Is local provenance important in habitat creation? A reply


N.R. Sackville Hamilton, Institute of Grassland and Environmental Research, Plas Gogerddan, Aberystwyth, Ceredigion SY23 3EB, UK (fax +44 1970821987; e-mail


  • 1Wilkinson (2001) argues that we cannot assume that hybrids between local and alien genotypes will have low fitness, and therefore, as low hybrid fitness has been presented as justification for using only locally provenanced material in habitat restoration schemes, provenance is not important.
  • 2His observations on fitness are important, correct and deserve wider recognition.
  • 3Nevertheless, I dispute his conclusion about the importance of provenance, for two main reasons. One is that his argument is based on questionable objectives for biodiversity conservation. The second is that, even if we accept these underlying objectives, the fitness of hybrids is only one of numerous relevant issues.
  • 4Use of locally provenanced seed should be standard practice, except where the introduction of non-local genotypes is specifically justified in terms of conservation genetics.

Local adaptation and hybrid fitness

Wilkinson (2001) argues that genotypes of local provenance do not necessarily have superior fitness, and hybrids between local and introduced genotypes do not necessarily have low fitness. This is fully supported by both theory and empirical evidence, although the point is not new. It was also made by Gould & Lewontin (1979), Clutton-Brock & Harvey (1979) and Harper (1982). These keystone papers criticized the evolutionary and ecological community for its uncritical Panglossian assumption that populations are optimally adapted to their environment and populations from other locations are less well adapted.

Much genetic variation is adaptive (in the retrospective sense in which this word is commonly but loosely used in an evolutionary context; Harper 1982). In a large literature studying adaptation by reciprocal transplant experiments, the majority of publications show that most populations have relatively high fitness in their home environment (Nagy & Rice 1997). Genetic introgression from populations with different combinations of co-adapted genes can have additional adverse effects on hybrid fitness (Keller, Kollmann & Edwards 2000). However, there are also numerous counter examples (Norton et al. 1999). Indeed, the worst cases of invasive species, e.g. prickly pear Opuntia spp. P. Mill., rabbits Oryctolagus cuniculus L. and Japanese knotweed Reynoutria japonica Houtt., have involved the introduction of exotics.

There are also abundant theoretical reasons for expecting adaptation to the local environment not to be optimal. Even in the earliest days of the formalization of natural selection theory, Fisher (1930) was able to demonstrate conditions under which natural selection does not improve fitness. Harper (1982) itemized seven distinct reasons for expecting genotypes not to be optimally adapted.

Unfortunately, it appears that, despite the publication of such hard-hitting papers by such eminent scientists, the ‘Panglossian paradigm’ (Gould & Lewontin 1979) still dominates much evolutionary thinking. Wilkinson is correct that it pervades the debate about the importance of provenance. In this respect, he is correct to focus on the issue of hybrid fitness. However, as recognized by Keller, Kollmann & Edwards (2000), hybrid fitness is only one of numerous relevant issues. I shall now consider these other issues, beginning with the basic objective of conservation.

What are we trying to conserve?

According to the Convention on Biological Diversity (CBD), ‘Biological diversity means the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems’ (CBD 1992, article 2). This definition, accepted at least by the 180 countries that signed up as parties to the CBD, explicitly ranks genetic diversity within species alongside species diversity as a component of biodiversity. The reasons given (CBD 1992, preamble) and the primary provisions made for conserving biodiversity in situ (CBD 1992, article 8) make no distinction between these components. That is, conservation of the patterns of genetic diversity within species is an objective in its own right, for the same reasons and with the same importance as conservation of species diversity.

With the exception of a handful of well-publicized cases, every species studied to date encompasses within it a wide range of genetic diversity, both within and between populations. Therefore, at least for the signatory countries to the CBD, the answer to the title question is indubitably and unarguably ‘yes’. If we ignore the provenance of seeds used for habitat creation, we risk disrupting the genetic structure of the species sown and we thereby risk failing to conserve biodiversity adequately.

Thus the primary reason for controlling provenance is to conserve genetic diversity in accordance with the CBD. The potentially detrimental effect of introduced seed on fitness is an additional but secondary reason. Wilkinson’s logic (that provenance is not important because the supposed effects of provenance on fitness are questionable) would apply only if the fitness effects were the sole reason for controlling provenance. Implicit in his argument, therefore, is the assumption that conserving genetic diversity is important only to the extent that it enhances the conservation of species. The same assumption is also apparent in the secondary importance he attaches to the fact that disrupting patterns of genetic diversity will ‘lead to the destruction of interesting genetic information’. Without accepting his assumption, I now turn to argue on his own grounds, by considering the relationship between genetic diversity and species survival.

Evolution, genetic diversity and extinction

An inability to evolve far enough or fast enough in response to a changed environment is, evolutionarily, the cause of extinction. The rate at which a population can evolve is, loosely, proportional to its genetic variability (Mather 1973). Progress towards extinction can be accelerated by inbreeding depression in an obligate outbreeder as it becomes rare. Outbreeding depression can also speed progress towards extinction if the remnants of the species comprise genetically very distinct populations.

