6 Genetic Diversity – Understanding Conservation at Genetic Levels

8/9/2019 6 Genetic Diversity Understanding Conservation at Genetic Levels

1/32

6Genetic Diversity Understanding Conservationat Genetic Levels

In the past two decades, a new field of conservation genetics has emerged with two general goals:(a)the precisedescription of genetic changes affecting population survival that occur during range and population contraction ;and (b)application of genetic insight to successful management of threatened populations.

Stephen J. OBrien 1996

153

In this chapter you will learn about:

1. Why genetic concerns are central to the study of

conservation biology

2. How to measure the genetic attributes of individuals

and populations

3. Why inbreeding can pose threats to population

persistence

4. The roles of hybridization and introgression in

conserving biodiversity

6.1. Genetics and Conservation:An Essential Integration

Conservation biology is a science concerned with the fateof populations, which are defined and identified by theirgenetic constituency. This unique genetic makeup not onlydistinguishes them from other populations, but also deter-mines their capacity to adapt to changing conditions and,potentially, to produce new species. Many conservationists

6.1. Genetics and Conservation: An Essential Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1536.2. Conservation Genetics and Conservation Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1546.3. Measuring Genetic Diversity in Populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.3.1. Foundational Measures of Genetic Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

6.3.2. The Loss of Genetic Diversity over Time: Bottlenecks and Genetic Drift. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1576.3.3. Genetic Drift and Effective Population Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1606.3.4. Bottlenecks, Small Populations and Rare Alleles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

6.4. The Problem of Inbreeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1626.4.1. What Do We Mean by Inbreeding and How Would We Measure It? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1626.4.2. The Problem of Inbreeding Depression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1636.4.3. Measures of Inbreeding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164

6.5. Inbreeding and Outbreeding in Population Subunits: Estimation of Gene Flow and Metapopulation Genetics. . . . . . . . . . . . . . . . . . . . . 1656.5.1. Historical Development of Gene Flow Theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1656.5.2. Current Models of Gene Flow: Predictions and Implications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1666.5.3. Models of Recolonization: Propagule Pools and Migrant Pools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

6.6. Can Inbreeding Cause Extinction?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1706.6.1. Laboratory Experiments and Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1706.6.2. Field Studies of Inbreeding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1716.6.3. Inbreeding was a Cause of Extinction in Butterfly Populations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

6.6.4. Inbreeding Effects Environmental and Demographic Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1736.7. Hybridization and Introgression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

6.7.1. Hybridization and Introgression in Animals: The Case of the Red Wolf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1746.7.2. Importing Genetic Diversity: Genetic Restoration of Inbred Populations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1756.7.3. Hybridization in Plants Conservation Threat or Conservation Asset?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1776.7.4. Introgression from Genetically Modified Organisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

6.8. Outbreeding Depression. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1806.9. Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181


2/32

154 6. Genetic Diversity Understanding Conservation at Genetic Levels

would argue that the conservation of genetic diversity isthe foundational basis of all conservation efforts becausegenetic diversity is requisite for evolutionary adaptation,and such adaptation is the key to the long-term survivalof any species (Schemske et al. 1994). To assure suchsurvival, conservation biologists have two primary goalsin the area of genetics. One is to preserve significant

amounts of heritable genetic variation, particularly insmall populations threatened with extinction. The otheris to prevent the fixation of deleterious alleles, a fixationthat can contribute to reduced fitness and accumulation ofharmful mutations (Lynch 1996). Preserving high levelsof heritable variation helps to retain a populations cur-rent reproductive fitness and maintain its evolutionarypotential, its capacity to adapt to environmental changeover the long term. Preventing the fixation of deleteri-ous alleles is intended to prevent declines in survivorshipand fecundity that often occur in small populations as aresult of reduced genetic diversity. Thus, the two goals areintimately related, and the overall aim and application of

conservation genetics is to preserve species not simply asstatic forms, but as dynamic entities capable of respondingto and coping with environmental change through time.Only when species possess this kind of adaptive poten-tial do they have a reasonable expectation of persistencein a changing world, and only through maintaining theirgenetic diversity can they hope to possess this potential.

To achieve these goals, conservation genetics todayencompasses three categories of activities: (1) geneticmanagement of small populations to maximize the reten-tion of genetic diversity and minimize inbreeding, (2)resolution of taxonomic uncertainties and delineation

of management units based on genetic characteristics ofpopulations, and (3) use of genetic analyses in forensics,especially in the enforcement of conservation laws andtreaties, and in understanding the biology of target species.We will examine each category in detail in this and thenext chapter to understand and appreciate the significanceof the kinds of approaches and techniques that can beused in each of these categories. In this chapter, we willdevelop a conceptual understanding of conservation genet-ics, including its history, development, and theoreticalframework. In the following chapter (Chapter 7), we willaddress more specific applications of genetic knowledge,

skills, and techniques in actual conservation management.We begin with an overview of the foundations of conserva-tion genetics.

6.2. Conservation Geneticsand Conservation Biology

In the 1960s and early 1970s, even before conservationbiology emerged as a distinct discipline, genetic concernsabout small populations were growing. Moore (1962)

and Hooper (1971) considered problems associated withinbreeding depression that could arise in populationsconfined to refuges. However, Hoopers conclusionsminimized the risk of inbreeding by asserting that smallamounts of immigration could stem inbreeding depres-sion and that inbreeding itself facilitated the adaptationof local population subunits to particular environments.

Small and declining populations also raised increasingconcern over the potentially deleterious effects ofgeneticdrift, the random fluctuations in gene frequencies thatoccur as a result of non-representative combinations ofgametes created during breeding. Geneticist R. J. Berrynoted that loss of genetic variation through drift couldlimit future adaptation to environmental change, butultimately concluded that natural selection would havesufficient strength to overcome the deleterious effects ofgenetic drift (Berry 1971).

Later investigators who examined the connectionsbetween genetics and conservation came to radically dif-ferent conclusions. In Michael Souls classic review of

potential genetic liabilities for small populations entitledThe Epistasis Cycle: A Theory of Marginal Populations(Soul 1973), Soul noted six factors that account for lossof genetic variation in marginal populations: inbreeding,reduced gene flow with other populations, genetic drift,problems associated with effective population size (thesize of an ideal population that would undergo thesame amount of genetic drift as the actual population),reduced variation in niche width, and directional selec-tion (Soul 1973). Soul demonstrated that, in popula-tions of fruit flies (Drosophila), marginal populationsrarely possessed novel or unique gene arrangements and

many had reduced allelic diversity. Although Soul madeno explicit connections to conservation, he laid the foun-dation for what would become the principle genetic para-digm for concerns about small populations. Marginalpopulations, wrote Soul, are prone to severe reduc-tion in numbers and can experience intermittent drift. Justa trickle of gene flow can, however, restore lost alleles;but, if the organism has poor dispersal powers, thenmarginal, isolated demes are expected to be allelicallydepauperate (Soul 1973).

The problems Soul described are exacerbated in ademographic event known as a bottleneck (Frankel and

Soul 1981; Figure 6.1) in which a population declinesto very low levels, which accompanying loss of geneticdiversity. After a bottleneck, the remaining individuals,and their remaining genes, represent only a sample of theoriginal source population. The smaller the sample, themore likely that it may not be representative of the sourcefrom which it was taken, and the more certain that somealleles, especially the rarer ones, may have been lost. Thisloss of genetic variation can mean a loss of heterozygosityin the population, which may be correlated with a loss ofoverall fitness (Frankel and Soul 1981) because it exposes


3/32

a greater proportion of recessive genes in a homozygouscondition, and traits previously masked are now expressed.Many of these recessive genes have deleterious or evenlethal effects on organisms. Little genetic variation is lostif the reduction in population size is temporary, but signifi-cant genetic variability will be removed if the populationremains small for many generations (Frankel and Soul1981; Table 6.1).

A second problem of small populations is that of

genetic drift. Recall that genetic drift represents therandom fluctuations in gene frequencies that occur asa result of non-representative combinations of gametescreated during breeding. Geneticist Ian Robert Franklinprovides one of the clearest definitions. In a finite popu-lation, wrote Franklin, the array of genotypes is formedby sampling gametes from the previous generation;virtually all of the genetic effects which arise in smallpopulations are an unrepresentative consequence of sam-pling, a process known as genetic drift (Franklin 1980).Thus, genetic drift is to genetics what sampling error is

to statistics. The smaller the population, the greater theprobability that the sample (random matings of individu-als) may represent neither the average nor the range ofpopulation characteristics.

