Abstract: 【Objective】 Red sprangletop (Leptochloa chinensis (L.) Nees.) is a troublesome weed seriously damaging rice growth and reducing rice yield in direct-seeding rice fields. The objective of this study is to investigate the effects of L. chinensis on rice growth in direct-seeding rice fields and its economic threshold, and provide a scientific basis for the formulation of control parameters on L. chinensis in direct-seeding rice fields and its control measures. 【Method】 A field experiment was carried out to study the effects of L. chinensis of different densities (0, 1, 3, 6, 9, 12, 15, 18 and 21 plant/m²) on the growth of middle, early and late rice of double seasons and their yields by the method of additive series experiment. The function relation between L. chinensis of different densities and forming factors of rice yields and their yield loss were investigated by the method of curve fits. Control economic thresholds on L. chinensis in different types of paddy fields were derived on the basis of the economic infestation level (EIL) formula. 【Result】 The plant height, tillers, effective spikes, filled grains per panicle, 1000-grain weight and grain yield of direct-seeding rice decreased gradually with the interference of L. chinensis as the weed density increased. Compared with the control, plant height of early, middle and late rice dropped by 11.73%, 4.80% and 7.82% at L. chinensis density of 21 plant/m², respectively, tillers dropped by 22.08%, 29.67% and 19.22%, respectively, effective spikes dropped by 12.00%, 25.57% and 21.43%, respectively, filled grains per panicle dropped by 13.28%, 15.07% and 14.82%, respectively, and 1000-grain weight dropped by 12.11%, 11.99% and 7.03%, respectively, which resulted in rice yield dropped by 39.22%, 44.36% and 37.80%, respectively. The quadratic regression model y=ax²+bx+c was relatively better for describing the relationship between L. chinensis density and plant height of early rice (y=-0.0037x²-0.1389x+57.2631, P＜0.001), middle rice (y=-0.0007x² -0.0816x+77.0617, P＜0.001) and late rice (y=-0.0006x²-0.2032x+78.0150, P＜0.001), tillers of early rice (y=0.0229x²-3.5944x+331.5809, P＜0.001), middle rice (y=-0.0529x²-5.3734x+464.9348, P＜0.001) and late rice (y=0.0302x² -3.7612x+399.7899, P＜0.001), and effective spikes of early rice (y=-0.0259x²-1.3288x+282.2607, P＜0.001), middle rice (y=0.0208x²-4.9362x+366.0443, P＜0.001) and late rice (y= 0.0412x²-3.6170x+326.3593, P＜0.001), while logarithmic regression model y=alnx+b was the best for filled grains per panicle of early rice (y=-2.8309lnx+87.8309, P＜0.001), quadratic regression model y=ax²+bx+c was the best for that of middle rice (y=0.0434x²-1.6046x+94.9940, P＜0.001) and late rice (y=0.0084x²-0.8042x+93.7427, P＜0.001). Meanwhile, the quadratic regression model y=ax²+bx+c was relatively better for describing the relationship between L. chinensis density and 1000-grain weight of early rice (y=0.0051x²-0.2633x+27.1196, P＜0.001), middle rice (y=-0.0003x²-0.1312x+25.7605, P＜0.001) and late rice (y=-0.0022x²-0.0335x+26.0215, P＜0.001), and grain yields of early rice (y=2.1426x²-144.7765x+6674.9628, P＜0.001), middle rice (y=5.3452x²- 298.2884x+8933.8063, P＜0.001) and late rice (y=1.4743x²-156.4516x+7906.4028, P＜0.001), while the power regression model y=ax^b was the best for yield loss of early rice (y=8.6313x^0.4757, P＜0.001), middle rice (y=2.5916x^0.9722, P＜0.001) and late rice (y=4.6306x^0.6864, P＜0.001). For yield levels of 7 500 kg•hm^-2 (early rice), 9 750 kg•hm^-2 (middle rice) and 8 625 kg•hm^-2 (late rice), the economic infestation levels for manual weeding were 12.73%, 8.80% and 10.07%, respectively, and the red sprangletop economic thresholds were 2.26, 3.52 and 3.56 plant/m², respectively. Also for yield levels of 7 500 kg•hm^-2 (early rice), 9 750 kg•hm^-2 (middle rice) and 8 625 kg•hm^-2 (late rice), the economic infestation levels for application of 10% cyhalofop-butyl EC or 30% oxaziclomefone SC were 2.26%-2.72%, 1.56%-1.88% and 1.77%-2.12%, respectively, and the L. chinensis economic thresholds were 0.06-0.09, 0.59-0.72 and 0.25-0.32 plant/m², respectively. 【Conclusion】The inhibitory effects of L. chinensis on rice were positively correlated with its density, which there were differences among early, middle and late rice. Also the economic thresholds of chemical control on L. chinensis showed early rice＜late rice＜middle rice based on the analysis of control economic thresholds.

Key words: red sprangletop (Leptochloa chinensis), paddy rice, curve fit, economic infestation level, economic threshold