The development of genetically improved cotton through molecular biotechnology will have a significant impact on the future of the cotton industry. However, studies evaluating the effects of genetic transformation on critical agronomic traits is limited. The objectives of this study were 1) to determine the tint yield potential of cotton strains genetically transformed to express both Buctril^® tolerance and production of bacillus thuringensis insect toxin, and 2) to compare the lint yields of these strains compared to the parent strain from which they were derived.

Yield trials were conducted in 1992 at three locations: Hartsville, SC, Stoneville, MS, and Maricopa, AZ. Trials were randomized complete block designs with four replications of 37 entries: 30 transgenic cotton strains, six non-transgenic check cultivars, and one non-transgenic parent strain. The 30 transgenic strains were derived from genetic transformation of 13 individual plants, or events, of the parent strain. Each transgenic strain carried the genes for Buctril tolerance and production of bacillus thuringensis insect toxin. Among the 13 events (analogous to 13 populations), three events were represented in the trial by at least four sister strains.

Analysis of variance for yield showed significant differences among the 37 entries at each location and across all locations ( < 0.01) Among the top 10 entries across all locations, five were check cultivars and five were transgenic strains. There was no statistical difference between any of the five transgenic strains and the highest yielding cultivar in the experiment, 'Stoneville LA 887'. These five transgenic strains also yielded from 25 to 43% more than the parent strain. These data indicate the yield potential of strains selected from genetically transformed plants is great enough to expect that many strains could compete favorably with current superior cultivars. Another series of analyses were conducted using only the parent strain and the three events which had at least four sister strains in the trials. The purpose of these analyses was to determine if one should expect that the choice of which plants to transform had an effect on yield of transgenic strains selected from them. If such effects existed, then event means should not be equal to the mean of the parent strain.

Linear contrasts showed that there were no differences between the mean yield of the parent strain and any of the three events. Also, there was no interaction between events and locations. However, there were significant yield differences ( < 0.05) among strains within events, Yields ranged from 711 to 954, 684 to 1068, and 807 to 1073 pounds per acre for each event, respectively, versus 819 pounds per acre for the parent strain. The possible causes of variation among strains within events could be 1) somaclonal variation or variation due to the point of gene insertion in the original transformed plant, 2) heterogeneity of the parent strain, or 3) random error associated with the trials.

Based on the results of these analyses, our preliminary conclusions are 1) yield potential of superior transgenic cotton strains carrying the genes for Buctril^® tolerance and production of bacillus thuringensis insect toxin were comparable to current superior cotton cultivars, 2) yielding ability of the parent strain was recovered successfully in transgenic cotton strains, and 3) potential exists to select within events for strains with yield superior to the parent strain.