Effective Population Size
To estimate the effects of bottlenecks and random genetic drift, as presented in Tables 5.2 and 5.3, we made some simplifying assumptions: that the organism is diploid, sexually reproducing, and has nonoverlapping generations; that the population is of constant size, has equal numbers of females and males, random mating, and no migration; that reproductive success of all individuals is the same; and that no mutation or natural selection occurs. Of course these assumptions are violated in any natural population. But making these simplifying assumptions allows us to avoid a major complexity: the difference between total or census population size (the actual number of individuals in a population) and the effective population size. To take a very simple example, let us return yet again to bison. Consider a population of 100 bison in which 25 are too young to breed and 15 adults are infertile, so 60 is the number of breeding adults and therefore the effective size of the population. In practice, the issue is more complicated. What if most of those breeding age adults are females (or males)? We have violated the assumption of equal sex ratios. Bison are of course large‐bodied and long‐lived creatures with overlapping generations. Bison moreover do not mate indiscriminately – some pairings are more successful than others, especially skewed among the bulls.
These biological realities mean that effective population size is usually much lower than actual (or census) population size. Effective population size is a very important concept in conservation biology. We will begin with a definition and then show two examples of how to calculate effective population size (see Frankham et al. 2009 for further details). First the definition: the effective population size (Ne ) of a population is the number of individuals in a theoretically ideal population (i.e. one that meets all the assumptions stated earlier) that would have the same magnitude of random genetic drift as the actual population. Now let us explore how biological realities reduce effective population size relative to census size.
Example 1. Population fluctuations. The effective size of a population that is fluctuating through time (as most do) is less than the actual population size. In this case, Ne is estimated to be the harmonic mean of the actual size of each generation (Hartl and Clark 1997). Mathematically,
In words, the harmonic mean is the reciprocal of the average of reciprocals of the population size for each of t generations. This method of estimating an effective population gives more weight to smaller population sizes. For example, the Ne for three generations (t = 3) in which N1 = 1000, N2 = 10, and N3 = 1000, would be
which is far less than 670, the arithmetic mean of 1000, 10, and 1000. (Also see Vucetich et al. 1997 and Lovatt and Hoelzel 2014 for the effect of population fluctuations.)
Example 2. Unequal numbers of females and males. If a population has an unbalanced sex ratio, the effective population size is less than the actual size and can be estimated as
where Nf is the number of breeding females and Nm is the number of breeding males (Hartl and Clark 1997). For example, if 96 females mated with four males,
This kind of imbalance, as extreme as it may seem, is fairly common. Genetic analyses that determine the mother and father of offspring frequently indicate that in many species relatively few individuals, especially among males, are responsible for a disproportionate share of a population’s reproduction (Parker and Waite 1997). Many apparently healthy adults do not leave any offspring. Such inequity is generally not a problem – it is the basis for natural and sexual selection – but it may lead to difficulties in populations that suddenly get small because of its effect on genetic diversity. Consider the endangered Española Island giant tortoise of the Galápagos (see Case Study 1.1, “Return of the Tortoises to Española”). These tortoises likely numbered in the thousands but after most were killed for food by early mariners the population plummeted to 15, consisting of 12 females and just three males, which fortunately were rescued, placed in captivity, and have produced more than 1200 offspring since 1950 that have been released back to the island where they now breed on their own (Gibbs et al. 2014). Sounds like a success? It is. But it has also been discovered that the unequal sex ratio of those 15 survivors along with the unequal reproductive activity among them had led to a genetic effective population size of just 5.7 tortoises, far smaller than the census size of 15 might suggest (Milinkovitch et al. 2004). What little genetic variation remains in this population is severely threatened by the genetic drift exerted during the very long and “tight” bottleneck this species endured for many decades. Carefully managed pairings of surviving tortoises are now being made to maximize “capture” of what little genetic diversity remains for the newer generations. The bottom line to remember is that the effective population size is often substantially less than the actual number of individuals in a population, often only 10–20% (Vucetich et al. 1997). Thus, if you want a “population” of bison with Ne = 100 (hopefully sufficient to retain 99.5% of its genetic variability through at least one generation; see Table 5.3), you actually need a census population of somewhere between 500 and 1000.
Inbreeding
Inbreeding refers to the mating between closely related individuals; such individuals are likely to share identical copies of some of their genes because they have ancestors in common. We measure inbreeding with the inbreeding coefficient, F, which is the probability that two copies of the same allele are identical by descent – in other words, derived from a common ancestor (Templeton and Read 1994). For example, in our bison example, if both MDH‐1 X alleles in the X/X homozygous buffalo calf were derived from its grandmother, those alleles would be considered identical by descent.
Among several methods to estimate F (Frankham et al. 2009), the simplest involves counting links in the pedigree chain: F = (½) n , where n is the number of individuals or links in the pedigree chain starting with one parent, going back to the common ancestor, and then going down the other branch to the other parent. Figure 5.12(a) shows the pedigree chain for the offspring (A) of a half‐sister (B) mating with her half‐brother (C) (i.e. B and C have the same mother, D, but different fathers). The inbreeding chain has three links – B, D, and C – and thus F is equal to (½)³ = 1/8 = 0.125. If B and C were full siblings (i.e. they had both the same mother D and the same father E) (Fig. 5.12b), then there would be two chains, one for each common ancestor (B, D, and C for the mother plus B, E, and C for the father). In this case the F values for each chain would be added: (½)³ + (½)³ = ¼ = 0.25.
Figure 5.12 Inbreeding