Field experiments were conducted in 1996 and 1997 to evaluate the effects of six johnsongrass densities on picker- vs. stripper-harvest efficiency, fiber properties, loan rate, and lint yield loss of cotton. The weed densities employed were 0 (the check), 3, 4, 5, 8, and 15 plants/15 m of row. With three or fewer weeds in 1996 and four or fewer in 1997, harvest efficiencies were 4.9 to 7.6% higher for stripper- than for picker-harvested cotton. At four and higher weed densities in 1996 and at five and higher in 1997, differences in harvest efficiency between the two machines were not significant. For each weed per 15 m of row, stripper-harvest efficiency in 1996 and 1997 was reduced 0.3 and 0.6%, respectively; picker-harvest efficiency was not affected by the johnsongrass densities included herein. Fiber fineness (i.e., micronaire) was significantly reduced at densities of 8 weeds/15 m of row in 1997 and at 15 weeds in both years. A questionable increase in staple length was detected at the 3-weed density in 1996. Reductions in fiber strength were noted in 1997 at densities of 3, 8, and 15 weeds/15 m of row. No influences on fiber length uniformity were shown. In 1996 the loan rate for picker-harvested lint was 570 points/kg higher than for stripper-harvested lint at 8 weeds/15 m of row. In 1997 it was 741, 801, 1,058, 1,225, 1,074, and 1,329 points/kg higher at weed densities of 0, 3, 4, 5, 8, and 15 plants/15 m of row, respectively. In 1997 picker-harvest loan rate was reduced 49 points/kg of lint, and stripper-harvest loan rate was reduced 85 points. Over both years, picker-harvest lint yield was reduced 32 to 43 kg/ha (3.9 to 5.5%) for each weed per 15 m of row, and stripper-harvest lint yield was reduced 29 to 43 kg/ha (3.5 to 5.2%).