untitled

which may alternatively be written:

Ht=H 0 e−t/2N

After t= 2 Ngenerations the population’s heterozygosity will have dropped to 0.37

(i.e. e−^1 ) of its initial value at time t=0. This holds as well for a single locus as it

does across all loci. The loss of additive genetic variance is exactly analogous and

conforms to the same equations. The process is called random genetic drift.

The rate of mutation at a single locus is about 10−^6 per gamete per generation.

However, most evolution is in terms of phenotypic characters that change in very

small steps. These quantitative characters are controlled by many genes. Mutation

within such a gene complex is much more frequent than at a single locus, closer to

10 −^3 or even 10−^2 per gamete per generation.

Heterozygosity (and hence additive genetic variance) will change over one gener-

ation by an amount ∆Haccording to:

∆H=−H/2N+m

where mis the input of heterozygosity by mutation. Its equilibrium is solved by set-

ting ∆Hto zero:

H= 2 Nm

which informs us that for any population size Nthere will be an equilibrium

between mutational input of additive genetic variance and loss of it by drift. What

varies, however, is the value of Hat equilibrium. It will be higher when Nis large

and lower when Nis small. The population size Nmust remain constant for many

generations for such an equilibrium to establish.

Selection generally comes in two forms: directional and stabilizing (or normalizing).

Directional selection, that which moves a trait in one direction, is the stuff of evo-

lution. Stabilizing selection, on the other hand, eliminates extremes of a trait and

holds the trait to its optimum for the current environment. There is a third form,

disruptive selection, where individuals with intermediate traits are selected against

whereas individuals with extreme forms for the traits are favored, but it is much rarer

than the other forms of selection.

At any particular time most selection will be of the stabilizing type. Any tendency

for the breeding season to expand, for example, will be attacked continuously by

stabilizing selection. An offspring born during a seasonally inappropriate time of the

year has little chance of survival and so that mistake, if it has a genetic basis, will

be swiftly corrected. The relaxation of that selection in captivity is presumably the

reason why the breeding season of captive populations tends to expand after several

generations.

Stabilizing selection is essential to maintaining the fitness of a wild population.

Ironically, it is one of the strongest forces reducing additive genetic variance and

heterozygosity. We must therefore avoid an obsession with maximizing variance to

the exclusion of maintaining fitness by reducing the number of deleterious alleles.

CONSERVATION IN THEORY 295

17.3.4Selection