i C ?  c ( I 
,> ,- . 
. < \  c \ t  
CONSERVATION SCIENCE AND ZOOS 
cinerea, Cheetah Acinonyx jubatus, Przewalski's horse Eqttus przewalskii, Iranian onager Equus hemionus 
onager, Sitatunga Tragelaphus spekei, Nilgai Boselaphus tragocamelus, Ellipsen waterbuck Kobus 
ellipsiprymnus, Scimitar-horned oryx Oryx dammah. 
Total = 15.  
Reviews and interim recommendations made: 
Luzon bleeding-heart pigeon Gallicolumba luzonica, Goodfellow's tree kangaroo Dendrolagus goodfellowi 
shawmayeri, Brush-tailed rock wallaby Petrogale p. penicillata, Emperor tamarin Saguinus imperator 
subgrisescens, Chimpanzee Pan troglodytes, Fennec fox Vulpes zerda, Guanaco Lama guanicoe, Giraffe Giraffa 
camelopardalis, Eland Taurotragus oryx. 
Total = 9. 
Other international studbook species for which Species Co-ordinators have been appointed to prepare SMPs: 
Brush-tailed bettong Bettongia penicillata ogilbyi, Black-and-white ruffed lemur Varecia v. variegata, Cotton- 
top tamarin Saguinus o. oedipus, Golden lion tamarin Leontopithecus rosalia, Red panda Ailurusfulgens, Snow 
leopard Panthera uncia, Southern white rhinoceros Ceratotherium s. simum. 
Total = 7 .  
Manuscript submitted 18 April 1988 
Population management: theory and practice 
Int. Zoo Yb. (1988) 27: 26-41 0 The Zoological Society of London 
Management of small populations 
THOMAS J. FOOSE1 & JONATHAN D. BALLOU2 
'AAZPA Conservation Co-ordinator, Minnesota Zoological Gardens, Apple Valley, 
Minnesota 55124 and 2~epartment  of Zoological Research, National Zoological Park, 
Smithsonian Institution, Washington, DC 20008, USA 
It appears that the survival of many 
species, especially larger vertebrates, will 
depend upon assistance from captive 
propagation over the next century or 
more. Zoos and aquaria can and must 
serve as arks for vanishing wildlife. 
However, captive propagation can 
truly assist conservation of endangered 
species only if zoo and aquarium 
populations are managed genetically and 
demographically in a manner to reinforce, 
not replace, wild populations. In the 
future, conservation strategies will ideally 
incorporate both captive and wild 
populations that are interactively 
managed for mutual support, that is, 
through regulated interchange of animals 
or at least genetic and demographic 
material (e.g. through sperm or embryos). 
Captive populations can serve as 
reservoirs of genetic and demographic 
material that can be infused periodically 
into remnant wild populations or re- 
established in vacant wildlands if 
conditions were suitable. Reciprocally, 
the wild populations, even if only 
remnants, will still be subjected to natural 
selection and thus maintain some 
CONSERVATION SCIENCE A N D  ZOOS 
semblance of the characteristic genetic 
makeup of the species in the wild. 
Species are vanishing because a 
combination of habitat destruction and 
unsustainable exploitation are reducing 
and fragmenting wildlife populations. 
However. the situation is often worse 
than is indicated by simple numbers 
because of the particular characteristics of 
small populations, which are subject to 
stochastic or random problems that 
imperil their survival even if the other 
dangers can be eliminated (Gilpin & 
Soul&, 1986; Soule, 1987). As a 
consequence, even when and where 
protection of remnant wild populations is 
feasible, it may not be sufficient to ensure 
the long-term survival of species. There 
are three general types of stochastic 
problems: environmental, demographic 
and genetic (Gilpin & Soulk, 1986). 
Stochastic environmental problems are 
of several kinds: e~idemic. disease. 
catastrophic events (natural disasters such 
as earthquakes or floods, fires, wars or, in 
the captive context, loss of financial 
or other support) (Dobson & May, 
1986). For larger populations, such 
environmental disturbances may be 
localised and part of the population 
might survive (Shaffer, 1987). In general 
one of the advantages of captivity 
might be to moderate environmental 
stochasticity, for example drastic 
fluctuations in food supply should be less 
common in captivity than in the wild. 
Demographically, stochastic problems 
afflicting small populations include 
unexpected failures in reproduction or 
survival, distortions of age distributions 
and biases in sex ratio. Severe random 
fluctuations in birth and death rates will 
be troublesome in very small populations. 
The failure of a few 99 to reproduce as 
expected may have little effect on larger 
populations, but could be disastrous for a 
small population with only a few 99. 
Age distributions will also be more 
vulnerable to stochastic destabilisation in 
small populations; extinction through 
- 
senescence has occurred in a number of 
27 
captive groups. Another particular 
problem may be distortions in sex ratio; 
every zoo manager has been confronted 
with the phenomenon of a substantial 
number of consecutive births of 
predominantly one sex. In a small 
population such distorted sex ratios can 
be very disruptive. 
Genetically, small populations tend to 
lose diversity rapidly through the 
stochastic process of genetic drift, as well 
as through inbreeding. Genetic diversity 
is important for both the adaptability of 
populations and the fitness (survival and 
fertility) of individuals. The smaller 
the population, the faster the loss 
(Appendix 1). 
Reduction and fragmentation of small 
populations, be they captive or wild, 
convert gene pools into gene puddles that 
are vulnerable to evaporation in an 
ecological and evolutionary sense. 
Genetic and demographic analysis and 
management can be applied to counteract 
these problems. In general it may be 
easier to apply such measures in captivity 
but the same types of intensive 
management will increasingly be required 
in wild sanctuaries which in reality are 
becoming 'megazoos'. 
Multi-institutional population propa- 
gation programmes for genetic and 
demographic management are being 
developed in several major regions: the 
Species Survival Plan (SSP) of the 
AAZPA in North America; the 
Europaisches Erhaltungszucht Pro- 
gramm (EEP) in Europe; the Joint 
Management of Species Programme of 
the British Zoo Federation and the 
Regional Studbook Programmes of the 
Anthropoid Ape Advisory Council in 
Great Britain; the Australasian Species 
Management Scheme in Australia and 
New Zealand; the SSP programmes in 
Japan. It is hoped that eventually these 
regional programmes can be integrated 
and co-ordinated internationally. The 
Captive Breeding Specialist Group of the 
IUCN Species Survival Commission, in 
co-operation with the International 
CONSERVATION SCIENCE A N D  ZOOS 
Union of Directors of Zoological 
Gardens, would be appropriate bodies to 
co-ordinate such co-operation. 
