Comparison of field methods for selection of protein and oil content in soybean

Soybean [Glycine max (L.) Merr.] is known for its high protein and oil content but a gradual decline has been reported in protein content of soybeans grown in the northern latitudes during the past 10 to 15 yr. Due to this decline, research investigating breeding methods aimed at improving protein content in soybean is needed. The objectives of this study were to compare soybean protein and oil content of: (i) row vs. hill plots, (ii) sampling at the end vs. center of the row, (iii) short, one-row plots at one location vs. long, four-row plots at two locations, and (iv) to assess the heritability of these two traits. Protein and oil content were measured among eighteen cultivars planted in 30-cm x 30-cm hill plots and 1.8-m row plots, and among four cultivars planted in separate experiments aimed at comparing the end vs. center of the row in 1.8-m row plots. Four F₅ and F₆ populations were planted in 1.8-m and 3.6-m row plots at five MN locations over 2 yr to determine the effect of plot size and the heritability of protein and oil content. Protein and oil content were not affected by plot type, supported by a high correlation between row and hill plots for both protein and oil content (r = 0.82, P > 0.01). Sampling position in the row had a large effect, with higher protein and lower oil content in seed collected at the end of the row. Selection for protein and oil content in short, single-row plots at one location appears to be equivalent to selection for protein and oil content from long, multiple-row plots at multiple locations, although year effects may have an impact. The heritability estimates for protein and oil content were 0.47 to 0.54 and 0.44 to 0.46, respectively. Our results support the continuation of short, single-row evaluation tests for protein and oil content with the consideration of using hill plots as an alternative to row plots.