Abstract
The effect of pre-harvest rodent damage on the yield of rice is difficult to demonstrate because other factors affecting yield tend to obscure the effect and because examination of the crop, unless carried out on the day of harvest, fails to give an accurate picture of damage. Here we show that tiller density is the main non-rodent factor affecting yield. By using a partial regression equation to hold tiller density constant a statistically significant relationship was established between yield and damage counts made at harvest. In ten rice fields the average yield reduction attributable to rodents was 19% and ranged from 2% to 43%. Losses to rodents may be assessed more easily by comparing actual yield with potential yield as estimated from tiller density, provided that the relationship between yield and tiller density is known, for a series of fields that differ primarily in the degree of rodent attack.