Barnyardgrass, large crabgrass, and Texas panicum were evaluated in field experiments over 3 yr to measure their duration of interference and density on grain sorghum yield. When grain yield data were converted to a percentage of the weed-free control, linear regression predicted a 3.6% yield loss for each week of weed interference regardless of year or grass species. Grain sorghum grown in a narrow (61-cm) row spacing was affected little by full-season interference; however, in wide (91-cm) rows, interference increased as grass density increased. Data from the wide-row spacing were described by linear regression following conversion of grain yield to percentages and weed density to log10. A separate nonlinear model also was derived which could predict the effect of weed density on grain sorghum yield.