In this context, important issues include metapopulation structure and gene migration between populations, and their effect on the magnitude and distribution of genetic diversity within and between populations. These in turn affect the potential ecogeographic range of the whole species, inbreeding depression and the probability of population extinction, and the potential for evolutionary adaptation to changing environments. For example, a small population with no gene flow between populations will lose genetic diversity by drift, and will lose its ability to adapt to local conditions. Zero gene flow between populations can thereby reduce the total genetic diversity in the species and the fitness of genotypes in each population. At the opposite extreme, unlimited gene flow between populations prevents the evolution of distinct locally adapted subpopulations, and so can also reduce the total diversity and fitness. Total diversity and fitness are maximized by limited gene flow that leads to locally adapted populations, genetically distinct from each other as well as being genetically diverse within populations.

Another important issue is the relationship of genetic distance between populations with geographical and ecological distance between their locations, and the fact that the relationship is expected to be different for neutral and non-neutral genes. A key quantity in this relationship is the genetic population area, i.e. the area within which mating occurs at random. No genetic differentiation, in neutral or non-neutral genes, can persist across generations within smaller areas. In plants, the size of this area is usually remarkably small, often only a few square metres (Hayward & Sackville Hamilton 1997).

At larger scales, the genetic distance between populations is expected to increase with the ecogeographic distance between them. For neutral genes, we expect no causal relationship with ecological distance but an asymptotic relationship with geographical distance. In populations so far apart that they are effectively completely isolated, with no gene flow between them even by ‘leap-frogging’ intermediate populations, there is no relationship between geographical and genetic distance. Because of leap-frogging and the skewed distribution of pollen and seed dispersal, the distance between populations required for complete isolation is several orders of magnitude greater than the genetic population area; for example, in perennial ryegrass Lolium perenne it appears to be over 100 km (Monestiez, Goulard & Charmet 1994).

In contrast, for non-neutral genes, genetic distance is primarily a function of difference in selection pressures, in turn a function of ecological distance. These effects can be seen over scales little larger than the genetic population area. Plants that superficially appear to form one physically continuous population can in fact be differentiated into genetically distinct and differently adapted subpopulations separated by only a few metres. There is then little relationship between genetic distance and geographical distance except as a secondary consequence of ecological distance increasing with geographical distance.

Also important is the concept that neutrality of genes depends on the situation. A functional gene that has no effect on fitness in some genetic backgrounds for some selection pressures may not be selectively neutral under other situations. This means that loss of currently neutral genes may reduce the potential to adapt to future changes in situations where the same genes are no longer neutral.

Yet another issue is that introducing exotic genotypes itself constitutes a change in environment that may affect survival of other components of the ecosystem. For example, Jones, Hayes & Sackville Hamilton (2001) demonstrated that imported hawthorn Crataegus monogyna can flower up to five weeks earlier than native UK hawthorn, potentially threatening the insects and birds whose reproductive cycles are timed to the phenology of the native genotypes.

Consequences for introductions

Introduction of genotypes from different populations is in one sense simply another form of gene flow. However, unlike artificial introductions, normal gene flow is highly skewed, with most seed or pollen typically moving only a few metres. Thus one introduction event can correspond to many hundreds of generations of normal gene flow. As such, artificial introduction can easily interfere with the process of evolution of locally adapted races, and we need to be very cautious about introducing seed.

Conversely, an insistence on using nothing but locally provenanced seed could also be detrimental, particularly for remnant species present as isolated populations where the natural level of gene flow is already below acceptable levels. Limited introduction of seed from other sources can produce new genotypes to be ‘tested’ by natural selection. Regardless of the provenance of introduced seed, if it hybridizes with the native genotypes there will be an increase in genetic variance that is likely to increase the rate of evolution. However, if an inappropriate provenance is used, this increased rate of evolution will simply involve elimination of the introduced genes rather than evolving new adaptation.

Therefore, choosing the provenance of material used for habitat recreation is vitally important. It will normally be safest to use the local provenance, although it may not always be the best choice. ‘Local’ must always be defined in terms of ecological proximity as well as geographical proximity.


Wilkinson (2001) is correct to challenge popular perceptions about the relationship between provenance and fitness, and to challenge the uncritical acceptance of using only locally provenanced material for habitat recreation. However, I argue that hybrid vigour is only one small part of the issue, and that the question of whether to use locally provenanced material for habitat recreation is vitally important. It should be used in all the situations where Wilkinson asserts it does not matter.

That is not to say that we must always use locally provenanced material. On the contrary, there are situations in which it may be necessary or even beneficial to introduce genotypes from other populations. The key message of this paper is that we must always consider the question carefully. Using locally provenanced material should be the normal practice in most situations, including the common situation where we do not have enough knowledge to assess the genetic consequences of introducing genotypes from elsewhere. We should use other genotypes only when there is good justification for doing so in terms of conserving or enhancing the genetic structure of the species, and even then we should choose the provenance carefully.