The application of spatial autocorrelation methods to the study of Calluna vulgaris population genetics

Spatial structure of Calluna vulgaris populations was investigated at various scales using autocorrelation statistical techniques. Intrapopulation structure has been studied by recording the genotypes at four polymorphic loci of 41 individual plants of Calluna vulgaris within a grid of 16 x 60 m, from the site of Sacrawe (High Ardenne, Belgium). The spatial correlograms gave no significant results of autocorrelation of genotypes within the grid. It is therefore concluded that at this scale there is panmictic mixing of genotypes. A short review of other similar studies show that patterns vary greatly between species. Interpopulation analyses were carried out by recording the allele frequencies of 18 populations from North Spain to Belgium in relation to their geographic position. This revealed different patterns for the different alleles studied including dines of frequencies, autocorrelation at regional scale or no detectable trends. It is concluded that autocorrelation usefully supplement other statistical methods, such a Wright's F-ST by giving information on the scales and processes involved in intra- and inter-population spatial structuring.