An Analysis of Yield Component Changes for New vs. Old Soybean Cultivars

Reasons for the gradual genetic yield improvement (10-30 kg ha(-1)yr(-1)) reported for soybean [Glycine max (L.) Merr.] during decades of cultivar development are not clearly understood. Identification of mechanisms for the yield improvement would aid in providing indirect selection criteria for streamlining cultivar development. Our objective was to identify yield components responsible for yield improvement in 18 public southern cultivars released between 1953 and 1999. The study was done at the Ben Hur Research Farm near Baton Rouge, LA (30 degrees N Lat) during 2007 and 2008, plus a validation study in 2009. Experimental design was a randomized complete block with four replications and one factor (cultivar). In the 2007-2008 study, 18 cultivars released across the 1953-1999 period were selected. Three old and three new cultivars were used for the 2009 validation study. Data were obtained on yield, seed m(-2), seed size, seed per pod, pod m(-2), pod per reproductive node (a reproductive node is one having at least one pod having at least one seed), reproductive node m(-2), percent reproductive nodes and node m(-2). Data were analyzed by ANOVAR and mean separation. Regression and path analyses were also done between yield and yield components, year of release and yield components, and among yield components themselves. Results of the 2007-2008 study indicated that yield differences were sequentially controlled by node m(-2), reproductive node m(-2), pod m(-2), and seed m(-2). However, node m(-2) was not as accurate at distinguishing low and high-yielding cultivars as the other three yield components and its role in yield formation was not substantiated in the validation study. A possible indirect selection criterion for yield during cultivar development is reproductive node m(-2).