BREEDING COTTONS FOR CONVENTIONAL AND LATE-PLANTED PRODUCTION SYSTEMS

Cotton (Gossypium hirsutum L.) growers in South Carolina are replacing some of their soybean [Glycine mar (L.) Merr.] hectareage planted aft er wheat (Triticum aestivum L.) with cotton, A breeder must decide whe ther a small but expanding hectareage of late-planted (LP) cotton will require a breeding scheme separate from one for the conventional (CN) full-season production system. The objective of this study was to ass ess selection strategies that will result in cottons with improved lin t yield and fiber quality for CN and LP production systems. Twenty-fiv e unselected Pee Dee cotton genotypes, along with a full-season and an early maturing cultivar, were evaluated in a 2-yr experiment conducte d at Florence, SC, in two production systems: CN, planted approximatel y the first week of May, and LP, planted 8 June. Lint yield, fiber, an d spinning properties were determined. Significant (P < 0.05) genotypi c variation existed for all traits, but there was no significant genot ype x production system interaction. Heritability of each trait plus t he genetic correlation between production systems was used to compare predicted and observed direct and correlated response in each producti on system. Predicted gain in LP yield when selection was conducted in the CN production system was a 2.1% increase in the population mean co mpared with a 3.1% observed response. Selection for 2.5% span length a nd yam strength in the CN system resulted in gains in both production systems, Antagonistic genetic correlations between some of the fiber t raits and lint yield will make concurrent improvements in both product ion systems difficult. Use of a low selection intensity in the CN prod uction system in the initial round of replicated yield testing was fou nd to be an efficient method to improve lint yield or yam strength.