Environmental and economic pressures on production agriculture have increased the interest in variable rate nitrogen fertilization as a method to improve nitrogen (N) use efficiency. Limited information exists on the spatial variability of corn yield and quality responses to N fertilizer which were examined in this study. Seven on-farm experiments were conducted during the 2003 and 2004 growing seasons in two growing areas in Illinois. Five (2003) and six (2004) N fertilizer rates ranging from 0 to 300 kg N ha-1 were applied with commercial applicators in five to ten areas within each field. At harvest corn ears were sampled at four georeferenced points within each plot. Yield, kernel number, individual kernel weight, and grain protein concentration were determined for each sample. The response of the variables to N fertilizer was analyzed using polynomial regression. Yield response to N varied both between fields and each field. Some regions within fields did not respond to N whereas others showed a large response to N. Within each field, the spatial variability of corn yield without N fertilizer (for example 3.3 to 13.4 Mg ha-1 in one field) was larger than the variability of maximum yield. Kernel number and individual kernel weight also varied in their response to N fertilizer within fields, although kernel number showed a higher correlation with yield than did kernel weight. Grain protein concentration increased with N fertilization and its response to N fertilizer varied within fields and tended to agree spatially with the yield response. Grain protein concentration showed a sigmoidal association with yield, indicating that within certain yield levels protein concentration increases rapidly with grain yield. These results indicate that variable rate N application could be a potentially useful tool to optimize grain yield and quality.