SUPPORTIVE BREEDING MAY REDUCE FITNESS IN WILD
 
Ford, Michael J., June 2002. Conservation Biology. 16:815-825.
 
ABSTRACT:
I used a quantitative genetic model to explore the effects of selection on the fitness of a wild population subject to supportive breeding.Ê Supportive breeding is the boosting of a wild fish populationâs size by breeding part of the population in captivity and releasing the captive progeny back into the wild.Ê The model assumes that a single trait is under selection with different optimum trait values in the captive and wild wild population will shift the wild populationâs mean phenotype so that it approaches the optimal phenotype in captivity.Ê If the captive population receives gene flow from the wild, the shift in the wild populationâs mean phenotype becomes less pronounced but can still be substantial.Ê The approach to the new mean phenotype can occur in less than 50 generations.Ê The fitness consequences of the phenotypic shift depends on the details of the model, but a >30% decline in fitness can occur over a broad range of parameter values.Ê The rate of gene flow between the two environments, and hence the outcome of the model, is sensitive to the wild environmentâs carrying capacity and the population growth rate it can support.Ê The results have two important implications for conservation efforts.Ê First, they show that selection in captivity may significantly reduce a wild populationâs fitness during supportive breeding and that even continually introducing wild individuals into the captive population will not eliminate this effect entirely.Ê Second, the sensitivity of the modelâs outcome to the wild environmentâs quality suggests that conserving or restoring a populationâs habitat is important for preventing fitness loss during supportive breeding.
 
CONSERVATION IMPLICATIONS:
For salmonids there are several examples of reductions in fitness caused by captive propagation ( e.g. Reisenbichler and McIntyre 1977; Leider et al. 1990; Fleming et al. 1996).Ê 
 
The model provides some insight not immediately obvious from existing empirical studies.Ê First, most experimental studies of the fitness effects of captive breeding focus on the survival or reproductive success of the captivity propagated fish when they are released into the wild.Ê The long-term effects of such releases on the mean fitness of a supplemented wild population have not been measured, but the model shows that these effects can be significant.
 
Second, most empirical studies to date have focused on genetic changes in captive populations that received few if any wild immigrants subsequent to the initial founding of the captive population.ÊÊ In contrast, nearly all the more recent supportive-breeding projects for Pacific salmonids involve deliberately managed migration between captive and wild populations.Ê The primary purpose of regularly bringing wild-origin broodstock into these programs is to avoid domestication of the captive stock.Ê I explicitly explored this two-population scenario and found that substantial phenotypic changes and fitness reductions can occur even if a large fraction of the captive broodstock is brought in from the wild every generation.Ê This suggests that regularly bringing wild-origin broodstock into captive populations cannot be relied upon to eliminate the effects of inadvertent domestication selection, although the rate and level of domestication will be reduced compared with those of a completely closed captive population. 
The sensitivity of the degree of phenotypic change to the relative proportions of captive and wild individuals·has two interesting implications.Ê First, it means that wild-origin breeders are important to a populationâs viability in the wild even in cases where the wild population is not able to sustain itself without the aid of supportive breeding.Ê Those wild populations with relatively high reproductive rates 0.9, however, were much less prone to phenotypic change during supportive breeding than less healthy populations.
 
The dependence on the potential reproductive rate in the wild environment means that conserving or restoring a populationâs habitat · may be the most effective method of preventing phenotypic change during supportive breeding, even if these improvements are not sufficient to allow the population to sustain itself entirely naturally.Ê The results also suggest that controlling the exchange rate between captive and wild populations would be an effective way of limiting domestication of wild populations.Ê In a supportive breeding situation, however, where the goal is to use captive individuals to increase the size of a declining wild population, it may be impossible to achieve the desired demographic boot while keeping the proportion of captive-origin individuals in the wild population low.Ê 
 
The sensitivity of the model to the wild reproductive rate also implies that those populations most in need of supportive breeding are also those most vulnerable to phenotypic change due to selection in captivity.Ê This suggests that in situations where improving the wild reproductive rate is not feasible, it is especially critical to avoid strong selection in captivity.Ê Representative sampling of the population for broodstock and the use of ãnaturalä methods of breeding and rearing may be particularly important in situations where wild reproductive rates are low, although the ability of these measures to adequately mimic the selection that occurs in the wild environment is unknown.