Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Bracts in upland cotton, Gossypium hirsutum L., have been implicated in byssionosis, a lung disorder associated with textile mill workers. Quantitative genetic studies were conducted to determine the inheritance of bract surface area and ratio of bract size/lint weight per boll. A diallel analysis was performed using seven parents and their 21 F(,1) and F(,2) populations. The parents were specifically chosen to represent a range of bract sizes, thus inferences were only made to the material included in this study. The two year-two location parent means ranged from 4.1 to 8.7 cm('2) in bract surface area and 2.7 to 5.6 ratio (cm('2)/g) of bract size/lint weight per boll. The experiments were conducted at Alexandria and Baton Rouge, Louisiana in 1978 and 1979. Heterosis was not detected for bract surface area. A partial failure of the assumptions of no epistasis, no multiple allelism, and independent gene distribution was observed in two of three general tests for bract surface area. Epistasis and multiple allelism were specifically tested and were considered negligible. Discrepancies were noted between generations for estimates of frequency of alleles and dominance gene effects. Partial dominance was expressed at most loci where dominant alleles governed bract surface area, and the majority of dominant alleles conditioned smaller bracts. Narrow-sense heritability estimates indicated that bract surface area heritability was primarily additive. Significant positive genotypic correlations were calculated between bract surface area and the traits, 2.5% span length, boll weight, and lint weight per boll. Also, T(,1) strength was significantly correlated, both genotypically and phenotypically a negative direction, with bract surface area. Since bract surface area was positively correlated with boll weight and lint weight per boll, the ratio of bract size/lint weight per boll was investigated as a more accurate measure of true changes in bract size relative to bolls. Highly significant negative heterosis was detected for bract size/lint weight per boll. Epistasis and multiple allelism were determined to be insignificant factors in the inheritance of this trait even though all assumptions were not fulfilled. Partial dominance was exhibited at most loci where dominance was found, and the major portion of dominant alleles conditioned smaller ratios of bract size/lint weight per boll. This ratio was positively correlated with 50% span length and negatively correlated with fiber micronaire. Apparent discrepancies between generations and negative values for various estimates were explained on the basis of assumption failures and the inability of the tests of assumptions to detect small violations of the assumptions.