Comparisons of Methods of Estimating the Seed Filling Period in Soybeans

An experiment was conducted at Lexington, K.Y., USA (38° N latitude) utilizing 18 soybean (Glycine max [L.] Merr.) genotypes in three stem termination groups (determinate, indeterminate, and segregating) to compare eight methods of estimating seed filling period in terms of 1) ability to differentiate among genotypes, 2) lack of bias towards any stem termination type, 3) ease of measurement, and 4) consistency across environments. Growth stage estimates R1 to R8, R4 to R7, R5 to R7 and the reproductive period duration (estimate of the time period of the linear increase in harvest index) detected significant differences among the seed tilling periods of the genotypes. Significant stem termination group measurement method interactions were present, but because the indeterminate stem termination group had a shorter seed filling period than the other two groups for most methods, we were unable to evaluate bias towards stem termination type. The growth stage methods required 4x to 10x less time for data collection than the effective tilling period and reproductive period duration methods. Only the R4 to R7 and R5 to R7 growth stage estimates had significant rank correlation coefficients among genotypes across years. From this and previously published evidence we recommend using the R5 to R7 estimate of seed filling period following the separation of genotypes into stem termination classes for the selection of soybean genotypes based on seed filling period duration.