Objective1?SDSU soybean CPT data for yield from 2001 to 2009 were provided by Dr. Robert Hall. Since some genotypes (cultivars) were not repeated in all locations every year and only a few cultivars were common between years, we analyzed these data by mixed linear model approaches from year to year.
Objective 2: Fifteen cultivars obtained from three different seed companies were grown in six locations in Eastern South Dakota. These six locations included: South East farm (2011 and 2012), Brookings (2011 and 2012), North East farm (2011 and 2012), Bancroft (2011 and 2012), Geddes (2011 and 2012), Warner (2011), and Bath (2012), where Warner and Bath are near each other. A randomized complete block design with three replications in each location was employed. Seed yield and protein and oil contents have been collected for two both years. In addition, yield components including pod number per plant, seed number per pod, and seed index were measured for all 12 environments.
We analyzed the data with different approaches: (1) environment-wide correlation analysis; (2) variance component analysis with mixed linear model approaches; (3) conditional variance component analysis with mixed linear model analysis; and (4) both conventional and conditional stability analysis.