Boll counts were taken in four entries in two separate on-farm variety trials in Oklahoma prior to boll opening and at harvest. Lint weight per boll was determined by removing lint from at least a 50 boll sample from each count area. The corresponding lint yield prediction based on early and at harvest boll counts, as described in literature, was compared to the actual yield of the entire plot as determined by commercial harvesting equipment and ginning on the Oklahoma State University Cotton Agronomy research gin.