These SSP-type programmes will each 
be based upon a Masterplan for 
population management. Fundamentally 
a Masterplan provides institution- 
by-institution and animal-by-animal 
recommendations for the entire popu- 
lation encompassed by the programme. 
Individual recommendations should be 
orientated to well-defined goals and 
objectives and should reflect attempts at 
demographic and genetic management. 
Although the emphasis in this paper is on 
genetic and demographic management, 
other considerations such as basic 
husbandry, behavioural aspects and 
veterinary care will also be vital to viable 
captive management programmes. A 
basic protocol has been developed for this 
kind of population management in the 
AAZPA (SSP) programmes (Appendix 3). 
Because genetic and demographic 
problems of small populations are a 
function of the population's size, a first 
step in developing a management 
programme for a captive population is to 
establish a captive carrying capacity. This 
will be the optimum number of 
individuals to be maintained over the 
long term and will represent a 
compromise between the minimum 
necessary for genetic and demographic 
viability and the maximum that can be 
accommodated without excluding other 
taxa from captive programmes. 
The lower limit for carrying capacity is 
the Minimum Viable Population (MVP) 
size of the captive population. An MVP 
depends on two sets of factors: the 
demographic and genetic objectives of 
the programme and the biological 
characteristics of the population. 
Demographic and genetic objectives 
include the probability of the population 
surviving; the kind and amount of genetic 
diversity to be preserved; the period of 
time over which this probability of 
survival and level of diversity are to be 
sustained. 
The perfect programme would ensure 
a 100% probability of survival and 
preservation of the total amount of all 
kinds of genetic diversity in the 
population for ever. In other words, it 
would protect against all the problems 
which confront small populations. 
Beyond the phi?osophical observation 
that nothing is certain or for ever, it is 
unrealistic to consider such a programme 
since the size of populations that can be 
maintained in captivity could never be 
large enough to achieve its objectives. 
A general guideline recommended by 
some conservation biologists is that 
captive programmes should attempt to 
preserve 90% of the average 
heterozygosity of the founders for 200 
years (SoulC et al., 1986). Figure 1 
illustrates several examples of minimum 
effective population sizes that would be 
necessary to preserve 90% average 
heterozygosity for 200 years. 
Biological characteristics The fact that 
the MVP size depends on the generation 
time of the species as well as the number 
of effective founders leads to 
consideration of the second set of factors 
that determine the MVP, the biological 
characteristics of the population. 
1. The generation time: the effect of 
generation time on the MVP size is based 
on the fact that genetic diversity is lost 
generation by generation not year by 
year. Thus for a given period of time, for 
example 200 years, a species with a 
shorter generation time will pass through 
more generations, have more opportunity 
to lose genetic diversity and hence require 
a larger MVP size than a species with a 
longer generation time. 
2. The number of founders: the number 
of effective founders that establishes a 
captive population also determines the 
MVP (Fig. 1). An effective founder is a 
wild-caught animal that has reproduced 
to have descendants in the living, 
managed population. Living animals out 
CONSERVATION SCIENCE A N D  ZOOS 
1 a b c  d 
Fig. 1. Effective population sizes required to preserve 90% average heterozygosity for 200 years for populations 
with various generation times and effective founders. T. generation time; N,. effective population size; a. no 
founder effect; b. 20 founders; c. eight founders; d. six founders. 
of the wild that have not reproduced may 
be potential founders but they are not 
effective until they reproduce. Basically, 
the fewer the number of founders the 
larger must be the MVP. There is, 
however, a point of diminishing returns. 
For most species increasing the number 
of effective founders beyond 20-30 (to 
preserve average heterozygosity) or 30-50 
(to preserve rare alleles from the wild 
population) will not significantly reduce 
the size of the MVP required for time 
periods in the order of 200 years 
(Allendorf, 1986). In most cases founders 
will not enter the population only at the 
start of the management programme; 
periodically, additional founders 
('immigrants') will enter the population. 
As few as one effective wild-caught 
founder per generation can keep the 
captive population representative of a 
wild gene pool. 
3. The N,/N ratio: an MVP prescribes a 
population size required for some set of 
objectives. The population size of 
relevance, however, is not the simple 
census number (N), rather it is the 
genetically effective size, denoted by 
N,. This is a measure of the way in which 
the population is reproducing to transmit 
its genes to the next generation. It is a 
function of a number of variables: the 
number of animals that reproduce; the 
sex ratio of these animals; the mean and 
variance of their relative number of 
offspring (lifetime family sizes) (Lande & 
Barrowclough, 1987). In general the more 
disparate the sex ratio and family size, the 
lower N, will be relative to N. In practical 
terms, N, is usually only a fraction of N. 
Common ratios of N, in genetically 
unmanaged populations are 0-2-0.5. As a 
consequence, actual carrying capacities 
may have to be several times larger than 
the N, necessary to achieve the genetic 
objectives of the programme. For 
example, if an N, of 250 were required for 
some set of objectives (e.g. maintaining 
90% heterozygosity for 200 years), but 
the N, ratio of the population was only 
0.25, the carrying capacity that would 
have to be maintained would be 1000 
individuals. By managing sex ratios and 
family sizes, N,/N ratios can be improved 
and actual (carrying-capacity) MVPs 
reduced. 
4. Growth rate: until a captive population 
attains its carrying capacity and while the 
population size remains small, genetic 
diversity will be lost at a relatively fast 
rate. The more rapidly the population can 
grow to carrying capacity the more it can 
30 CONSERVATION SCIENCE A N D  ZOOS 
N, and carrying capacity necessary for maintaining the specified amount of genetic diversity for a specified time 
period 
YEARS PER GENERATION: 10 NO. GEN. DURING PERIOD: 20 
YEARLY % GROWTH RATE: 1.080 GEN. GROWTH RATE: 2.16 
EFFECTIVE NO. FOUNDERS: 25 GEN. EXPON. GROWTH 0.77 
ESTIMATED Ne/N RATIO: 0.5 
DESIRED % HET. RETAINED: 0.90 
LENGTH OF TIME PERIOD: 200 YEARS 
Effective Size required to maintain desired amount of original variation for the specified 
length of time: 199 
Carrying capacity necessary to maintain desired amount of the original variation over the 
time: 238 
Table 1. An example of population viability analysis. 
reduce the rate at which it loses genetic 
variation. Maximising the rate of growth 
will therefore minimise loss of diversity 
during the growth phase. 
