Genetic variation and local provenance
John Warren, Farmland Ecology Unit, Scottish Agricultural College
The decline of many habitats and the species they support, has caused increasing demand, now incorporated in conservation policy, to re-introduce species and re-create habitats. Both these activities may, of necessity, involve the introduction of non-local genotypes. This possibility has itself raised concerns regarding whether such introductions will be genetically adapted to their new homes, and beyond that whether they will impact negatively on the ecosystem into which they are introduced or on any remnant local populations.
If genetic variation within species is randomly dispersed over space, then provenance would not be an issue. Thus, any attempt to assess the risks associated with the introduction of non-local genotypes, needs to understand the nature of the distribution of genes within species. In the following review I will consider the relative importance of neutral and adaptive variation and the potential consequences of disturbing natural patterns in this variation.
Understanding the scale over which such genetic variation occurs in nature is vital if we are to understand what local really means in terms of provenance. Such information is also likely to throw light upon which species are likely to be most affected by the introduction of 'alien' ecotypes.
Many plant species occur in nature as more or less discrete populations, this is increasingly the case as a result of habitat fragmentation. Gene flow may occur between such populations via the movement of seeds, fragments of plants or pollen. In the absence of any selection, how genetically distinct populations are is related to the balance between gene flow among populations and genetic drift within them. In the absence of gene flow populations are expected to differ, due to the random selection of the small number of individuals usually involved in founding of a new population.
Modern molecular genetics has shown that the vast majority of DNA within a cell is non-coding and is thought to be invisible to the forces of natural selection. For the most part it is this non-coding DNA which is the target of most studies of molecular genetics. For this reason we expect to find, simple positive relationships between geographic distance between populations and genetic distance between populations, when measured via molecular methods. Populations which are close together are more likely to exchange genes than are population separated by greater distances. Such a relationship between geographic distance and genetic distance can be seen in the grass Agrostis curtisii, (Bristle Bent). This species occurs on lowland heaths of southern England in isolated populations. Gene flow within this species has not been greatly directly influenced by man, allowing the expected natural distribution of neutral genetic variation to be observed. In contrast, when neutral genetic variation is observed in populations of the common agricultural species Lolium perenne (Perennial ryegrass), by chance, the exact opposite relationship is found with geographic distance. The most likely explanation for this unexpected observation, relates to the considerable amount of historic movement of ryegrass that has occurred related to its agricultural usage. Typically we would expect, massive random gene flow within a species to result in no relationship at all between genetic distance and geographic distance. In ryegrass any natural patterns of genetic variation within the UK have long since been disrupted. But does this matter? This native species is also a component of many natural habitats. Yet the total disruption of the natural pattern of genetic variation which has occurred even in old unimproved pastures and meadows seems to have occurred within this most common of species without any apparent impact on the species itself or to the natural ecosystems in which it is found. However, this is not to say that historic patterns of neutral genetic variation are unimportant. Such patterns can have much of interest to tell us about the history of colonisation of the species, its breeding system, patterns of gene flow, effective population size and the amount of extant genetic variation. This information is of value to both theorists and conservation practitioners alike. The existence of a band on a gel associated with Scots Pine from Shieldaig does not necessarily mean that trees without this band will not survive in the area, but it may provide a fascinating insight into how our flora survived the last glaciation. Such information could easily be lost forever.
As early as the beginning of the eighteenth century botanists began to recognise that many species did not flower at the same time in different populations and could be described as separate types. Ludwig in 1901, coined the term 'local race' following his studies which showed that plants of Ranunculus ficaria (Lesser celandine), from different populations differed in the numbers of carpels and stamens within their flowers. The situation was however, more complex than originally assumed, as many such differences between populations are entirely plastic and disappear when plants are grown together in a uniform environment. Indeed Ludwig's original observation could also be explained as seasonal variation, since carpel number decreases as spring progresses within successive flowers on the same plant. Controversy surrounded the reality or otherwise of ecotypes for the first two decades of this century. By 1925, greatly due to the pioneering work of Turesson, a genetic component had been demonstrated to adaptive inter-population variation within more than fifty species. Turesson divided ecotypes into two classes, those adapted to climatic factors (climatic ecotypes) and those adapted to soil type (edaphic ecotypes). The idea that local populations are adapted to local conditions was further supported by Stapledon (1928) observations of Daclylis glomerata (Cock's-foot). He found that early flowering plants were associated with meadows cut for hay, whereas plants originating from grazed meadows tended to produce more tillers. This observation was the first example of a biotype, that is a local variant of a species, adapted to biotic effects rather than climatic or soil variables.
