向日葵籽实产量预测数学模型的研究

根据2年10个点次的田间试验结果建立了由数粒法、盘重法和盘径法预测向日葵籽实产量的回归数学模型，并根据1994年4个点次的产量实测结果对各模型的可靠性进行了验证．结果表明，三种方法建立的模型均有效、可靠．但在我国向日葵科研和生产实践中，以盘径法最为实用，建议推广应用．

MATHEMATICAL MODELS FOR PREDICTING GRAIN YIELD IN SUNFLOWER

Field experiments were conducted for 3 years at different sites in Jilin Province to investigate into the usefulness of three techniques for predicting grain yield at harvesting stage in sunflower.They involve in prediction of grain yield from (1) head dry weight,(2) number of full seeds per head,and (3) effective head diameter.Results showed that all of them were effective,but the mathematical model involving measurements of head diameter proved the most valuable in forecasting grain yield before harvest in view of practical application at the present stage of sunflower production in China.The predicted grain yield highly agreed with the observed yield for several sunflower cultivars in different sites.