It must be emphasised that there is no 
single or magic number that represents an 
MVP for all species at any time or for any 
species all of the time (Soulk, 1987). 
The MVP size will vary with the 
circumstances of the programme. 
Computer software is available to 
perform the population viability analysis 
(PVA) that will prescribe MVPs required 
for various circumstances; some examples 
of such analysis are presented in Table 1. 
A secondary consideration for 
determination of the MVP is 
demographic stochasticity which is 
significant if the MVP prescribed by 
genetic considerations is fewer than 
50-100. Populations smaller than 50 and 
possibly even 100 may be particularly 
vulnerable to 'crashes' or extinctions due 
to random demographic causes such as 
epidemic diseases, natural disasters or sex 
ratio distortions (Shaffer, 1987; Soulk, 
1987). 
The MVP establishes the lower limit 
for the carrying capacity of a managed 
captive population. It should be evident 
from the preceding discussion that it is 
advantageous for a population to be 
larger than the MVP size; indeed, in this 
respect, more is always better. Enlarging 
any one captive population, however, 
might well exclude other taxa from the 
zoo ark and therefore there has to be an 
upper limit on the carrying capacity for 
any one taxon in captivity. 
The upper limit can be established by: 
1. Assessing how much captive habitat 
(space and resources) is available for 
taxa with similar captive ecologies (i.e. 
forms that have equivalent enclosure 
requirements, resemble each other in 
terms of public expectation of what 
should be in a zoo, etc.). A crude measure 
of the captive habitat available is 
provided by the number of 'ecologically 
similar' specimens currently being 
maintained (Foose & Seal, 1986). 
2. Ascertaining how many taxa with 
similar captive ecologies are in need of 
assistance by propagation in captivity. In 
this regard information from the IUCN 
SSC Specialist Groups will be important. 
The CBSG is already trying to develop 
recommendations for captive priorities 
among several broad groups, including 
psittacines and primates (Oates, 1985). 
3. Allocating the captive habitat to as 
many taxa as possible while still 
maintaining an acceptable MVP for each. 
Table 2 illustrates an attempt to apply 
this type of analysis to one group of 
animals, the large felids. As is the case for 
every broad category examined in this 
way so far, there is not enough captive 
habitat to accommodate acceptable 
MVPs for all the taxa that will need 
assistance to survive. This severe 
limitation of captive habitat argues 
CONSERVATION SCIENCE A N D  ZOOS 3 1 
SPECIES 
EXTANT 
SPP SPP IN RDB 
CAPTIVE 
POPULATION 
NO. O F  SPP IF 
POPULATION 
100 250 500 
Panthera leo 
Panihera iigri.7 
TOTAL lions, tigers 
Panthera onca 
Panihera pardus 
Panthera uncia 
Felis concolor 
Neofelis nebulosa 
Acinonyx jubatus 
TOTAL other large felids 
TOTAL 
Table 2. Capacity of captive facilities for the larger felids calculated from data in ISIS and the appropriate 
studbooks. (An example of analysis to determine capacity of zoos for taxa with similar captive ecologies.) 
strongly for the participation of as much 
of the zoo world as possible in the 
population management programme and 
for international co-ordination of the 
regional efforts. 
Once the carrying capacity is 
established the institution-by-institution 
and animal-by-animal recommendations 
must be formulated. Based on genetic and 
demographic guidelines or criteria for 
management, these recommendations 
form the basis of the SSP-type 
Masterplan. Appendices 1 and 2 describe 
the kinds of genetic and demographic 
analysis and models which are needed for 
population management. Generally the 
objectives will be to develop a genetically 
diverse and demographically stable 
population. 
Genetic management objectives will 
normally be to: (1) adjust the 
representation of founder lineages to 
rectify past disparities, that is, there will 
be an attempt to adjust the existing 
founder distribution in the population to 
the target founder distribution that has 
been established for the population; 
(2) regulate family sizes and sex ratios 
to maximise effective size of the popula- 
tion. Until the disparities in 
founder representation are rectified the 
management programme will deliberately 
reproduce from some animals more than 
from others. When these adjustments are 
completed, the objective will be for every 
animal to produce the same number of 
offspring. An exception to this guideline 
would be if there were a deliberate 
decision to have an unequal sex ratio for 
gregarious species when one family size 
objective will apply to 99 and another to 
33. At carrying capacity each animal 
during its lifetime will be expected to 
produce on average two offspring, 
preferably one of each sex, which in turn 
survive to reproduce; (3) manage 
inbreeding coefficients to ensure that 
survival and fertility are not declining 
significantly. 
Demography management objectives 
will normally be to: (1) expand the 
population from its founder or initial size 
to the carrying capacity as rapidly as 
possible within the constraints of the 
genetic guidelines, that is, regulation of 
family sizes and adjustment of founder 
representation. In cases of very small 
populations, demographic considerations 
will usually override genetic ones (Seal, in 
press); (2) stabilise the population at the 
carrying capacity by some combination of 
regulation of fertility (birth control) 
and survival (removal or culling). 
Programmes for regulation of fertility and 
32 CONSERVATION SCIENCE A N D  ZOOS 
survival are based on analysis of life table 
data. 
The institution-by-institution and ani- 
mal-by-animal recommendations should 
specify which animals should reproduce 
when and with which mate to achieve the 
genetic and demographic objectives. 
These specifications will normally entail 
some relocation of animals between 
institutions to produce better genetic and 
demographic combinations of mates. 
The Masterplan can then also 
determine what the genetic and 
demographic expectations are for each 
individual in the population. Once an 
individual has fulfilled what is expected or 
required of it in demographic and genetic 
terms, it becomes 'surplus' to the 
management programme. When this 
occurs, the individual should not 
reproduce again in or for the managed 
population. 