A third concern regarding the genetics of small popu-lations is inbreeding, the mating of individuals havingany degree of genetic relatedness. Inbreeding in a popu-

lation often results in inbreedingdepression, a patternof reduced reproduction and survival that occurs onaccount of inbreeding (Frankham et al. 2002:24). Basedon theoretical arguments regarding inbreeding, Frankeland Soul derived thebasic rule of conservation genet-ics, expressed as a percent change in the inbreedingcoefficient (the probability that two alleles at the samelocus in an individual are both identical by descent). Therule asserted by Frankel and Soul (1981:73) is thatnatural selection for performance and fertility can bal-ance inbreeding depression if the change in the inbreed-ing coefficient (F) is no more than 1% per generation(Figure 6.2). The significance of this rule is well stated

by the authors themselves: We refer to the 1% rule asthebasic rule of conservation geneticsbecause it servesas the basis for calculating the irreducible minimumpopulation size consistent with the short-term preserva-tion of fitness (Frankel and Soul 1981:73, emphasismine). Such short-term fitness preservation was consid-ered safely achieved in most populations with an effec-tive population size of 50 (Franklin 1980; Frankel andSoul 1981). In contrast, long-term fitness was based onadaptation, measured most objectively in the ability ofa population to speciate. Comparing long-term adaptivepotential to short-term considerations regarding inbreed-

ing, Franklin (1980) wrote, In the long-term, geneticvariability will be maintained only if population sizes arean order of magnitude higher, (i.e., 500). Franklin basedhis 500 rule on data compiled by geneticist Russell

6.2. Conservation Genetics and Conservation Biology 155

Figure6.1. A graphical representation of population size before,during, and after a population bottleneck. Tindicates time andNindicates relative population size. (Drawing by M. J. Bigelow.)

Figure6.2. Percent change in the inbreeding coefficient (F) atdifferent population sizes (N). Note that the value of the inbreed-ing coefficient increases as population size declines.

Table 6.1. Percent change in genetic variation and proportionof rare alleles lost from a population at bottlenecks (minimumsizes) of different magnitudes. Values ofprepresent proportions

of each of four alleles.Number ofIndividualsin Sample(N)

PercentChange in

GeneticVariation

Proportion of Rare Alleles Lost

p1= 0.70, p1= 0.94,p2=p3=p4= 0.10 p2=p3=p4= 0.02

1 50.0 0.6300 0.7200 2 25.0 0.4950 0.6925 6 8.3 0.2125 0.590010 5.0 0.0925 0.500050 1.0 0.0025 0.1000

0.0 0.0000 0.0000

Source: Developed from equations from Frankel and Soul (1981).Table design by M. J. Bigelow.


4/32


Lande from mutation rates associated with Drosophila,maize, and mice. The figure of 500, noted Lande, may beroughly correct to maintain typical amounts of heritablevariation in selectively neutral quantitative characters,but should never be used as a blanket application tospecies conservation and should not be incorporatedinto species survival plans when other factors have not

been considered that might require larger populationsfor effective persistence (Lande 1988). Even as early as1981, Frankel and Soul, while citing Franklins rule,noted that to accommodate continuing evolution, theactual number of individuals that satisfy this criterionmay be several times greater than 500 for a variety ofreasons. They went on to argue against rule-of-thumbestimates and for continued genetic monitoring of endan-gered populations to determine if genetic variation wasremaining at or above critical minimum levels (Frankeland Soul 1981).

Despite its limitations, the 50/500 rule was impor-tant to the development of conservation biology because

it provided one of the first specific estimates of whatconstituted a minimum viable population or MVP.Although the term would not appear in published litera-ture until a year after Franklins estimates (Shaffer 1981),Franklins efforts represented a first attempt to answerthe question of minimum numbers needed for populationpersistence. Subsequent studies revealed increasinglycomplex relationships between genetics and demography,preventing any single rule of thumb from being usedwith certainty. In genetics, long-term inbreeding depres-sion has been demonstrated in populations with effectivesizes of 50500, and may occur in larger populations as

well (Latter et al. 1995; Frankham 1995a). Despite theselimitations, conservation genetics provided the first para-digms, albeit initially little more than rough guidelines,for estimating minimum populations needed for speciespersistence.

Genetics was not established as a preeminent compo-nent of conservation biology only on the basis of theo-retical argument. Souls earlier (1973) concern aboutsome demes becoming allelically depauperate wasdramatically supported with the publication of OBrienet al.s 1983 Sciencepaper, The Cheetah is Depauperatein Genetic Variation (Figure 6.3). Most articles in sci-

entific periodicals begin with objective, descriptive (andrather dull) titles, but the choice of such a vivid declara-tive sentence to entitle this investigation illustrated theauthors convictions about the veracity and significanceof their findings. Specifically, an examination of 47allozyme loci in 55 cheetahs (Acinonyx jubatus) fromtwo isolated populations revealed no polymorphic lociand an average heterozygosity of 0.0 (OBrien et al.1983). The authors attributed this genetic uniformityin cheetahs to past population bottlenecks followed bysevere inbreeding, and supported their explanation with

data showing that sperm counts were ten times lower incheetahs than in related felid species, and that 70% ofthe sperm were morphologically aberrant. Because thecheetah was an endangered species with a wild popula-tion estimated at less than 25,000 individuals, scientificinterest was high and concern grew about loss of geneticdiversity in wild populations.

In skin tissue grafts among 14 cheetahs, 12 of which

were between unrelated animals, all grafts were acceptedbeyond the rapid rejection stage, suggesting that themajor histocompatibilility complexes (MHC) of indi-vidual cheetahs were identical (i.e., the grafts were theequivalent of receiving tissue from a genetically identicalindividual) (OBrien et al. 1985). The authors assertedthat lack of genetic variation contributed to increasedsusceptibility to disease in cheetahs (OBrien et al. 1985).When an inbred population of lions (Panthera leo) in theNgorongoro Crater in Tanzania underwent a populationcrash due to poor reproductive performance and highsusceptibility to epizootics, it raised further concernthat inbreeding depression was the cause of populationdecline (Packer et al. 1991).

Populations of concern to conservation biology, such aspopulations on islands, populations in fragmented habitats,and populations in zoos were examples that made con-servation genetics important because genetic theory andmeasurement helped define critical applications as well asimportant theoretical puzzles to solve. We now turn to thescience of such measurement, specifically targeted to solvethe problem of how to measure the genetic diversity of aplant or animal population.

Figure6.3. Genetic samples of cheetahs (Acinonyx jubatus) haveshown little genetic variation at sampled loci. Various studieshave shown the cheetah to suffer higher-than-average rates ofinfant mortality, infertility, sperm abnormalities, and susceptibil-ity to disease, all characteristics associated with high rates ofinbreeding and low genetic variability. (Photo courtesy of DawnPatrick and Cheetah Conservation Botswana.)


5/32

6.3. Measuring Genetic Diversityin Populations

6.3.1. Foundational Measures of GeneticDiversity

Precisely because genetic diversity is so important to

population conservation, one must have reliable quanti-tative means of measuring it. Some measures of geneticdiversity are identical to the measures of communitydiversity described in Chapter 4. For example, theShannon Index (a measure of species diversity in a com-munity) can be used with equal efficacy as a measure ofgenetic diversity if the proportional abundance of allelesis substituted for the proportional abundance of species.Likewise, the Simpson Index (Chapter 4), used as ameasure of dominance in the assessment of communitystructure, can be used as a measure of expected hetero-zygosity (H

e) with the same substitution of allelic fre-

quencies for species abundance (Vida 1994). Currently,there are three commonly used quantitative measuresof genetic diversity. These are polymorphism, averageheterozygosity, and allelic diversity.

Polymorphism refers to a genetic locus that has twoor more forms (alleles). In a population or populationsubunit, polymorphism is expressed as the probability(P) of encountering a polymorphic loci among all lociin the population. To begin with a simple example, if apopulation has 100 genes and 50 of these have two or morealleles, the level of polymorphism is 0.50 (50%). Moregenerally, we could say that P (polymorphism), for anindividual or any larger group or unit, can be determined

from the expression

P = number of polymorphic loci/total number of loci.

Although polymorphism is a concept that is easy tounderstand, it is not the most frequently used measureof population genetic diversity. That measure is averageheterozygosity (H). Individual heterozygosity describesthe observed proportion of heterozygous loci in an indi-vidual. Thus, individual heterozygosity is a measure ofsingle-locus diversity. Average heterozygosity refers tothe average proportion of individuals in a population thatare heterozygous (carrying two different alleles) for a par-

ticular trait. This metric reflects the proportion of hetero-zygous individuals measured across several loci. We cancalculate average hetereozygosity with the expression

H H Ni

= /

whereHis the average heterozyosity at locus iandNis thetotal number of loci used in the estimate. Suppose there arefour loci in a population. We will designate them (for lackof imagination) as 1, 2, 3, and 4. Now suppose that thefrequency of heterozygotes for locus 1 is 0 (all individuals

are homozygous for this locus), 0.3 for 2, 0.5 for 3, and1.0 for 4. Then

H= + + + = =( . . ) / . / . .0 0 3 0 5 1 4 1 8 4 0 45

Finally, the third measure,allelic diversity(A) refers to theaverage number of alleles per locus. It can be calculated at

A A A A Nn= + + +[ ]1 2

/whereA1is the number of alleles at locus 1,A2the numberof alleles at locus 2, and so on through all Nloci.