Since that time much more has been discovered about the nature of the spread of adaptive variation within species. The largest data set of geographic/genetic variation exists for the flower colour polymorphism in Lotus coriculatus (Birdsfoot trefoil). As can be seen from the work of Crawford & Jones (1988), populations in the south and west of the UK predominantly consist of plants with flowers with yellow keels. In the north and east the proportion of plants with dark keels rises to above 75 percent. Although the exact nature of the selection acting upon this variation is uncertain, it seems likely that dark keels are favoured in slightly cooler climates because they are more efficient at absorbing solar radiation than are plants with yellow coloured keels. This example raises two important issues.
Firstly, genetic change can occur over both short and long distances. The change in keel colour frequency along the 6 km length of the Spurn peninsula is similar to the change over the 300 km distance between north Yorkshire and Birmingham. This has important implications in defining what is local in terms of provenance. Frequently, different ecotypes separated by a mile or two can be more different than populations from a similar habitat, but separated by vast distances. Obvious examples of this are maritime and inland variants, metal tolerant and non-tolerant populations.
The fact that virtually all populations of Lotus contain both yellow and dark keeled individuals, demonstrates a second important point. The spread of variation found within each population of a species is often great enough to include genotypes associated with other ecotypes. Locally adapted populations may differ from each other only in the frequency of the genotypes types within them. This is good news for those wishing to translocate plant material in the cause of conservation, for it implies that populations will rapidly be acted upon by selection, until locally adapted genotypes again predominate.
By the 1960s our understanding of many of the early studies of adaptive genetic variation within species had become complicated by the discovery of cytotypes. It became apparent that many species occurred as groups of individuals which differed from other populations cytologically (most commonly in chromosome number). Major genetic differences between populations of this type are not surprisingly reflected in morphological and ecological differences. This phenomenon is known in many species, some likely to be the target of genetic translocations as part of habitat creation attempts. These include: Achillea millefolium, Centaurea nigra, Leucanthemum vulgare, Caltha palustris, Ranunculus ficaria, Dactylis glomerata and Agrostis stolonifera.
Unfortunately, the study of cytotypes has become rather unfashionable in recent years and there is a lack of detailed knowledge regarding their geographical and ecological distributions. Unlike other adaptive variation, populations differences based on cytology are likely to be discrete, thus offering no opportunity for introduced populations to be selected to the extent that they converge on the original local form. Furthermore, although remote the introduction of plant material differing in chromosome number from the remnant local population raises the possibility of hybridisation and consequently sterility problems. While the chances of this are not high, it would be embarrassing if a conservation project were to have such consequences.
The flora of the UK has experienced much interference at the hand of man. However, never before have non-agricultural species deliberately experienced such a magnitude of gene flow. The use of non-local genotypes as part of conservation efforts may result in anything between the take over of superior new genotypes to the failure of unadapted types to thrive. In reality somewhere between these two extremes seems most likely. Introduced 'alien' populations are most likely to be acted upon by natural selection until they come to resemble the original local population. This may unfortunately happen at the expense of information hidden within neutral genetic markers, but this must be balanced against the need to act to conserve the species in which the information is buried.
Some species are known to contain distinct ecotypes which may be built on the back of cytological differences between population. Greater thought should be given translocating material of these species around the country. More information is needed regarding how widespread cytological variation within species is and over what scale it is distributed. In the words of Sir Robert May (1997) "Eminent Personages scattering wildflower mixtures along roadsides is all very well if your view of 'wildflowers' is shaped by Victorian chocolate box tops, but a great deal less sensible if you are aware of the realities of local varieties and the structure of local gene pools".
Crawford, T.J. & Jones, D.A. 1988. Variation in the colour of keel petals in Lotus coriculatus. Heredity 61, 175-188.
Ludwig, F. 1901. Variationsstatistische probleme und materialen. Biometrika, 1, 11-29.
May, R.M. 1997. Forward. In, The role of genetics in conserving small populations. Eds. T.J. Crawford, J. Spencer, D. Stevens, M.B. Usher, T.E. Tew & J. Warren. JNCC symposium volume. 5-6.
Stapledon, R.G. 1928. Cocksfoot grass (Dactylis glomerata L.): ecotypes in relation to the biotic factor. Journal of Ecology, 16, 72-104.
Turesson, G. 1925. The plant species in relation to habitat and climate. Hereditas, 6, 47-236. March 1997
Since preparing this paper, John Warren has relocated to Aberystwyth University. Contact:jhw at aber.ac.uk">firstname.lastname@example.org