Finally, it must be stated that genetic 
and demographic analysis are possible 
only if adequate data are available. 
Compilation of such data is the purpose 
of studbooks and the mission of ISIS and 
of the regional inventory systems around 
the world, all of which are thus vital to 
the conservation programmes of zoos. 
heterogeneous population 
QQ 
APPENDIX 1 
BASIC POPULATION GENETICS FOR 
CAPTIVE MANAGEMENT 
Genetic diversity The many thousands 
of genes carried by an animal, interacting 
with the environment, ultimately control 
the structure and function of the 
organism. The site of a gene on a 
chromosome or elsewhere in the cell is 
known as its locus and the terms gene and 
locus tend to be used interchangeably. 
Almost all animals are diploid, carrying 
two copies of each gene which can occur 
in alternative forms or alleles. Different 
alleles may produce totally different 
effects in the organism. Where more than 
one allele exists there is genetic diversity. 
The genes of an individual are known as 
its genome. Collectively all the alleles for 
all the genes of all the individuals of a 
population or species constitute its gene 
pool. 
Genetic diversity can occur at both the 
level of the individual and of the 
population (Fig. 2). If the two copies of a 
gene carried by an individual represent 
different alleles, the individual has genetic 
diversity and is known as heterozygous. If 
both copies are the same allele, there is no 
homogeneous population 
0-2--7 
homorygous tndivlduai heterozygous lndlvldual 
Fig. 2. Genetic diversity at both the individual and population level. 
CONSERVATION SCIENCE AND ZOOS 33 
diversity and the individual is known as 
homozygous. Populations containing 
genetic diversity will have some 
heterozygous and some homozygous 
individuals. The terms heterogenous or 
polymorphic are sometimes used to 
describe a population with genetic 
diversity; homogenous is the analogous 
term for a population without diversity. 
At population level genetic diversity can 
be measured as allelic diversity or as 
average heterozygosity. 
Genetic diversity is important for both 
individuals and populations. In 
populations it is needed in order to 
adapt to changing environments. For 
individuals heterozygosity is important 
for maintaining fitness, that is, the ability 
to survive and reproduce (Allendorf, 
1986: Hedrick et al.. 1986). 
The gene pools of a captive population 
can best be visualised through the genetic 
lineages or bloodlines that descend from 
the founder animals, that is, the animals 
from the wild that reproduce to have 
descendants in the living population. For 
captive populations, the original gene 
pool consists of all the genes that are 
carried by the founders. Preserving 
genetic diversity is thus tantamount to 
preserving as many as possible of these 
founder genes in the population for as 
long as possible. 
Genetic drift In small populations, 
alleles may be lost entirely from the gene 
pool through genetic drift, that is, the 
random or stochastic process which 
results when a limited and therefore 
incomplete sample of genes from one 
generation is selected for transmission to 
the next. In other words, reproduction 
constitutes a random draw of some of the 
alleles in the population for perpetuation 
in the offspring; alleles which are not 
passed on to any of the offspring are lost 
to genetic drift. 
Founder representation An animal 
transmits a copy of only one of each pair 
of genes to its offspring which thus 
receives half of its genes from its sire and 
half from its dam. Thus by constructing 
an individual's pedigree back to its 
founders it is possible to determine what 
founder genes the living individual 
contribution 
0.4 
212 39 211 4 0  1 231 5 2  5 18 17 11 dom 12 
founder 
Fig. 3. Founder representation in the Przewalski horse Equus przewalskii SSP population at 1 March 1988. 
Cross-hatching = founder contribution; shading = target founder contribution; dotted line = parity. 
CONSERVATION S C I E N C E  A N D  ZOOS 
Fig. 4. An example of bottlenecks in a pedigree. 
Circles represent QQ, squares 3 3 ;  the black borders 
show deceased individuals; numbers 1-5 represent 
wild-bred founders; numbers 6-16 are their captive- 
bred descendants. 
actually carries, that is, its founder 
representation. Summation of the founder 
representation over all founders provides 
the distribution of each founder genome 
in the population (Fig. 3). If the situation 
were this simple, maximising 
representation of genetic diversity over 
time would be a matter of equalising 
founder representation but the process is 
more complicated. Simple founder 
representation does not acknowledge that 
each founder actually carried two copies 
of each gene, that is, the founders are 
diploid not haploid. (For simplicity these 
two copies are referred to as two alleles in 
each founder even though they may not 
be different forms of the gene. Thus it is 
conventional to discuss founder alleles, 
not founder copies of genes.) 
Because only one copy of each gene is 
transmitted from parent to offspring and 
because which of the two alleles carried 
by a parent is actually transmitted is a 
matter of chance, problems can be caused 
when a line of descent passes through 
very few individuals or perhaps even one 
animal. Such bottlenecks may prevent the 
perpetuation of one of the two founder 
alleles from that point onwards. As a 
result only part of the genome of some 
founders may actually survive in the 
living population while all the alleles of 
other founders may still be present. In the 
extreme example of a bottleneck of one in 
the F, generation from a founder, at most 
only one allele from each locus, or 50% 
of the genome if all loci are considered, 
could actually be transmitted to future 
generations (Fig. 4). 
Genetic management In order to 
maximise preservation of genetic diversity 
(i.e. maintain as many of the founder's 
alleles as possible), genetic management 
should attempt to develop representation 
of founder lineages that are proportional 
to the percentage of the founder genomes 
that still survive. Again, referring to the 
extreme example of a bottleneck of one, 
the ideal representation of this founder in 
the population should be half as much as 
for founders whose entire genomes 
survive. 
It is possible from computer 
algorithms, such as gene drop programs, 
to estimate the probable loss of founder 
genes through pedigrees (MacCluer et al., 
1986). For any living population it is then 
possible to compute a distribution of the 
proportion of each founder's genome that 
survives. Based on this distribution, a 
target or desired distribution of founder 
representations can be calculated. 
Matings that will adjust the existing 
founder lineage representations towards 
the target distribution can then be 
recommended. 
Loss of alleles will obviously reduce the 
diversity of the population as a whole. 