Other measures can be used to describe populations,some derived from these foundations and others that areindependent of them. Throughout the chapter, we willexamine additional measures and their applications todescribing genetic diversity. Keep these three introductorymeasures in mind as a foundation for new, more complexconcepts and measurements that follow. Now we put theconcept of heterozygosity (H) and allelic diversity (A)to immediate use in understanding two critical concepts

affecting genetic diversity in small populations, bottle-necks and genetic drift.

6.3.2. The Loss of Genetic Diversity overTime: Bottlenecks and Genetic Drift

A population bottleneck is a minimum population sizeas a result of a crash (Frankel and Soul 1981). Theremaining individuals possess only a sample of the geneticvariation present in the original source population. Abottleneck that lasts for only a short time has only minor

effects on overall genetic variation, but one that persistsfor many generations will deplete genetic variability.Once depleted, genetic variation is slow to be restoredeven after the population recovers to a much larger size.Thus, current population sizes do not always correlatepositively with the genetic diversity of a population if ithas suffered one or more bottlenecks in the past. This lossof genetic variation can lead to a loss of heterozygosity,which, in some studies, has been correlated with a loss inoverall fitness (Frankel and Soul 1981). The correlationexists, as previously noted, because loss of heterozygosityallows a greater proportion of recessive alleles to occur in a

homozygous condition so that traits previously masked areexpressed. Many of these recessive genes have deleteriousor even lethal effects on an organism.

Small populations that suffer a prolonged bottleneckmay experience genetic drift. Genetic drift can occur inpopulations of any size, and is a normal evolutionaryforce that changes population gene frequencies throughtime. However, genetic drift usually has a greater effectin small populations because the proportion of such non-representative matings tend to increase when the actualnumber of matings is low. The smaller the population, the

6.3. Measuring Genetic Diversity in Populations 157


6/32


greater the probability that the sample (random matingsof individuals) may represent neither the average nor therange of characteristics found in the population. Geneticdrift, which can be indexed by the change in frequencyof a randomly selected allele, q, in one generation, can beestimated by the expression

q q q

Ne=

( )12

where Ne is the variance effective population size, a

concept we will examine in detail shortly. Thus, if thefrequency of qis 0.2 and the effective population size is100 the expected change in q(i.e. q) is 0.2 (1 0.2)/2(100) = 0.16/200 = 0.0008, or eight one-hundredths of1%. In contrast, if effective population size is ten, theeffect on the same allele would be equal to 0.16/20 =0.008 (eight-tenths of 1%). Both results represent verysmall effects, but notice the order of magnitude increasein the smaller population. If the population remains

small and this relationship is reiterated for many genera-tions, the effect of genetic drift on gene frequencies canbecome large.

Genetic drift can lead to a loss of heterozygosity ora fixation of deleterious alleles. These outcomes cancause random changes in the phenotype, and can leadto a decline in genetic variability (Franklin 1980).Such effects are exacerbated in small populations, par-ticularly if they are closed to migration. In such a state,there is a decrease in the number of different alleles ata single locus in the population and in heterozygosity(Caughley 1994). The degree of decline in heterozy-gosity is a function of the population size, N, over the

number of generations, t (Wright 1931). For example,over one generation, the amount of heterozygosity, H,changes this way:

H H N1 0 1 1 2= [ /( )],

where H0 represents the original level of heterozygosity(usually expressed as a proportion) and H1 representsthe new level of heterozygosity after one generation.Generalizing the equation for any number of tgenerations,heterozygosity declines as

H H Nt

t= 0 1 1 2[ /( )] .

Note that the smaller the value ofN, the greater the declinein heterozygosity. For example, a population of 50 indi-viduals that began with a 0.5 level of heterozygosity wouldlose 1% of its heterozygosity in each generation (from0.5 to 0.495). In contrast, a population of ten individualswith the same initial heterozygosity would lose 5% of itsheterozygosity (from 0.5 to 0.475) (Table 6.2). Mutationscan and do occur, and they increase genetic variability andheterozygosity, but this change in heterozygosity, H, alsois affected by population size:

H H N mH H mN

= + =

/( ) ,21

2

where m is the addition of heterozygosity through muta-tion, typically expressed as a rate. Populations reach anequilibrium level of heterozygosity (H= 0) at

H Nm* .= 2

The smaller the size of the population, the lower its equi-librium heterozygosity.

Genetic drift and population bottlenecks can combineto produce long-lasting effects on populations, even afternumerical recovery. Consider the Mauritius kestrel (Falco

punctatus) (Figure 6.4), a small, rare falcon found onlyon the island of Mauritius in the southwestern part of the

Indian Ocean, east of Madagascar. Following declinesassociated with the use of pesticides and the destructionof its habitat from deforestation throughout the 1960sand early 1970s, this population is believed to have beenreduced to a single breeding pair in 1974. Through carefulprotection, habitat restoration, and managed breeding thatwas part of an intensive conservation and recovery pro-gram from 1983 to 1993, the population had been restoredto over 200 individuals by 1994 (Figure 6.5a), and to400500 individuals, and over 200 breeding pairs, by thelate 1990s (Groombridge et al. 2000). By examining DNAfrom living kestrels and comparing it to the same DNA loci

in museum skins of kestrels from ancestral populations,Jim Groombridge and his colleagues documented the lossof genetic diversity that has accompanied this populationbottleneck. Overall allelic diversity has declined by 57%.A number of unique alleles present in the ancestral popu-lation are no longer extant in living kestrels (Figure 6.5b)and the restored population shows a 57% reduction inheterozygosity (Table 6.3) Despite these genetic scars, thepopulation has continued to increase, and its productivity,indexed by the average number of fledglings per nest, rose31% from 1994 to 1998, even after intensive management

Table6.2. Heterozygosity and the effect of population size. Theheterozygosity (H) of smaller populations declines at a faster ratethan that of larger populations. Shown here are two populations:Population A with a starting size of 50 and Population B witha starting size of 10. Within one generation, Population B hasdeclined to a level of 0.475. In contrast, it takes Population A5 generations to decline to that level.

Ht Population A (50) Population B (10)

H0 0.500 0.500H1 0.495 0.475H2 0.490 0.451H3 0.485 0.429H4 0.480 0.407H5 0.475 0.387

Source: Developed from data from Caughley (1994). Table design byM. J. Bigelow.


7/32

had ceased. Although such increase is a hopeful sign, onemust remember that, in every population, there are somegenes, known as lethal genes, which, although reces-sive and unexpressed in a heterozygous state, will, in ahomozygous condition, always result in the death of the

individual. The proportion of such genes in a population,its lethal load, often rises when alleles are lost during aperiod of population reduction. As alleles are lost, hetero-zygosity is reduced, as in the Mauritius kestrel, and theprobability of homozygous expression of such lethal genesis higher if they are present. The sampling of genetic mate-rial present in the founders represents a random sample ofthe populations genetic variability, but, in small foundergroups, the genetic constituency of the founders may ormay not be representative of the population. In this case, ifthe last remaining pair of kestrels possessed lethal genes,their expression is likely to emerge in future generations

and depress population survival rates. But, if such lethalswere not present in these founding individuals, the popula-tion is not exposed to this risk.

The case of the Mauritius kestrel suggests that wildpopulations may be highly resilient to genetic loss, andprovides hope that conservation efforts can be success-ful even after severe reductions in numbers and loss ofgenetic diversity. In fact, recent genetic studies of multiplespecies suggest that low levels of genetic variation arenot necessarily an indication of population endangerment(Zhang et al. 2002). Nevertheless, the case of the Mauritiuskestrel reveals that population reductions can have effects

Figure 6.4. The Mauritius kestrel (Falco punctatus), a spe-

cies found only on the island of Mauritius in the Indian Ocean,displays classic evidence of loss of genetic diversity followinga severe reduction in population size (population bottleneck)during the 1960s and 1970s. (Photo courtesy of The MauritiusWildlife Foundation.)

