Analysis of grain-yield components and inflorescence levels in winter-type white lupin

In winter-type white lupin (Lupinus albus L.), an improved understanding of the importance of inflorescence levels and yield components among environments could be useful when considering selection criteria for breeding. We examined inflorescence levels and yield components of indeterminate, determinate, and dwarf accessions among a broad range of environments. Data were acquired from date-of-seeding (DOS) and dry matter partitioning (DMP) studies. The DOS studies were conducted from 1991 to 1993 in north, central, and south Alabama, USA. The DMP studies were conducted for 2 yr in western France and central Alabama. We used principal-component analysis to reduce the number of yield components for further consideration in future breeding. The first four principal components had Eigenvalues >1 and contributed greater than or equal to 74% of the variance. Pod number m(-2), seed number m(-2), pod yield m(-2), and seed yield m(-2) were highly correlated (r greater than or equal to 0.80, P less than or equal to 0.01) with most principal components. The first two to four principal components also had biological meaning because yield components that were highly correlated with each principal component were associated with a specific inflorescence level. Mainstem and primary-branch inflorescence levels were most important for yield and yield-component development. Basal branches were important in the southeastern USA, but these inflorescence levels should be selected against in France. Pod number m(-2) and seed weight had the greatest effect on grain yield. Breeding for higher seed weight will improve cold tolerance and stabilize grain yield in western France, but further testing is needed in Alabama. Selecting germplasm with low pod abortion will help increase pod number potential.