The development of Upland cotton (Gossypium hirsutum L.) varieties is facilitated by knowledge about the heritability of the traits concerned. Previous heritability estimates for yield components and fiber quality traits in cotton have been not included bias or standard error calculations and have also been determined for a limited number of genotypes. Using 10 randomly selected genotypes from a variety trial experiment involving three harvest dates, heritabilities, bias', standard errors and genetic coefficients of variation were calculated across dates and for individual harvest dates for yield components and fiber properties. Across harvest dates, broadsense heritabilities and their associated standard errors were 0.29=0.16, 0.18=0.19, 0.22=0.14, 0.14=0.13, 0.30=0.18, 0.09=0.08, 0.34=0.22, and 0.47=0.16 for lint percentage, 100 seed weight, boll weight, lint yield, 2.5% span length, uniformity index, strength and elongation, respectively. Heritability biases were 0.32, 0.37, 0.22, 0.15, 0.446, 0.05, 0.81, and 0.81, respectively. Micronaire had a negative heritability and was therefor assumed to be zero. Single harvest date heritability estimates showed considerable fluctuation. Heritability estimates from this experiment tended to be lower than those previously reported which may be a reflection of their status as elite commercial varieties. Nevertheless, sufficient variation for further progress in breeding for improved yield and fiber through the crossing of elite varieties exists though progress may be slower than previous studies might have suggested.