Figure6.5. (a) The population size of the Mauritius kestrel (Falcopunctatus) from 1940 through 1994. Note the severe reductionsbeginning in the 1960s when the population declined and remainedat less than 50 individuals. (b) DNA fingerprints (microsatellitegenotypes) from Mauritius kestrel museum skins (top) comparedto DNA from the same region in birds from the restored popula-tion (bottom). b.p. refers to specific DNA base pairs. Dark bandsrepresent the presence of specific alleles. Note the reduction inthe number of bands in the restored population, indicating reduc-tion in allelic (genetic) diversity. (Reprinted by permission fromMacmillan Publishers Ltd:Nature, Groombridge, J. J., C. G. Jones,M. W. Bruford, and R. A. Nichols 2000. Ghost alleles ofthe Mauritius kestrel.Nature403:616. Copyright 2000.)

Table6.3. Genetic diversity of the Mauritius kestrel and otherkestrel populations. Mean numbers of alleles (A) and averageheterozygosity (He) of the restored (post-bottleneck) Mauritiuskestrel population compared with those of the ancestral (pre-bottleneck) population and with other kestrel population. Notethat with the exception of the Seychelles population, which alsosuffered severe population reduction, the ancestral Mauritiuskestrel population had less genetic diversity than other Africanand European kestrel populations. With further reductions in therestored population, the differences are now even greater.

Species A He Sample Size

Endangered

Mauritius kestrelRestored 1.41 0.10 350Ancestral 3.10 0.23 26Seychelles kestrel 1.25 0.12 8Non-endangered

European kestrel 5.50 0.68 10Canary Island kestrel 4.41 0.64 8South African rock kestrel 5.00 0.63 10Greater kestrel 4.50 0.59 10Lesser kestrel 5.41 0.70 8

Source: Reprinted by permission from Macmillan Publishers Ltd:Nature,Groombridge, J. J., C. G. Jones, M. W. Bruford, and R. A. Nichols 2000.Ghost alleles of the Mauritius kestrel. Nature403:616. 2000.



8/32


on genetic diversity which do not rapidly disappear withincreases in population size alone.

6.3.3. Genetic Drift and EffectivePopulation Size

The theoretical consequences of genetic drift are normallycalculated for an ideal population in which each indi-vidual contributes gametes equally to a genetic pool fromwhich the next generation is formed. Real populationsrarely conform to this happy genetic vision. Instead, wemust be able to estimate the effective population size,N

e,

which represents the size of a randomly mating popula-tion that is subject to the same degree of genetic drift as aparticular real population. Another way to say this is thatthe effective population size of a real population is equal tothe size of an ideal population that has the same amount ofvariance in allelic frequencies. Hence, the effective popu-lation size, defined in this way, is more correctly called

the variance effective size. In contrast, the effectivepopulation size of a real population also can be defined,and estimated, as size of an ideal population that has thesame level of inbreeding, which is then more preciselyreferred to as inbreeding effective size (Loew 2002:242).Regardless of which measure is used, the effective popu-lation size of any population is affected by a number ofvariables, including variance in progeny number (broodor litter size), differential sex ratios, fluctuations in totalnumbers, and deviations from random mating systems(Frankham 1980). The first three problems can be evalu-ated separately if we assume that mating is random.

To examine the effect on Neof variation in the numberof progeny, letNequal the populations actual size (censussize) and 2the variance in progeny number. Then

N N

e=

+4

2 2.

Thus, if the size of the population is 100, brood size rangesfrom 0 to 8 and the variance is 4, the effective popula-tion size (N

e) is 400 divided by 6, or 67, which is one-

third less than the census population size. Effectively, anequalization of family size in a population should lead toan approximate doubling of the effective population size.This prediction has proven true in experimental tests. In

Drosophila, populations subjected to equalization of fam-ily size had greater genetic variation and greater reproduc-tive fitness than populations in which family size was notequalized (Boriase et al. 1993).

For populations with unequal sex ratios, the effectivepopulation size is

NN

e

f

f

=+

4N

N Nm

m

where Nm is the number of males and N

f is the number

of females. Consider a population of 100 elk (Cervus

elaphus) (Figure 6.6). If there are 50 reproductive bulls

and 50 reproductive cows, each bull mates with one cow,and each pair represents a unique association of individu-als, then the effective population size is 10,000 divided by100, or 100. But there is no wild elk population anywherewith such a sex ratio, nor are there any that use such amating system. Through natural selection and the effectsof sexually differential hunting pressure, wild elk popula-tions have more females than males. In autumn, during thebreeding period or rut, males gather groups of females(harems) that they defend against other males for exclu-sive breeding privileges. Suppose, in such a setting, that thebreeding population of 100 elk is actually composed of 10

males and 90 cows. Each male takes a harem of 9 femalesand successfully defends it from other males. This scenariois a gross simplification of what really happens, but it is alittle closer to real elk life. In harem-mating systems, therelatedness of offspring born to females within a harem ishigher than the relatedness of offspring from females ofdifferent harems. Thus, in this revised scenario, the effec-tive population size is 4 10 90 divided by 10 + 90, or3,600 divided by 100, producing a result of 36.

Here the effective population size of a population witha biased sex ratio is about one-third that of a monogamouspopulation with a balanced sex ratio. Thus, the sampling

error (genetic drift) associated with random mating ina population of 36 individuals is equivalent to the sam-pling error associated with mating in a population of100 individuals with the sex ratio and mating system justdescribed, and would lead to increased rates of inbreeding.Studies by Briton et al. (1994) have confirmed this predic-tion. Polygamous mating systems associated with unequalsex ratios increase rates of inbreeding and loss of geneticvariation, leading Frankham (1995a) to assert that harembreeding structures should be avoided whenever possiblein captive breeding programs.

Figure 6.6. Elk (Cervus elaphus) are an example of a specieswith a harem mating system that reduces the effective populationsize. (Courtesy of U.S. National Park Service.)


9/32


Effective population size also changes when populationsfluctuate. If population size varies from generation to gen-eration, then the effective number is the harmonic mean(the reciprocal of the arithmetic mean of the reciprocals ofa finite set of numbers):

1 1 1 1 1

1 2

N t N N Ne t

= + +

... ,

whereNtis the effective size of the population at genera-

tion t. It can also be expressed as:

N t Ne ei= / ( / ),1

whereNeiis the effective population size in generation i.

Like effects of unequal family sizes and unequal sex ratios,unequal population sizes should lead to increased levelsof genetic drift and loss of heterozygosity. These predic-tions have been verified experimentally (Woodworthet al. 1994).

All of these formulas assume random mating, but thatassumption is often violated in real populations. Morecomplex mathematics are required to determine effectivepopulation sizes where there is significant deviation from

random mating. Even if random mating is approximated,however, most populations will have a lower geneticallyeffective population size than their census size. Theproblem of genetic drift becomes especially importantwhen the effective population size is small. When theeffective population size is large, the expected variationin a typical genetic character is determined mainly bythe strength of selection for or against that character (i.e.genetic variation is determined by mathematical prob-ability). When effective population size becomes lessthan a few hundred individuals, expected variation of thecharacter becomes largely independent of the strengthof selection and is determined primarily by the balancebetween mutation and drift (i.e. variation is determinedby random events).

6.3.4. Bottlenecks, Small Populationsand Rare Alleles

Although bottlenecks have little effect on genetic vari-ability in a population unless the population remains smallfor a long time, the effect of size reductions on rare allelesis a different matter. Rare alleles can be lost quickly in

small populations that experience a sudden decline or thatremain at low levels for extended periods. The expectednumber of alleles, E(n), remaining after a genetic bottle-neck is equal to

m pj

Ne ( )1 2

where m is the number of alleles prior to the bottleneck,

p is the frequency of the jth allele, and Ne is the effec-tive number of individuals at the bottleneck. Suppose thatm= 4 and that one allele is common, but the other three arerare. Look what happens to the average number of alleles(Table 6.4) as the effective number of individuals dropsfrom 50 to 1. The rarer an allele is, the more likely that itwill be lost (Frankel and Soul 1981).

The earliest paradigms of modern conservation biologyarose from concerns about long-term loss of genetic vari-ation in small populations, leading to one of the earlieststated goal of conservation biology: the retention of 90%of a populations genetic variability for 200 years (Soul

et al. 1986). The loss of genetic variation is reduced (andthe probability of meeting this goal improves) as theeffective population size grows to an effective populationsize of about 1,000 individuals. Beyond this level, fur-ther increases in effective population size do not usuallyincrease the amount of genetic variability in the population(Lynch 1996). But effective population size is often onlyone-tenth to one-third the number of breeding adults in thepopulation for reasons noted previously, including unequalfamily sizes, unequal sex ratios, and unequal populationsizes over time. Thus the N

e> 1,000 criterion suggests

the need for a stable population of 3,00010,000 breed-ing adults in each generation to prevent long-term loss ofgenetic variation.

