The establishment of minimum viable population sizes is one of the principal goals of preservation genetics. In arriving at such minimum sizes it is necessary to consider all aspects of the biology of the species involved, not just the genetic. Other important criteria would include the demography and life history of the species, and certain ecological variables, for example, the probability and severity of catastrophe.
Very little in general can be said about the latter subject because the nature and consequences of catastrophe are highly dependent on the life history of the species and particularly on the kind of environment in which it lives. For instance, fishes living in shallow bodies of water in regions subject to extreme drought are likely to have a high probability of extinction. Also, they may go through severe bottlenecks with significant frequency (see Section 4.3). At the opposite extreme, deep sea species are unlikely to be exposed to physical events which bring about a collapse in population size.
Given sufficient demographic information, it is sometimes possible to produce estimates of minimum viable population sizes for a species. But even when the demographic information is at hand, the genetic approach is relevant and could be the dominating consideration, assuming that the minimum sizes based on a genetic criterion are smaller than those based on demographic or ecological criteria.
The “time scale of survival” is a useful device for structuring a discussion of preserving genetic variation. Somewhat arbitrarily, there are three problems or issues:
a short-term issue is immediate fitness - the maintenance of vigour and fecundity during an interim holding operation, usually in an artifical environment, such as when breeding domesticated or semi-domesticated fish stocks. (If, however, breeding is expected to continue for more than Ne (see Section 4.2) generations, the programme, in effect, becomes a long-term operation.);
the long-term issue is adaptation - the persistance of vigour and evolutionary adapatability of a population in the face of a changing environment;
the third issue is evolution in the broadest sense, i.e., speciation, or the creation of evolutionary novelty (Soule, 1980). For our purposes, the first and second issues are the most relevant and the third is the least relevant.
From the above discussions on heterozygosity and inbreeding in both natural and captive populations (see Section 3), we can conclude that excessive loss of genetic variability, particularly inbreeding, can and must be avoided. The issue, however, is a quantitative one and the above qualitative discussions are not very helpful in providing specific guidelines to minimum viable population size. We must examine this point in greater detail.
Captive populations will tend to be small and potentially subject to the deleterious effects of inbreeding. By trial and error, animal breeders have discovered the magnitude of inbreeding that can be tolerated by domestic animals before the lines begin to decline in fitness. (In discussing inbreeding, it is convenient to use the inbreeding coefficient, F, which is a quantitative measure of the magnitude of inbreeding. In a population that is totally outbreeding, F = 0. For a population that is totally inbreeding, F = 1.0.)
A general rule is that the per generation rate of inbreeding should not be higher than one to three percent (Franklin, 1980; Soule, 1980). Higher rates of inbreeding fix deleterious recessive genes too rapidly for selection to eliminate them, and the vigour and fertility of the line decreases.
The lower inbreeding rate of one percent (F = 0.01) is preferred because:
poultry and mammal stocks have been partially purged of deleterious genes over the milennia which allows them to tolerate higher rates of inbreeding than wild outbreeding species;
animal breeders can safely ignore some inbreeding and random loss of genes. In constrast, conservationists wish to preserve the “wildtype”.
How does this basic “one percent rule” translate into population size? The rate of loss per generation of heterozygosity due to inbreeding as measured by F is equal to 1/(2 Ne), where Ne is the effective population size. (“Effective population size” is the size of an idealized population. The definition of Ne is cumbersome, but the population must have an equal sex-ratio and individuals must mate at random. A number of additional “ideal” characteristics could be stated. For our purposes, it is important to note that in practice Ne is nearly always smaller than the actual number of breeding individuals.) Thus, Ne must equal at least 50 if the inbreeding rate is to be kept below the one percent level.
However, even when F = 0.01, the loss of genetic variation is appreciable after a few generations, and a gradual attrition of genetic variation cannot be prevented. Eventually, the population will become virtually homozygous, the time depending on Ne. This is why the one percent rule must be viewed as short-term criterion. A population held in check at Ne = 50, will lose about one-fourth of its genetic variation after 20 to 30 generations, and along with it, much of its capacity to adapt to changing conditions. Thus, if it is desired to maintain a particular stock for longer than this, it will be necessary to increase its Ne. A rough rule of thumb is that G is approximately equal to Ne, G being the number of generations the stock is likely to retain its fitness at a relatively high level.
The above information is necessary but not a sufficient basis for the conservationist or the fisheries' biologist to conserve short-term fitness, or to maintain short-term fitness in captive populations of fish. The reason is that the effective population size is not a simple phenomenon and is affected greatly by variation in sex ratio, population size through time, by a non-random distribution of progeny among families, and other aspects of the breeding system. To the extent that any of these effects occur, a larger absolute population size must be maintained to achieve a desired Ne. Section 4.3 summarizes these complicating factors.
