Atlanta, Georgia
January 17, 2002

D&PL database plays key role in analyzing choices

Variety selection by cotton growers and the crop advisors, consultants, extension personnel and company representatives who assist them, has become a more difficult task each year. This is due in large part to the number of varieties and the transgenic traits offered. Determining the best choice may be aided by using a regression analysis that includes more information than the traditional averaging of yields by variety over a wide geographic area.

"What a farmer needs is a better method of looking at his own fields and analyzing which variety is the best choice for each one," says Dr. David Albers, Delta and Pine Land Company's director of technical services for the western U.S. "Farmers know the yield history of their fields may be quite different from statewide or regional averages. Summarizing information by averaging tests across a region will determine which variety has the highest yield over all the locations, but may not determine the best variety to plant on a particular farm or field."

A regression analysis can compare two varieties in head-to-head regional trials and determine the yield and crop value response over a rangne of environments. A farmer can then apply this information to his own fields. "Properly analyzed, a farmer may find that the highest yielding variety on a regional basis may or may not be the best selection for his farm, either in yield or crop value," says Albers.

The research demonstrated variety comparisons using D&PL's Agronomic Information System performance database, which is comprised of data sources from D&PL, university and Extension variety tests from the Mid-South states of Missouri, Tennessee, Arkansas, Mississippi and Louisiana. He compared varieties that are adapted within the region and avoided data from regions for which the selected varieties are not adapted. Albers said the Mid-South region was selected because it has a large database, which makes for a more meaningful analysis.

The analysis method used includes yield scatterplots of two different varieties and a simple linear regression comparison of yields by variety. A slope and intercept point is established and a crossover point is determined where the two varieties are equal, either in yield or crop value, whichever is being analyzed.

Albers illustrated the following examples of regression analysis to determine variety selection. In one example, he compared DP 451 B/RR to ST 4892 BR. In this head-to-head comparison the crossover point was 764 pounds per acre based on yield and $571 per acre based on crop value. The DP 451 B/RR was favored where yields were less than 764 pounds per acre and ST 4892 BR was favored
where yields were more than 746 pounds per acre. Therefore, based on yield only, DP 451 B/RR would be the preferred choice for a farmer whose field historically yielded one to 1.5 bales per acre.

In another example PM 1218 BG/RR is compared to ST 4892 BR for both yield and crop value. In this case PM 1218 BG/RR had a 95 lb./acre lint advantage and $49/acre crop value advantage over ST 4892 BR. The regression analysis of the lint yields for this head-to-head comparison showed yield environments up to 1337 lb./acre favored PM 1218 BG/RR and crop values up to $817/acre.

Crop values are based on the 2001 USDA crop loan chart of $.0.52 per pound.

Albers says that this approach does not remove the need to assess varieties for local soils, diseases and other factors, but does allow a broader view of performance that typical multi-location and multi-year summaries do not permit.

For more information, see 2002 Final Beltwide Presentations on the Delta and Pine Land Company's website.

Delta and Pine Land Company is a commercial breeder, producer and marketer of cotton planting seed, as well as soybean seed in the Cotton Belt. For more than 80 years, the Mississippi-based company has used its extensive plant breeding programs drawing from a diverse germplasm base to develop superior varieties. Delta and Pine Land (NYSE: DLP) has offices in eight states and facilities in several foreign countries.