A border effect and yield overestimation due to a border effect have been studied for rice plants grown in small plots with the square geometry of transplanting. The objectives of this study were to measure the border effect of a plot transplanted in rectangular geometry with wide and narrow hill spacing and to quantify the effects of plot size and shape on yield overestimation due to a border effect. Two hybrid rice varieties (Zheyou 3 and II-you 838) were grown in a farmer's field with a hill spacing of 0.133m×0.267m in 2011 in Hubei province, China. Grain yield and yield components were measured for outmost row, second and third outmost rows, and center rows. A significant border effect on grain yield was observed in the outmost row, but not in the second and third outmost rows in comparison with the center rows for both varieties. Higher biomass production, more panicles per m 2 and spikelets per panicle, and higher grain-filling percentage were responsible for the border effect. A larger border effect was observed on sides with wide hill spacing than with narrow hill spacing (142% vs. 62%). An equation was developed to calculate yield overestimation by considering the border effect of sides with wide and narrow hill spacing separately. According to this equation, minimum yield overestimation rate due to a border effect was 2.7% for a plot with an area of 1mu (1mu=1/15ha). Yield overestimation would be substantially higher for a plot with more rectangular shape and smaller plot size.