Once the alleles have disappeared they 
can be recovered only through the very 
slow process of mutation or through 
migration from another population (if 
there is one). Attempts to prevent the loss 
of alleles from the population therefore 
have priority in genetic management. 
The way in which alleles are arranged 
in individuals is also important. In small 
populations, animals are more likely to 
breed with relatives resulting in increased 
CONSERVATION SCIENCE A N D  ZOOS 
levels of inbreeding. Inbred organisms 
have higher levels of homozygosity since 
there is a greater probability that an 
identical copy of an allele at any one 
locus will be received from its sire and 
dam. These alleles are said to be identical 
by descent. Thus the alleles that do 
remain in the population tend to become 
organised into more homozygous than 
heterozygous individuals. Since increased 
levels of homozygosity can expose 
potentially deleterious recessive alleles, 
the result can be so-called inbreeding 
depression, that is, a reduction in the 
animal's ability to survive and reproduce 
compared with that of the non-inbred 
animal. Inbreeding depression has been 
demonstrated in many captive popu- 
lations (Ralls & Ballou, 1983) and 
management must therefore consider 
inbreeding levels in captive populations. 
Preventing loss of alleles and avoiding 
inbreeding are not always equivalent; they 
may even be in conflict. In most such 
cases the prevention of loss of alleles will 
have the higher priority unless inbreeding 
depression is severe enough to endanger 
the continued survival of the population 
(Templeton & Read, 1984). 
Recommended reading on genetic 
management and conservation biology 
includes Soule & Wilcox, 1980; Frankel & 
SoulC, 1981; Foose, 1983; Allendorf & 
Leary, 1986; Ralls & Ballou, 1986. 
APPENDIX 2 
BASIC DEMOGRAPHY FOR CAPTIVE 
MANAGEMENT 
Population management requires the 
integration of both genetic and 
demographic analyses. For example, 
generation length, which is calculated 
from survival and fecundity rates, is 
essential for estimates of MVP size 
requirements while age distribution, sex 
ratios and reproductive rates are used to 
calculate effective population sizes. 
This Appendix is essentially a basic 
demographic primer covering some of the 
fundamental concepts necessary for 
conducting demographic analyses of 
captive populations. Of primary 
importance are life tables, generation 
length, population growth rates and 
stable age distributions. 
LIFE TABLES 
Life tables contain information on age- 
specific survival and fecundity rates in the 
population. Usually these rates are 
calculated separately for each sex. Age 
classes are typically represented as yearly 
intervals. For each age class, the 
following demographic parameters are 
calculated: 
Mortality rate (4,) is the proportion of 
individuals that die during age class x. It 
is calculated from the number of animals 
that die during an age class divided by the 
number of animals that were alive at the 
beginning of the age class (i.e. the number 
of animals 'at risk' during the age class) 
(Table 3). Individuals still alive in the age 
class are not included in the q, calculation 
because they have not yet lived through 
the entire age class. 
Age-speclfic survival rate (p,) is the 
proportion of individuals surviving from 
the beginning of the age class (x) to the 
beginning of the next age class (x + 1). It 
is simply 1 - q, (Table 3). 
Age-specific survivorship (I,) is the 
proportion of individuals surviving from 
birth to the beginning of the age class. 
The 1, for the first age class (denoted age 
class 0 because it includes animals aged 
0-1 years) is 1.00; 100% of the 
individuals survive to the beginning of 
age class 0. The I, values are most simply 
calculated from the p, values and are the 
product of all p, values from age class 0 
up to, but not including, the age class for 
which the I, is being calculated (Table 3). 
Fecundity rate (m,) is the average 
number of same-sexed young born to (or 
sired by) animals in that age class. For 
example, in a life table for 99 the m, 
36 CONSERVATION S C I E N C E  A N D  ZOOS 
CLASS 
NO OF 
NO.OF NO. OF 9 
?P DEATHS BIRTHS 
AGE-SPECIFIC 
MORTALITY SURVIVAL 
RATE RATE 
4. P, 
AGE-SPECIFIC 
SURVlVORSHlP 
RATE 
f x  
FECUNDITY 
RATE 
m. fxm, 
Net Reproductive Rate (R,) = 1.21 
Table 3. Calculation of 9 life-table data from information on 9 births and deaths in a hypothetical population. 
would refer to the average number of 9 
offspring born to 99 in age class x. The 
m, values are calculated by dividing the 
number of 9 (or 3 )  births by the number 
of 99 (or d d )  alive at the beginning of an 
age class (Table 3). The fecundity rates 
provide information on the age of first 
and last reproduction, and ages of 
maximum reproduction. 
Table 3 shows a simple life table for a 
population consisting of five age classes, 
aged 0-5. Values of q,, p,, 1, and rn, are 
calculated for each age class from data on 
numbers of deaths and births in a 
hypothetical population. Values of I ,  and 
m, can be plotted to illustrate graphically 
the survivorship and reproductive rates in 
the population (Figs 5 and 6). 
These data are used to calculate values 
of generation time, population growth 
rate and stable age structure for the 
population. 
survivorship rate 
GENERATION LENGTH 
Conceptually, the generation length is the 
age at which an individual 'replaces' itself 
in the population. Technically, it is 
defined as the average age at which a 9 
(or 3 )  produces offspring (Caughley, 
1977). If data were available on the age of 
each parent when it produced young, the 
average of these ages would be the 
generation length of the population. 
Generation length is often incorrectly 
thought of as the age of first 
reproduction; however, this under- 
estimates generation lengths in most 
cases. Males and 99 often have different 
generation lengths. 
The generation length can also be 
calculated direct from the I ,  and rn, life 
table data. The 1, values provide the 
proportion of animals surviving to each 
age class while the rn, values provide how 
many young these survivors produce. The 
product of lXm, is therefore the 
fecundity rate 
' -OF  
a g e  c l a s s  a g e  c l a s s  
Fig. 5. Age-specific survivorship, I,. Fig. 6. Age-specific fecundity, m,. 
CONSERVATlON SCIENCE A N D  ZOOS 
age class 
Fig. 7. The product of I p , .  
contribution of age class x to total 
reproduction. Figure 7 shows the 
distribution of Ixmx values for the 
hypothetical population in Table 3. The 
generation length is the mean age of this 
distribution and is calculated by: 
C xExmx Generation Length (T) = -C lxmx 
Therefore, the generation length is the 
weighted average of reproductive age 
classes where the weights are the lxmx 
values. 
