Data gathered from the Ministry of Agriculture, Fisheries and Food (MAFF) National Rabbit Survey sites in the 1980s showed substantial site-to-site and year-to-year variation in rabbit Oryctolagus cuniculus abundance. Data on physical and biological factors from 450 sites were collected to seek explanations for this variation. The rabbit abundance index on the 2.25 km transects (walked along the edge of fields, woods, hedges, etc.) was used as the dependent variable for multiple regression and principal components analyses. Additionally, a preliminary attempt at mapping rabbit abundance and distribution in England and Wales using a GIS system was made. Each of the analyses produced a list of the important variables enhancing or reducing rabbit indices. Common to each list were aspects of woodland – area, number of pieces and length of edge; the dryness and patchiness of soils on site; rainfall in late winter, or variables associated with wetness of soil (e.g. number of days with rain, altitude). Other important variables included predator removal policy and the site location nationally. The two strongest components, woodland parameters and predator removal policy, are under human influence. The effects of region, county and year were small, but year-to-year variation within sites was larger, suggesting that rabbit populations were reacting markedly to unmeasured variables, e.g. myxomatosis outbreaks. The analysis provides a useful explanation of the factors that affect rabbit populations and a basis for how these factors may be manipulated.