 4.00 (0) 4.15 7.00

HO 0.63 0.86 0.69 0.67 0.70�

HE 0.72 0.74 0.72 0.53 0.84

Overall

FIS 0.065 �0.058 �0.069 �0.036 0.142

Mean AR 3.57 3.84 2.77 2.17 5.33

HO 0.51 0.60 0.53 0.37 0.51�

HE 0.54 0.57 0.50 0.36 0.60

Tabulated are the number of alleles scored for each locus, the allelic richness (AR), the number of private alleles (PA), the observed heterozygosity

(HO), the expected heterozygosity (HE), inbreeding coefficient (FIS) and results of test for deviation from Hardy–Weinberg equilibrium are given by

an asterisk. Significance levels were applied using the overall Bonferroni’s technique (Rice, 1989). Po0.00417 for the total population and

Po0.00104 for the subgroups.

Table 2 Posterior mean estimates of the assignment level (highest

values in bold) on the inferred clusters for 60 individuals of captive

onagers

Subgroups

Inferred clusters

1 2 3 4

A (red) 0.864 0.096 0.010 0.030

B (blue) 0.038 0.878 0.050 0.034

C (green) 0.016 0.023 0.955 0.005

D (yellow) 0.009 0.008 0.004 0.978

The results by STRUCTURE 2.0 revealed four clusters as the most

consistent result (burn-in period; 100 000; simulations; 1 000 000).

1.00

0.80

0.60

0.40

0.20

0.001 2 3 4

Figure 1 Estimated population structure (K=4) derived using the software STRUCTURE 2.0 on the total captive onager population. Each of the

60 individuals is represented by a vertical bar, which is partitioned into K colored segments that represent the individual’s estimated membership

proportions in each of the four clusters. Numbers one to four refer to subgroup A (red), B (blue), C (green) and D (yellow), respectively.

Q11

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 5

Conservation of onagersR. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

Population demographic changes

The test to evaluate the effect of a recent bottleneck was

conducted for the four subgroups separately in BOTTLE-

NECK 1.2. When the IAM was assumed, the software

revealed evidence for a recent reduction in Ne in subgroups

A, C and D (Po0.01). In MSVAR 1.3, an exponential

model was used for simulations to detect demographic

changes over a short time period. The software MSVAR

1.3 gives four important output estimates: current and

ancestral Ne, mutation rate and time since population

decline or expansion. The model indicated a reduction in

Ne starting c. 3591 years ago for the total onager population.

Assuming a generation time of 7 years per generation, which

is an estimate for Equids (Kruger et al., 2005), the popula-

tion decline should have started 513 generations ago. Recent

and historical Ne was estimated to be 26 and 4677 indivi-

duals, respectively. The mutation rate was estimated to be

1.2� 10�3. The limits of recent and historical Ne are dis-

played in Fig. 3.

Discussion

Maintenance of genetic variation

The captive onager population was founded between 1954

and 1973 (Denzau & Denzau, 1999). It can be difficult to

identify losses of genetic variation in the captive onager

population as levels before population decline and captivity

are unknown. To evaluate the genetic state of endangered

populations, empirical data of genetic variation obtained in

the same or related undisturbed species are often used for

comparison (Spielman, Brook & Frankham, 2004). How-

ever, in a phylogenetic research of equids, the mean HE in

seven of the same examined loci as in this investigation was

0.54 for onagers, which indicate that the level of genetic

variation found in the onager population is comparable to

other studies (Kruger, 2003).

The theoretical loss of genetic variation is 1/2Ne per

generation. The current Ne for the captive onager popula-

tion was estimated to be 26 individuals. With a generation

time of 7 years per generation for equids, the calculated

proportion of initial heterozygosity retained in 100 years

(14.3 generations) is 76% for the total captive onager

population. This is not in agreement with the traditionally

accepted goal in captive breeding programs to maintain

Figure 2 Plot of factorial component analysis. The first factor explains

14% and the second 9% of the genetic variation between the four

subgroups. Red, blue, green and yellow symbols represent subgroup

A, ^(n=24); B, ’(n=14); C, m(n=13) and D, �(n=9), respectively.