2.16 
For the example: T = --- = 1.79 
1.21 
which corresponds with the apparent 
average in Fig. 6. 
POPULATION GROWTH RATES 
The life table data provide accurate 
estimates of the future growth rate of the 
population only if survival and 
reproductive rates remain constant over 
time. Since population growth rates are a 
function of birth and survival rates, the 
reproductive values again provide the 
basic data needed to calculate growth 
rates. 
If a generation of 90 (or $3) lived out 
their lifetimes according to the survival 
and fecundity rates shown in the life 
table, how many Q (or $) offspring would 
they leave to replace themselves? In other 
words, how would the population size 
change from one generation to the next? 
The sum of lxmx values over all ages is the 
average number of 9 (or 6) offspring 
surviving to replace each individual and 
therefore provides an estimate of the 
population's growth rate per generation. 
The sum of Ixmx values in the example is 
1.21 (Table 3). Therefore, on the average, 
each Q leaves 1-21 Q young in the next 
generation and the population grows at 
21% per generation. This sum is called 
the net reproductive rate (R,). 
If each Q were only to replace herself 
each generation, the net reproductive rate 
would be 1.00 and the population would 
remain the same size. The R, in growing 
populations is greater than one; declining 
populations have an R, of less than one. 
A more useful measure is the growth 
rate per year (A). This is calculated direct 
from the growth rate per generation: 
For the example, with R, at 1.21 and T at 
1.8, the yearly growth rate is : 
A = 1.21(1/1'8)1.11, or 11% per year. 
Another commonly used measure of 
yearly growth rate is the intrinsic rate of 
increase, or r. It is the exponential yearly 
growth rate of the population and is 
related to A in the following way: 
For the example: 
STABLE AGE DISTRIBUTION 
If survival and reproductive rates remain 
constant over time, the proportion of 
animals in each age class will become 
constant from one year to the next and 
the population will achieve a Stable Age 
Distribution (SAD); the population is 
stable. The SAD is also determined 
entirely from the survival and fecundity 
rates and each unique combination of 
rates will result in a unique SAD. In a 
stable population, the proportion of the 
population in age class x is: 
C O N S E R V A T l O N  S C I E N C E  A N D  Z O O S  
AGE CLASS K ' I x  cx 
0- 1 1 .O 48% 
1-2 0.54 26% 
2-3 0.4 1 20% 
3-4 0.15 7% 
4-5 0 0% 
Total = 2.1 101% 
Table 4. 9 stable age distribution (C,) calculated from 
life table values shown in Table 3 and a yearly growth 
rate of 11%. 
SAD for the example is shown in Table 4. 
Since survival and reproductive rates can 
differ between sexes, so can their SAD. 
OTHER APPLICATIONS AND POPULATION 
MANAGEMENT 
Several other demographic parameters 
can be calculated from life-table data. 
One demographic question of particular 
interest to population management is how 
to manage reproductive and survival rates 
to achieve zero population growth (i.e. 2 
= 1.00) for a population at carrying 
capacity. 
One simple option is to examine the 
effect of delaying age of first reproduction 
to a later age class. In the example, if age 
of first reproduction were delayed to age 
class 2, the R, would be reduced to 0.85 
(the sum of the l,m, values for ages 2, 3 
and 4). This would reduce the generation 
growth rate to below self-sustainment; the 
population would be reduced by 15% per 
generation; the generation length would 
be extended to 2.1 years and the per year 
growth rate would be 0.92 (8% reduction 
per year). Clearly this is too much of a 
reduction in reproduction. However, if 
the age of first reproduction were delayed 
for only 50% of the individuals, the m, 
value in the example for age class 1 would 
be 0.3, the R, reduced to 1.03, the 
generation length increased to 1.9 years, 
and the yearly growth rate reduced to 
1.02 (a 2% increase per year). This is 
obviously a more appropriate alternative 
than delaying age of first reproduction for 
all individuals. 
Similar types of manipulation are 
possible for estimating survival rates 
necessary for zero population growth. 
Rather than guess what levels of 
reproduction or survival rates will achieve 
;1 of 1.00, however, methods for 
calculating exact m, and 1, values are 
available (Goodman 1980). 
These basic demographic concepts 
summarise some of the primary 
calculations used for demographic 
management. When analysing real 
populations, however, modifications of 
basic methods are often necessary. For 
example, calculations of generation length 
need to be modified for populations with 
largely overlapping generations and life 
tables might be calculated differently for 
seasonal versus aseasonal breeders 
(Caughley, 1977). Several computer 
programs are currently available for 
calculating and modeling demographic 
data for captive populations (Ballou & 
Bingaman, 1986; Bingaman & Ballou, 
1986; Rockwell & Teare, 1986; Flesness & 
Scobie, 1987). Most of these require 
studbook data in pre-specified format 
(ISIS ARKS format or Omaha Studbook 
format). Methods for statistical analyses 
of life-table data are included and 
discussed in Lee (1980). 
Further useful references include: 
Caughley, 1977; Keyfitz, 1968; Goodman, 
1980; Mertz, 1970; Beddington & Taylor, 
1973; Foose, 1983. 
APPENDIX 3 
THE AAZPA SSP PROTOCOL FOR 
DEVELOPMENT OF A POPULATION 
MANAGEMENT MASTERPLAN 
DATA COMPILATION 
The first step in the development of an 
SSP Masterplan is to compile the basic 
data required for population analysis. 
This compilation will often be in the form 
of a studbook. However, ISIS should be 
involved in the compilation process: 
initially as a source of some of the data 
for studbook development; ultimately as 
CONSERVATION SCIENCE A N D  ZOOS 
a repository of the assembled data. An 
important part of the compilation process 
is a 'clean up' of the ISIS data. 
The basic data required on each animal 
for population analysis and management 
are: (a) individual identification (a simple 
numeric lifetime identity); (b) sex; (c) 
birth date; (d) death date; (e) parentage 
(if captive born); ( f )  place of capture (if 
wild caught); (g) institutions/facilities 
where it has been held, with dates; (To 
achieve this identification, it may be 
necessary to link a series of different ID 
numbers the animal has had as it moved 
from one institution to another in its 
captive history, e.g. the local ISIS 
specimen ID numbers.) (h) available 
information on circumstances of death. 