Table 6.4. Decreasing population size influences the averagenumber of alleles. In this case, four alleles are observed onewith a high frequency and three with lower frequencies. Rare,less common alleles are more likely to be lost during a bottle-neck. These rare alleles are typically not essential in the initialenvironment. However, as the environment changes they mightbe crucial for survival.

Effective Number

Average Number of Alleles Retained,

of Individuals (Ne)

Given the Original Frequency of Allelep1= 0.70, p1= 0.94,

p2=p3=p4= 0.10 p2=p3=p4= 0.02

4.00 4.0050 3.99 3.6010 3.63 2.00 6 3.15 1.64 2 2.02 1.23 1 1.48 1.12

Source: Developed from equations by Frankel and Soul (1981). Tabledesign by M. J. Bigelow.

POINTS OF ENGAGEMENT QUESTION 1

Work out the mathematics of the effective populationsize for 10 generations with a population of 100 in everygeneration, then repeat the calculation a second time,

letting one generation crash to 10 individuals. Whathappens to the effective population size?


10/32


6.4. The Problem of Inbreeding

6.4.1. What Do We Mean by Inbreedingand How Would We Measure It?

As if the problem of finding a mate was not enough,individuals in small populations may suffer just as much

or more as a result of finding the wrongmate. More spe-cifically, they are likely to mate with close relatives withwhom they share many genes. This situation is known asinbreeding, a problem we have already alluded to in thischapter. Inbreeding can be defined as the production ofoffspring related by descent. To make our understandingmore precise, consider three different biological meaningsof inbreeding in terms of its measurement. A populationslevel of inbreeding can be assessed by: (1) a measure ofshared ancestry in the maternal and paternal lineages ofan individual; (2) a measure of genetic drift in a finitepopulation; or (3) a measure of a system of mating in areproducing population. Each of these three dimensions

of inbreeding grows stronger in its effects as populationsize declines, and such effects can and must be measuredto make informed and appropriate management deci-sions regarding breeding strategies for small populations(Templeton and Read 1994).

The first concept of inbreeding, the measure of sharedancestry of an individual in its maternal and paternal lines,has been called inbreeding by descent orpedigree inbreed-ing(Templeton and Read 1994). This type of inbreeding isquantified as the inbreeding coefficient, symbolized by Fp(pedigree inbreeding). The value of F

p, which varies from 0

to 1, can be calculated, as we will do in a subsequent sec-

tion, only for an individual of known pedigree. It measuresthe amount of ancestry an individual shares with its mater-nal and paternal lines. Pedigree inbreeding intensifies as thesize of a population decreases.

The second concept of inbreeding is that of inbreeding asa measure of genetic drift in a population. If we knew theindividual values of F

pfor every individual in a population,

added these values together, and divided the sum by thenumber of individuals, the resulting quotient would be theaverage probability of inbreeding by descent, symbolized asF

d. Here the subscript d is meant to signify that this value

of Fis a measure of the averaged inbreeding by descent of

all members of the local population, or deme (Templetonand Read 1994). This value represents the average prob-ability of inbreeding-by-descent, a measure of the effect ofgenetic drift on a population relative to an ideal popula-tion experiencing completely random mating.

Remember that the first type of inbreeding, pedigreeinbreeding, increased in magnitude as population sizedeclined. In the second type of inbreeding, inbreeding bydescent, the value of F

dalso increases as population size

decreases and, for a given population size, Fd increases

over time. This is expressed by the relationship

Fd(t)

t= 1 1 1 2[ / ,N]

where tis equal to time in generations. Note that the largertbecomes, the closer F

d(t)comes to 1 (a completely inbred

population). Thus, Fd(t)

will eventually reach a value of1, and how fast it does so is a function of population size(Figure 6.7). The smaller the population, the faster it willbecome inbred. In this scenario, genetic drift causes theaverage probability of inbreeding by descent to increase andgenetic variation to decrease. This means that inbreedingand loss of genetic variation are correlates, but inbreedingof this type is not the causeof a loss of genetic variation.

Finally, inbreeding can be used as a measure of a systemof mating in a population, quantified as a value called thepanmictic index,f. The panmictic index measures inbreed-ing as a deviation from a reference population, which hasa system of mating in which alleles at a locus are pairedin proportion to their frequencies in the overall population(by definition, random mating). The panmictic index thusevaluates deviations from the heterozygosity frequencies

expected under random mating, sof H H

o e= 1 / ,

where He is, as defined earlier, the expected hetero-

zygosity under random mating and Ho is the observed

heterozygosity. Recall that in a randomly mating popula-tion, the frequency of heterozygosity is defined by theHardy-Weinburg equation. For two alleles, pand q, thatfrequency is

( ) .p q p pq q+ = + +2 2 22

Figure 6.7. The relationship between the proportion of inbredindividuals in a population and population size. The smaller thepopulation, the less time it will take to become completely inbred.(Van Dyke, Conservation Biology: Foundations, Concepts,Applications, Copyright 2003, McGraw-Hill Publishers.Reproduced with permission of the McGraw-Hill Companies.)


11/32

Thus, the expected frequency of the heterozygote is2pq. For example, if the frequency of allele p is 0.6and the frequency of allele q is 0.4, then the H

e is

2 0.6 0.4 = 0.48. Observed heterozygosity (Ho) can

be calculated from genetic measurements of sampledindividuals. If observed heterozygosity is greater thanexpected, f< 0, and the population has a reproductive

system that avoids inbreeding. If observed heterozygosityis less than expected,f > 0, and inbreeding is not avoided.The value of the panmictic index can be used to quantifythe degree of avoidance of inbreeding in a population.

Measuring these three aspects of inbreeding separatelyand accurately provides a powerful array of informationfrom which to make intelligent management decisions forany population, but these measurements can be especiallyimportant in managing breeding in a captive population.Using the values of F

p in a captive breeding program,

for example, a manager can determine which potentialbreeding pairs would produce inbred versus non-inbredoffspring. If the goal is to minimize inbreeding, a manager

could choose to mate individuals who are least related toone another, thereby avoiding the production of inbredoffspring. An animal which is itself inbred should not beexcluded from the breeding pool, but breeding pairs shouldbe selected to avoid the creation of inbred offspring. Usingthe value of F

d, a manager could determine the effect of

genetic drift on a population and the degree of hetero-zygosity present in that population and, from these data,make an intelligent decision about whether the currentpopulation size is sufficient to maintain an acceptable levelof heterozygosity. Using the value of f, a manager coulddetermine if current mating systems in the population lead

to avoidance or encouragement of inbreeding, and then actaccordingly.

6.4.2. The Problem of Inbreeding Depression

When populations become inbred, genotypic frequenciesare skewed toward increased proportions of homozygousindividuals and heterozygosity declines. As the proportionof homozygous individuals increases, so will the manifes-tation of recessive traits, which can only be expressed ina homozygous condition, but which are maintained in the

population by heterozygous carriers. In environments thatselect against recessives, inbreeding can then lead predict-ably to inbreeding depression.

Inbreeding depression, as noted earlier, is a pattern ofreduced reproduction and survival that occurs on accountof inbreeding (Frankham et al. 2002:24), and can happenwhen historically large, outcrossing populations sud-denly decline to only a few individuals. These remainingindividuals may or may not be related. However, with thelimited mate choices now available to them, high averagerelatedness will result in just a few generations. The population

then experiences reduced survival and fecundity. As matechoice is now restricted to related individuals, inbreedingdepression may increase as relatedness increases (Figure 6.8).When the degree of relatedness of individuals in the popu-lation (inbreeding by descent) is regressed against oneor more traits affecting fecundity or survival, the resultingregression can be used to calculate the degree to whichincreased mortality or lower fecundity is associatedwith increased relatedness.

Inbreeding depression is an especially well-documentedproblem in captive populations of vertebrates (Frankham1995b). Forty-two of 45 captive, inbred vertebrate popula-

tions examined by Ralls and Ballou (1983) had reducedjuvenile survival compared to outbreeding populations ofthe same species. The most comprehensive experimentalstudies of inbreeding have been conducted by conserva-tion geneticist Richard Frankham and his colleagues.Using captive populations of the fruit fly, Drosophilamelanogaster, Frankham created experimental popula-tions in which he manipulated density, rates of inbreeding,and levels of environmental stress. In these experiments,several recurring trends appeared: (1) inbreeding andconsequent loss of genetic diversity reduced the resist-ance of the flies to disease; (2) in inbred populations,

extinction rates rose as the level of environmental stresswas increased by adding additional stress factors (Figure6.9); (3) the adaptive evolutionary potential (capacity forlong-term genetic change) was reduced in small popula-tions as environmental stress was increased; and (4) ratesof inbreeding increased under stressful conditions (Reedet al. 2002; Frankham 2005). Among populations of delib-erately inbred domestic animals and plants, up to 95%became extinct after eight generations of brother-sistermatings (animals) or three generations of self-fertilization(plants) (Frankel and Soul 1981). In wild populations,

Figure 6.8. Relationship of relatedness in mating to levels ofinbreeding depression, which is the decline in fitness (reducedsurvival and fecundity) associated with increased frequencies ofmating among closely related individuals. The probability of anindividual mating with a relation increases as the population sizedecreases. (Figure by M. J. Bigelow.)

