The effect of various infestation levels of Chilo partellus Swinhoe on resistant and susceptible sorghum genotypes was assessed under artificial infestation in two field trials, conducted over two seasons. In the first season, infestation levels were 0, 10, 20, 40, 60 and 80 larvae per 2 m plot row. In the second season, additional treatments of one and two larvae per plot were included. A series of all infestation levels on susceptible plants was also included with an insecticide application 5 days after inoculation. Non-linear relationships between infestation level and yield loss were observed. In susceptible plants, small increases in initial larval infestation levels resulted in sharp increases in yield loss. An opposite tendency was observed in resistant plants where a tolerant reaction at lower infestation levels was observed. Gain threshold (GT) value, indicating the yield increase necessary to compensate for control costs, economic injury level (EIL) and economic threshold level (ETL) at which control actions should be initiated were determined for each genotype. Insecticide application on susceptible plants resulted in a 10% reduction in yield loss. However, the yield was still less than that of unsprayed resistant plants. The GT values indicated that insecticide application was economically warranted at an ETL of 20% and 2% of plants with visible symptoms of whorl damage, for the resistant and susceptible genotypes respectively.