With these data, genetic and 
demographic analyses can be performed. 
GENETIC ANALYSES 
(a) construct the pedigree for each animal 
in the population; (This process may be 
the construction of a pedigree chart; more 
often it will be an inherent part of various 
algorithms and computer programs, e.g. 
the additive relationship matrix or 
various 'gene drop' computations.) 
(b) identify all the founders of the 
population; (A founder is an animal 
which is from outside the population, 
usuallv the wild. and which has no known 
relationship to any other individual at its 
time of entry into the population and has 
. descendants in the living population.) 
(c) compute the representation of 
founders ('bloodlines'), or preferably the 
probable distribution of founder alleles, 
in living individuals and the present 
population as a whole; (d) locate any 
extreme bottlenecks in the history of 
particular founder lineages or bloodlines 
and compute the proportion of each 
founder's genome that has survived to the 
living population (see Appendix 1); (This 
step may be an inherent part of more 
sophisticated algorithms that calculate 
probable distributions of founder alleles 
rather than iust crude founder 
representation.) (e) calculate the founder 
representation or founder allele 
distribution in offspring of the possible 
matings of living members of the 
population; (f) determine the number and 
sex ratio of animals that actually 
reproduce in the population; (g) calculate 
the number of offspring of each living 
individual in the population and hence 
the mean and variance of lifetime family 
sizes; (h) estimate the genetically effective 
population size (N,) of the population 
and then the N,/N ratio, where N is the 
total number of animals in the 
population; (i) calculate the inbreeding 
coefficients of existing individuals in the 
population and of the potential offspring 
of possible matings between these 
animals; Cj) conduct various biochemical 
analyses as needed that measure genetic 
variability, genetic distance and identity 
(e.g. electrophoretic, DNA and karyo- 
typic studies). 
DEMOGRAPHIC ANALYSES 
(a) determine the size of the current 
population and the number of institutions 
over which it is distributed; (Usually it 
will also be necessary to obtain the same 
kind of information for other taxa with 
similar 'captive ecologies', that is, space 
and resource requirements, but in less 
detail, for example the Siberian tiger SSP 
needs to be cognisant of the other tiger 
and large felid populations.) (b) 
determine the age and sex structure of the 
population; (c) compute the age-specific 
survivorships and fertilities of the 
population, that is, construct a life table 
(see Appendix 2); (d) establish a carrying 
capacity that is a compromise between a 
minimum viable population (MVP) for 
genetic and demographic viability and a 
maximum number that will not preclude 
other taxa from the zoo ark; (This 
carrying capacity should be based on 
the programme's goals as well as the 
biological characteristics of the popu- 
lation and should specify the number 
not only of animals but of the facilities 
over which they should best be 
distributed. In the absence of more 
CONSERVATION SCIENCE A N D  ZOOS 
refined or species-specific recommen- 
dations on the long-term genetic 
objectives, the guideline of maintaining 
90% of the founders' heterozygosity 
for 200 years may be used as a crude 
starting point.) (e) using the survival and 
reproductive rates from (c), calculate: the 
rate of change, that is, the growth or 
decline, of the population; the capacity of 
the population for self-sustainment; 
whether the population is at, or when it 
will be at, the carrying capacity; 
how the fertilities and survivorships 
can be managed by 'removals' of 
animals and regulation of reproduction 
(birth control) to stabilise the popu- 
lation at the desired carrying capacity; 
(This process may entail much 'what if..?' 
analysis to determine how management's 
modifications to the patterns of 
survivorship and fertilities will affect popu- 
lation size, growth rate, age distribution, 
etc.) (f) if survivorships and fertilities 
are not adequate for the population 
to be self-sustaining, devise appropriate 
research and husbandry programmes 
to resolve the problems. 
POPULATION MANAGEMENT 
Once genetic and demographic analyses 
are performed, an SSP Masterplan for 
propagation and management of the 
population can be formulated. The SSP 
Masterplan should provide institution-by- 
institution and animal-by-animal recom- 
mendations for every individual in 
the population maintained by SSP 
participants. Specifically, the Masterplan 
should: (a) designate which animals are 
surplus because they are: from over- 
represented bloodlines or lineages, too 
old to reproduce, have already produced 
their share of offspring and have attained 
the oldest age class necessary or allowable 
for a stable age distribution in the SSP 
population; (b) state explicitly that 
surplus animals should not be allowed to 
reproduce again; (Further recommen- 
dations on disposal of surplus will vary 
from programme to programme, time to 
time and institution to institution. In this 
connection the issue of euthanasia will 
have to be confronted.) (c) recommend 
which animals should reproduce, when (a 
schedule over at least the next one to five 
years is needed) and with which mate 
(identify specific individuals and any 
recommended shipments of animals); (d) 
explain the genetic and demographic 
analyses and objectives on which the 
surplus and reproduction recommen- 
dations are based; (There should also be 
an explanation of how the Masterplan 
arrived at the particular carrying capacity 
established.) 
Normally, these genetic and 
demographic guidelines will include: an 
attempt to expand rapidly and stabilise 
the population at its established carrying 
capacity and a strategy to maximise 
preservation of genetic diversity. 
Currently, the best methods to achieve 
these objectives seem to be to: (1) adjust 
representation of founder lineages to be 
proportional to the probable distribution 
of alleles surviving from founders at the 
initiation of the programme; (2) equalise 
lifetime family sizes; (This process will 
become fully operative only when the past 
inequalities in founder representation 
have been corrected.) (3) manage 
inbreeding coefficients; (4) perhaps sub- 
divide the population into several parts 
or demes between which gene flow (i.e. 
usually exchange of animals but also 
increasingly of gametes or embryos) is 
regulated. 
HUSBANDRY STANDARDS 
Husbandry standards for the taxon 
should be developed, culminating in a 
Handbook which can be kept current as 
new advances occur. 
REVIEW AND RATIFICATION 
Once the SSP Masterplan is formulated, 
it should be reviewed, revised if 
appropriate and ratified by the AAZPA 
Propagation Group. The Masterplan 
should then be submitted to the SSP 
Subcommittee for their evaluation, 
recommendations and endorsement. 