The subgroups consist of the following individuals noted by their

international studbook number and their city of origin at the time of

tissue sampling: (H, Hamburg; A, Augsburg; S, Stuttgart; Ob, Oberwil;

Si, Sigean; M, Montpellier P, Pis; C, Cologne; R, Rotterdam; Wh,

Whipsnade; Ch, Chester and O, Overloon). A: Germany: H620, H656,

H741, H723, A428, A513, A710, A558 and A313; Switzerland: Ob686,

Ob451, Ob540, Ob618, Ob744, Ob754, Ob757. England: Ch591,

Wh585 and Wh418; the Netherlands: O607, O622, O635, O653 and

R611. B: Germany: H594, S509, S508, S596, S689, C402, C497 and

C559; France: M477and P287; the Netherlands: R310 and R529;

England: Wh506 and Wh551. C: France: Si372, Si610, Si633, Si738,

Si749, Si751, Si592, Si691, Si679, Si703 and Si718; the Netherlands:

R608 and R493. D: France: M515, M701, M602, M616, M698, M719,

M713, M733 and M737.

Table 3 Pairwise FST values (above diagonal) between the four

subgroups of captive onagers and level of significance (below diag-

onal) using the overall Bonferroni’s technique (Rice, 1989) with

Po0.00833

A (red) B (blue) C (green) D (yellow)

A (red) 0.10 0.21 0.21

B (blue) � 0.19 0.19

C (green) � � 0.35

D (yellow) � � �

Current population size

Anc

estr

al p

opul

atio

n si

ze

−2 −1 0 1 2 3 4

7

6

5

4

3

2

1

Figure 3 Joint posterior distribution of current and ancestral effective

population size for the total captive onager population. The plots show

90, 50 and 10% highest probability density on a log10 scale.

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London6

Conservation of onagers R. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

90% of the initial genetic variation within 100 years. To

achieve this goal, an Ne of least 68 individuals is required.

Departure from random mating

Maintenance of genetic variation in captive populations

requires large Ne to eliminate the effect of genetic drift and

inbreeding. Consequently, the goal of most captive breeding

programs is to rotate animals among institutions in order to

keep threatened species as a single random mating popula-

tion. The significant heterozygote deficiency found within

the total onager population indicates a deviation from

random mating. An explanation could be inbreeding and/

or a Wahlund effect (substructure) due to inappropriate

rotation of animals between onager institutions and/or due

to technical problems during PCR amplification such as

null-alleles. Because of the fact that only five loci deviated

from HWE in the total onager population and the non-

significant presence of null-alleles, a Wahlund effect rather

than inbreeding is suggested as an explanation.

The hypothesis of a genetic substructure was supported

by the four inferred clusters derived from STRUCTURE 2.0

and the high levels of genetic differentiation between sub-

groups A, B, C and D. Furthermore, the presence of PA

found in nearly all subgroups with relatively high frequen-

cies also supported the evidence of a genetic substructure in

the captive onager population. The findings of substructure

indicate a need to increase transfers of individuals between

the four subgroups in order to achieve a random mating

captive onager population. However, translocations of large

mammals are usually risky and expensive. In this context, an

artificial insemination strategy could limit the costs of

transfers and reduce the spread of diseases between captive

groups (Conway, 1995). An increased gene flow between

captive groups may lead to a reduced genetic substructure,

and therefore to an increased Ne. Artificial insemination has

already succeeded in endangered species kept in captivity,

and could additionally be a potential strategy in the future

to increase the number of effective breeding animals in the

captive onager population (Masui et al., 1988; Conway,

1995).

Evaluation of the studbook

The significant correlation between estimates of relatedness

and inbreeding coefficients calculated from microsatellites

and the studbook, respectively, indicates that the recorded

studbook for the captive onagers is an appropriate tool for

genetic management of this species. Still, the relatedness

coefficients estimated from microsatellites provide further

information on relatedness coefficients that are of equal

values in the studbook, and therefore contribute to the

choice of optimal breeding pairs. However, the relatedness

and inbreeding coefficients did not show perfect correlations

when compared with the studbook estimates. Discrepancies

can be explained by the subdivision of onagers into four

separate groups. The most reliable estimates will be achieved

when the genetic structure of the onager population is

known as the assumption of random mating in populations

is important to achieve the most precise values of relatedness

(Queller & Goodnight, 1989; Lynch & Ritland, 1999).