6.4. The Problem of Inbreeding 163


12/32


inbreeding depression has been documented in fish, snails,lions, shrews, white-footed mice (Peromyscus leucopus),and plants (Frankham 1995a and references therein).

Some inbreeding occurs in all populations, no matterhow large. However, inbreeding has disproportionatelydetrimental effects on small populations. Caughley (1994)provides a summary of the sequence of events that inbreed-ing can initiate:

1. The frequency of mating between close relatives rises2. Heterozygosity is reduced in offspring, reducing the ability

of the population to respond to environmental change3. Semi-lethal recessive alleles are expressed in a

homozygous condition4. As a result of this expression, fecundity is reduced and

mortality is increased5. The population becomes even smaller, amplifying the

sequence initiated in step 1

Caughley referred to this sequence as the extinctionvortex of a positive feedback loop, or as Caughley put it,the worse it gets, the worse it gets (Caughley 1994). Thus,inbreeding can begin and sustain a pattern of significantdecline in a small population.

6.4.3. Measures of Inbreeding

The most basic measure of inbreeding, as noted previously,is the inbreeding coefficient, F, which is a measure of theloss of heterozygosity in a population due to the effectsof inbreeding. Therefore, one can estimate the inbreed-ing coefficient from changes in heterozygosity over time.Earlier, we defined the concepts associated with thismeasure, and now we will undertake its calculation. Thismeasure of the inbreeding coefficient, known as the effec-tive inbreeding coefficient (F

e) can be estimated as

F H He t= 1 0( / ),

where, as you recall, Ht is the level of heterozygosity at

time or generation tandH0is the level of heterozygosity inthe previous or base comparison generation. For example,if the level of heterozygosity in generation tis 0.4 and thelevel of heterozygosity in the previous generation 0was

0.8, then the value of Feis1 0 4 0 8 1 0 5 0 5 = =( . / . ) . . ,

which means that inbreeding has reduced heterozygosityover this period of time in this population by 50%. Onealso can use the same relationship to make comparisonsbetween related populations. For example, suppose thelevels of heterozygosity of a small island population ofindividuals and that of the larger mainland population theyoriginated from, are both known and are respectivelyH

island

= 0.35 and Hmainland

= 0.81. Then the effective inbreedingcoefficient indirectly estimated from these values is

Fe= = =1 0 35 0 81 1 0 43 0 57( . / . ) . . .

We would thus conclude that inbreeding within the smallerisland population has reduced its heterozygosity by 57%compared to the mainland population.

Alternatively, some effects of inbreeding can be evalu-ated by measuring the rate of juvenile survival, which iscalculated theoretically as

ln( )=S A BF +

where Sis the juvenile survival rate,Athe instantaneous rateof juvenile mortality in progeny of unrelated parents,Bthe

same rate when the line is completely inbred (H = 0) and Fthe inbreeding coefficient (Ralls et al. 1979). This equationcan be expressed as the line S=ABF.Acan then be esti-mated ifBis known, and vice versa (Caughley 1994).

If one knows the Fcoefficient and a measure of the indi-viduals fitness in the population, it is possible to estimatethe severity of inbreeding depression. A general measureof inbreeding depression is normally designated by thesymbol , which represents the proportionate decline inmean fitness due to a given amount of inbreeding. In gen-eral, the relationship could be expressed as

d= 1(fitness of inbred offspring/

fitness of outbred offspring).However, to be useful, we must make the measurement morespecific. For example, let us replace fitness in the aboveequation with survival rate. Conventionally, the magnitudeof inbreeding depression is expressed as the average reductionin mean fitness value per 10% increase in the F coefficient(Van Oosterhout et al. 2000). Given this information, wecould estimate the value of inbreeding depression as

10 25

0 0

==

Survival rate at

Survival rate at

F

F

.

..

Figure 6.9. Proportion of populations of Drosophila mela-nogaster surviving at different inbreeding coefficients (F) forinbred (full-sib mating) populations in benign (no stress), singlestress factor, and variable stress factor environments. (Frankham2005. Stress and adaptation in conservation genetics. Journal ofEvolutionary Biology. Copyright 2005 by Blackwell Publishing.)


13/32

For example, suppose that the survival rate at F0.25 = 0.2and at F0.0 = 0.8. Then the cost of inbreeding is1(0.2/0.8) or 1 0.25 = 0.75. Conceptually, the cost ofinbreeding is the proportional decline in survival that canbe attributed to inbreeding of a given magnitude. Suchdeclines are well documented. In 38 species of mammals,Ralls et al. (1988) estimated the actual average cost of

inbreeding to be 0.33.The estimates we have examined so far are applicableto populations in which all individuals can be considereda single population unit. What would happen to the genet-ics of a population that was both small and subdivided?To answer that question, and explore another importantdimension of inbreeding effects, we need to add one moretheoretical component to our understanding of conserva-tion genetics. That component is the concept of gene flowand metapopulation structure.

6.5. Inbreeding and Outbreeding in

Population Subunits: Estimation of GeneFlow and Metapopulation Genetics

6.5.1. Historical Development of GeneFlow Theory

In 1954, the Australian ecologists H. G. Andrewartha andL. C. Birch noted that A natural population occupyingany considerable area will be made up of a number of local populations or colonies (Andrewartha and Birch1954:657). This is perhaps the first published expression ofthe idea of ametapopulation, a concept that Andrewartha

and Birch represented clearly with elegant illustrationsremarkably similar to those used in metapopulation litera-ture today (Figure 6.10). However it was not until the late1960s and early 1970s that the idea of metapopulationsbecame an explicit model in population biology (den Boer1968; Levins 1968, 1969, 1970). Levins offered perhapsthe first intentional definition of a metapopulation as anyreal population [that] is a population of local populations

which are established by colonists, survive for a while,

send out migrants, and eventually disappear (Levins1970).

Spatial subdivision affects the genetic structure of popu-lations, and can influence the persistence of such popula-tions in a landscape. Before the articulation of the concept ofspatially divided populations by Andrewartha and Birch,a genetic basis for the same concept had been proposedby geneticist Sewall Wright in the 1930s (Wright 1931).Wright proposed a shifting balance theory of naturalselection in which small, subdivided populations (demes)achieved high levels of local adaptation to changing envi-ronments through (local) natural selection, genetic drift,migration (among subunits), and interdemic selection (i.e.,local extinctions of less-fit demes and colonization of newor vacated areas by more fit demes). The problem that

Wright was attempting to solve was how novelty arose ina constantly changing environment. The small, subdividedpopulations served as natures many small experiments(Wade and Goodnight 1998), and speciation resulted asa by-product of local adaptation when it produced repro-ductive isolating mechanisms. It was this subdivision thatprevented the averaging of environmental variation into asingle genetic optimum (one size fits all) for the entirepopulation. Thus, genetic optima shifted in different popu-

lation subunits.Historically, the scientific alternative to Wrights shift-ing balance theory was R. A. Fishers theory of largepopulation size (Fisher 1958), which was an attemptto explain how existing adaptations were refined in aslowly changing environment (Wade and Goodnight1998; Table 6.5). Fisher saw the world as a collectionof large, interbreeding (panmictic) populations in whichadaptation occurred primarily through mutation andnatural selection. Genetic drift and migration were atbest inconsequential, and at worst counter-productive toadaptation. In contrast, Wright saw genetically subdi-vided populations of multiple and varied fitness peaks,Fisher envisioned fitness as a global average for theentire population.

Although both theories found support, the weight ofopinion traditionally favored Fishers theory as the simplerexplanation for empirical data from natural populations,which were assumed to be panmictic. If populations hadapparent separation and spatial diversity, they still facedno real barriers to migration, exchange, or gene flow.However, by the 1960s, the realities of habitat destructionand fragmentation, combined with increasing concern overgrowing rates of species extinctions, led researchers to

Figure 6.10. Diagrammatic representation of an arrangementof local populations of the same species (a metapopulation).Empty circles represent favorable habitats that populations donot occupy. Partially or completely filled circles represent favo-rable habitats and relative densities of populations in them as aproportion of the habitats maximum capacity. Crosses indicatehabitats in which local populations recently became extinct.(Adapted from Andrewartha and Birch 1954. Illustration byM. J. Bigelow.)

