The objectives of this project are to (1) determine the relationship between cotton canopy reflectance and cotton plant parameters that are used to determine PIX recommendations, and (2)create variable rate PIX application maps from digital aerial imagery using the model from objective (1). Two separate fields with a history of highly variable cotton growth were studied during the 2005 growing season. The first field of approximately 150 acres was divided into 0.5-acre grids with 81 sample points. A second larger field of approximately 200 acres was divided into approximately 2.5-acre grids with 83 sample points. These sample points were scouted weekly for growth parameters related to PIX recommendations. Scouting began on June 20 and continued until August 1. The parameters recorded were: plant height, number of nodes, first fruiting node, number of fruiting nodes, and internode length. Color infrared images were taken at the fields corresponding to each scouting date. These images were processed to develop an NDVI index based on the colors in the photography. Areas of interest of three meters in diameter were created around each scouting point and pixel counts were extracted from the NDVI image at each area. Simple linear regressions were run comparing these pixel counts to the plant parameter data taken at each point. Initial results show significant correlations between NDVI and cotton height and height to node ratio. ith this data it is then possible to create a variable-rate Mepiquat Chloride map for each cotton field using the Modified Early Bloom Strategy as outlined in the North Carolina Cotton Production Guide.