Good plant establishment is the critical first step in growing a crop. To achieve this, the correct seed rate must be calculate. This is done by assessing the optimum target plant population per m² and then making an estimate of any  losses over winter. Losses will depend on the quality of seedbed created which is related to texture, stoniness and compaction of the soil. If there is any variation in these field characteristics then the correct seed rate will also change across the field. This study aims to assess the benefits of varying seed rates. In September 2009, twelve fields were surveyed using emi techniques. Homogenous emi zones were created and then ground truthed to record soil texture, stone content and compaction. An assessment of the seedbed quality in each zone and plant losses overwinter was made. This information was then used to create a variable seed plan which was used in a GPS enabled drill. After plant emergence an assessment was made of the seed bed quality based around clod size and distribution. The correlation between emi measurements and seed bed quality and clod characteristics was assessed. Plant counts per m² were made in each seed rate zone to establish the variable element of the seed drill had been successful and that full germination had taken place. In early spring further plant counts in the same locations were made. This allowed measurement of the differences in actual plant survival rates between seedbed quality zones compared with the estimate. The variation of plant counts within zones is also made. The relationship between emi values and plant losses is examined. The success of the variable seed rates in creating a more even plant population is made, along with any yield benefit over a uniform seed rate.