An understanding of the physiological changes that have occurred in cultivar development in the past would be helpful in the continued production of high yielding cultivars of the various crop species. Field experiments using conventional cultural practices with supplemental irrigation were conducted in 1976 to determine the physiological basis for yield differences between old, low yielding soybean [Glycine max (L.) Merr.] cultivars (Lincoln' and ‘Dorman’) and new high yielding cultivars (‘Williams’ and ‘Essex’). Mechanical support was used to prevent lodging for comparison to control plots with no mechanical support. There was no lodging in Lincoln and Williams (Maturity Group III); whereas, in maturity group V, Dorman lodged significantly more than Essex in the control plots. However, there was no relationship between lodging and yield. Williams produced 34% more yield than Lincoln, primarily because of larger seed. Essex produced 88% more yield than Dorman, primarily because of more seed per unit area. Yield differences were not correlated with total shoot weights or CO2 uptake rates of a single leaf during vegetative or early reproductive growth. Rates of acetylene reduction were similar for Williams and Lincoln, but Essex had a higher rate than Dorman. The yield advantage of Williams resulted partially from a longer filling period (12%), which was associated with higher CO2 uptake and acetylene reduction rates late in the filling period. Essex and Dorman produced similar total shoot weights late in the filling period, so the yield advantage for Essex resulted from partitioning more photosynthate to the seed. The data suggest that yield improvement in the future may be possible by lengthening the filling period. It may not be possible to increase seed yields by increasing the partitioning of photosynthate to the seed if the partitioning in current cultivars is approaching the maximum.