Furthermore, examination of more variable loci but also

the right choice of estimator of relatedness would increase

the confidentiality of the relatedness coefficients (Lynch &

Ritland, 1999; Van de Casteele, Galbusera & Matthysen,

2001). Furthermore, the discrepancy can also be partially

explained by the small sample size (n=60), which may not

reflect the real allelic frequencies. However, nearly all

onagers in the capive breeding program were sampled, and

therefore these are the best estimates given. According to the

‘founder assumption,’ an underestimation of relatedness

and inbreeding among founders when the studbook was

started could be another reason for discrepancy (Hedrick &

Kalinowski, 2000; Russello & Amato, 2004), but also due to

the fact that individuals are misrecorded in the studbook.

Thus, it is critical to calculate relatedness coefficients both

by means of microstallites and the studbook.

Population demographic changes

A recent decline in population size in three out of the four

subgroups (A, C and D) was detected, suggesting a non-

equilibrium state. Onagers were brought into captivity c.

50 years ago (7.4 generations), and therefore an explanation

of a recent decline in population size could be equivalent to a

founder effect with all its consequences. It is crucial to detect

population bottlenecks in managed species because a reduc-

tion inNe may enhance the rate of inbreeding, loss of genetic

variation and fixation of deleterious alleles considerably and

thereby increase the risk of population extinction (Luikart

et al., 1998).

In this investigation, it was found that the wild onager

population has gone through a bottleneck 3591 years ago

(513 generations), which indicates a gradual decline in Ne

and a gradual increase of inbreeding over several genera-

tions. The result of a population decline is supported by

archeological evidence that an extensive hunting of onagers

started c. 3–4000 years ago (Clutton-Brock, 1992, 1999). It

has been hypothesized that inbreeding depression is primar-

ily a consequence of (partially) recessive deleterious alleles

occurring in a population (Charlesworth & Charlesworth,

1987). The mechanism to remove the genetic load in a

population is purging, which is natural selection against

deleterious alleles (Frankham et al., 2001). Different rates of

inbreeding influence the effectiveness of purging. If inbreed-

ing is fast, the Ne is reduced and random genetic drift will

become stronger relative to selection, but in contrast, when

inbreeding is slow it provides the opportunity for selection

to act (Pedersen, Kristensen & Loeschcke, 2005). Therefore,

it can be hypothesized that selection has acted continuously

and that the genetic load was low when the captive onager

population was established in 1954. Knowledge about the

magnitude of the genetic load in a population plays a central

role in the genetic management of captive species (Pertoldi

et al., 2006).

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 7

Conservation of onagersR. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

Recommendations for future geneticmanagement

With the evidence of a slow decline in Ne and thus a

supposed low genetic load in the captive onager population,

the risk of inbreeding depression should currently not be

severe. Consequently, all four subgroups of the captive

onager population should be valuable both as source and

recipient populations. It is recommendable to establish a

higher connectivity between the four subgroups that will

increaseNe. The accumulation of deleterious alleles may not

only be a consequence of smallNe in which they can increase

in frequency or become fixed, but can also be due to benign

conditions in captivity that result in relaxed selection against

harmful alleles (Gilligan & Frankham, 2003; Theodorou &

Couvet, 2004). Therefore, a higher Ne will give selection an

opportunity to act and prevent an increase of a genetic load

in the captive onager population. To increase Ne as fast as

possible, all breeding individuals should be actively involved

in the captive breeding program.

The relatedness matrix between males and females (Sup-

plementary Material Table S1) was calculated on the basis

of the non-structured onager population, which can be

problematic as the estimator of relatedness requires random

mating in the population to calculate correct values (Queller

& Goodnight, 1989). Thus, if mating between two indivi-

duals within the same subgroup is proposed, the relatedness

matrix inferred for the given subgroup would be more

precise to use (data available on request). Additionally, the

relatedness coefficients calculated from molecular markers

have an advantage over the relatedness estimated from the

studbook for relatedness coefficients that are of equal values

in the studbook.

In this investigation, a Bayesian approach was used to

infer the genetic structure in the captive onager population

and detect the demographic history in terms of Ne. The

importance of such an approach should be underlined in

order to be critical to the current genetic management of

captive species. By applying molecular markers to the

management of captive species, it is possible to evaluate the

success of breeding programs that may be relevant in

planning future management strategies as for example in

actions of translocations, reintroductions and repopula-

tions. However, before repopulation efforts by means of

releasing captive onagers into wild populations in Iran is

undertaken, a genetic screening of wild populations is

recommendable.