When populations decline or “crash”, the survivors constitute a genetic “bottleneck” in the history and evolution of the population. Any deviation in the genetic makeup of these survivors from the gene pool of the original population will be reflected in future generations. More particularly, if the progenitor's gene pool is less diverse than that which existed in the original population, future generations will have a corresponding deficit in genetic diversity.
If the minimum population size is very small, due either to normal fluctuations or to an environmental changes or catastrophes, it is tantamount to squeezing the genetic variability of the source population through a very narrow channel and eliminating a significant amount of this variability. Bottlenecks inevitably accompany the establishment of a captive stock for breeding purposes.
Prevention or further genetic erosion or a recovery to the original level of genetic variation depends greatly on how fast the population grows to a moderate size or several hundred or more. If a preserved population is subject to fluctuations in numbers (as it most probably will be), the influence of the minimum absolute size on effective population size is more relevant to preservation of genetic diversity than the average absolute size.
The loss of genetic variability concomittant with the bottleneck event has both qualitative and quantitative aspects. Qualitatively, specific alleles may be lost and if they are lost it is very unlikely that they will be replaced by mutation as long as the population remains small. Quantitatively, the variability for specific traits will be reduced and the mathematics in the loss of the variance of quantitative traits have been described by Falconer (1960) and others. The qualitative effect is usually greater than the quantitative one; that is, the loss of alleles, especially low-frequency alleles, is much greater than is the loss of genetic variance per se. Incidentally, several workers have pointed out that the number of founders in a colony, so long as it is greater than about five individuals, is not nearly as important as the long-term maintenance size of the colony (Nei et al., 1975; Denniston, 1978). That is, a single bottleneck event followed by rapid growth to a large size, say 2Ne greater than 500, does relatively little damage, compared, that is, to a chronically small Ne.
We know from experience with resistance to pesticides in anthropods that some alleles that occur at very low frequency in natural populations (and which are likely to be lost during a bottleneck) can be very important. Such alleles can mean the difference between survival and extinction. The same thing probably applies to resistance genes in general. Therefore, it would be expected that populations of fish passing through bottlenecks might not be noticeably affected until a disease epidemic swept through the population. Only then would the loss of these resistance genes be detectable.
Several other factors determine Ne. Among these are the sex ratio; Ne is lowered by deviations from an equal sex ratio. Another of the characteristics of a genetically ideal population is that the number of progeny are randomly distributed among families. When this condition does not hold, for example when the reproductive output of a few families is especially great, Ne will be lowered. It is incumbent on persons dealing with captive populations, for purposes of either preservation or culturing, to be aware of these effects and to maintain Ne at a level that will maintain the fitness of their stocks. In some cases, it will be desirable to consult with a population geneticist, especially if the breeder is in doubt about the estimation of Ne.
Long-term preservation requires rather large population sizes, large enough so that an equilibrium will be maintained between the loss of genetic variability due to drift and selection and its generation from mutation. When 2Ne is a large number, say greater than 500 or 1 000, the effect of drift will be negligible compared to that of weak selection. When 2Ne is small, say less than 100, the randomization of gene frequencies between generations will not only fix many loci, it will also counteract all but the strongest deterministic forces, particularly directional selection, thus, to a large degree precluding adaption by natural selection.
The consequences of small population sizes might be thought to be irrelevant when considering the genetics of fishes in an aquatic reserve, or similar ecosystem. However, there will always be some populations in a natural ecosystem, particularly large predators, which have quite small numbers (such as groupers on an atoll, pike or muskies in a lake). The loss of such keystone predators from a natural community can have serious effects on the diversity of prey species, as has been documented by many workers, particularly with marine invertebrate systems (Paine, 1966; Harper, 1969).
Franklin (1980) argues that a minimum effective size of 500 is needed to preserve useful genetic variation, because:
The relevant phenotypic traits in conservation are quantitative (polygenic). For such traits the average effect of a gene is small, and most of the genetic variation is additive.
Weak directional or stabilizing selection does not erode additive genetic variation at a significant rate;
The significant evolutionary forces, therefore, are mutation, and genetic drift. That is, if a population is below some threshold size, it loses variation by drift at a faster rate than it gains variation by mutation.
Franklin derives his number from the work of Lande (1976) on bristle number variation in Drosophila. The evidence is meagre but Franklin believes his number (500) is about the right order of magnitude. Simple theory also yields this number as mentioned above.
It is necessary to caution again that the employment of any number is subject to all the same qualifications given in the preceding section for short-term preservation, namely that any effective size translates into a much larger number of breeding adults, when dealing with real, not ideal, populations. In addition, this recommendation ignores genetic differences between species. Conservation questions of this kind must also be considered on a case by case basis.