Genotype X environment interactions over seven years for yield yield components fiber quality and gossypol traits in the regional high quality tests

Authors: Meredith Jr, William; Boykin, Deborah; BOURLAND, F - University Of Arkansas; CALDWELL, W - Louisiana State University; Campbell, Benjamin - Todd; GANNAWAY, J - Texas A&M University; GLASS, K - Auburn University; JONES, ANDREA - University Of Missouri; LORD, MAY - Delta & Pine Land Company; SMITH, C - Texas A&M University; ZHANG, JINFA - New Mexico State University

Submitted to: Journal of Cotton Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/24/2012
Publication Date: 6/15/2012
Citation: Meredith Jr, W.R., Boykin, D.L., Bourland, F.M., Caldwell, W.D., Campbell, B.T., Gannaway, J.R., Glass, K.M., Jones, A.P., Lord, M.O., Smith, C.W., Zhang, J. 2012. Genotype X environment interactions over seven years for yield yield components fiber quality and gossypol traits in the regional high quality tests. Journal of Cotton Science. 16:16-169.

Interpretive Summary: Cotton is an important crop whose fiber is used in the production of clothing and other industrial uses. Its second crop is its seed which can be reused for human and animal feed. To successfully compete with synthetic fibers and world competition, cotton breeders and growers need to determine what varieties/strains perform the best in a specific area. Interactions occur when the comparative performance of any two varieties are not consistent across a wide range of test environments. Decisions on best breeding approaches to take or which variety to grow are made based on the size of the variety variability and its interaction with environments variability. When the size of the interaction variability is equal or greater than the variety variability, then growers need to either partition the environments into smaller groups or increase the number of tests. Either solution is very costly and time consuming. However, if the variety variability is greater than twice the interaction variability, less testing and time is required to select the best performing varieties. Over time, changes in variety types, such as maturity and transgenics, occur. Therefore, researchers should periodically reevaluate variety-interaction of new variety types. The objective of this study involving 98 genotypes grown in 56 environments was to determine the variability associated with varieties and their interactions. Despite the many changes in breeding, management systems, and test environments, the interactions have changed little in the last 50 years. Yield is the most economically important trait followed by fiber traits. Its interaction variability is still greater than variety variability, therefore considerable testing is needed to effectively determine the best yielding varieties. However, for fiber traits such as length, length uniformity, strength and micronaire, the variety variability is more than twice that for the interactions. Gossypol has even greater variety variability superiority over the interaction variability that that of fiber traits. These data support the concept of identifying optimum testing environments for specific varieties. The need for regionalization yield evaluation is still needed. However, fewer environments are needed for fiber and gossypol evaluations.

Technical Abstract: Since the 1960s, many changes in cotton (Gossypium hirsutum L.) cultivar tests have been made. This study partitions the total variation for 26 traits into environments (E), genotype (G) and GE variance components for the 2001 through 2007 Regional High Quality (RHQ) tests with 98 genotypes. It evaluates 26 traits and consists of 56 year-location environments. There were four yield traits, five yield components, six traditional breeder-geneticists (BG) fiber traits, seven High Volume Instrumentation (HVI) fiber traits, and four gossypol traits. Yield variance components for lint, seed, oil, and N were similar with an average of 87, 5, and 8% of the total variance due to E, G, and GE, respectively. Lint %’s E, G, and GE were 57, 33, and 10%, respectively and was similar to oil % which was 53, 37, and 10%, respectively. Length, strength, and micronaire’s G components for BG fiber was 28, 52, and 16%, respectively. For the HVI samples, G was similar with 36, 48, and 18%, respectively. Average G for total gossypol and its two isomers, plus (+) and minus (-) was 36, 47, and 29%, respectively. The plus (+) percent of total gossypol was 17, 72, and 11% for E, G, and GE, respectively. This is the lowest E% and highest G% of all the 26 traits. The results of this study suggests in the last 50 years, little changes in E, G, and GE variance components have occurred.