Acknowledgements

We are very grateful to all onager institutions for their

assistance in collecting tissue samples. Furthermore, we

thank Ebeltoft Zoo & Safari for lending us sampling equip-

ments. For technical assistance in the laboratory, we thank

Nørlund Horse Hospital. We also thank Ian Davies for

writing the R script to analyze the MSVAR 1.3 output.

Finally, many thanks are due to Editor Juliet Clutton-Brock

and two anonymous reviewers for suggestions and com-

ments on a previous version of this paper. This investigation

was financed by grants from WWF Denmark and Oticon.

References

Beaumont, M.A. (1999). Detecting population expansion and

decline using microsatellites. Genetics 153, 2013–2029.

Belkhir, K., Borsa, P., Chikhi, L., Raufaste, N. & Bonhomme,

F. (2004). GENETIX version 4.05 Q6.

Belkhir, K., Castric, V. & Bonhomme, F. (2002). IDENTIX, a

software to test for relatedness in a population using

permutation methods. Mol. Ecol. Notes 2, 611–614.

Charlesworth, B. & Charlesworth, D. (1987). Inbreeding

depression and its evolutionary consequences. Ann. Rev.

Ecol. System. 18, 237–268.

Clutton-Brock, J. (1992). Horse power: a history of the horse

and donkey in human societies (Chapter 6) Q7. Natural History

Museum Publication & Harvard University Press.

Clutton-Brock, J. (1999). A natural history of domesticated

mammals Q8. 2nd edn. Cambridge University Press/The Nat-

ural History Museum. (Equus hemionus onager, p. 118).

Conway, W. (1995). Wild and zoo animal interactive man-

agement and habitat conservation. Biodivers. Conserv. 4,

573–594.

Cornuet, J.M. & Luikart, G. (1996). Description and power

analysis of two tests for detecting recent population bottle-

necks from allele frequency data. Genetics 144, 2001–2014.

Denzau, G. & Denzau, H. (1999).Wildesel, Thorbecke species

3. Stuttgart: J. Thorbecke Verlag.

Doyle, J.J. & Doyle, J.L. (1987). A rapid DNA isolation

procedure for small quantities of fresh leaf tissue. Phyto-

chem.l Bull. 19, 11–15.

Feh, C., Shah, N., Rowen, M., Reading, R. & Goyal, S.P.

(2002). Status and action plan for the Asiatic wild ass

(Equus hemionus Q9). In Equids: zebras, asses, and horses.

Moehlman, P.E. (Ed.). Gland, Switzerland: IUCN/SCC

Equid Specialist Group, IUCN. Available at: http://www.

iucn.org/themes/ssc/sgs/equid/docs/part2chap ter5.pdf.

Frankham, R., Gilligan, D.M., Morris, D. & Briscoe, D.A.

(2001). Inbreeding and extinction: effects of purging. Con-

serv. Genet. 2, 279–285.

Gilligan, D.M. & Frankham, R. (2003). Dynamics of genetic

adaptation to captivity. Conserv. Genet. 4, 189–197.

Goudet, J. (1995). FSTAT (Version 1.2): a computer program

to calculate F-statistics. J. Hered. 86, 485–486.

Hedrick, P.W. & Kalinowski, S.T. (2000). Inbreeding depres-

sion in conservation biology. Ann. Rev. Ecol. System. 31,

139–162.

IUCN (2004). IUCN Red List of threatened species. www.

redlist.org.

Jones, K.L., Glenn, T.C., Lacy, R.C., Pierce, J.R., Unruh, N.,

Mirande, C.M. & Chavez-Ramirez, F. (2002). Refining the

whooping crane studbook by incorporating microsatellite

DNA and leg-banding analyses.Conserv. Biol. 16, 789–799.

Kruger, K. (2003). Vergleichende molekulargenetische Unter-

suchungen zur Phylogenese der Equiden unter besonderer

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London8

Conservation of onagers R. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

UNCORRECTED PROOF

Beruksichtigung von E. hemionus kulan und E. hemionus

onager. PhD thesis, Zurich University.

Kruger, K., Gaillard, C., Stranzinger, G. & Rieder, S. (2005).

Phylogenetic analysis and species allocation of individual

equids using microsatellite data. J. Anim. Breed. Genet.

122, 78–86.

Luikart, G., Sherwin, W.B., Steele, B.M. & Allendorf, F.W.

(1998). Usefulness of molecular markers for detecting

population bottlenecks via monitoring genetic change.

Mol. Ecol. 7, 963–974.