6.5. Inbreeding and Outbreeding in Population Subunits: Estimation of Gene Flow and Metapopulation Genetics 165


14/32


question the generality of contiguously distributed, pan-mictic populations.

Wrights shifting balance theory has profound implica-tions for the maintenance of biodiversity as well as forthe dynamics of metapopulations. Wright proposed that

evolution might proceed rapidly in spatially structuredpopulations, especially if local extinctions and re-coloniza-tions occurred (Hanski and Simberloff 1997). Wade andGoodnight (1998) note that Wright imagined that the mem-bership of most species was distributed into small, semi-isolated breeding groups. In such a system, random geneticdrift and selection within demes become strong evolution-ary forces. Far from being detrimental to the population,random genetic drift and epistasisbecome potential sourcesof additional genetic variance by fueling additional adap-tation to local conditions within metapopulations (Wadeand Goodnight 1998) and potentially could create moregenetic variance per generation than mutation (Wade 1996).

Wrights Shifting Balance Theory also made the processesof gene flow, migration, and recolonization of temporarilyvacant habitats essential to understanding genetic change inpopulations. We now take up the study of these processes,and their effects on population genetics, in greater detail.

6.5.2. Current Models of Gene Flow:Predictions and Implications

Gene flow, previously discussed in Chapter 4, is the move-ment of genes between populations or population subdivi-sions and can be determined from the expression

N mF

Fe

ST

ST

= 14

where Neis the effective population size, mis the rate of

immigration and FST

is the total genetic diversity foundamong all populations. The result is expressed in immi-grants per generation. But how do genes really get fromhere to there in natural populations, and how will this

affect the actual genetic diversity of these populations?Three models of gene flow have been proposed to providea conceptual understanding of the process and its effectson genetic diversity.

Traditionally, ecologists have believed that populations,

and the genes they carried, tended to move or disperse asingle step among population subunits each generation. Thisview, known as the Stepping-stone Model, would mean itwould take many generations for an introduced allele to movethrough all segments of a large population (Figure 6.11A).Wright argued, in contrast, that gene flow would be equaland constant among all subdivisions in a population. We canvisualize Wrights model as one in which new alleles comefrom a single source (mainland) and move to all new areas(islands) at the same time. Hence it is often referred to as theMainlandIsland Model or simply as the Island Model(Figure 6.11B) A third, more recently developed view is thatthe probability of an allele from one source point decreases

with distance to potential destination points, and the so-calledIsolation-by-distance Model (Figure 6.11C). Thus, in com-parative terms, the Stepping-stone Model would predict theslowest rates of gene flow, the Island Model the highest, andthe Isolation-by-distance Model intermediate (and highlyvariable) rates (Hamrick and Nason 1996:205206).

A reasonable assumption of all three models is thatgenetic heterogeneity among populations should besolely a function of gene flow (i.e. you cant haveshared genes unless you exchange individuals). As partof the development of his Shifting Balance Theory ofpopulation genetics, Wright developed three measures,

often referred to as Wrights F statistics, for measuringand partitioning genetic variation within and amongindividuals of a population that was subdivided intoseparate genetic units, a necessary part of his theoryselaboration if one were to understand how spatiallysubdivided populations were structured genetically.The first, known as Wrights F

ST, is a measure of the

correlation of genes of individuals within subpopulationsor, more precisely, the proportion of the total genetic

Table6.5. Essential differences between Wrights Shifting Balance Theory and Fishers Theory of Large Population Size. WrightsShifting Balance Theory provides a conceptual basis for understanding how metapopulation structure (spatial division of populationsubunits) might affect genetic and evolutionary change in a spatially divided population.

Wright Fisher

Central problem of evolutionary theory Origin of adaptive novelty in a constantly changingenvironment

Refinement of existing adaptation in a stableor slowly changing environment

Major processes of evolutionary change Combination of local natural selection, random geneticdrift, migration, and interdemic selection

Mutation and natural selection

Ecological context of evolution Small, subdivided populations Large, panmictic populationsGenetic basis of evolutionary change Epistasis and pleiotropy; context-dependence of allelic

effectsAdditive genetic effects; context-independence

of allelic effectsProcess of speciation Inevitable by-product of local adaptation in epistatic systems Disruptive or locally divergent selection

Source: Wadeand Goodnight (1998). Perspective: the theories of Fisher and Wright in the context of metapopulations: when nature does many smallexperiments.Evolution. Copyright 1998 by Blackwell Publishing.


15/32


variation of the population that is found among sub-populations within individuals. The second statistic,known as Wrights F

IS, is a measure of the correlation of

genes within individuals relative to the gene frequencieswithin the subpopulation, in other words, a measure ofsubpopulation genetic variability. Finally, Wrights F

IT

reflects the correlation of genes within individuals relative

to gene frequencies of the entire population. The threemeasures can be quantified as

FST

= (q )/(1 )

FIS

= (F )/(1q)

FIT

= (F)/(1 )

where q is the correlation of genes between randomlyselected individuals in the same subpopulations, is thecorrelation of genes between random individuals fromdifferent subpopulations, and Fis the correlation of geneswithin individuals (Chesser et al. 1996).

Wrights F statistics are used to determine the breeding

structure of populations (i.e. was genetic variation randomlydistributed in subunits, or were subunits genetically differentfrom population means?) and to determine indirect indicesof dispersal among subunits. The second use is based onthat fact that, the more closely the genetic characteristics ofsubunits match average genetic population characteristics,the more exchange of genes (gene flow) must be takingplace between the subunits. Wrights ideas have been easilyincorporated into models and theories of metapopulationsand their genetics, such that we can restate the value ofWrights F

ST more simply. Treating the genetic subunits

as subpopulations of a metapopulation, we can compare a

familiar genetic measure, heterozygosity,H, within popula-tion subunits,Hs, to the average heterozygosity of the total

metapopulation,HT, and rewrite Wrights F

STexpression as

FH H

HST

T s

T

=

,

which tells us the ratio of between subunit heterozygos-ity to total metapopulation heterozygosity, a measure ofgenetic diversity among subpopulations.

Although we have already examined one expression ofgene flow as a measure of immigrants per generation,a more genetically precise and widely used measure is

Neis GST (Nei 1972), which is the proportion of totalgenetic diversity in a population attributable to differentiationamong subpopulations. In other words, Neis G

STis simply

a multi-allelic equivalent of Wrights FST

(Hamrick andNason 1996:207). The value of G

STcan be determined if

we can measure the total genetic diversity of the popula-tion at a given locus (total heterozygosity orH

T) and parti-

tion it into genetic diversity within populations and amongpopulations. The ratio of the among-population component

Figure 6.11. Three models of gene flow in populations ofa single species spatially separated from one another (metap-opulation). In each model a novel allele (*) is assumed to be

initially introduced into one population. In the Stepping-stonemodel (A), gene movement (m) is in a given generation (t) isonly among adjacent, evenly-spaced populations. In WrightsIsland model (B), the rate of gene movement in any givengeneration is equal among all populations regardless of spatiallocation. In the Isolation-by-distance model (C), gene flow ratesare a function of distance, with gene flow being higher amongneighboring populations and lower among populations moreremote from one another. (After Hamrick and Nason 1996.Illustration by M. J. Bigelow.)


16/32


of diversity (DST

) to the total diversity is a measure of geneflow, or Neis G

ST, expressed as

GD

HST

ST

T

=

The model of dispersal will affect the value of GST

. TheStepping-stone Model, with its low rates of gene flow, shouldshow the greatest differences between population subunits(highest G

ST). The high levels of gene flow predicted by the

Island Model should more thoroughly mix the genetics ofthe population and produce a low value of G

ST.

6.5.3. Models of Recolonization: PropagulePools and Migrant Pools

In nature, populations do suffer local extinction, often tobe recolonized later from other sources. As with the moregeneral models of dispersal, the degree of subunit dif-ferentiation and associated value of G

STwill affected by

the pattern of recolonization (Figure 6.12). The patternsof recurrent local extinctions and recolonizations charac-teristic of metapopulations significantly affect the geneticstructure within and among local population subunits(demes), and additional genetic models of metapopula-tions have been formulated to explain and predict the intri-cacies of metapopulation genetics.

The nature of the genetic effect depends upon specific rec-olonization patterns. The Propagule-Pool Model (Figure6.12A) is a genetic analogy to the previously discussedMainland-Island Model of metapopulations because itassumes that all colonists are drawn from a single extant

deme in the metapopulation. In Propagule-Pool Model ofrecolonization, the vacant habitat associated with a localextinction is recolonized by individuals (propagules) thatall come from the same subunit (pool).

