Assessing the condition of an agricultural crop through remote sensing and crop modeling requires knowledge of crop characteristics related to plant canopy structure and reflectance. A study was conducted to characterize the structure and reflectance of cotton (Gossypium hirsutum L.) leaf canopies over 3 yr using detailed physical and spectroradiometric measurements. Three canopy characteristics (leaf canopy reflectance, width, and height) were measured in several fields in southern Texas and in one field in Arizona. Canopy reflectance in the near-infrared was independent of canopy density, while canopy reflectance at red wavelengths was only slightly affected by canopy density. Reflectance appeared to involve primarily the top layer of leaves in the cotton canopy. Canopy width, when normalized by leaf area per plant and plant spacing, could be fit with a single relationship for all locations and years. A single relationship was also found to apply to canopy height. The consistency of these characteristics between locations and years suggests that they might be incorporated into physiology-based cotton models to allow explicit simulation of canopy reflectance.