Lynch, M. & Ritland, K. (1999). Estimation of pairwise

relatedness with molecular markers. Genetics 152,

1753–1766.

Masui, M., Hiramatsu, H., Nose, N., Nakazato, R., Sagawa,

Y., Tajima & Saito, K. (1988). Successful artificial insemi-

nation in the giant panda (Ailuropoda melanoleuca) at ueno

zooQ10 . Zoo Biol. 8, 17–26.

Montgomery, M.E., Ballou, J.D., Nurthen, R.K., England,

P.R., Briscoe, D.A. & Frankham, R. (1997). Minimizing

kinship in captive breeding programs. Zoo Biol. 16,

377–389.

Olech, W. & Perzanowski, K. (2002). A genetic background

for reintroduction program of the European bison

(Bison bonasus) in the Carpathians. Biol. Conserv. 108,

221–228.

Pedersen, K.S., Kristensen, T.N. & Loeschcke, V. (2005).

Effects of inbreeding and rate of inbreeding in Drosophila

melanogaster – Hsp70 expression and fitness. J. Evol. Biol.

18, 756–762.

Pertoldi, C., Breyne, P., Cabria, M.T., Halfmaerten, D.,

Jansman, H.A.H., Berge, K.V.D.B., Madsen, A.B. &

Loeschcke, V. (2006). Genetic structure of the European

polecat (Mustela putorius) and its implication for conser-

vation strategies. J. Zool. (Lond.) 270, 102–115.

Piry, S., Alapetite, A., Cornuet, J.M., Paetkau, D., Baudouin,

L. & Estoup, A. (2004). GENECLASS2: a software for

genetic assignment and first-generation migrant detection.

J. Hered. 95, 536–539.

Pritchard, J.K., Stephens, M. & Donnelly, P. (2000). Infer-

ence of population structure using multilocus genotype

data. Genetics 155, 945–959.

Queller, D.C. & Goodnight, K.F. (1989). Estimating related-

ness using genetic markers. Evolution 43, 258–275.

Reed, D.H. & Frankham, R. (2003). Correlation between

fitness and genetic diversity. Conserv. Biol. 17, 230–237.

Rice, W.R. (1989). Analyzing tables of statistical tests. Evolu-

tion 43, 223–225.

Russello, M.A. & Amato, G. (2004). Ex situ population

management in the absence of pedigree information. Mol.

Ecol. 13, 2829–2840.

Seymour, A.M., Montgomery, M.E., Costello, B.H., Ihle, S.,

Johnsson, G., St John, B., Taggart, D. & Houlden, B.A.

(2001). High effective inbreeding coefficients correlate with

morphological abnormalities in populations of south

Australian koalas (Phascolarctos cinereus). Anim. Conserv.

4, 211–219.

Spielman, D., Brook, B.W. & Frankham, R. (2004). Most

species are not driven to extinction before genetic factors

impact them. Proc. Nat. Acad. Sci. USA 101, 15261–15264.

Storz, J.F., Beaumont, M.A. & Alberts, S.C. (2002). Genetic

evidence for long-term population decline in a savannah-

dwelling primate: inferences from a hierarchical Bayesian

model. Mol. Biol. Evol. 19, 1981–1990.

Tatin, L., Darreh-Shoori, B.F., Tourenq, C., Tatin, D. &

Azmayesh, B. (2003). The last populations of the critically

endangered onager, Equus hemionus onager in Iran: urgent

requirements for protection and study. Oryx 37, 488–491.

Theodorou, K. & Couvet, D. (2004). Introduction of captive

breeders to the wild: harmful or beneficial? Conserv. Genet.

5, 1–12.

Van de Casteele, T., Galbusera, P. & Matthysen, E. (2001). A

comparison of microsatellite-based pairwise relatedness

estimators. Mol. Ecol. 10, 1539–1549.

Van Oosterhout, C., Hutchinson, W.F., Wills, D.P.M. &

Shipley, P. (2004). MICRO-CHECKER: software for

identifying and correcting genotyping errors in microsatel-

lite data. Mol. Ecol. Notes 4, 535–538.

Willis, K. (2001). Unpedigreed populations and worst-case

scenarios. Zoo Biol. 20, 305–314.

Wilson, G.A., Nishi, J.S., Elkin, B.T. & Strobeck, C. (2005).

Effects of a recent founding event and intrinsic population

dynamics on genetic diversity in an ungulate population.

Conserv. Genet. 6, 905–916.