In contrast to the Propagule-Pool Model of gene flowin metapopulations, the Migrant-Pool Model assumesthat colonists to a new deme are drawn randomly fromthe entire metapopulation (Slatkin 1977) (Figure 6.12B).In both models, local extinctions cause decreases in thegenetic diversity of the metapopulation, both withinand among demes, but genetic diversity is maintainedat higher levels in the Migrant-Pool Model (Pannell andCharlesworth 1999). Such higher levels of diversity occurbecause sites colonized by individuals from differentdemes approximate the genetic diversity of the entiremetapopulation, rather than the genetic diversity of onlyone deme.

We will not work out the details of the mathematicalrelationships here (Wade and McCauley 1988), but wewill examine the general trends and implications of thesetwo models for changes in population genetic structure.Specifically, if the Propagule-Pool Model is a recurrentpattern, recolonized subunits will each have unique genetic

traits heavily influenced by their founders (the originalpropagules), and their genetic differentiation (G

ST

), com-pared to the overall population, will be greater. In contrast,the Migrant-Pool Model assumes that colonists for vacanthabitat are recruited randomly from many pools. In thisscenario, genetics of the new population subunit are repre-sentative and may not be differentiated from other subunitsif the number of colonists is large relative to the number ofmigrants in the population (Wade and McCauley 1988). Ifthe number of colonists is small, colonists are less likely tobe representative of the population and the predicted valueof G

STwill be higher.

Figure 6.12. Two models of extinction and recolonization inspatially separated populations of the same species. The opencircle represents a site occupied by a population that has recently

become extinct. In the Propagule-Pool model (A), colonists to thevacant site are recruited from a single extant population, resultingin more genetic differentiation among populations (increasingG

ST). In the Migrant-Pool model (B), colonists are recruited ran-

domly from all populations. As a result of recolonization by thisrandom sample, genetic differentiation may increase or decrease.If the number of colonists is large relative to the number ofmigrants, G

ST will decline, but if the number of colonists is

small relative to the number of migrants, GST

will increase. AfterHamrich and Nason 1996:209. Illustration by M. J. Bigelow.


17/32


The predictions of these models have been supported bystudies of real metapopulations, reinforcing the assumptionthat genetic diversity within and among demes is influ-enced by characteristics and methods of dispersal, as wellas distances between populations. Undersea hydrothermalvents offer a classic metapopulation example. The ventsare highly specialized, discrete, patchily distributed, and

surrounded by non-habitat of more typical ocean floor.Deep-sea invertebrates often disperse along rifts in theocean floor where vent activity is prominent. Vrijenhoek(1994) found that indices of gene flow and genetic differ-entiation declined with increasing distance between popu-lations in some species that had poor dispersive abilities,such as the giant tube worm (Riftia pachyptila) (Figure6.13), a result consistent with the Isolation-by-distanceModel of gene flow previously described. Distance hadless effect on species with greater dispersive abilities, suchas the musselBathymodiolus thermophilus, whose mobilelarvae disperse widely via ocean currents.

These models also have implications regarding the

effect of habitat fragmentation. If a population has rela-tively little differentiation between subunits (low geneticstructure), fragmentation is unlikely to result in significantloss of genetic diversity because each subunit is reasonablyrepresentative of the entire population (Figure 6.14A). If,on the other hand, a population has a high level of differ-entiation between subunits (high genetic structure), thenthere is the potential for significant loss of populationgenetic diversity (Figure 6.14B). Conservation geneticistsJames Hamrick and John Nason note a further dimensionof these implications. As a result of the expected loss ofgenetic diversity, the tendency is to abandon fragments as

being of little value. This may be a mistake, because frag-ments, if maintained and preserved, can serve as reservoirsof genetic diversity (Hamrick and Nason 1996:228).

Although such models are useful in making predictionsin field experiments, real populations also are influencedby the type of breeding system employed, especiallyin plants, where breeding systems strongly influencedispersal distances and genetic exchange. Hamrick and

Godt (1989) demonstrated that, both within species andwithin populations, outcrossing plant populations usuallyhad higher levels of polymorphic loci, a larger number

Figure6.13. A colony of giant tubeworms (Riftia pachyptila),denizens of deep-sea hydrothermal vents and a species withpoor dispersive abilities. Genetic differentiation (G

ST) between

populations is high and increases with distance, supporting theIsolation-by-Distance Model of genetics in metapopulations.The genetic structure of mussel populations, represented by someindividuals around the tubeworms, will be less affected by dis-tance between population subunits because of their highly mobilelarvae. (Photo courtesy of Karen Von Damm.)

Figure 6.14. The effect of habitat fragmenta-tion on maintenance of genetic diversity in ametapopulation. Under conditions of low geneticstructure (A), there is little loss of genetic vari-ation following fragmentation because result-ing populations in the fragments are randomsamples of the populations original geneticdiversity, and there is little genetic differentiation(low G

ST) among fragments. If a population has

high genetic structure (B) (populations in differ-ent parts of the original habitat are geneticallydifferent), fragmentation can result in greaterdifferentiation of remaining fragmented popula-tions (high G

ST), with concurrent loss of genetic

diversity because lost habitat contained geneti-cally unique populations. (After Hamrick andNason 1996. Illustration by M. J. Bigelow.)


18/32


of alleles at each polymorphic locus, and higher levels ofheterozygosity than selfing plant populations, and windpollinated species tended to be higher in all these meas-ures of genetic diversity than animal pollinated species(Hamrick and Godt 1989).

In both individual population units and metapopulations,geneticists today emphasize the dangers associated with

inbreeding (Frankham 1995a), but some conservation biolo-gists have asserted that genetic concerns in small popula-tions are of less importance than historically believed andthat loss of diversity is more likely to be a symptom ofendangerment than its cause (Holsinger et al. 1999). AsHolsinger et al. (1999) put it, those alleles most likelyto be lost as a result of genetic drift rare alleles are alsothe least likely to contribute to any immediate response tonatural selection. Further, there are some small popula-tions that show excellent long-term viability with no sign ofnegative genetic effects (Simberloff 1988; Walter 1990), andthus cast doubt on the assertion that inbreeding is alwaysdetrimental to population persistence and viability. In fact,

some wild populations are notoriously inbred, yet persistent,such as the blue duck (Hymenolaimus malacorhynchos), anendemic species of New Zealand, whose inbreeding seemsto be a product of its isolated populations and low dispersalabilities (Triggs et al. 1992). As more studies are completed,the effects of inbreeding depression have become the subjectof intense debate. The critical question is: can we identifyany specific cases in the wild, in populations or populationsubunits, in which inbreeding is the identifiable causativefactor in extinction? This is a critical question for assessingthe relevance of conservation genetics to conservation biol-ogy. We will examine the question in detail, learning what

we can from captive and laboratory populations, theoreticalmodels of populations, and natural populations themselves.

6.6. Can Inbreeding Cause Extinction?

6.6.1. Laboratory Experiments and Models

In laboratory experiments on inbreeding in the fruit fly,geneticist Richard Frankham and his colleagues havebeen able to demonstrate that inbreeding and associatedloss of genetic variation (1) reduced resistance to disease,

(2) led to increased extinction rates in environmentallystressed populations, and (3) reduced adaptive evolution-ary potential (heritability) in small populations subjectedto stressful environments (Frankham 2005). Further,Frankhams studies demonstrated that, when populationswere subjected to environmental stress, rates of inbreed-ing increased.

In a controlled experiment involving both field andlaboratory environments, Jimnez et al. (1994) estimatedthe survivorship of inbred and non-inbred white-footedmice in a mixed deciduous forest in Illinois (USA). Mice

in both treatments had been raised in a laboratory but wereall descendents of wild mice. Mice were released into thefield and recaptured at regular intervals, and trapping wasdone on surrounding adjacent habitat to estimate the pro-portions of mice that left the release site. Most mice werecaptured within 50 m of their release sites, and emigrationrates were not different in the two treatments. Based on

releaserecapture ratios of marked mice in both inbred andnon-bred treatments, Jimenez et al. determined that, over a10-week period, non-inbred mice had consistently highersurvivorship than non-inbred mice (Figure 6.15). Some dif-ferences were sex-specific. For example, male mice in bothtreatments lost weight in the first few days after release.Non-inbred males regained their weight loss, but inbredmales continued to lose weight throughout the experiment.Overall, a recent analysis of 33 population data sets revealedthat approximately 40% of the variation in fitness in thesepopulations could be explained by variance in heterozy-gosity, which is a reliable index of inbreeding (Reed andFrankham 2001).

Low survivorship is one explanatory factor in extinc-tion, and there is a clear threshold relationship betweenrates of inbreeding and rates of e

6 Genetic Diversity – Understanding Conservation at Genetic Levels

Documents

Transcript of 6 Genetic Diversity – Understanding Conservation at Genetic Levels