Cotton harvest losses due to delayed picking after defoliation and out-of-adjustment pickers can be as high as 20%. This study was conducted to determine the effects of a new harvest aid attachment, called the Boll Saver (which was developed by the Hydrapak Corp., Morgan, GA), on harvest losses as affected by cotton variety and adverse harvesting conditions in South Carolina. Replicated tests were conducted for 3 yr (1994–1996) during harvest seasons at the Edisto Research and Education Center near Blackville, SC. In all 3 yr, cotton was picked later than an ideal time for harvest. The two middle rows of each plot were machine harvested for yield determinations of eight varieties of cotton, either with or without the Boll Saver attachment. In all 3 yr, stalk and ground losses, percent lint turnout, and yield from each plot were measured. Use of a Boll Saver attachment significantly reduced ground and total harvest losses during the three harvests for all cotton varieties. Savings in lint cotton ranged from 27 to 59 kg ha-1. There were no significant differences in trash content between samples picked with and without the Boll Saver attachment in 1996, except in Stoneville 474 cotton which had lower trash content without the attachment. The Boll Saver attachment had no effect on lint turn out within a given cotton variety. In 1996, stalk losses were significantly less with the Boll Saver attachment except for Suregrow 125 and Deltapine 51.