Wisely, S.M., McDonald, D.B. & Buskirk, S.W. (2003). Eva-

luation of the genetic management of the endangered black-

footed ferret (Mustela nigripes). Zoo Biol. 22, 287–298.

Supplementary material

The following material is available for this paper online:

Table S1. Relatedness matrix for males (n=21) and

females (n=39) given by their studbook number and letters

for their city of origin at the time of tissue sampling:

(H=Hamburg, A=Augsburg, S=Stuttgart,

Ob=Oberwil, Si=Sigean, M=Montpellier, P=Paris,

C=Cologne, R=Rotterdam, Wh=Whipsnade,

Ch=Chester and O=Overloon). Results of pairwise relat-

edness coefficients, rxy (Queller and Goodnight 1989), are

based on the non-structured captive onager population.

Most negative pairwise relatedness coefficients indicate

most unrelated individuals. Most positive pairwise related-

ness coefficients indicate most related individuals.

This material is available as part of the online article

from http://www.blackwell-synergy.com/doi/abs/10.1111/

j.1469-7998.2006.00294.x primary_article

Please note: Blackwell Publishing are not responsible

for the content or functionality of any supplementary

materials supplied by the authors. Any queries (other than

missing material) should be directed to the corresponding

author for the paper.

Journal of Zoology (2007) c� 2007 The Authors. Journal compilation c� 2007 The Zoological Society of London 9

Conservation of onagersR. K. Nielsen, C. Pertoldi and V. Loeschcke

JZO 294(BW

UK

JZ

O 2

94.P

DF

19-J

an-0

7 19

:55

1965

27 B

ytes

9 P

AG

ES

n op

erat

or=

Sure

sh B

abu)

Author Query Form

_______________________________________________________

_______________________________________________________

Dear Author,

During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary changes/additions. Please write your answers clearly on the query sheet if there is insufficient space on the page proofs. If returning the proof by fax do not write too close to the paper's edge. Please remember that illegible mark-ups may delay publication.

Journal       JZOArticle       294

Query No. Description Author Response

Q1AQ: Please provide manufacturer information for Amplitaq Gold: town, state (if USA), and country.

Q2AQ: Please provide manufacturer information for 9700 GeneAmp: town, state (if USA), and country.

Q3AQ: Please provide manufacturer information for ABI 310 Genetic: town, state (if USA), and country.

Q4AQ: Please provide manufacturer information for Applied Biosystem: town, state (if USA), and country.

Q5AQ: Please check the change of spelling from Bohomme to Bonhomme in the reference Belkhir et al. (2002).  

Q6AQ: Please update the reference Belkhir et al. (2004).

Q7AQ: Please provide place of publication for the reference Clutton-Brock (1992).

Q8AQ: Please provide place of publication for the reference Clutton-Brock (1999).

Q9AQ: Please provide page range for the reference Feh et al. (2002).

Q10AQ: Please provide initials for the author Tajima in the reference Masui et al (1988).

Q11

AQ: Supplied Figure1 is not of print quality, image unsharp. Please supply print quality fig (more than 300 dpi) at next stage.   

MARKED PROOF

Please correct and return this set

Instruction to printer

Leave unchanged under matter to remain

through single character, rule or underline

New matter followed byor

or

or

or

or

or

or

or

or

and/or

and/or

e.g.

e.g.

under character

over character

new character new characters

through all characters to be deleted

through letter orthrough characters

under matter to be changedunder matter to be changedunder matter to be changedunder matter to be changedunder matter to be changed

Encircle matter to be changed

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

(As above)

linking characters

through character orwhere required

between characters orwords affected

through character orwhere required

or

indicated in the marginDelete

Substitute character orsubstitute part of one ormore word(s)

Change to italicsChange to capitalsChange to small capitalsChange to bold typeChange to bold italicChange to lower case

Change italic to upright type

Change bold to non-bold type

Insert ‘superior’ character

Insert ‘inferior’ character

Insert full stop

Insert comma

Insert single quotation marks

Insert double quotation marks

Insert hyphenStart new paragraph

No new paragraph

Transpose

Close up

Insert or substitute spacebetween characters or words

Reduce space betweencharacters or words

Insert in text the matter

Textual mark Marginal mark

Please use the proof correction marks shown below for all alterations and corrections. If you

in dark ink and are made well within the page margins.wish to return your proof by fax you should ensure that all amendments are written clearly