C O N S E R V A T I O N  S C I E N C E  A N D  Z O O S  
IMPLEMENTATION 
Once approved, the Masterplan should be 
distributed to each of the participating 
institutions through its institutional 
representatives. The Species Co- 
ordinators and Propagation Group 
should provide follow-up to encourage 
and facilitate implementation. 
REFERENCES 
ALLENDORF, F. (1986): Genetic drift and the loss of 
alleles versus heterozygosity. Zoo Biol. 5: 18 1-190. 
ALLENDORF. F. W. & LEARY, R. F. (1986): 
Heterozygosity and fitness in natural populations of 
animals. In Conservation biology: the science of 
scarcity and diversity: 57-76 Soul&, M. E. (Ed.). 
Sunderland, MA: Sinauer Associates. 
BALLOU, J.D. & BINGAMAN, L. (1986): ?'EDDEMO: 
dBase III life-table generator. Washington, DC: 
National Zoological Park. 
BEDDINGTON, J. R. & TAYLOR, D. B. (1973): Optimum 
age specific harvesting of a population. Biometries 
29: 801-809. 
BINGAMAN, L. & BALLOU, J. D .  (1986): DEMOGPHY: 
Lotus 1-2-3 spreadsheet demographic model. 
Washington, D.C.: National Zoological Park. 
CAUGHLEY, G. (1977): Analysis of vertebrate 
populations. New York: John Wiley & Sons. 
DOBSON. A. P. & MAY, R. M. (1986): Disease and 
conservation. In Conservation biology: the science of 
scarcity and diversity. 345-365. SoulC, M. E. (Ed.). 
Sunderland, MA: Sinauer Associates. 
FLESNESS, N. R.  & SCOBIE, P. N. (1987): User manualfor 
ARKS Animal Records Keeping System. Apple 
Valley, MN: International Species Inventory System 
(ISIS). 
FOOSE, T. J. (1983): The relevance of captive 
propagation to the conservation of biotic diversity. 
In Genetics and conservation: 374-401. Schonewald- 
Cox, C. M., Chambers, S. M., MacBryde, B. & 
Thomas, L. (Eds). Menlo Park, CA: 
Benjamin/Cummings. 
FOOSE, T. J. & SEAL, U.  S. (1986): Species Survival Plans 
for large cats in North American zoos. In Cats of 
the world 173-198. Miller, S. D. & Everett, D. D. 
(Eds). Washington, DC: National Wildlife 
Federation. 
FRANKEL. O. H.  & SOULE, M. E. (1981): Conservation 
and evolution. Cambridge: Cambridge University 
Press. 
GILPIN,  M.  E. & SOULE, M.  E. (1986): Minimum viable 
populations: processes of species extinction. In 
Conservation biology: the science of scarcity and 
diversity: 19-34. Soul&, M. E. (Ed.). Sunderland, 
MA: Sinauer Associates. 
GOODMAN, D. (1980): Demographic intervention for 
closely managed populations. In Conservation 
biology: 171-195. Soule, M. E. & Wilcox, B. A. 
(Eds). Sunderland, MA: Sinauer Associates. 
HEDRICK, P. W., BRUSSARD, P. F., ALLENDORF, F. W., 
BEARDMORE, J. A. & ORZACK, S. (1986): Protein 
variation, fitness and captive propagation. Zoo Biol. 
5: 91-99. 
KEYFITZ, N.  (1968): Introduction to the mathematics 
of populations. Reading, MA: Addison-Wesley. 
LANDE, R. & BARROWCLOUGH, G.  (1987): Effective 
population size, genetic variation, and their use in 
population management. In Viable populations for 
conservation: 87-124. Soule, M. E. (Ed.). 
Cambridge: Cambridge University Press. 
LEE, E. T. (1980): Statistical methodsfor survival data 
analysis. Belmont, CA: Lifetime Learning 
Publications. 
MACCLUER, J. W., VANDEBERG, J. L., READ, B. & RYDER, 
o. A. (1986): Pedigree analysis by computer 
simulation. Zoo Biol. 5: 147-160. 
MERTZ, D. B. (1970): Notes on methods used in life- 
history studies. In Ecology and ecological genetics: 
4-17. Connell, J. H., Mertz, D. B. & Murdoch, W. 
W. (Eds). New York: Harper & Row. 
OATES, J. F. (1985): Action plan for African primate 
conservation: 1986-1990. Washington, DC: WWF 
Primate Program and IUCN/SSC Primate Specialist 
Group. 
RALLS, K. & BALLOU, J. D. (1983): Extinction: lessons 
from zoos. In Genetics and conservation: 164-184. 
Schonewald-Cox, C. M., Chambers, S. M. 
MacBryde, B. & Thomas, L. (Eds). Menlo Park, 
CA: Benjamin/Cummings. 
RALLS, K .  & BALLOU, J .  D. (1986): Preface to the 
proceedings of the workshop on genetic 
management of captive populations. Zoo Biol. 5: 
81-86. 
ROCKWELL, R. & TEARE, A. (1986): Studbook 
management system. Omaha, NE: Henry Doorly 
zoo. 
SEAL. U.  S. (In press). Proceedings of workshop on 
reproductive biology of black-footed ferrets and small 
population biology as they relate to conservation. 
New Haven, CN: Yale University Press. 
SHAFFER, M (1987): Minimum viable populations: 
coping with uncertainty. In Viable populations for 
conservation: 69-86. SoulC, M. E. (Ed.). Cambridge: 
Cambridge University Press. 
SOULE, M. E. (Ed.) (1987): Viable populations for 
conservation. Cambridge: Cambridge University 
Press. 
SOULE, M.. GILPIN, M., CONWAY, W & FOOSE, T .  (1986): 
The millenium ark: how long a voyage, how many 
staterooms, how many passengers? Zoo Biol. 5: 
101-1 13. 
SOULE, M. E. & WILCOX, B. A. (1980): Conservation 
biology. Sunderland, MA: Sinauer Associates. 
TEMPLETON, A. R. & READ, B. (1984): Factors 
eliminating inbreeding depression in a captive herd 
of Speke's gazelle. Zoo Biol. 3: 177-199. 
Manuscript submitted 15 April 1988